Structural Control of Fully Condensed Polysilsesquioxanes: Ladderlike vs Cage Structured Polyphenylsilsesquioxanes

Article abstract of DOI:10.1021/acs.macromol.5b01539

Through fine-tuning of the myriad of reaction conditions for an aqueous base-catalyzed hydrolysis-polycondensation reaction, a facile synthesis of structurally controlled polyphenylsilsesquioxanes was developed. Mechanism and kinetic studies indicated that the condensation reaction proceeded through a T¹ structured dimer, which was quantitatively and in situ formed through mild hydrolysis of a phenyltrimethoxysilane (PTMS) monomer, to give either the cage-structured polyhedral oligomeric silsesquioxanes (POSS) or the corresponding ladderlike silsesquioxane (LPSQ) with excellent yields. Ladderlike and POSS materials were selectively achieved at higher and lower initial concentrations of PTMS, respectively, and an in-depth spectroscopic analysis of both compounds clearly revealed their structural differences with different molecular weights.

Full text of DOI:10.1021/acs.macromol.5b01539

Article
pubs.acs.org/Macromolecules
Structural Control of Fully Condensed Polysilsesquioxanes:
Ladderlike vs Cage Structured Polyphenylsilsesquioxanes
†
†
†,‡
,†,‡
Seung-Sock Choi, Albert S. Lee, Seung Sang Hwang, and Kyung-Youl Baek*
†Materials Architecturing Research Center, Korea Institute of Science and Technology, Seoul 136-791, Korea
‡Nanomaterials Science and Engineering, University of Science and Technology, Daejeon 305-333, Korea
*
S Supporting Information
ABSTRACT: Through fine-tuning of the myriad of reaction
conditions for an aqueous base-catalyzed hydrolysis−poly-
condensation reaction, a facile synthesis of structurally
controlled polyphenylsilsesquioxanes was developed. Mecha-
nism and kinetic studies indicated that the condensation
1
reaction proceeded through a T structured dimer, which was
quantitatively and in situ formed through mild hydrolysis of a
phenyltrimethoxysilane (PTMS) monomer, to give either the
cage-structured polyhedral oligomeric silsesquioxanes (POSS)
or the corresponding ladderlike silsesquioxane (LPSQ) with
excellent yields. Ladderlike and POSS materials were
selectively achieved at higher and lower initial concentrations
of PTMS, respectively, and an in-depth spectroscopic analysis of both compounds clearly revealed their structural differences
with different molecular weights.
INTRODUCTION
too complicated to illustrate on paper. As such, the structural
feasibility of the obtained products by Brown has been widely
■
Polysilsesquioxanes (PSQs), of chemical formula [RSiO ] ,
comprise a class of inorganic−organic hybrid materials that
exhibit unique physical and chemical properties unrealized in
purely organic or inorganic sources. These materials have
shown a myriad of advantageous properties including excellent
thermal stability, low dielectric constant, good mechanical
1
.5 n
8
debated and eventually disproved by Frey. Yamamoto also
9
studied the acid-catalyzed hydrolysis of phenyltrichlorosilane,
1
with insights into the formation of a ladderlike structured
oligomeric silsesquioxanes, but difficulties in obtaining high
molecular weights, and significant uncondensed silanol moieties
have rendered this synthesis impractical for obtaining polymeric
LPSQ materials. Nevertheless, since Brown’s hypothesis of an
idealistic perfect ladder structure, LPSQs have been a goal
many organosilicon scientists have been striving toward, as a
polymeric analogue of cage structured POSS would allow for
new applications and properties for which polysilsesquioxanes
have been limited to as fillers or composites with organic
polymers.
1
,2
properties, chemical resistance, and even biocompatibility. Of
the three known structural classes of PSQsrandom branched
sols, polyhedral oligomeric silsesquioxanes (POSS), and
ladderlike polysilsesquioxanes (LPSQ)only POSS and
LPSQs can be characterized to be of controlled structures,
where POSS generally possess oligomeric three-dimensional
cage structures while a hypothetical LPSQ is a polymeric3
analogue with a linear, double-strained siloxane backbone.
Although POSS compounds have been rigorously inves-
More recently, Zhang et al. developed a unique synthetic
method to prepare LPSQ with relatively high molecular weight
4
,5
tigated, relatively few reports of LPSQ have been reported,
especially for perfect ladder structured polysilsesquioxanes with
high molecular weight, because of its difficult synthesis in spite
of their superior thermal stability, solubility, and improved film
3
,10
using template alkoxysilane monomers.
However, several
cumbersome synthetic steps for preparing the template
alkoxysilane monomers such as aromatic amine derivatives for
π−π stacking and hydrogen-bonding interactions were
required, and the templating moieties between the siloxane
bonds in the product were removed to obtain LPSQ by
6
properties.
Synthesis of ladder-structured polysilsesquioxanes was first
reported in 1960 by Brown et al. using a phenyltrichlorosilane
monomer in the presence of potassium hydroxide (KOH)
1
0
additional steps such as thermal or acid treatments. A similar
synthetic concept for rodlike LPSQs with hexagonally stacked
structures has also been reported by Kaneko et al. using ionic
7
catalyst at high temperatures (∼200 °C). While polymeric,
highly condensed siloxane structures may have been obtained
by Brown, the highly reactive trichlorosilanes used for
hydrolysis and polycondensation most probably did not lead
Received: July 12, 2015
Revised: August 18, 2015
2
9
to linear siloxane structures as hinted by the broad Si NMR
peaks, but rather branched and random structures which are
©
XXXX American Chemical Society
A
DOI: 10.1021/acs.macromol.5b01539
Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
Scheme 1. Syntheses of LPPSQ (4) and T -Phenyl POSS (5) with K CO in H O/THF at 25 °C at Different Initial PTMS (1)
1
2
2
3
2
Concentrations: (A) Higher [PTMS] for LPPSQ (4) and (B) Lower [PTMS] for T -Phenyl POSS (5)
1
2
interactions of alkoxysilane monomers; however, these
the in situ hydrolysis−polycondensation reaction with our
11,12
materials were only soluble in water.
Another approach
previously reported T -Phenyl POSS, where in-depth charac-
12
for the synthesis of perfect ladder structures has also been
examined by Unno by sequential condensations of dimeric
species, but low molecular weight and the HPLC methods used
for isolation of the products were unsuitable for scaled-up mass
terizations were also carried out to figure out their structural
differences. Moreover, in this study, we report the first
investigation of the size and the shape of LPPSQ in solution
as characterized by small-angle X-ray scattering (SAXS).
1
3
production.
A common theme among the recent strategies of Zhang,
Kaneko, and Unno for the synthesis of LPSQ materials has
entailed the use of templating silanol moieties in such a manner
that regioregular condensations may occur through hydrogen-
bonding interactions, ionic interactions, or fractionation of
irregular structures. However, the possibility remains in that,
within a conventional sol−gel reaction in which the hydrolysis−
polycondensation equilibrium is dynamic, if an intermediary
structure may form under conditions in which hydrolysis is fast
and condensation polymerization occurs in a controlled
manner, LPSQ may form in one-pot. In this study, we sought
to reinvestigate the synthetic feasibility of LPSQs by a facile
synthetic method utilizing less reactive phenyltrimethoxysilane
EXPERIMENTAL SECTION
■
Materials. Phenyltrimethoxysilane (PTMS) (Shin Etsu, 98%) was
vacuum-distilled before use. Tetrahydrofuran (THF) (J.T. Baker,
9.8%) was distilled from metal sodium with benzophenone before
9
use. Potassium carbonate (K CO ) (Sam-Jun) was dried overnight
2
3
under vacuum at 50 °C. Deuterium oxide (D O) (Aldrich, 99.9 atom
2
% D), tetrahydrofuran-d (THF-d ) (Aldrich, 99.9 atom % D), and
8
8
dichloromethane (J.T. Baker, 99.8%) were used as received.
Synthesis of Ladderlike Polyphenylsilsesquioxane (LPPSQ). In a 1
L round-bottomed flask, deionized water (24 g, 1.33 mol) and K CO
0.2 g, 1.45 mmol) were charged and stirred for 10 min. Dry THF (40
g, 0.56 mol) was added and stirred for additional 30 min. Afterward,
phenyltrimethoxysilane (PTMS) (79.32 g, 0.4 mol) was added
2
3
(
dropwise via syringe under atmospheric N , and the reaction solution
2
(
PTMS) monomer coupled with a mild base catalyst, K CO ,
2 3
was stirred for 36 h at room temperature. After 36 h, the crude was
divided into two phase as colorless and white phases. Crude white
viscous products were obtained by decantation of the colorless mixed
solvent. Crude white viscous product were dissolved in dichloro-
methane (150 mL) and extracted with deionized water (200 mL)
several times. Thee organic layers were collected and dried over
MgSO overnight. The solution was filtered to remove MgSO , and
in water/THF mixture solvent at room temperature to give
1
4
either the preferential formation T -Phenyl POSS or
1
2
ladderlike structured polyphenylsilsesquioxane (LPPSQ)
Scheme 1). While our previous study examined the in situ
(
formation of a dimeric, intermediary species for the formation
4
4
of T -Phenyl POSS, by incrementally increasing the initial
12
the volatiles removed by rotary evaporator to obtain LPPSQ. (49.1 g,
yield = 95%).
monomer concentration (up to 8.9 M), fully condensed, high
15,16
molecular weight polyphenylsilsesquioxanes
were obtained
Hydrolysis and Polymerization Conversion Ratio and Time for
Kinetics Using NMR. In a 50 mL round-bottomed flask, deuterium
oxide (2.66 g, 0.133 mol) and K CO (0.02 g, 0.15 mmol) were
which we ascertained to be of ladderlike structure through
previously reported analytical methods. However, the structure
and characterization of the ladder structure of polysilsesquiox-
anes has not been universally accepted yet in the academic
community due to the lack of conclusive evidence to elucidate
the linearity of a hypothetical high molecular weight LPSQ,
which was often confused with that of POSS. Therefore, a new,
simpler, one-pot reaction for the preparation of LPSQ materials
as well as the investigation into the analysis of the molecular
shape and size for LPSQ should be explored. In this study, we
examined the synthetic feasibility of the obtained ladderlike
poly(phenylsilsesquioxane) (LPPSQ) through monitoring of
2
3
charged and stirred for 10 min. THF-d (4.49 g, 0.056 mmol) was
added and stirred for an additional 30 min. PTMS (7.932 g, 0.04 mol)
was added dropwise via syringe under atmospheric N , and the
8
2
reaction solution was stirred at room temperature. The sample
products were removed from the reaction flask 0.5 mL per every hour,
from reaction start time to polymerization end point. All samples were
1
29
analyzed for reaction kinetics using H NMR, Si NMR, and FT-IR.
Characterization. The weight-averaged molecular weight (Mw)
and molecular weight distributions (M /M ) of the polymers were
w
n
measured by JASCO PU-2080 plus SEC system equipped with
refractive index detector (RI-2031 plus) and a UV detector (λ = 254
B
DOI: 10.1021/acs.macromol.5b01539
Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
nm, UV-2075 plus) using THF as the mobile phase at 40 °C with a
flow rate of 1 mL/min. The samples were separated through four
the reaction time increased, the methoxy proton peaks of
PTMS (1) at δ 3.6 ppm gradually became smaller and almost
completely disappeared after 4 h, while an additional peak at δ
1
columns (Shodex-GPC KF-802, KF-803, KF-804, KF-805). H NMR,
1
3
29
C NMR, and Si NMR spectra were recorded in CDCl at 25 °C on
3
3
.3 ppm appeared. This additional peak was originated from
1
13
29
a Varian Unity INOVA ( H: 300 MHz; C NMR: 75 MHz; Si: 59.6
MHz). Solid state CP-MAS Si NMR spectra were obtained on a
Varian Inova 400 MHz ( Si: 79.5 MHz) with a 7.5 mm CP-MAS
probe in air at 298 K, with a contact time of 4 ms, spun at a spin rate of
CH OD, a byproduct after hydrolysis of PTMS (1) with D O.
29
3
2
However, resultant PTMS hydrolysate (2) [PhSi(OD)3] was
not detected probably due to strong interaction of the hydroxyl
29
groups with D O. This result indicated that PTMS (1) was
1
0 kHz with an automated spinner. FT-IR spectra were measured
2
quantitatively hydrolyzed to give the PTMS hydrolysate
precursor (2) within 3−4 h. This was a unique result, as
conventional sol−gel reactions with base catalysts do not
using PerkinElmer FT-IR system (Spectrum-GX) using solvent-cast
films on KBr pellets. Thermal gravimetric analysis (TGA) was
performed by TA Instruments TGA 2950 under N . The X-ray film
2
19
diffraction and small-angle X-ray scattering (SAXS) were examined at
the beamline 3C2, 10B1, and 4C1 of Pohang light source (PLS) in the
Pohang Accelerator Laboratory (PAL). Simulation by the MM2
method was conducted by Chem3D Pro 12.0 software.
quantatively hydrolyze, but rather condense during hydrol-
ysis. Using this method, we were able to determine that the
reaction rate orders for PTMS and H O were 1 and 2.3,
2
respectively, by measuring the initial reaction rates at various
[
PTMS] and [H O] concentrations (Table 1), giving an overall
2
RESULTS AND DISCUSSION
■
Table 1. Selected Sample Reaction Conditions for K CO -
Catalyzed Hydrolysis of PTMS at 25 °C
The selection of catalyst began with an overview of base
catalysts. As base-catalyzed sol−gel reactions often suffer from
low rates of hydrolysis and high rates of condensations, several
hydroxides such as LiOH, NaOH, and KOH were disregarded
and less reactive carbonates was investigated. And while
Brown’s selection of potassium hydroxide was understandable7
due to the use of the highly hydrolyzable phenyltrichlorosilane,
the use of industrially feasible trialkoxysilane derivatives under
KOH-catalyzed sol−gel reactions were unsuccessful in obtained
fully hydrolyzed precursors. Therefore, investigation of
carbonates led us to potassium carbonate (K CO ), which
2
3
run no.
[PTMS]
[H O]
2
rate (M/s)
1
2
3
4
5
2.25
4.45
8.9
29.6
29.6
1.55 × 10−
4
−4
3.10 × 10
6.18 × 10
3.04 × 10
1.51 × 10
−4
−3
−3
29.6
8.9
59.2
8.9
118.4
1
2.3
reaction rate equation of rate = k[PTMS] [H O] , where the
2
2
3
−
8
reaction rate constant (k) was found to be (2.87 ± 0.14) × 10
gives a pH on par with the more corrosive base, potassium
hydroxide, KOH, while providing a surplus of potassium ions
which facilitate condensations between silanol groups, without
stabilizing the siloxide anion, a common phenomenon
−2.3
M
/s. The rate orders of 1 and 2.3 for [PTMS] and [H O],
2
respectively, offer insight into the factors that most greatly
affect hydrolysis rate in these set of conditions. It is worthwhile
to note that the hydrolysis rates for PTMS are significantly
lower (∼2−3 orders) compared with acid-catalyzed hydrolysis
17,18
encountered for sodium and lithium cations.
The kinetic study of phenyltrimethoxysilane (PTMS) was
first carried out to monitor the intermediary products formed
during the initial stages of hydrolysis. Figures 1A and 1B−E
20,21
of other trialkoxysilane derivatives reported previously,
allowing for facile in situ monitoring using NMR and the ability
to model the hydrolysis reaction as elementary pseudo-first-
order, which was not possible for fast acid-catalyzed hydrolysis
reactions.
Using the above reaction parameters for the hydrolysis of
PTMS (1), the kinetic rate conversion plots were plotted as a
1
show the H NMR spectra of PTMS (1) and the products
obtained during the first 4 h, where the reaction was carried out
in deuterated mixture solvents of D O and THF-d , such that
the sample was directly received from the reaction solution. As
2
8
function of time for various [PTMS] . As shown in Figure 2, the
i
hydrolysis of PTMS greatly varied depending on [PTMS]i
concentration, and the conversion of PTMS (1) to the
hydrolysate (2) increased linearly with increasing reaction
time, as indicative of the pseudo-first-order rate kinetics. By
increasing the [PTMS] up to 13.3 M and decreasing [PTMS]
i
i
1
Figure 1. H NMR spectra of PTMS (A) and the products (B−E)
obtained for the first 4 h in D O/THF-d at 25 °C with K CO at
2
8
2
3
[
PTMS] = 8.9 M. Obtained PTMS hydrolysate (2) was then
Figure 2. Kinetic rate plots for hydrolysis of PTMS at various initial
i
polymerized by condensation reaction.
PTMS concentrations.
C
DOI: 10.1021/acs.macromol.5b01539
Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
down to 0.9 M, were able to show that the hydrolysis rate of
thought to be of the PTMS hydrolysate (2) as shown in Figure
1
PTMS was pseudo-first-order with respect to [PTMS] over a
1E, but under careful review was assigned to the T structure of
i
−Si−O−Si(OH) R, as the T0 structure of RSi(OH)3
17
wide range of concentrations. This is an important result as the
complete and impeccable hydrolysis can be said to be a key
prerequisite into the formation of the highly condensed Si−O−
2
generally gives a peak around −51 ppm. This result indicated
that the product obtained after 4 h was not the PTMS
hydrolysate (2) as we concluded in Figure 1 but rather was the
22
Si structures without defect silanol groups. Moreover, the
critical [PTMS] for which the preferential formation of LPPSQ
dihydroxyphenylsiloxy dimer structure (Ph(OH) −Si−O−Si−
i
2
22
was at or above 4.5 M, while that of T -Phenyl POSS was
(OH) Ph) (3). This unexpected result has been already
2
1
2
below 4.5 M.
theoretically explained by Kudo et al., who reported that the
energy barriers of hydrolysis of trimethoxysilanes and one
single condensation (dimerization) were almost equal, because
2
9
Figures 3A(i−vi) and 3B(i−v) show the Si NMR spectra of
PTMS (1) and evolution of the condensed products obtained
at predetermined time intervals for the synthesis of LPPSQ and
T -Phenyl POSS, respectively. A sharp and single peak (a) at
1
T structured dimer was stabilized by substantial hydrogen-
24
bonding interaction. This indicated that the dihydroxy-
phenylsiloxy dimer (3) was almost simultaneously generated
as PTMS hydrolyzed. After 8 h, this sharp peak (b) broadened
and disappeared while additional broad peaks (c and d) newly
appeared at −70 and −79 ppm (Figure 3A(iii)), which
1
2
0
−
54.8 ppm originating from PTMS (1) indicative of a T
structure in Figure 3A(i) mostly shifted to −63.7 ppm (b) after
h in Figure 3A(ii). At first, this peak at −63.7 ppm was
4
2
3
indicated the formation of T [Ph−Si(OSi−) (OH)] and T
2
1
2
[
Ph−Si(OSi−) ] structures, respectively (T = 4%, T = 26%,
T = 70%). This result indicated that the dihydroxyphenylsiloxy
dimer (3) was mostly consumed by condensation reaction to
give a phenylsilsesquioxane oligomer. After 13 h, no more T
structure was found, and the T peak substantially decreased in
3
3
1
2
Figure 3A(iv), which was almost not detected after 36 h in
2
9
Figure 3A(v). After final purification steps, the Si NMR
spectra of LPPSQ are shown in Figure 3A(vi). Obtained
LPPSQ only showed a monomodal, somewhat broad peak
centered at −80 ppm, with a width of 2.7 ppm. Such high
3
fraction of T siloxane structure is generally seen only in well-
defined POSS and highly regularly structured LPSQ, where the
3
peak shape of T in LPSQ is broader than that of POSS due to
10−16
3
29
its high molecular weight.
The T peak width in Si NMR
for LPSQs has been a critical analytical tool for the formation of
ladderlike siloxane structures, usually reported as the full width
25
at half-maximum (Δ1/2). The LPPSQ obtained in our study
showed a Δ of 172 Hz actually very comparable with
1
/2
3
previous reports of LPPSQs synthesized by Zhang, which have
a Δ of 149 Hz. Moreover, the solid-state CP-MAS Si NMR
spectrum (Figure 3C) for LPPSQ showed an exceeding high
2
9
1/2
3
degree of condensation (∼98%), with the T Si-Phenyl peak
centered at −78 ppm having a Δ of 89 Hz, which was even
1/2
3
lower than that synthesized by Zhang. Thus, the structure of
the synthesized LPPSQ can be said to have a high degree of
regularity.
When the mechanism for formation of LPPSQ in Figure 3A
was compared with T -Phenyl POSS in Figure 3B, two distinct
12
differences were observed. First, while the hydrolysis of PTMS
to the dimer (Ph(OH) −Si−O−Si−(OH) Ph) (3) was
2
2
identical, the manner in which further condensations proceeded
was drastically different. For T -Phenyl POSS, the condensa-
1
2
3
tions between dimeric species yielded T [Ph−Si(OSi−) ]
3
structures directly as shown in Figure 3B(iii), indicating that
most of the silanols derived from the dimer condensed all at
once. However, for the formation of LPPSQ, the condensations
between the dimer proceeded in a much more continuous
manner, as condensations between dimers yielded oligomeric
polysilsesquioxanes and, consequently, high molecular weight
9
,26
LPPSQ. As been previously reported,
the dimer (3)
Figure 3. 29Si NMR spectra of PTMS monomer and condensation
condenses to form the T -tetraol, or so-called “half-cage”, and
4
subsequently may form ladder-oligomers under acid-catalyzed
hydrolysis−condensation. However, the base-catalyzed system
used in this study yielded high molecular weight ladderlike
intermediary products for the synthesis of (A) LPPSQ, (B) T -Phenyl
12
POSS at various times in D O/THF-d , and (C) solid state CP-MAS
2
8
2
9
Si NMR spectra for LPPSQ.
D
DOI: 10.1021/acs.macromol.5b01539
Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
polysilsesquioxanes, most likely due to the surplus of potassium
ions, which facilitate a high degree of condensation.
FT-IR analysis has been also known as strong tool to evaluate
4
,5,26
well-defined siloxane structures such as POSS
and
LPSQs. LPSQ structures are known to show two absorption
1
3
−1
peaks at 1150 and 1050 cm because of the asymmetrical
horizontal (−Si−O−Si−) and vertical (−Si−O−Si−R) direc-
13
tions of the siloxane bond. Thus, as the molecular weight of
LPSQ increases, the asymmetrical horizontal peak was expected
to increase in intensity. Figure 4 shows FT-IR spectra of the
Figure 5. SEC curves of the obtained products from 13 to 72 h for
LPPSQ (A−C) and T -Phenyl POSS (D).
12
structure with no POSS or POSS-derived structures mixed
within the product. The high molecular weight of LPPSQ was
1
also supported by H NMR (Supporting Information Figure
S2), as substantial differences in the aromatic phenyl groups
were observed, while LPPSQ gave highly broadened peaks
attributed to the high molecular weight and amorphous
crystalline structure of the polymeric LPPSQ backbone, to be
discussed later.
The selection of solvent in sol−gel reactions is a critical
component in inducing a hydrolysis−condensation equilibrium
tailored toward the formation of highly condensed structures.
Up until this point the solvent selected was THF, but solvents
of varying polarity may either increase or decrease hydrolysis
rates as shown for PTMS hydrolysis kinetic rate plots (Figure
Figure 4. FT-IR spectra of the obtained LPPSQs from 4 to 72 h.
LPPSQs as shown in Figure 3A. As the reaction time increased
from 4 to 36 h, the shape of the absorption peak ranged from
1
9
60 to 1250 cm− became sharply separated to be bimodal (a
6A). As expected, solvents of higher polarity or greater
−1
19,20
and b) and the peak intensity at 1150 cm (b) became higher
dielectric constant gave faster hydrolysis rates,
ranging 1−
−1
than that at 1050 cm (a). In addition, the characteristic
absorption peaks (c and d) of the silanol group (−Si−OH) at
30 and 3500 cm
2 orders faster. However, the correlation with solvent polarity
on molecular weight is not always proportional. While polar
solvents give faster hydrolysis rates, the highest molecular
weights for obtained LPPSQ were obtained when polar aprotic
solvents were used. This can be attributed to the fact that protic
solvents contribute to the sol−gel hydrolysis reaction by
solvolysis, which can depreciate the condensation reaction. And
as polycondensation proceeds, the condensed siloxane
structures become more hydrophobic and become less soluble
in polar media. Thus, highly polar solvents such as DMSO and
DMF produce LPPSQs with somewhat lower molecular weight
compared with less polar aprotic solvents such as THF.
Moreover, the fact that LPPSQs are completely insoluble in
these highly polar solvents renders steady and continuous
polycondensations rather difficult, due to premature precip-
itation of incompletely condensed siloxane structures. For the
case of nonpolar solvents such as nonsubstituted aromatics and
chloroform, these solvents are immiscible with the aqueous
media forming phase separated emulsion-like solvent mixtures,
which lead to greatly decreased hydrolysis rates and subsequent
lower molecular weights (Figure 6B).
−
1
9
also disappeared as condensation
progressed. The spectra for the product, LPPSQ (4) shown
in Figure 4E, showed sharp and doubly split peaks at 1050 and
−1
−1
1
150 cm , without any silanol groups at 3500 and 960 cm .
Compared with the FT-IR spectra for T -Phenyl POSS
1
2
(
Supporting Information Figure S1) which showed only one
−1
single peak, Si−O−Si vibration peak at 1100 cm , the doubly
split Si−O−Si bond vibrations were indicative of high
molecular weight.
In addition, we monitored the SEC derived molecular
weights of the intermediary structures at 13 and 36 h as well the
final products LPPSQ and T -Phenyl shown in Figure 5. While
1
2
the weight-averaged molecular weight (M ) of the intermediary
w
products that were extracted from the reaction mixture before
1
3 h were too small to be distinguished from the eluent, the
products extracted at 13 and 36 h in Figure 5A,B were found to
have a Mw of 2.2K and 9.3K, respectively, showing steady
molecular weight increase. This is indicative of slow and steady
polycondensation, of which is a prerequisite for LPPSQ
formation. The M of LPPSQ obtained after 72 h was found
to be 15.3K with a polydispersity of 2.02 (Figure 5C), which is
Ladderlike structured polysilsesquioxanes and POSS com-
pounds also vary in their thermal properties. Figure 7A,B shows
w
in stark contrast to T -Phenyl POSS, which showed an M of
the TGA thermograms for LPPSQ and T -Phenyl POSS,
1
2
w
12
900 with narrow polydispersity (PDI = 1.1) in Figure 5D.
respectively. Both exhibit superior thermal properties with no
weight loss at temperatures exceeding 380 °C, indicative of fully
condensed siloxane structures with indistinguishable amounts
While this molecular weight of T -Phenyl POSS was different
1
2
from the absolute molecular weight derived from MALDI-TOF
14
19,20
(
FW = 1550.26 g/mol) due to the standardization of our SEC
of silanol groups which condense around 150 °C.
system, obtained LPPSQ was found to have a fully polymeric
Furthermore, the manner in which the degradation occurs is
E
DOI: 10.1021/acs.macromol.5b01539
Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
temperatures. Moreover, the initial degradation temperature in
which 5 wt % is lost is about 60 °C higher for LPPSQ
compared with T -Phenyl POSS. This result is also in line with
12
previous reports of LPPSQ materials. Also noteworthy is that
LPPSQ retained its solubility after thermal treatments at 400
°
C for several hours, which is a key distinguishing characteristic
with other traditional methods of preparation of LPPSQ
3
,11−13
materials
indicative of thermoplastic behavior, and that
LPPSQ showed no discernible glass transition, melting, or
28
crystallization temperature by DSC analysis.
The bulk structures of LPPSQ and T -Phenyl POSS were
12
analyzed by X-ray diffraction (XRD) (Figure 8). For T -Phenyl
1
2
Figure 6. (A) Kinetic rate plots for hydrolysis of PTMS at [PTMS] =
i
8
.9 M under various solvent conditions and (B) Mw of LPPSQs
obtained under various solvent conditions.
Figure 8. XRD patterns for (A) T -Phenyl POSS and (B) LPPSQ and
12
their coating properties on glass (C) T -Phenyl POSS and (D)
12
LPPSQ before rubbing with fingers (left) and after (right).
POSS (Figure 8A), single crystal peaks with the peak at 7.2°
4
assigned to the cage face (c) were found, but for LPPSQ
(
Figure 8B), two characteristic diffraction peaks appeared. The
first sharp peak (a) at 7.3° (d1 = 12.1 Å) and the second
diffusion peak (b) at 18.8° (d2 = 4.7 Å) were assigned to the
intramolecular periodic chain-to-chain distance and the average
thickness of the amorphous LPPSQ, respectively. This XRD
Figure 7. TGA thermograms for (A) LPPSQ and (B) T -Phenyl
POSS.
12
result was also comparable with the spectrum of the previous
3,15,16
LPPSQ studies.
As coupled with the above TGA result,
XRD patterns showed no change after treatments at temper-
atures exceeding 400 °C (Supporting Information Figure S3),
which indicated that the obtained LPPSQ maintained its
structure at such high temperature without defects. This
temperature-independent structure can be said to be
distinguished from other polymeric silsesquioxanes with
random structures, as secondary condensations between silanol
different. While the phenyl groups in POSS compounds are
27
known to degrade in intermittent steps, the degradation of
3
the phenyl groups in LPPSQ occurs all at once, more like a
polymer. It is interesting to note that while alkyl-POSS
compounds sublimate under high-temperature conditions,
phenyl-POSS leave behind a higher residual silica and carbon
2
7
1,9,20,21
content, which is contributed to the higher oxidative
resistance of phenyl-POSS. This can also explain the residual
weight difference between LPPSQ and T -Phenyl POSS, as
groups lead to curing and inevitable structural change.
Compared to T -Phenyl POSS, LPPSQ was found to be a
12
noncrystalline, amorphous polymer. These differences in
1
2
polymeric materials exhibit enhanced oxidative stability at high
crystallinity were further shown by examining the coating
F
DOI: 10.1021/acs.macromol.5b01539
Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
properties of T -Phenyl POSS and LPPSQ, as shown in Figure
12
8C,D, respectively. While LPPSQ was able to be coated as a
transparent film at thickness ∼10 um via the drop-casting
technique, T -Phenyl POSS exhibited inhomogeneous coating
12
layers due to the crystalline domains exhibiting similar sizes to
the visible light wavelength, while the LPPSQ was completely
2
9
amorphous.
In evaluating the feasibility of the proposed ladderlike
2
9
1
polyphenylsilsequioxane (LPPSQ) by Si NMR, H NMR,
FTIR, TGA, and XRD methods up to this point, LPPSQ
compares favorably to more recent reports of LPPSQ materials
2
9
3
29
as the Si NMR T peak was shown to be narrow (solution Si
2
9
NMR Δ1 = 172 Hz, solid state Si NMR Δ = 89 Hz), with
/2
1/2
indistinguishable amounts of uncondensed silanol groups. Also,
SEC, FT-IR, TGA, and XRD analyses indicated that LPPSQ
was of polymeric nature. However, none of these analytical
tools were able to prove the linearity of LPPSQ. This led us to
investigate the size and shape of LPPSQ in solution (0.1 mg/
mL in THF) by small-angle X-ray scattering (SAXS) at 25 °C,
which to our knowledge is the first of its kind (Figure 9). We
stipulated that solutions SAXS experiments would not only be
able to prove the linearity of LPPSQ but also undoubtedly
distinguish LPPSQ from POSS-based materials. And while the
Zhang group has previously reported the helical nature of
2
9
30
LPPSQ materials through Si NMR, SAXS analyses of the
shape of LPPSQ can be said to be a more direct measurement
31
of the molecular shape.
The Guinier analysis of the obtained SAXS profiles gave a
radius of gyration (R ) of 14.2 Å as calculated from the
g
2
g
2
31,32
equation I(q) = I(0) exp(−1/3)(R q ) (Figure 9A).
shape of LPPSQ was first preliminarily estimated from the
The
SAXS profile, q = 1/R , because the relationship of I(q) and
g
−α
q , where the value of α = 1, 2, and 4 is that of thin rod (1D),
3
3
thin disk (2D), and sphere (3D) shapes, respectively. In this
case, the α value was 0.997, indicating thin rod shape of
LPPSQ, which was also confirmed by correlation of the form
34,35
factor of a thin rod model (open circles in Figure 9A).
Thin
rod shape of a molecule was generally well-corresponded to
2
Kratky plot, where I(q)q was plotted as a function of q; thus,
the straight line of the scattering curve was interpreted as a
Figure 9. Model fitting of the SAXS profiles of LPPSQ in THF at 25
°C: (A) thin rod model, (B) Kratky model, and (C) the most stable
structure of LPPSQ by MM2 simulation.
34
rigid-rod-like structured molecule. Our experimental result
with LPPSQ clearly showed directly proportioned to q−
1
(
Figure 9B), indicating that the obtained LPPSQ resided as a
thin rodlike molecule in dilute solution. From these results,
cross-sectional radius of gyration (R ) and the radius (r) of the
polymer chain of LPPSQ was formed as rod shape with
ladderlike siloxane structure.
c
LPPSQ was calculated from the following equations: (1) I(q) =
2
2
1/2 32
CONCLUSION
I(0) exp(−1/2)(R q ) and (2) r = 2 R , which gave R and
c
c
c
■
r values of 9.5 and 13.3 Å, respectively. In addition, the length
A one-pot, facile synthesis of structurally controlled polyphenyl-
silsesquioxanes was developed. Mechanism and kinetic studies
indicated that the condensation reaction proceeded through a
2
g
of the LPPSQ (L) was also calculated from the equation R =
2
2
33
(
r /2) + (L /12), showing a length (L) of 49.1 Å. The shape
1
of LPPSQ assuming M = 15 000 was estimated by simulation
T structured dimer, which was quantitatively and in situ
w
using the MM2 method. Figure 9C shows the resulting
simulated shape, which was long and helical structured with
asymmetrical high aspect ratio. This simulated model gave a
length (L) of 50.3 Å, which was nearly identical to the
calculated experimental length of 49.1 Å, further supporting our
claim that LPPSQ was of thin rod shape with helical structure.
And while dilute solution SAXS results for T -Pheyl POSS
were unable to be obtained due to its size, the thin rod shape of
LPPSQ was clearly differentiated from POSS materials in terms
of its size and shape. This molecular conformation of LPPSQ
was quite different from that of conventional random coil
polymer such as polystyrene, etc., which indicated that the
formed through hydrolysis of a mild phenyltrimethoxysilane
(PTMS) monomer, to give either the phenyl-substituted, 12-
membered, cage-structured polyhedral oligomeric silsesquiox-
anes (T -Phenyl POSS) or the corresponding ladderlike
polyphenylsilsesquioxanes (LPPSQ) with excellent yields.
Ladderlike and POSS materials were selectively achieved at
higher and lower initial concentrations of PTMS, respectively.
An in-depth spectroscopic comparison of both compounds
clearly revealed differences in structure, molecular weight,
thermal properties, crystallinity, and coating properties.
Furthermore, the linearity of LPPSQ analyzed by dilute
solution small-angle SAXS method revealed that LPPSQ
12
1
2
G
DOI: 10.1021/acs.macromol.5b01539
Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
formed as rod shape with ladderlike siloxane backbone
structure.
(17) Hyde, J. F.; Johannson, O. K.; Daudt, W. H.; Fleming, R. F.;
Laudenslager, H. B.; Roche, M. P. J. Am. Chem. Soc. 1953, 75, 5615−
5
618.
(
5
(
18) Chinnam, P. R.; Wunder, S. L. Chem. Mater. 2011, 23, 5111−
121.
19) Sun, X.; Xu, Y.; Jiang, D.; Yang, D.; Wu, D.; Sun, Y.; Yang, Y.;
ASSOCIATED CONTENT
Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.macro-
mol.5b01539.
■
*
S
Yuan, H.; Yuan, H.; Deng, F. Colloids Surf., A 2006, 289, 149−157.
(20) Chambers, R. C.; Jones, W. E., Jr.; Haruvy, Y.; Webber, S. E.;
Fox, M. A. Chem. Mater. 1993, 5, 1481−1486.
(21) Paquet, O.; Salon, M.-C. B.; Zeno, E.; Belgacem, M. N. Mater.
1
FTIR spectrum of T -Phenyl POSS, H NMR spectra
1
2
Sci. Eng., C 2012, 32, 487−493.
for LPPSQ and T -Phenyl POSS, and XRD patterns for
1
2
(22) Lee, H. S.; Choi, S.-S.; Baek, K.-Y.; Hong, S. M.; Lee, E. C.; Lee,
J.-C.; Hwang, S. S. Eur. Polym. J. 2012, 48, 1073−1081.
LPPSQ at various temperatures (PDF)
(23) Seto, I.; Gunji, T.; Kumagai, K.; Arimitsu, K.; Abe, Y. Bull. Chem.
Soc. Jpn. 2003, 76, 1983−1987.
AUTHOR INFORMATION
Corresponding Author
E-mail baek@kist.re.kr (K.-Y.B.).
■
*
(
24) Kudo, T.; Gordon, M. S. J. Phys. Chem. A 2002, 106, 11347−
1
1353.
(25) Seki, H.; Kajiwara, T.; Abe, Y.; Gunji, T. J. Organomet. Chem.
2
(
(
010, 695, 1363−1369.
Author Contributions
S.-S.C. and A.S.L. contributed equally.
26) Laine, R. M.; Roll, M. F. Macromolecules 2011, 44, 1073−1109.
27) Fina, A.; Tabuani, D.; Carniato, F.; Frache, A.; Boccaleri, E.;
Notes
Camino, G. Thermochim. Acta 2006, 440, 36−42.
(28) Kang, W. R.; Lee, A. S.; Park, S.; Park, S.-H.; Baek, K.-Y.; Lee, K.
B.; Lee, S.-H.; Lee, J.-H.; Hwang, S. S.; Lee, J. S. J. Membr. Sci. 2015,
The authors declare no competing financial interest.
4
(
75, 384−394.
29) Tokunaga, T.; Shoiriki, M.; Mizumo, T.; Kaneko, Y. J. Mater.
Chem. C 2014, 2, 2496−2501.
30) Zhang, X.; Xie, P.; Shen, Z.; Jiang, J.; Zhu, C.; Li, H.; Zhang, T.;
ACKNOWLEDGMENTS
■
This work was financially supported by a grant from the
Fundamental R&D for Core Technology of Materials funded
by the Ministry of Knowledge Economy, Republic of Korea,
and partially from a grant from the Materials Architecturing
Research Center of Korea Institute of Science and Technology
KIST). Synchrotron solution SAXS and WAXD measure-
ments were performed at the Pohang Light Source (PLS).
(
Han, C. C.; Wan, L.; Yan, S.; Zhang, R. Angew. Chem., Int. Ed. 2006,
45, 3112−3188.
(31) Mizukoshi, M.; Norisuye, T. Polym. Bull. 1998, 40, 555−562.
(
(
32) Yoon, D. Y.; Flory, P. J. Macromolecules 1976, 9, 294−299.
33) Roe, R. J., Ed.; Methods of X-ray and Neutron Scattering in
(
Polymer Science; Oxford University: New York, 2000.
34) Pedersen, J. S.; Schurtenberger, P. Macromolecules 1996, 29,
(
ABBREVIATIONS
■
7602−7612.
LPSQ, ladderlike polysilsesquioxane; POSS, polyhedral oligo-
meric silsesquioxanes.
(35) Potschke, D.; Hickl, P.; Ballauff, M.; Astrand, P. O.; Pedersen, J.
S. Macromol. Theory Simul. 2000, 9, 345−353.
REFERENCES
■
(
1
1) Baney, R. H.; Itoh, M.; Sakakibara, A.; Suzuki, T. Chem. Rev.
995, 95, 1409−1430.
(
(
2) Loy, D. A.; Shea, K. J. Chem. Rev. 1995, 95, 1431−1442.
3) Zhang, Z. X.; Hao, J.; Xie, P.; Zhang, X.; Han, C. C.; Zhang, R.
Chem. Mater. 2008, 20, 1322−1330.
4) Cordes, D.; Lickiss, P.; Rataboul, F. Chem. Rev. 2010, 110, 2081−
173.
(
2
(
(
5) Laine, R. M.; Roll, M. F. Macromolecules 2011, 44, 1073−1109.
6) Unno, M.; Suto, A.; Matsumoto, H. J. Am. Chem. Soc. 2002, 124,
1
574−1575.
7) Brown, F.; Vogt, L. H.; Katchman, A.; Eustance, J. W.; Kiser, K.
M.; Krantz, K. W. J. Am. Chem. Soc. 1960, 82, 6194−6195.
8) Frye, C. L.; Klosowski, J. M. J. Am. Chem. Soc. 1971, 93, 4599−
601.
9) Yamamoto, S.; Yasuda, N.; Ueyama, A.; Adachi, H.; Ishikawa, M.
Macromolecules 2004, 37, 2775−2778.
10) Ren, Z.; Cao, X.; Xie, P.; Zhang, R.; Yan, S.; Ma, Y. Chem.
(
(
4
(
(
Commun. 2009, 24, 4079−4081.
(
(
11) Kaneko, Y.; Iyi, N. J. Mater. Chem. 2009, 19, 7106−7111.
12) Kaneko, Y.; Toyodome, H.; Sato, H. J. Mater. Chem. 2011, 21,
1
(
2
(
6638−16641.
13) Unno, M.; Matsumoto, T.; Matsumoto, H. J. Organomet. Chem.
007, 692, 307−312.
14) Lee, A. S.; Choi, S. S.; Lee, H. S.; Baek, K.-Y.; Hwang, S. S.
Dalton Trans. 2012, 41, 10585−10588.
15) Choi, S. S.; Lee, H. S.; Hwang, S. S.; Choi, D. H.; Baek, K.-Y. J.
Mater. Chem. 2010, 20, 9852−9854.
16) Choi, S. S.; Lee, A. S.; Jeon, H. Y.; Baek, K. Y.; Choi, D. H.;
Hwang, S. S. J. Polym. Sci., Part A: Polym. Chem. 2011, 49, 5012−5018.
(
(
H
DOI: 10.1021/acs.macromol.5b01539
Macromolecules XXXX, XXX, XXX−XXX