. .ux...

1.5? . ‘

hwfig .r. x

n: A.- 13v o

. raga?
4,. 2..
n?

«it

. . S. u. i...
.1:
1w 1
V 1.5%. .

. anfia

‘
.3 . a.
.. .3. .3

1 VI .
‘35:.

35%.»:

5 a):

81%

X
i

,r

339 I
M’W
. \

~ ‘. I Z ‘- '1:
Ida . H.513...”
$315.31.: . . (.3...

wfiflhfivf nwé..HJ.nanv3.¢t

x .33.)...35 .' y 2; \AQ:

\Ql’lnzua. .4 kits: 05) \I.

. V 33.13.43.554:
J‘flvzbbnn... ’3'

I: I! .r .13 :11
313‘}. a
(1. $3351.11);
Ili'xq .I- I.

. 1|:le 0.3
. liHlylalln .

I»... 0.3.

t.\c.rvt l. ..

ylr\lu y it}
.1 o

flfirtfi...§m mmmflnm

2.1.

 

a... LIBRARY .
h; _ Michigan State
44):) University

This is to certify that the
dissertation entitled

MESOSTRUCTURED SILICA FOR THE REINFORCEMENT
OF THERMOSET EPOXY POLYMERS

presented by
IN PARK

has been accepted towards fulfillment
of the requirements for the

Ph.D. degree in Department of Chemistry

 

 

\

 

 

r Proféssor’s Signature

1% /& 07W!

i [Q

  

MSU is an Afiinnative Action/Equal Opportunity Institution

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K./Proj/Acc&Pres/ClRC/DateDue indd

MESOSTRUCTURED SILICA FOR THE REINFORCEMENT OF
THERMOSET EPOXY POLYMERS

By

In Park

A DISSERTATION
Submitted to
Michigan State University

in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Chemistry
2006

ABSTRACT

MESOSTRUCTURED SILICA FOR THE REINFORCEMENT OF
THERMOSET EPOXY POLYMERS

By
In Park

Large-pore mesoporous silicas with wormhole framework structures have
been assembled through hydrogen-bonding pathways from sodium silicate or
tetraethylorthosilicate as the silica source and amine-terminated Jeffamine
surfactants of the type H2NCH(CH3)CH2[OCH2CH(CH3)]XNH2 as the structure-
directing porogen. Depending on the molecular weight of the a,w-diamine
surfactant (x = 33 and 68 for Jeffamine D2000 and D4000, respectively) and the
synthesis temperature (ZS—65°C), the mean pore size distributions of the
mesostructured silicas (denoted MSU-J) were centered between 4.9 and 14.3 nm.

The as-made and calcined forms of MSU-J (5.3 nm pore size) have been
used to form rubbery epoxy mesocomposites containing 1.0—12 wt°/o silica. The
tensile modulus, strength, elongation, and toughness for the mesocomposites
formed from as-made and calcined forms of MSU-J silica are systematically
reinforced by up to 4.8, 5.7, 1.6, and 8.5 times, respectively, in comparison to the
pure epoxy polymer. Moreover, the reinforcement benefits are realized without
the need for organic modification of the silica surface. The oxygen permeability of
the mesocomposites prepared form as-made MSU-J silica increases dramatically
at loadings 25.0 wt%, whereas the compositions made from the calcined form of

the mesostructure show no permeation dependence on silica loading.

Mechanically improved therrnoset epoxy mesocomposites have also been
prepared from calcined silica foam, MSU-F which exhibits large cell size of 26.5
nm, window size of 14.9 nm and high pore volume of 2.2 cm3/g. Despite no
organic modifier in its pore the calcined MSU-F silica foam is well dispersed in
the epoxy matrix due to its very large pore diameter and low intrinsic density. The
tensile modulus, strength, elongation, and toughness for the resulting rubbery
epoxy mesocomposites are systematically enhanced by 3.7, 6.8, 2.2, and 20.6
times, respectively, at relatively low silica loading (9 wt%) compared to the silica-
free rubbery epoxy polymer. The remarkable mechanical enhancements are
contributed by the large window and cell sizes of the MSU-F silica foam which
are comparable to the layer separation (>10 nm) of the exfoliated clay layers in
the polymer-epoxy nanocomposites. Moreover, little or no sacrifice of thermal
stability and transparency were observed in the rubbery—MSU-F

mesocomposites.

Copyright by
IN PARK
2006

ACKNOWLEDGEMENTS

First, I would like to express my appreciation to Professor Thomas J.
Pinnavaia for his thoughtful guidance for the last 4 and half years. I thank the
Lord for being his student and for learning from him a lot of things that I need to
be a Ph. D. or an independent researcher. I’d also tell former and current
members in our group that you are so lucky because you had (or have) an
intelligent and kind advisor.

Also, I would like to give thanks to the former and current group members.
Costas, my mentor, taught me how to use a polisher and the MOCON instrument
and how to prepare epoxy polymers and clay papers. Pete took me good pizza
places and fishing stores in Lansing area. Dr. Kim helped me synthesize
mesostructures. DK has always been willing to help me a lot when I got troubles
especially my early years in the US. Sebastian has not hesitated to answer my
English questions. Siqi is a really good partner to discuss tensile properties and
preparation process with. Junho and l have shared good and bad things
happening to us. Marc is one of the most diligent and kindest people I have ever
seen. Zharong, Hui, and Sun are so nice to me. I am also grateful to Soong and
Kyoungsoo for answering my organic questions.

I fully understand that this thesis could not have been completed without
my wife and son’s sacrifice. Thus I would like to dedicate this humble thesis to
Haikyung and Jihwan. Do you guys know how much I love you? You can never

know it!!!

TABLE OF CONTENTS

LIST OF TABLES ................................................................................................ viii
LIST OF FIGURES ............................................................................................... ix
CHAPTER 1
INTRODUCTION ................................................................................................... 1
1.1 Supramolecular Assembly of Mesoporous Materials ........................ 1
1.1.1 Electronic Pathways ............................................................... 1
1.1.2 Neutral Templating Pathways ................................................ 6
1.1.3 Mesocellular Silica Foams ..................................................... 8
1.2 Polymer-Based Nanocomposites .................................................... 11
121 Smectite Clays... ...11
1 ..2 2 Polymer-Clay Nanocomposntes ............................................ 13
1.2.3 Themoset Epoxy-Clay Nanocomposites .............................. 16
1.2.4 Mesocomposites .................................................................. 19
1.2.5 Gas Permeability .................................................................. 20
1.3 Research Goals and Significance .................................................. 24
1.4 References ..................................................................................... 27
CHAPTER 2
ASSEMBLY OF LARGE-PORE MESOPHASES WITH WORMHOLE
FRAMEWORK STRUCTURES FROM A, Q-DlAMINE POROGENS .................. 31
2.1 Introduction ..................................................................................... 31
2.2 Experimental ................................................................................... 32
2.3 Results and Discussion .................................................................. 34
2.4 Summary ........................................................................................ 44
2.5 References46
CHAPTER 3
EPOXY-SILICA MESOCOMPOSITES WITH ENHANCED TENSILE
PROPERTIES AND OXYGEN PERMEABILITY .................................................. 48
3.1 Introduction ..................................................................................... 48
3.2 Experimental ................................................................................... 50
3.3 Results ............................................................................................ 54
3.4 Discussion ...................................................................................... 64
3.5 Summary ........................................................................................ 70
3.6 References ..................................................................................... 73
CHAPTER 4
SILICA FOAM REINFORCED EPOXY NANOCOMPOSITES ............................. 76

vi

4.1 Introduction ..................................................................................... 76

4.2 Experimental ................................................................................... 80
4.3 Results ............................................................................................ 83
4.4 Discussion ...................................................................................... 92
4.5 Summary ........................................................................................ 99
4.6 References ................................................................................... 1 02

CHAPTER 5

MESOPOROUS SILICA STRUCTURES ASSEMBLED FROM JEFFAMINE AND

TOHMAR3 SURFACTANTS .............................................................................. 105
5.1 Wormhole Mesostructures ............................................................ 105
5.2 MSU-G Type Silicas ..................................................................... 112
5.3 References ................................................................................... 120

APPENDIX

HIGH OXYGEN GAS BARRIER CLAY-EPOXY COMPOSITE FILMS .............. 121
A.1 Introduction ................................................................................... 121
A2 Sample Prepartion ........................................................................ 122
A3 Results and Discussion ................................................................ 125
A4 Conclusion .................................................................................... 132
A5 References ................................................................................... 1 33

vii

LIST OF TABLES

Table 1.1 Physical properties for selected mesostructured silicas prepared by
supramolecular assembly pathways .................................................................... 10

Table 1.2 Elemental compositions of common 2:1 type smectites. In dioctahedral
smectite clays two-third of the octahedral sites are occupied by metal ions.
Trioctahedral smectites possess all octahedral positions filled by metal ions.
lnterlayer cations are all assumed as monovalent but various valent cations may
occupy the intenayer position .............................................................................. 13

Table 2.1 Textural properties of calcined MSU-J silicas with wormhole framework
structures ............................................................................................................. 40

Table 3.1 Tensile, thermal stability and oxygen permeability properties of pristine
epoxy polymer and epoxy/MSU-J mesocomposites ............................................ 72

Table 4.1 Tensile and thermal properties of pristine epoxy polymer and epoxy—
MSU-F mesocomposites ...................................................................................... 89

Table 5.1 Physical properties for wormhole mesostructured silicas prepared from
sodium silicate and Jeffamine or Surfonamine surfactants ................................ 107

Table 5.2 Textural properties for lamellar silica vesicles assembled by TEOS in
the presence of Tomah gemini surfactants ........................................................ 118

Table A.1 Oxygen Permeability data of the epoxy—clay composite films .......... 131

viii

LIST OF FIGURES

Figure 1.1 Three structure types observed for silica-surfactant mesophases: (a)
MOM-41 (hexagonal), (b) MGM-48 (bicontinuous cubic), (c) MOM-50 (lamellar).
In each type there is a periodic arrangement of pores (or layers), but the
inorganic walls (or sheets) are amorphous ............................................................ 2

Figure 1.2 Schematic representation of the general formation of mesostructures
from inorganic precursors and surfactants. Inorganic precursor species are
associated with surface of surfactant micellar structures to form an ordered
inorganic framework by condensation and polymerization of inorganic precursors.
............................................................................................................................... 4

Figure 1.3 Schematic view of a surfactant and values for calculation of the local
effective packing parameter, 9 = V/aol. (Inset table) the critical 9 values at which
mesophase transitions occur ................................................................................. 5

Figure 1.4 Morphological transitions for P123-templated materials progressively
swollen by TMB (1,3,5-trimethylbenzene) of the silica products templated by the
micellar arrays: micelle arrays (a, d, f); silica pore morphologies (b, g); and TEM
images (0, e, h). In the absence of TMB, cylindrical silica pores (b) are formed
from cylindrical micelles (a), as confirmed by the TEM image (c). Spherical
micelles (f) and the resulting cellular silica foam with windows (9, h) are prepared
at a TMB/P123 ratio of 5.0. Intermediate phases with buckling pore morphology
are obtained at a 0.21 ratio (d, e) ........................................................................... 9

Figure 1.5 Schematic representation of the idealized structure of a smectite clay
mineral. Normally, silicon atoms primarily occupy tetrahedral positions. In
dioctahedral 2:1 structures, the octahedral interspaces are either two-thirds
occupied by M3+ ions (especially AI3+) whereas in trioctahedral structures they
are completely occupied by M2+ cations, especially Mg”. Aluminum may
substitute for silicon in tetrahedral positions and magnesium, iron, or lithium
atoms may replace metal ions in octahedral sites ............................................... 12

Figure 1.6 Schematic illustrations of polymer-clay composite structures.
Rectangular bars represent clay layers with ~1 nm thickness; (a) conventional
composites, (b) intercalated nanocomposites. and (c) exfoliated (delaminated)
nanocomposites. The corresponding powder XRD patterns for composites at low
angle region (20 = 1—10°) are also illustrated ...................................................... 15

Figure 1.7 A model (a) for the tortuous path of a diffusing molecule through a
polymer filled with oriented impermeable disks (b) with aspect ratio, a = L / W...22

Figure 2.1 X-ray diffraction patterns of as-synthesized ( ----- ) and calcined (—)
MSU-J silica molecular sieves using D2000 as templating agents and sodium
silicate as a silica source under neutral pH conditions at 25, 45, and 65 °C. The
numbers are the pore-pore correlation distances in nanometer (nm) units ......... 35

Figure 2.2 TEM images of calcined MSU-J silicas synthesized with the Jeffamine
surfactants and the silica sources; (a) 02000 and sodium silicate at 25°C, (b)
02000 and TEOS at 45 °C, (c) D2000 and TEOS at 60 °C, and (d) D4000 and
sodium silicate at 65 °C. Average pore sizes are 5.3, 7.5, 11.9, and 14.3 nm for
(a), (b), (c), and (d), respectively .......................................................................... 37

Figure 2.3 N2 adsorption-desorption isotherms of MSU-J silica molecular sieves
assembled from TEOS in the presence of D2000 as a structure director at
different synthesis temperatures and calcined at 600 °C for 4 h. The insert
provides BJH framework pore size distributions determined from the adsorption
branches of the N2 isotherms (o, o, and A for 25, 45, and 60°C, respectively). 38

Figure 2.4 N2 adsorption-desorption isotherms of MSU-J silica molecular sieves
assembled from sodium silicate in the presence of D2000 as a structure director
at different synthesis temperatures and calcined at 600°C for 4 h. The insert
provides BJH framework pore size distributions determined from the adsorption
branches of the N2 isotherms (0, c1, and A for 25,45, and 65 °C, respectively). 41

Figure 3.1 Small angle powder X-ray diffraction patterns: (a) as-made MSU-J
mesostructure containing 58 wt% intercalated Jeffamine D2000; (b) epoxy—silica
mesocomposite (12 wt% silica) formed from as-made MSU-J; (c) mesostructured
silica recovered through calcination at 600 °C of a mesocomposite containing 12
wt% silica ............................................................................................................. 55

Figure 3.2 TEM images of (a) calcined MSU-J silica and (b) the silica obtained
through calcination at 600 °C a cured epoxy-silica mesocomposite (12 wt%
silica) formed from as-made MSU-J silica ........................................................... 57

Figure 3.3 Comparison of stress-strain curves for pristine epoxy and MSU-
J/epoxy mesocomposites containing different loadings of (a) as-made MSU-J
silica and (b) calcined MSU-J silica ..................................................................... 58

Figure 3.4 Loading dependence of the (a) tensile moduli (b) tensile strengths (c)
strain-at-break and (d) toughnesses for epoxy mesocomposites prepared from
MSU-J silica. The solid and open data points are for the composites prepared
from as-made and calcined MSU-J silica, respectively ........................................ 59

Figure 3.5 Dependence of oxygen permeability on silica loadings for epoxy
mesocomposites prepared from (a) as-made MSU-J silica and (b) calcined MSU-
J silica .................................................................................................................. 62

Figure 3.6 Optical image illustrating differences in the transparency of (a) pristine
epoxy polymer and epoxy composites prepared from (b) as-made MSU-J silica at
10 wt% loading and (c) octadecylammonium-modified montmorillonite (ODA-
MMT) at 10 wt% silicate loading .......................................................................... 63

Figure 3.7 Therrnogravimetric analysis curves in air for (a) the epoxy
mesocomposite prepared from as—made MSU- J silica at a loading of 12 wt%
silica and (b) the pristine epoxy polymer .............................................................. 65

Figure 3.8 Schematic representations of (a) the proposed decrease in chain
cross-linking density around the as-made MSU-J silica particles and (b) the
proposed penneant gas pathways below and above ~5 wt% loading of as—made
MSU-J silica ......................................................................................................... 69

Figure 4.1 (a) Nitrogen adsorption-desorption isotherm for calcined MSU-F silica
foam prepared from postsynthesis hydrothermal treatment of the as-synthesized
MSU-F for 24 h at 100 °C. (b) Corresponding cell size (solid line) and window
size (dash line) distributions obtained from adsorption and desorption branches,
respectively. The distributions were determined by applying the BdB-FHH2
model ................................................................................................................... 84

Figure 4.2 Transmission electron microscope (T EM) images of a calcined MSU-
F silica foam synthesized through postsynthesis treatment at 100 °C for 24 h.
Images (a) and (b) are taken with high and low magnifications, respectively ...... 85

Figure 4.3 Comparison of stress-strain curves for pristine epoxy ( """ ) and MSU-
F-epoxy mesocomposites containing different silica loadings ............................. 86

Figure 4.4 Loading dependence of the (a) tensile moduli, (b) tensile strengths,
(c) strain-at-break and (d) toughness for rubbery epoxy mesocomposites
prepared from calcined MSU-F silica ................................................................... 87

Figure 4.5 Optical images illustrating differences in the transparency of (a)
pristine rubbery epoxy polymer and epoxy composites prepared from calcined
MSU-F silica at (b) 5 and(c) 10 wt% .................................................................... 90

Figure 4.6 Therrnogravimetric analysis curves in nitrogen for the silica-free
epoxy polymer (open circles) and the epoxy mesocomposites (solid circles)
prepared from calcined MSU-F silica at a loading of 9 wt% ................................ 91

Figure 4.7 Experimental (symbols) and computed (lines) values of Young’s
modulus as a function of silica loadings. Solid circles: composites made from
calcined MSU-J and open squares: composites made from calcined MSU-F. Dot
line: Computed dependence taking only the silica wall volume as the particulate
volume fraction; dash line: assuming the particulate volume to be the silica wall
volume plus the pore volume of MSU-J; solid line: assuming the particulate

xi

volume to be the silica wall volume plus the pore volume of MSU-F. The
prediction (dot-dash line) based on a simple rule of mixture is also plotted ........ 96

Figure 5.1 (a) XRD patterns, (b) N2 sorption isotherms, and (0) pore size
distributions for MSU-J silicas prepared from sodium silicate and M2005. The
dotted (°"') and solid (—) XRD scans represent as-made and calcined MSU-J
silicas at different synthesis temperatures. lsotherrns are measured for calcined
MSU-J silicas synthesized at 25 (o), 45 (a), and 65 (A) °C ................................ 108

Figure 5.2 (a) XRD patterns, (b) a TEM image (synthesized at 25 °C), (0) N2
sorption isotherms, and (d) pore size distributions for MSU-J silicas prepared
from sodium silicate and T3000. The synthesis temperatures are indicated as 25
(o), 45 (I), and 65 °C (A) (c and d) .................................................................... 109

Figure 5.3 (a) N2 sorption isotherms and (b) pore size distributions for MSU-J
silicas prepared from sodium silicate and T5000. The synthesis temperatures are
indicated as 25 (o), and 45 (:1) °C ...................................................................... 110

Figure 5.4 (a) XRD patterns (as-made; and calcined; -), (b) N2 sorption
isotherms, and (0) pore size distributions for MSU-J silicas prepared from sodium
silicate and Surfoamine surfactants (ML300; o and MNPA1000; a in b and
c) ........................................................................................................................ 111

Figure 5.5 (a) XRD patterns (as-made; and calcined; —), (b) N2 sorption
isotherms, and (c) pore size distributions for silica vesicles prepared from TEOS
and partially protonated Tomah3 DA surfactants (DA1214; o and DA14; I in b
and c) ................................................................................................................. 115

Figure 5.6 (a) TEM images with low and (b) high magnification of MSU-G type
lamellar vesicles formed from TEOS and neutral DA1214 surfactant (as-
synthesized). (c) A TEM image of the calcined lamellar silica structure assembled
from partially protonated DA1214 surfactant ..................................................... 116

Figure A.1 Optical images of a self-supported Na+-montmorillonite clay film with
a thickness of 35 um prepared from a 3 wt% aqueous suspension ................... 126

Figure A.2 X-ray diffraction patterns of (a) Na*—montmorillonite powder and (b)
the Na+-montmorillonite film prepared from a 3 wt% water suspension. The
intensity of the patterns from 12° to 80° was multiplied three-fold and shown as
an insert for clarity. Offsets are 500 counts for both sets of scans .................... 127

Figure A.3 Proposed clay nanolayer structure of an oriented clay film in (a)
pristine and (b) epoxy-impregnated form. The mixture of epoxy resin and 0230
represented by hatched pattern is impregnated in voids of the clay films and
cured at 125 °C to form high gas barrier composite films. Sodium ions in the
gallery regions of the clay are omitted for clarity ................................................ 129

xii

CHAPTER 1

INTRODUCTION

1.1 Supramolecular Assembly of Mesoporous Materials
1.1.1 Electrostatic Pathways

In 1992, Mobil researchers reported the discovery of the novel family of
molecular sieves (M41S) prepared by supramolecular assembly through
electrostatic interactions between cationic surfactant micelles and anionic
inorganic precursors.1 These mesoporous (alumino)silicates exhibits regular and
well-defined pore sizes of 15—100 A, very high surface area (>1000 m2/g) and
hydrocarbon sorption capacities (>0.7 00/9). The applications of these materials
are based on their microstructures which allow molecules access to large internal
surfaces that enhance catalytic activity and adsorptive capacity.2 MCM-41, one of
the members of the M418 family, possesses a two-dimensionally ordered
hexagonal array (p6m) of uniform mesopores. Its pore size is easily controlled by
surfactant chain lengths or/and adding a pore expander (1 ,3,5-trimethylbenzene,
TMB). Also a cubic (MGM-48, Ia3d) and a lamellar (MOM-50, p2) phase were
prepared by varying the surfactant to silicon molar ratio (Figure 1.1).

A proposed mechanism for the formation of M418 is a liquid crystal
templating (LCT) which has two possible pathways.2 One is that the inorganic
precursor species are deposited on a preexisting hexagonal liquid crystal phase

and the other is that the inorganic species and the surfactant micellar rods are

(a) 2-d hexagonal (MGM-41)

(b) Cubic (MGM—48)

 

(c) Lamellar (MCM-50)

 

Figure 1.1 Three structure types observed for silica-surfactant mesophases:
(a) MGM-41 (hexagonal), (b) MOM-48 (bicontinuous cubic), and (c) MOM-50
(lamellar). In each type there is a periodic arrangement of pores (or layers)

while the inorganic walls (or sheets) are amorphous.

combined to render a hexagonal structure. However a number of mechanisms,
such as charge density matching3 and folding sheet“ 5, have been proposed to
elucidate the supramolecular assembly of the mesoporous materials. Generally,
surfactants containing a hydrophilic head group and a hydrophobic tail group are
assembled to form micelles or micellar arrays which directs inorganic
mesostructures on the surface of the surfactant structure6 (Figure 1.27).

Various synthetic routes for inorganic mesostructures can be categorized
by reaction conditions such as interactions between the structure-directing agent
(surfactant or porogen) and the inorganic precursor, pH conditions, types of
surfactants, etc. In the case of M41S mesostructured molecular sieves, cationic
quaternary ammonium surfactants (3+) were used under basic pH conditions.
Electrostatic interaction between S+ and anionic inorganic precursors (I') is
involved in this pathway.

Huo and co-workers reported another electrostatic route (S‘X‘F) with
intermediate anions (X', e.g. Cl', Br') under acidic conditions below the isoelectric
point of silica (pH ~2).7'9 Through this pathway, a new family (SBA-1 (cubic,
Pm3n), SBA-2 (3d hexagonal, P63/mmc) and SBA-3 (MGM-41 type)) was
synthesized from gemini surfactants in acidic media. The formation of amphiphilic
liquid-crystal arrays was described in terms of the local effective 9 packing
parameter9 (defined in Figure 1.3) in order to explain and predict product
structures and phase transitions. As the Q value increases, the surface curvature
of micelles decreases so that a lamellar phase is favorable when the g value

approaches unity.

..,f?7f,,.7y“
"’ $§ —* %/
Surfactant Micell Micellar rod \

Remove

template

   

Mesopours material

Figure 1.2 Schematic representation of the general formation of
mesostructures from inorganic precursors and surfactants. Inorganic precursor
species are associated with surface of surfactant micellar structures to form an
ordered inorganic framework by condensation and polymerization of inorganic

precursors.

V: the total volume of the surfactant chains

 

a0: the effective head group area at the micelle surface

I I: the kinetic surfactant tail length

 

 

g mesophase

V3. cubic (Pm3n)

1/2 hexagonal (p6m)
1/2 — % cubic (Ia3d)

1 lamellar

 

 

Figure 1.3 Schematic view of a surfactant and values for calculation of the
local effective packing parameter, 9 = V/aol. (Inset table) the critical 9 values at

which mesophase transitions occur.

1.1.2 Neutral Templating Pathways

A neutral templating route for preparing mesoporous molecular sieves
(HMS) is demonstrated based on hydrogen-bonding interactions and self-
assembly between neutral primary amine micelles (3°) and neutral inorganic
precursors (l°).10 Unlike cylinder-like framework of MGM-41, HMS materials
possess wormhole framework morphology with nearly uniform pore diameters.
The S°l° templating pathway produces mesoporous materials with thicker
framework walls and substantially higher thermal stability11 compared to charge-
matching derived silicates. Another benefit of the HMS materials is improved
textural mesoporosities12 owing to the very small elementary particle domain size
in comparison with M418 and SBA materials templated by quaternary
ammonium cations of equivalent chain length. This synthetic strategy using
relatively weak S°l° interaction also allows for the facile, environmentally benign
recovery of the cost-intensive template by simple solvent extraction methods.
Unlike the tubular hexagonal pores (MOM-41 type) the wormhole pores of HMS
are three-dimensionally interconnected, confirmed by a strong peak in the XRD
pattern of SNU-2 carbon prepared by using AI-HMS as a template.” 1‘

The neutral amine surfactant templating pathway under neutral pH
conditions generates silica molecular sieves with lamellar frameworks and
hierarchical vesicular particle morphology (MSU-V15 and MSU-G16). MSU-V
silicas with gallery-confined micropores/mesopores (1.3—2.7 nm) have been

assembled in the interlayered regions of multilamellar vesicles formed from a

neutral diamine bolaamphiphiles (H2N(CH2)nNH2, n = 12—22). Electrically neutral

gemini surfactants of the type CnH2n-1NH(CH2)2NH2 (n = 10—14) and silica
precursors produces another family of silica vesicles (MSU-G). The high SiO4
crosslinking of MSU-G silicas allows for unprecedented thermal (up to 1000°C)
and hydrothermal (more than 150 hours at 100°C) stabilities. Both of the
molecular sieves with vesicular silica morphology retain the lamellar framework
structures after surfactant are removed due to lnterlayer pillars.

The hydrogen-bonding pathway was expanded to nonionic surfactants
(N°). MSU-X silica molecular sieves have been prepared from
tetraethylorthosilicate (TEOS) in the presence of low-cost, nontoxic, and
biodegradable polyethylene oxide (PEO) surfactants, which act as the structure-
directing agents.17 Cost-effective and water-soluble sodium silicates have also
been used as an inorganic precursor for preparation of MSU-X type
mesostructures (designated MSU-X').18 This N°l° templating pathway to
mesostructures uses hydrogen bonding interactions between the hydrophilic
surfaces of a flexible rod- or worm-like inorganic oxide framework. Disordered
channel structures with uniform diameters ranging from 2.0 to 5.8 nanometers
have been obtained by varying the size and structure of the surfactant molecules.
Increasing synthesis temperature weakens the hydrogen-bonding interactions
and consequently the curvature of the micelle decreases so that the framework
pore size is expanded up to approximately two times (65°C compared to 25°C).19
Metal-substituted silica and pure alumina mesostructures have also been
prepared by the hydrolysis of the corresponding alkoxides in the presence of

PEO surfactants.”

1.1.3 Mesocellular Silica Foams

Another group of nonionic surfactants, amphiphilic triblock copolymers
(typically P123, (PEO)20(PPO)70(PEO)20, polyethylene oxide—polypropylene
oxide—polyethylene oxide) have been used to prepare well-ordered hexagonal
mesoporous silica structures (SBA-15) in acidic aqueous solutions (pH <1 ).21 The
pore size of SBA-15 is expanded up to 30 nm by adding a swelling agent (TMB).
Under nearly neutral pH conditions, hexagonally ordered mesoporous silicas
(denoted MSU-H) have also been synthesized from the same surfactant and
water-soluble sodium silicates.22 The assembly pathway under neutral pH allows
MSU-H to possess a higher degree of crosslinking due to little or no charge
present on the silica framework walls.

Increasing the TMB-P123 ratio sufficiently induces a phase transformation
from SBA-15 to mesocellular foams (MCF).23 The MCF phase possesses
interconnected spherical pore system with windows or openings. Detail analysis
on nitrogen sorption isotherms and transmission electron microscopy (TEM)
demonstrated that the cylinder pore system at the low TMB-polymer ratios (0.1—
0.2, mass ratio) transforms to the spherical pores at the high ratios (>0.3) through
intermediate phases with spherical nodes.24 Increasing the TMB content leads to
the formation of microemulsion droplets with a mean surface curvature similar to
that of the cylindrical micelles (Figure 1.424). Transition from MSU-H to MSU-F
(silica foams) prepared under nearly neutral pH conditions has also been
reported.22 Physical properties of selected mesostructured materials are listed in

Table 1.1.

(b)

   

Cylindrical

 

    

Spherical

 

. .. , ;...,. 6‘33, mm
Figure 1.4 Morphological transitions for P123-templated materials
progressively swollen by TMB (1,3,5-trimethylbenzene) of the silica products
templated by the micellar arrays: micelle arrays (a, d, f); silica pore
morphologies (b, 9); and TEM images (c, e, h). In the absence of TMB,
cylindrical silica pores (b) are formed from cylindrical micelles (a), as
confirmed by the TEM image (0). Spherical micelles (f) and the resulting
cellular silica foam with windows (9, h) are prepared at a TMB/P123 ratio of

5.0. lnterrnediate phases with buckling pore morphology are obtained at a 0.21

ratio (d, e).

.NIz...:No.IzaIeo Q .mcmomuooogsmaég o .omAOm.-.m-...~_._m.-..o .. .8...

309.3... 05.... ..oo 5 890a moi 29:9on .65 0. can mEmtma omx Boson. Eo: mcozooco. >m.-x 5.58 Co: on 2.50.
3.2.8 9: . .9595 8:939. 9: 3.... women. .0 So 829 cows. 9m mo:_m> och .. .6 .o .m .N u m .3 .o .3 .3 .3 u 5
AmIOVZENImOAMIOVZ.-chch . 6902.2 m. EEomtam 9: 5% 38.260 m. 932:6 .m__oEm_ 8-5.0.2 of: .8555.
on. c. 8255 Co: 0.5 mo.:_m> .. 98:3 Eauow s ......AMIOVZBISO . .oE:_o> Eon. a .35 oomtam me ... .36 Son.

0 823026 03:0 .0. :Nn ”mo_o_mm> .m__oEm_ ncm .o_ocE.o.s ._m:ommxoc UN .2 86 .050QO .33 a .mco_._vcoo In .959.

”A5 new .228 x8 .2me ”EV ... .m_m._.ofiE 8528 5.3 02:93.... Be; oE:_o> Boo new .53 .o~_m Boo .96QO U or... ..

 

 

mom 8.. ....mm. mm . ESE ...NBN mm .msE RE .5 o_..z EB“. 3.2.8 aujms.
43 mm... we now 2.8 mm RE .5 o...z accuse; um NN1.3m...
.. «3. mm B 352 mom: .9300 ......om 20.8., 5.2.3 20.8....
B... one E ...m 3.3 mom: .20 3.0.8, 2.0.8., 5.8.3 836....
.. 0.: m... S ta... mom: Saws; .5 o...z 29.5.0.5 :xjws.
m3 5: ma em SEN woe. £22120 .5 .....m 2255.... ems:
...N mum ea: 8 . 85$ .33 mom: .msE RE E .53 ES“. 5.2.8 ...“...o...
A: 02 a... 02 ER: 6%..” mom: mNE ...... ..-xb .2896; cm smgmw
...? .80.. .82 .. 8.2: 83 ...-.-co ...... ...-xb .2898; um Ream
.. .. .. em .52: mm .50 3. -...m 5.053 a8...,5...
so .. 3 am $.52 mm .50 .3 -..m 0.2.0 mmmvzos.
ed 9.2 o... o... 2.3.: ..mm ...Eo 3. -...w .2892 um a3120.).
BEES AQNEV A.E.5 A85 $30.50.» .838... ccozomoov mfav ommzaomos.
a > n. sham o w n. no 0E... ..th 2.590... ..cmuomtam 230m ..

 

.m>m>>£mo 29:33 .m_:oo_oEm.o...m .5 8533 28:6 no.3ogmomoE umfiflow .2 83.805 .mofiEn. 9... Eco...

10

1.2 Polymer-Based Nanocomposites
1.2.1 Smectite Clays

The hydrous layer silicates commonly known as clay minerals are part of
the larger family of phyllosilicates.30 Clay minerals are composed of two-
dimensional octahedron layers and tetrahedral layers. Depending on the number
of layers, clay minerals are commonly separated into two subgroups, 1:1 layered
and 2:1 layered silicates.31 In case of the 1:1 type, the layers are formed by
coupling of only one tetrahedral sheet to an octahedral sheet. Layers of the 2:1
layered type clay minerals consist of a central octahedral sheet and two
tetrahedral sheets, as shown in Figure 1.531. Substitutions by ions of lesser
charge produce resultant negative charges on the layers which are balanced by
gallery cations. The layers of talc and pyrophyllite (See Table 1.2) are electrically
neutral. Substitution of Al3+ by Mg2+ generates negative charges on the
octahedral sheets in layers of montmorillonite which is most common member of
smectite clays. Other smectite clays shown in Table 1.2 have negative charges
on the octahedral sheets (beidelite and saponite) or the tetrahedral sheet
(hectorite) in similar manner to that of montmorillonite. The overall charge is
balanced with hydrated cations in the clay gallery and the cations can be
replaced with other cations by utilizing ion-exchange reactions which allow for
modification of the interlayer region.

The smectite clay minerals are capable of osmotic swelling in polar
solvent (usually water) to give very high spacings (>>100 A) between the layers.

The osmotic swelling is depending on factors such as layer charge density, type

11

       
     

   

  
 

 

   

 

-:::.!ég‘v::t'._ T
Wt'dfll'
V ‘V \ ......

 

 

 

Figure 1.5 Schematic representation of the idealized structure of a smectite I
clay mineral. Normally, silicon atoms primarily occupy tetrahedral positions. In
dioctahedral 2:1 structures, the octahedral interspaces are either two-thirds
occupied by M” ions (especially Al”) whereas in trioctahedral structures they
are completely occupied by M” cations, especially Mg”. Aluminum may
substitute for silicon in tetrahedral positions and magnesium, iron, or lithium

atoms may replace metal ions in octahedral sites.

12

of interlayer cations, and position of layer charge.32 Under appropriate conditions
individually swollen smectite layers (exfoliation or delamination) are completely

dispersed in a certain matrix.

Table 1.2 Elemental compositions of common 2:1 type smectites. In dioctahedral
smectite clays two-third of the octahedral sites are occupied by metal ions.
Trioctahedral smectites possess all octahedral positions filled by metal ions.
lnterlayer cations are all assumed to be monovalent but cations of various charge

may occupy the interlayer position.30

 

Mineral group Dioctahedral Trioctahedral

Talc-pyrophyllite Pyrophyllite:AI2Si4O1o(OH)2 Talc: MgGSi4O1o(OH)2
Montmorillonite: (Mx+nH2O) Hectorite: (Mx+nH2O)

 

 

Smectites (Al2-ngx)Si4O1o(OH)2 (Mgg-xLix)Si4O1o(OH)2
(x ~ 0.2—0.6)'r Beidellite: (Mx+nH2O) Saponite: (Mx-y+nH2O)((Mgg-y)
AI2(Si4-xAlx)O1o(OH)2 (Al,Fe)x)(Si4-xAlx)O10(OH)2

 

T x is charge per formula unit (010(OH)2).

1.2.2 Polymer-Clay Nanocompsites

Layered silicates have been widely used to improve the mechanical and
barrier properties of organic polymers. Toyota researchers first developed
polymer-smectite clay nanocomposites by uniformly dispersed individual silicate

layers in a nylon-6 matrix.” 34 Protonated e-caprolactam was intercalated and

13

polymerized in the galleries of montmorillonte. The unlimited swelling of the clay
layers in the polymer matrix provided reinforced mechanical properties”, such as
modulus (which increased by 200%), strength (increased by 50%) and heat
distortion temperature (increased by 80 °C). The reinforcement realized for
nylon-6 was attributed to the large surface area (~750 m2/g) and high aspect ratio
(>200) of the clay.36 Such properties allow the clay nanolayers, when dispersed
in polymer matrix, to carry out a significant component of the applied load.37 A
recent theoretical study demonstrated that high aspect ratio and high intrinsic
modulus (400 GPa) of the clay nanolayers facilitate the reinforcement in
mechanical properties.38 Moreover the high aspect ratio of the impermeable clay
nanolayers provides a tortuous pathway (details discussed in 1.2.5 section) for
decomposed versatile components, to enhance heat distortion temperature of the
polymer.

In order to generalize the exfoliation of clay nanolayers in other polymers
the inorganic exchange ions in the galleries of the native clay are generally
replaced by a long-chain organic modifier (usually, alkylammonium surfactants)
to provide the surface of the clay with hydrophobic environment, which can
compatibilize the clay surface and the hydrophobic polymer matrix.36 The
nanocomposite concept has been applied to a variety of organic polymer
systems.

Polymer-clay nanocomposites can be generally classified into three types
39 In

based on the structures of clay layers in the polymer matrix (Figure 1.6).

contrast to conventional composites (Figure 1.6 a) formed from clay tactoids with

14

(a) Conventional

 

 

- , 4"
-.. 2'“, A
A A

'v
5&5.

 

*°IIIII\\\

Polymer

 

 

 

/ - - V" Vie/5:019"
‘fié’. , REVSWJ;

 

 

 

 

Figure 1.6 Schematic illustrations of polymer-clay composite structures.
Rectangular bars represent clay layers with ~1 nm thickness; (a) conventional
composites, (b) intercalated nanocomposites, and (c) exfoliated (delaminated)
nanocomposites. The corresponding powder XRD patterns for composites at

low angle region (29 = 1—10°) are also illustrated.

15

no intercalation of the polymer into the clay structure, intercalated
nanocomposites (Figure 1.6 b) are prepared when one or a few polymer chains
are inserted into the clay galleries with retained layer ordering. In the exfoliated
nanocomposites (Figure 1.6 c) the clay nanolayers are completely dispersed and
the separation between nanolayers may be uniform or disordered in the
continuous polymer matrix. Exfoliated nanocomposites allow for higher
homogeneity and more improved interfacial interactions between the layers and
the matrix, unlike conventional and intercalated composites which have greatly
reduced interfacial interactions with the polymer.
1.2.3 Thermoset Epoxy-Clay Nanocomposties

Crosslinked epoxy polymers are known for their high chemical resistance,
good adhesion, attractive optical properties and flexibility.40 The term “resin or
thermoset plastic” is applied to both prepolymer and its cured polymer. The most
common structure of epoxy resins are based on bisphenol A (structure 1-1).41 To
generate thermoset epoxy polymer network, the epoxy resins are cured (or
crosslinked) by curing agents. Polyamines are the most common curing agent
with reaction involving ring-opening addition of amine (reaction 1-2). Carboxylic
acids and anhydrides have also been used as curing agents. In this thesis two
types of epoxy resins (diglycidyl ether of bisphenol As, EPON 826 and 828,
(structures 1-3)) were used in combination with diamine curing agents (a,w-
diamine-terminated polypropylene oxide, Jeffamine D230 and 02000, (structures

14)).

16

CH3
1 1-1
CH3

9H
0 9H $H2‘CH-CH2W
VVV‘CH2-CH-CH2-N (1-2)
CHz-CH-CH2~A + NHz-R-NHZ —> F', OH

I l
N-CH2-CH-CH2W
WCH2-?H-CH2
OH

—

0 CH3 OH '
/ \ | l
CHz-CH-CHz-O‘ (I: O-CH2-CH-CH2

CH3

0
l / \
o (I: O-CH2-CH-CH2

I

 

 

CH3 _ n CH3
EPON 826; n = 0 (90%), 1 (10%) '
EPON 828; n = 0 (88%), 1 (10%), 2 (2%) (1'3)

NHz-CH-CH21'O-CH2-CH--NH2 Jeffamine D230; n = 5.6, Mw ~200
l I Jeffamine 02000; n = 33.1, Mw ~2000
CH3 CH3

.. un

(14)

 

 

The first example of an exfoliated clay-epoxy nanocomposite structure42
was prepared using protonated alkylammonium ions intercalated in
montmorillonite. The intercalate facilitates the polymerization—delamination
process. In the absence of an amine curing agent the acidity of the exchanged
onium ions plays the role of a catalyst for the cationic ring-opening self-
polymerization of the epoxy to form a polyether (reaction 1-5).43 In the presence
of amine curing agent the acidic intergallery onium ions also catalyze the
polymerization reaction in the clay gallery at rates that are competitive with the

extragallery polymerization rate.39 The competitive polymerization rates facilitate

17

clay exfoliation, which is confirmed by the observation of only intercalated
composites formed from nonacidic ammonium ions. Other factors that affects the
exfoliation of clays are the chain-length of the organic modifiers and the layer

charge density of the pristine clays.39

o: . - *OH

O + RNH3 - /\ + RNH2
o: +oH

{ E’mf § l> +—CH2-CH2-OH

{ E + [>o+—CH2-CH2-0H > (1.5)

Exfoliated silicate—epoxy nanocomposites can show remarkable

 

 

V

 

V

improvement in mechanical properties. A glassy thermoset epoxy nanocomposite
containing 4 v/v% exfoliated montmorillonite shows a 60% increase in storage
modulus compared to the pure epoxy.44 A rubbery epoxy nanocomposite
containing 23 wt% montmorillonite exhibits more than 10-fold improvement in
tensile modulus and strength and 40-60% elongation-at-break.45
Generally ,intercalated clay nanocomposites are less effective than exfoliated
analogues in improving mechanical properties39 and transparency“. Recently,
protonated a, w—diamines were used as clay surface modifiers and curing agents
and exfoliated montmorillonite—epoxy nanocomposites were successfully
prepared through the acid-catalyzed curing reaction shown in the reaction
scheme (1.6).47 This approach reduces the need for an organic modifier and

limits the plasticizing effect caused by long-chain organic modifier. Carbon

18

nanofibers48 and exfoliated graphite49 have also been used as reinforcement
agents. However, the resulting nanocomposites manifest less improvement in

mechanical properties than that of clay counterparts at the same loading of the

agents.
*NR1R2R3
'1
: NR1R2R3
A (O H OH
:./H 5+/\ , +I l
H 1'.
cm
——> R—N—CHz-CH—R' + H*NR1R2R3 (1.6)

H
1.2.4 Mesocomposites

Mesoporous silica materials with large pore size are promising hosts for
the construction of nanocomposites with organic polymers.50 For example, one
dimensional channels of hexagonal MGM-41 allow conjugated polymer chains
such as polydiacetylenes‘, polyvinylene52 and polythiophene53 to align within the
pores so that the resulting nanocomposites exhibit unusual chromatic changes,
polarized luminescence, and high stability after 20 voltametric cycles,
respectively. However, only a few studies have been published describing
polymer-based nanocomposites containing mesoporous silicas to improve

l54‘56 and gas permeability properties”.

mechanica
Kojima et al.54 reported that Nylon 6,6/FSM silica composites (35 wt%

FSM silica loading) could be prepared above 2 MPa pressure, which provided a

19

driving force for nylon 6,6 chains to insert into the pores of FSM. The resulting
composites exhibited about a two-fold larger storage modulus in the viscoelastic
experiments than that of pure nylon 66 processed under the same conditions.
Collapse of FSM framework structure was not observed under 30 MPa pressure,
as confirmed by XRD and TEM. However, even higher loadings of FSM particles
showed only 2-fold enhancement of storage modulus from the composites. A
mesostructured organosilica55 and a silylated MOM-4856 were used as reinforcing
agents for poly((3-trimethoxysilyl)propylmethacrylate) (PTMSPMA) and poly(vinyl
acetate) (PVAc), respectively. Approximately 140% and 700% increases were
observed in tensile strength and tensile modulus, respectively, for the 28.2 wt%
silica-loaded PTMSPMA composite, but the elongation and the toughness of the
composite were sacrificed. In case of the PVAc-based composite containing 4.3
wt% silylated MGM-48, tensile modulus, strength, and toughness were increased
by 300%, 80%, and 40 %, respectively along with a 33% decrease in elongation.
1.2.5 Gas Permeability

The high aspect ratio (200-2000) of clay nanolayers provides enhanced
barrier characteristics as well as reduced solvent uptake, chemical resistance
and flame retardance.36 A simple model58 (shown in Figure 1.7 a) has been used
to explain the reduction of gas permeability by the impermeable exfoliated clay
platelets (Figure 1.7 b) which present a tortuous path (Figure 1.7) to the
permeant gases. In the tortuous path mechanism the relative permeability is Pc/
P0 = 1 / r, where Pc and P0 are the penneabilities of the composite and unfilled

polymer and ,ris the tortuosity factor defined by the ratio of distance a molecule

20

must travel to get through the polymer film to thickness of the film. The
expression for the tortuosity factor in the case where all the platelets are aligned

are shown in Figure 1.7 is
1
r =1+§a¢ (1.6)

where (:5 is the volume fraction of the platelets and a is the aspect ratio of the

plates. This mechanism neglects horizontal variations and disordered
arrangement of the platelets.59 The more developed version of the tortuosity
factor60 appropriate for the three-dimensional, disordered arrange of platelets has

been given as
r =1+%pa2¢2(1—¢) (1.7)

where p is a geometric factor depending on the size and shape of the platelets
and on the extent of positional disorder.

Polymer membranes for gas separation require high permeability
coefficients as well as high selectivity coefficients.61 Highly selective zeolites
have been used as absorbents in rubbery or glassy polymer-based membranes
to improve gas separation performance.°2‘65 Approximately a two-fold increase of
oxygen gas permeability was observed for zeolite/polymer composite
membranes.” 65 High zeolite loadings (53 v/v% and 50 wt%), however, are
required to achieve the enhancement of oxygen permeabilities. For lower zeolite
loadings (0—30 wt%), the oxygen permeability coefficients of the composite
62, 64, 65

materials generally decrease along with the increase of zeolite loadings

due to insufficient pore channel networks. Non-selective interfacial voids caused

21

 

 

 

 

 

 

(b)

 

 

Figure 1.7 A model (a) for the tortuous path of a diffusing molecule through a

polymer filled with oriented impermeable disks (b) with aspect ratio, a = L / W.

22

by the weak interaction between the zeolite surface and the polymer matrix do
not provide significant perrnselectivity improvements.66 The meager interfacial
interaction between the polymer chains and the inorganic additive surface also
results in poor dimensional properties for the polymer composites.

A study of the mechanical properties of natural zeolite-epoxy composites
showed that the tensile strength, Young’s modulus and impact strength decrease
with increase in the zeolite loadings, confirming the poor adhesion between the
zeolite surface and polymer chains.67 Wetting polymer chains at the surface of
the reinforcing agents are needed in order to achieve the high mechanical
stability that is demanded for good membrane performance.68

Only a few examples describing polymer-based membranes containing
mesostructured silicas have been reported to show improved performance for
gas separation membranes. The Addition of 30 wt% MGM-41 silica to a
polysulfone (PSF) resulted in a 155% increase in permeability for oxygen gas
without loss of the O2/N2 selectivity due to the increase in diffusivity from
nonselective voids.57 As MGM-41 loading increased, however, the MCM-41/PSF
composite membranes became more brittle, despite FT-IR evidence for
hydrogen-bonding interaction between PSF and MCM-41. The mesoporous
silica-polymer membrane concept was expanded to include amine-functionalized
DAM-1(MCM-41 type silica)«polyimide membranes“. Approximately 260% and
45% increases in oxygen permeability and O2/N2 selectivity were observed at 30

wt% and 20 wt% loadings, respectively. However the weak mechanical

23

properties of the DAM-1-polyimide membranes were a drawback, indicating the

poor interaction between the polymer and the mesoporous silica.

1.3 Research Goals and Significance

Although intercalated/exfoliated clay-polymer nanocomposites exhibit
improved properties such as mechanical, thermal and barrier properties the
hydrophilic nature of the clay surface must be modified with organic modifiers in
order to compatibilize the hydrophobic polymer chains. Despite using organic
modifiers on the surfaces of clays to improve the compatibility of the clay with the
polymer, primarily intercalated or partially exfoliated clay structures have been
obtained for many polymer systems"): 71 Also, many efforts, such as changing
processing methods“, using solvents”, and selecting organic modifiers39 have
been made to facilitate induce exfoliation of the smectite clays in polymer
matrices. Despite these efforts, it is still extremely difficult to routinely exfoliate
clay nanolayers in a polymer matrix.

Mesoporous silicas may be promising candidates reinforcing organic
polymers owing to their high surface area and high pore volume. However,
earlier studies aimed at using mesoporous silica as a reagent have not been
successful and the characterizations of the mesocomposites are not clear. Thus
we will herein propose that relatively large pore size (5—30 nm) can provide
polymer chains with facilitated access to the inner silica wall without the need for
organic surfactants in the pores. The framework pore diameters of

mesostructured silicas are comparable to the gallery heights of the clay layer

24

structures in polymer media. Moreover the pore diameters are easily expandable
by temperature, surfactants, co-surfactant and pore expanders to distances that
exceed the basal spacing of exfoliated clay layers (>10 nm). As a result,
mesostructured silica can afford cost- and labor-effective process for preparation
of reinforced polymer nanocomposites.

In this thesis MGM-41 type mesostructures are excluded as viable
reinforcing agents due to the limiting size and one-dimensional nature of the
cylindrical pores, which is not suitable for filling by polymer chains. For analogous
reasons cubic mesoporous silica (MGM-48, 3.5 nm pore size) with bicontinous
pore networks is not expected to be a useful candidate because the MOM-48
does not disperse in thermoset epoxy matrices”. The relatively small pore size
and low pore volume (high bulk density) of the MGM-48 may be the cause of the
poor dispersion of the silica. Thus, we must take into account large pore size and
high pore volume as well as three-dimensional pore structure to select beneficial
candidates for the improvement of mesostructure dispersion and polymer
properties. Moreover mesostructures prepared under electro-statically neutral
assembly conditions are preferred because the surfactants used to template the
mesostructures can be easily removed by ethanol extraction.

One of the objectives in this work is to synthesize of large-pore wormhole
mesostructures and study their use as reinforcing agents for improvement of
mechanical properties of thermoset epoxy polymers, which have been widely
used in a variety of applications including coatings, adhesives and electronics. A

mesocellular foam (MSU-F) prepared under a nearly neutral pH condition will be

25

also explored as a reinforcement agent of thermoset epoxy system.

26

1.4 References

(1) Kresge, C. T.; Leonowicz, M. E.; Roth, W. J.; Vartuli, J. C.; Beck, J. S. Nature
1992, 359, 710.

(2) Beck, J. S.; Vartuli, J. C.; Roth, W. J.; Leonowicz, M. E.; Kresge, C. T.;
Schmitt, K. D.; Chu, C. T. W.; Olson, D. H.; Sheppard, E. W.; Mccullen, S. B.;
Higgins, J. B.; Schlenker, J. L. J. Am. Chem. Soc. 1992, 114, 10834.

(3) Monnier, A.; Schuth, F.; Huo, Q.; Kumar, D.; Margolese, D.; Maxwell, R. S.;
Stucky, G. D.; Krishnamurty, M.; Petroff, P.; Firouzi, A.; Janicke, M.; Chmelka, B.
F. Science 1993, 261, 1299.

(4) lnagaki, S.; Fukushima, Y.; Kuroda, K. J. Chem. Soc. Chem. Comm. 1993,
680.

(5) Yanagisawa, T.; Shimizu, T.; Kuroda, K.; Kato, C. Bull. Chem. Soc. Jpn. 1990,
63, 988.

(6) Ying, J. Y.; Mehnert, C. P.; Wong, M. S. Angew. Chem. Int. Ed. 1999, 38, 56.
(7) Huo, Q. S.; Margolese, D. l.; Stucky, G. D. Chem. Mater. 1996, 8, 1147.

(8) Huo, Q. S.; Margolese, D. l.; Ciesla, U.; Demuth, D. G.; Feng, P. Y.; Gier, T.
E.; Sieger, P.; Firouzi, A.; Chmelka, B. F.; Schuth, F.; Stucky, G. D. Chem. Mater.
1994, 6, 1176.

(9) Huo, Q. 8.; Leon, R.; Petroff, P. M.; Stucky, G. D. Science 1995, 268, 1324.
(10) Tanev, P. T.; Pinnavaia, T. J. Science 1995, 267, 865.

(11) Tanev, P. T.; Pinnavaia, T. J. Chem. Mater. 1996, 8, 2068.

(12) Zhang, W. Z.; Pauly, T. R.; Pinnavaia, T. J. Chem. Mater. 1997, 9, 2491.

(13) Lee, J.; Yoon, 8.; Oh, S. M.; Shin, C. H.; Hyeon, T. Adv. Mater. 2000, 12,
359.

(14) Lee, J.; Han, S.; Hyeon, T. J. Mater. Chem. 2004, 14, 478.
(15) Tanev, P. T.; Pinnavaia, T. J. Science 1996, 271, 1267.
(16) Kim, S. S.; Zhang, W. Z.; Pinnavaia, T. J. Science 1998, 282, 1302.

(17) Bagshaw, S. A.; Prouzet, E.; Pinnavaia, T. J. Science 1995, 269, 1242.

27

(18) Kim, S. S.; Pauly, T. R.; Pinnavaia, T. J. Chem. Comm. 2000, 835.
(19) Prouzet, E.; Pinnavaia, T. J. Angew. Chem. Int. Ed. 1997, 36, 516.
(20) Bagshaw, S. A.; Pinnavaia, T. J. Angew. Chem. Int. Ed. 1996, 35, 1102.

(21) Zhao, D. Y.; Feng, J. L.; Huo, Q. S.; Melosh, N.; Fredrickson, G. H.;
Chmelka, B. F.; Stucky, G. D. Science 1998, 279, 548.

(22) Kim, S. S.; Pauly, T. R.; Pinnavaia, T. J. Chem. Comm. 2000, 1661.

(23) Schmidt-Winkel, P.; Lukens, W. W.; Zhao, D. Y.; Yang, P. D.; Chmelka, B.
F.; Stucky, G. D. J. Am. Chem. Soc. 1999, 121, 254.

(24) Lettow, J. 8.; Han, Y. J.; Schmidt-Winkel, P.; Yang, P. D.; Zhao, D. Y.;
Stucky, G. D.; Ying, J. Y. Langmuir 2000, 16, 8291.

(25) Vartuli, J. C.; Schmitt, K. D.; Kresge, C. T.; Roth, W. J.; Leonowicz, M. E.;
Mccullen, S. B.; Hellring, S. D.; Beck, J. S.; Schlenker, J. L.; Olson, D. H.;
Sheppard, E. W. Chem. Mater. 1994, 6, 2317.

(26) Zhao, D. Y.; Huo, Q. S.; Feng, J. L.; Chmelka, B. F.; Stucky, G. D. J. Am.
Chem. Soc. 1998, 120, 6024.

(27) Pauly, T. R.; Pinnavaia, T. J. Chem. Mater. 2001, 13, 987.

(28) Tanev, P. T.; Liang, Y.; Pinnavaia, T. J. J. Am. Chem. Soc. 1997, 119, 8616.
(29) Karkamkar, A.; Kim, S. S.; Pinnavaia, T. J. Chem. Mater. 2003, 15, 11.

(30) Bailey, S. W. In Crystal Structures of Clay Minerals and Their X-Ray
Identification, Edited by Brindley, G. W. Brown, G., Mineralogical Society, London,
1980.

(31) Pinnavaia, T. J. Science 1983, 220, 365.

(32) McEwan, D. M. C.; Wilson, M. J. In Crystal Structures of Clay Minerals and
Their X—Ray Identification, Edited by Brindley, G. W. Brown, G., Mineralogical
Society, London, 1980.

(33) Fukushima, Y.; Okada, A.; Kawasumi, M.; Kurauchi, T.; Kamigaito, 0. Clay
Miner. 1988, 23, 27.

(34) Okada, A.; Kawasumi, M.; Usuki, A. Mat. Res. Soc. Symp. Proc. 1990, 171,
45.

28

(35) Kojima, Y.; Usuki, A.; Kawasumi, M.; Okada, A.; Fukushima, Y.; Kurauchi,
T.; Kamigaito, O. J. Mater. Res. 1993, 8, 1185.

(36) LeBaron, P. C.; Wang, Z.; Pinnavaia, T. J. Appl. Clay Sci. 1999, 15, 11.
(37) Weon, J. l.; Sue, H. J. Polymer 2005, 46, 6325.

(38) Fomes, T. D.; Paul, D. R. Polymer 2003, 44, 4993.

(39) Lan, T.; Kaviratna, P. D.; Pinnavaia, T. J. Chem. Mater. 1995, 7, 2144.
(40) Lange, J.; Nicolas, B.; Galy, J.; Gerard, J. F. Polymer 2002, 43, 5985.
(41) Odian, G. Principles of Polymerization. 3rd ed. Wiley, New York, 1993.
(42) Wang, M. S.; Pinnavaia, T. J. Chem. Mater. 1994, 6, 468.

(43) Lan, T.; Kaviratna, D.; Pinnavaia, T. J. J. Phys. Chem. Solids 1996, 57, 1005.
(44) Messersmith, P. B.; Giannelis, E. P. Chem. Mater. 1994, 6, 1719.

(45) Lan, T.; Pinnavaia, T. J. Chem. Mater. 1994, 6, 2216.

(46) Brown, J. M.; Curliss, D.; Vaia, R. A. Chem. Mater. 2000, 12, 3376.

(47) Triantafillidis, C. S.; LeBaron, P. C.; Pinnavaia, T. J. J. Solid State Chem.
2002, 167, 354.

(48) Richard, P.; Prasse, T.; Cavaille, J. Y.; Chazeau, L.; Gauthier, C.; Duchet, J.
Mater. Sci. Eng. A 2003, 352, 344.

(49) Fukushima, H. Ph. D. Thesis, Michigan State University, East Lansing, 2001.
(50) Shi, J. L.; Hua, Z. L.; Zhang, L. X. J. Mater. Chem. 2004, 14, 795.

(51) Lu, Y. F.; Yang, Y.; Sellinger, A.; Lu, M. C.; Huang, J. M.; Fan, H. Y.;
Haddad, R.; Lopez, G.; Burns, A. R.; Sasaki, D. Y.; Shelnutt, J.; Brinker, C. J.
Nature 2001, 410, 913.

(52) Nguyen, T. 0.; Wu, J. J.; Doan, V.; Schwartz, B. J.; Tolbert, S. H. Science
2000, 288, 652.

(53) Li, G. T.; Bhosale, S.; Wang, T. Y.; Zhang, Y.; Zhu, H. S.; Fuhrhop, K. H.
Angew. Chem. Int. Ed. 2003, 42, 3818.

(54) Kojima, Y.; Matsuoka, T.; Takahashi, H. J. Appl. Polymer Sci. 1999, 74,
3254.

29

(55) Ji, X. L.; Hampsey, J. E.; Hu, O. Y.; He, J. B.; Yang, Z. 2.; Lu, Y. F. Chem.
Mater. 2003, 15,3656.

(56) He, J.; Shen, Y. B.; Yang, J.; Evans, D. G.; Duan, X. Chem. Mater. 2003, 15,
3894.

(57) Reid, B. D.; Ruiz-Trevino, A.; Musselman, l. H.; Balkus, K. J.; Ferraris, J. P.
Chem. Mater. 2001, 13, 2366.

(58) Nielsen, L. E. J. Macromol. Sci. (Chem) 1967, A1 , 929.
(59) Fredrickson, G. H.; Bicerano, J. J Chem Phys 1999, 110, 2181.

(60) Cussler, E. L.; Hughes, S. E.; Ward, W. J.; Aris, R. J. Membrane Sci. 1988,
38, 161.

(61) Maier, G. Angew. Chem. Int. Ed. 1998, 37, 2961.
(62) Jia, M. D.; Peinemann, K. V.; Behling, R. D. J. Membrane Sci. 1991, 57, 289.

(63) Duval, J. M.; Folkers, B.; Mulder, M. H. V.; Desgrandchamps, G.; Smolders,
C. A. J. Membrane Sci. 1993, 80, 189.

(64) Gur, T. M. J. Membrane Sci. 1994, 93, 283.

(65) Suer, M. G.; Bac, N.; Yilmaz, L. J. Membrane Sci. 1994, 91, 77.
(66) Zimmerman, C. M.; Singh, A.; Koros, W. J. J. Membrane Sci. 1997, 137, 145.
(67) Lee, J. Y.; Shim, M. J.; Kim, S. W. Polym. Eng. Sci. 1999, 39, 1993.
(68) Stem, S. A. J. Membrane Sci. 1994, 94, 1.

(69) Balkus, K. J.; Cattanach, K.; Musselman, I. H.; Ferraris, J. P. Mater. Res.
Soc. Symp. Proc. 2003, 752, 91.

(70) Robello, D. R.; Yamaguchi, N.; Blanton, T.; Barnes, C. J. Am. Chem. Soc.
2004, 126, 8118.

(71) Shen, M. M.; Lu, M. G.; Chen, Y. L.; Ha, C. Y. Polym lnt2005, 54, 1163.
(72) Choi, Y. 8.; Ham, H. T.; Chung, l. J. Chem. Mater. 2004, 16, 2522.

(73) Park, I.; Pinnavaia, T. J. Unpublished data, 2004.

30

CHAPTER 2

ASEMBLY OF LARGE-PORE SILICA MESOPHASES WITH WORMHOLE

FRAMEWORK STRUCTURES FROM a,w-DIAMINEPOROGENS

2.1 Introduction

Supramolecular assembly pathways based on hydrogen-bonding
interactions between electrically neutral amine surfactants and electrically neutral
silica precursors typically afford mesostructures with wormhole framework
motifs.1 Compared to the highly ordered monolithic mesophases prepared
through electrostatic templating pathways, wormhole mesostructures typically
exhibit improved activity as heterogeneous catalysts for condensed phase
reactions due to the three-dimensional pore network and an improved textural
mesoporosity that minimize diffusion and facilitate access to reaction sites on the
pore surfaces.”6 Neutral amine templating offers the additional advantage of
facile recovery of surfactants by solvent extraction methods.”9 Also, the use of
amine surfactant micelles as mesostructure-directing agents is not limited to
silicate frameworks. Mesoporous forms of certain transition-metal oxides also
have been prepared using amine surfactants."9

Despite the advantages of amine surfactants for the hydrogen-bonding
assembly of highly accessible mesostructured silicas, the average framework

pore sizes have been limited to the range 1.6-3.1 nm.1 An expansion of

framework pore sizes up to 4.5 nm can be achieved for some wormhole silicas

31

through the incorporation of a cosolvent, such as 1,3,5-trimethylbenzene, in the

hydrophobic interior region of the micelle.”11

Even greater mesopore
accessibility and catalytic activity can be anticipated for these wormhole
framework structures if the framework pore sizes can be expanded beyond these
limiting values and into the >10 nm range.

We report herein a new group of large pore mesoporous silicas, denoted
MSU-J, with three-dimensional wormhole framework structures prepared through
hydrogen-bonding pathways. Our approach is based on the use of
amineterrninated polypropylene oxide (PPO) Jeffamine surfactants of the type
H2NCH(CH3)CH2[OCH2CH(CH3)]XNH2 as the structure-directing porogen. We
show that pore sizes and pore volumes up to 14.3 nm and 2.29 cm3/g,
respectively, can be achieved, depending on the synthesis conditions and the
molecular weight of the surfactant. The resulting pore sizes and BET surface

areas are comparable to those of SBA-15 silicas, but unlike SBA-15, the

mesopores are three-dimensionally connected.

2.2 Experimental

Jeffamine D2000 and D4000 surfactants (abbreviated to D2000 and
D4000, respectively) were obtained from Huntsman Corp. and used without
further purification as structure-directing porogens. Commercially available
sodium silicate (Aldrich, 27% SiO2, 14% NaOH) solution and TEOS (Aldrich)

were used as the silica precursors.

32

In a typical synthesis of MSU-J from sodium silicate, an amount of acid
(HCI) equivalent to the formal hydroxide content of the sodium silicate precursor
was added to an aqueous solution of the surfactant. The desired quantity of
sodium silicate was added to the surfactant solution under vigorous stirring at
ambient temperature, and the mixture with a pH of 8.0-8.5 was allowed to age in
a heated water bath at the desired synthesis temperature for 20 h with stirring.
The surfactant was then removed from the washed, air-dried solids by calcination
at 600°C for 4 h in air.

MSU-J molecular sieves also were prepared with TEOS used as a silica
precursor in water/ethanol solvent mixture. The surfactant was dissolved in
ethanol, and then the desired amount of water was added under stirring. TEOS
was added to the surfactant solution. The molar composition of the reaction
mixture was 1.0:0.125:220:17 TEOS:D2000:H2O:EtOH. The assembly conditions
were the same as those used in the preparation of MSU-J mesostructures made
from sodium silicate.

Xaray diffraction (XRD) patterns were obtained on a Rigaku Rotaflex 2008
diffractometer equipped with Cu K2, X-ray radiation and a curved crystal graphite
monochromator operating at 45 kV and 100 mA. Transmission electron
micrOSCOpy (TEM) images were taken on a JEOL JEM-1OOCX ll microscope with
a CeBe filament and an accelerating voltage of 120 kV. Sample grids of calcined
mesoporous silicas were prepared via sonication of powdered sample in EtOH
for 10 min and evaporating 2 drops of the suspension onto a carbon-coated,

holey film supported on a 3 mm, 300 mesh copper grid.

33

2.3 Results and discussion

Tetraethyl orthosilicate (TEOS) and alkylamine surfactants are known to
be effective reagents for the supramolecular assembly of wormhole framework
structures through hydrogen- bonding interactions between the amine
headgroups of the micelles and the silanol groups of the TEOS hydrolysis
products.1'1°'11 Sodium silicate also can be used for the preparation of silica
mesophases through supramolecular hydrogen-bonding mechanisms, provided
that the formal hydroxide ion content of the solution is balanced through the
addition of an acid to achieve a reaction mixture at a near-neutral pH.12 In the
present work both reagents were used as the silica source for the amine-directed
assembly of very large pore wormhole mesostructures.

To optimize the framework pore size of the mesostructured products, we
selected two high molecular weight a,a)-diamine surfactants of the type
H2NCH(CH3)CH2[OCH2CH(CH3)]XNH2 as the structure-directing porogens. The
derivatives with hydrophobic propylene oxide segments corresponding to x ~ 33
and 68 have approximate molecular weights of 2000 and 4000, respectively.
Reflecting these average molecular weights, the surfactants are marketed under
the trade names Jeffamine D2000 and Jeffamine D4000.

Numerous survey experiments indicated the optimum reaction
stoichiometry for the assembly of MSU-J mesostructures from sodium silicate to
be 1.0:0.83:0.125:0.83:230 SiO2zNaOHzDZOOO (or 0.062 for D4000):HC|:H2O, as
judged by N2 adsorption isotherms and the corresponding textural properties of

the assembly products. For TEOS as the silica precursor, the optimal reaction

34

 

 

 

Intensity (arbitrary units)

 

 

 

 

 

I I T l l l I rfi I I l I

12 3 4 5 6 7 8
29

Figure 2.1 X-ray diffraction patterns of as-synthesized ( ----- ) and calcined (—)
MSU-J silica molecular sieves using D2000 as templating agents and sodium
silicate as a silica source under neutral pH conditions at 25, 45, and 65°C. The

numbers are the pore-pore correlation distances in nanometer (nm) units.

35

stoichiometry was 1.0:0.125:220:17 TEOS:D2000:H2O:EtOH.

Figure 2.1 illustrates X-ray diffraction (XRD) patterns of as-synthesized
and calcined MSU-J silicas prepared from the sodium silicate and D2000
surfactant at temperatures of 25, 45, and 65°C. Each sample exhibits an intense
and relatively broad reflection in the 26 range 1-1.5° with a weak, broad shoulder
appearing at 2-3°. The absence of multiple hkl reflections indicates that the
framework pore system is lacking in long-range order, but the well-expressed
low-angle reflection is consistent with the average correlation distance between
pores in a wormhole framework structure. Relatively broad XRD reflections (not
shown) also were observed for MSU-J silicas synthesized from TEOS and D2000
at 60°C and from sodium silicate and D4000 at 45 and 65°C. In case of MSU-J
silicas with pores > 8 nm, no XRD reflections could be detected at a 219 angle
higher than 1° due to the detection limit of our wide-angle X-ray diffractometer.

The presence of a wormhole framework structure for each of the
mesostructured products was verified by TEM images, examples of which are
shown in Figure 2.2. Although we were not able to observe a low-angle XRD
peak for mesostructures with a pore size > 8 nm, the pore structures shown in
the TEM images of this product (Figure 2.20,d) confirm the wormhole structure
and the relatively regular porosity.

N2 adsorption-desorption isotherms and Barrett-Joyner-Halenda (BJH)
framework pore-size distributions (inset) for calcined MSU-J phases synthesized

from TEOS are presented in Figure 2.3. Filling of the framework-confined

36

* .
: 1i *3 * " ii "
.1). "a... go. I i. ‘
, ‘1 II c a.)
. :0- 1'
.80.? ? "KI/r

”.07 .03: g:-

 

Figure 2.2 TEM images of calcined MSU-J silicas synthesized with the
Jeffamine surfactants and the silica sources; (a) 02000 and sodium silicate at
25°C, (b) D2000 and TEOS at 45°C, (c) D2000 and TEOS at 60°C, and (d)
D4000 and sodium silicate at 65°C. Average pore sizes are 5.3, 7.5, 11.9, and

14.3 nm for (a), (b), (c), and (d), respectively.

37

1600‘

E
I- 1400-
U)
”2’ 1200—
E
0
V 1000-

$

 

 

 

N Volume Adsorbed
$

  

respectively).

 

 
  
  
  
  

 

10 100 10

 

 

I I I I I I I I I I I I I

I 0!2 0.4 0.6 0.8
Relative Pressure (PIPo)

Figure 2.3 N2 adsorption-desorption isotherms of MSU-J silica molecular
sieves assembled from TEOS in the presence of 02000 as a structure director
at different synthesis temperatures and calcined at 600 °C for 4 h. The insert
provides BJH framework pore size distributions determined from the

adsorption branches of the N2 isotherms (o, a, and A for 25, 45, and 60°C,

38

mesopores occurred at P/Po) 0.5—0.9. As seen from the BJH plots, the pore size
increases from 4.9 at 25°C to 11.9 nm at 60°C. The TEM images of the calcined

MSU-J mesostructures assembled from TEOS and D2000 at 45 and 60°C in (see
Figure 2.2, parts b and c, respectively) reveal pore sizes consistent with those
obtained from the pore-size distributions calculated on the adsorption branches
of the nitrogen adsorption isotherms.

As shown by the nitrogen isotherms and pore-size distributions in Figure
2.4, the pore volumes and pore sizes of MSU-J materials assembled from
sodium silicate and D2000 surfactant are less sensitive to the assembly
temperature in comparison to the mesostructures assembled from TEOS. The
silicate oligomers generated from these two precursors are presumed to differ in
molecular weight and in the extent of cross linking. Also, differences in the ionic
strengths of the sodium silicate and TEOS reaction mixtures are likely to
contribute to differences in pore sizes of the resulting mesostructures. These
differences in precursor structure are manifested in differences in the
temperature dependence of the framework pore sizes, as well as the porepore
correlation distance.

As expected on the basis of polypropylene oxide chain length, the pore
sizes of mesostructures assembled from D4000 are larger than those assemble
from D2000 at the same synthesis temperature. Table 1 provides the BET
surface areas and pore volumes of calcined MSU-J wormhole mesostructures.
The pore diameter expands with increasing synthesis temperature over the range

25-65°C from 5.3 to 6.7 nm with sodium silicate as the silica source. The use of

39

TEOS as the silica source results in even larger average pore sizes of 7.5 and
11.9 nm at assembly temperatures of 45 and 60°C, respectively. The increases
are accompanied by expected decreases in the BET surface areas. The largest
average pore size (14.3 nm) is observed for the mesostructure assembled at

60°C from D4000 and sodium silicate.

Table 2.1 Textural properties of calcined MSU-J silicas with wormhole framework

 

 

structures.
. . S nthesis BJH Wall Pore
33?; Surfactant 3temp. (:32) d (nm)” 2%: thickness vo|u3me
(’0 (nm) (nm) (cm /9)
$58 02000 25 947 6.1 (6.0) 5.3 0.8 1.41
38 D2000 45 899 6.6(6.2) 6.0 0.6 1.49
$3 02000 65 805 7.0(6.7) 6.6 0.4 1 .49
TEOS 02000 25 1127 6.6(6.0) 4.9 1.7 1.53
TEOS 02000 45 962 8.2(7.2) 7.5 0.7 1.96
TEOS 02000 60 798 -c 1 1.9 - 2.29
$3 04000 45 756 -c 8.0 - 1 .56
$5 04000 65 408 -c 14.3 - 1 .37

 

‘3 SS = sodium silicate. b The d values in parentheses are for as-synthesized
MSU-J mesostructures. ° XRD reflections were not resolved for these samples. d
The wall thickness was determined by subtracting the BJH pore size from the
pore-pore correlation distance.

40

 

1000-

 

 

 

8001 1 1o -

     

l

 

N2 Volume Adsorbed (cm3lg, STP)

 

 

1 l I T I

O .0121 0:41 I 0:67 'OIBt 1

Relative Pressure (PIPo)
Figure 2.4 Mg adsorption-desorption isotherms of MSU-J silica molecular
sieves assembled from sodium silicate in the presence of 02000 as a
structure director at different synthesis temperatures and calcined at 600 °C
for 4 h. The insert provides BJH framework pore size distributions determined
from the adsorption branches of the N2 isotherms (o, a, and A for 25, 45, and

65°C, respectively).

41

The temperature dependence of MSU-J pore sizes is related to a
decrease in H-bonding at the silica-surfactant interface with increasing
temperature. As the framework cross-linking is enhanced through further
condensation of silanol groups at higher temperature, the polarity of the interface
is decreased.11 Consequently, the loss of hydrogen bonding structure between

the surfactant and silica aggregates at the I°S° interface will decrease the

effective headgroup area (a0) and increase the surfactant packing parameter (9 )
V/(Iao), where V is the volume of the surfactant molecule and I is the surfactant
length). The resulting decrease in micelle curvature results in an increase in
thepore size. The maximum chain lengths of 02000 and 04000 are
approximately 12 and 25 nm, respectively, when the propylene oxide units of the
structure-directing agents are placed in a “stretched" all-trans conformation, as
indicated by reverse Monte Carlo simulation.” The more likely conformations of
the PPO chains include kinks that allow for hairpin configurations in the structure-
directing micelles. In accord with this expectation, the largest pore sizes
observed for MSU-J silicas are smaller than would be predicted by the maximum
surfactant chain lengths.

The pore-size maxima for hexagonal MCM-4114 and SBA-1515'16
mesostructures has been widely examined and characterized. In the case of
MGM-41 electrostatically assembled from quaternary ammonium ion surfactants,
the framework pore sizes can be increased from values in the 34 nm range up to
values approaching ~ 7 nm by increasing the assembly temperature over the

range 100-150°C.17 Also, SBA-15 silicas prepared under strongly acidic

42

conditions in the presence of a triblock copolymer (Pluronic 123) exhibit pore
sizes of 4.7-10.6 nm, depending on reaction and postsynthesis reaction
conditions.“19 Thus, the average pore sizes observed for the wormhole
framework structures in the present work span the values found for large-pore
MCM-41 and SBA-15. Only mesostructured micellar foam structures exhibit
larger pore sizes. However, foam structures require the use of much more

‘8'19 and latex polymers20 as porogens. Also, the BET

complex microemulsions
surface areas of MSU-J silicas (408-1127 m2/g) are comparable to those of SBA-

15 (630-1040 m2/g). Moreover, MSU-J silicas have larger framework pore
volumes (1 .37-2.29 cm3/g) in comparison to SBA-15 materials (0.56-1.23 cm3/g).

In addition to spanning the pore sizes characteristic of large-pore MCM-41
and SBA-15, the mesostructures reported herein add a new dimensionality to the
family of large pore mesostructures. Being wormhole frameworks, the mesopores
in MSU-J materials are fully three-dimensional, which can be an advantage over
the one-dimensional pore system of MGM-41, particularly in facilitating access to
the framework walls under diffusion-controlled conditions. SBA-15 formally has a
three-dimensional pore system, but the mesopores are one-dimensional and
linked primarily through micropores. Secondary mesopores that link the primary
mesopores can be formed in some SBA-15 derivatives, but the secondary
mesopores represent less than 6% of the total pore volume.21 Because MOM-41
and SBA-15 are assembled through electrostatic assembly pathways, the pore-
size distributions for these mesostructures are typically narrower than the

distributions for wormhole framework structures assembled through hydrogen-

43

bonding pathways. The broader pore-size distributions for MSU-J silicas,
however, do not represent a substantial disadvantage for most catalytic or
adsorption applications, save perhaps for separations of macromolecules on the

basis of size.

2.4 Summary

Wormhole mesostructures assembled through neutral pH pathways
typically exhibit improved activity as heterogeneous catalysts, an improved
textural mesoporosity and facile recovery of surfactants by solvent extraction
methods. Despite the advantages of wormhole mesostructured materials, the

average framework pore sizes have been limited to the range 1.6—4.5 nm.

Large-pore mesoporous silicas with wormhole framework structures have
been assembled through hydrogen-bonding pathways from sodium silicate or
tetraethylorthosilicate as the silica source and amine-terminated Jeffamine
surfactants of the type H2NCH(CH3)CH2[OCH2CH(CH3)]XNH2 as the structure-
directing porogen. Depending on the molecular weight of the a,w-diamine
surfactant (x = 33 and 68 for Jeffamine 02000 and 04000, respectively) and the
synthesis temperature (25—65°C), the mean pore size distributions of the
mesostructured silicas (denoted MSU-J) were centered between 4.9 and 14.3 nm.
The resulting pore sizes and BET surface areas are comparable to those of SBA-
15 silicas, but unlike SBA-15, the mesopores are three—dimensionally connected.
Also, the BET surface areas of MSU-J silicas (408-1127 m2/g) are comparable to

those of SBA-15 (630-1040 m2/g). Moreover, MSU-J silicas have larger

44

framework pore volumes (1.37-2.29 cm3/g) in comparison to SBA-15 materials

(0.56-1.23 cm3/g). MSU-J wormhole mesostructures represent the largest pore
sizes observed to date for a fully three-dimensional mesoporous framework

assembled from a single micellar porogen.

45

2.5 References

(1) Tanev, P. T.; Pinnavaia, T. J. Science 1995, 267, 865.

(2) Sayari, A. Chem. Mater. 1996, 8, 1840.

(3) Kloetstra, K. R.; vanBekkum, H.; Jansen, J. C. Chem. Commun. 1997, 2281.,
(4) Reddy, J. S.; Sayari, A. J. Chem. Soc. Chem. Commun. 1995, 2231.

(5) Lee, J.; Yoon, 8.; Oh, S. M.; Shin, C. H.; Hyeon, T. AdV. Mater. 2000, 12, 359.
(6) Lee, J.; Han, S.; Hyeon, T. J. Mater. Chem. 2004, 14,478.

(7) Antonelli, 0. M.; Ying, J. Y. Angew. Chem. Int. Ed. 1996, 35, 426.
(8) Sun, T.; Ying, J. Y. Nature 1997, 389, 704.

(9) Kondo, J. N.; Takahara, Y.; Lu, 0. L.; Domen, K. Chem. Mater. 2001, 13,
1200.

(10) Zhang, W. 2.; Pauly, T. R.; Pinnavaia, T. J. Chem. Mater. 1997, 9, 2491.
(11) Pauly, T. R.; Pinnavaia, T. J. Chem. Mater. 2001, 13, 987.
(12) Kim, S. S.; Pauly, T. R.; Pinnavaia, T. J. Chem. Commun. 2000, 835.

(13) Carlsson, P.; Swenson, J.; Borjesson, L.; Torell, L. M.; McGreevy, R. L.;
Howells, W. S. J. Chem. Phys. 1998, 109, 8719.

(14) Kresge, C. T.; Leonowicz, M. E.; Roth, W. J.; Vartuli, J. 0.; Beck, J. S.
Nature 1992, 359, 710.

(15) Zhao, 0. Y.; Feng, J. L.; Huo, Q. S.; Melosh, N.; Fredrickson, G. H.;
Chmelka, B. F.; Stucky, G. 0. Science 1998, 279, 548.

(16) Zhao, 0. Y.; Huo, Q. S.; Feng, J. L.; Chmelka, B. F.; Stucky, G. 0. J. Am.
Chem. Soc. 1998, 120, 6024.

(17) Corma, A.; Kan, Q. B.; Navarro, M. T.; PerezPariente, J.; Rey, F. Chem.
Mater. 1997, 9, 2123.

(18) Yang, C. M.; Zibrowius, B.; Schmidt, W.; Schuth, F. Chem. Mater. 2003, 15,
3739.

46

(19) Matos, J. R.; Mercuri, L. P.; Kruk, M.; Jaroniec, M. Chem. Mater. 2001, 13,
1726.

(20) Lukens, W. W.; Yang, P. 0.; Stucky, G. 0. Chem. Mater. 2001, 13, 28-34.

(21) Galameau, A.; Cambon, H.; 0i Renzo, F.; Ryoo, R.; Choi, M.; Fajula, F. New
J. Chem. 2003, 27, 73.

47

CHAPTER 3

EPOXY -S|L|CA MESOCOMPOSITES WITH ENHANCED TENSILE

PROPERTIES AND OXYGEN PERMEABILITY

3.1 Introduction

Inorganic particles with a domain size less than 100 nm in at least one
dimension can enhance the tensile properties of a polymer through beneficial
interactions at the polymer—particle interface.1 Layered silicate clays in exfoliated
form have been widely investigated in recent years as reinforcing agents, as well
as barrier and flame retardation agents, for a variety of organic polymer
systems.“ Unlike conventional composites formed from monolithic particles, clay
nanocomposites are formed through the complete dispersion (exfoliation) of the
nanolayers that comprise the clay tactoids. This exfoliation process provides a
large surface area (~750 mzlg) and a high particle aspect ratio (2 200)5 that
normally cannot be achieved with monolithic particles.

Mesostructured silica and other metal oxides6 represent another class of
inorganic material that more recently has been considered for polymer composite
formation. These solids have controllable nanopores (2-10 nm) and surface
areas that are even higher than exfoliated clays, though the particle morphology
tends to be isotropic. Mesostructured silica with a large mesopore size has
shown promise for the formation of novel polymer composites.7 For example, the

one dimensional channels of hexagonal MCM-41 have been used to direct the

48

polymerization of conjugated polymers such as polydiacetylenea, polyvinyleneg'10
and polythiophene”. The alignment of the polymers within the pores can result in
unusual chromatic changes, polarized luminescence, and stability after multiple
cyclovoltametric cycles. Nevertheless, there has been only a few attempts to
improve the mechanical”‘14 and permeability properties15 of a polymer using
mesoporous silica as an adjuvant.

Although embedded nanoparticles can improve both the barrier and the
mechanical properties of a polymer for potential applications in packaging,16 it
generally is much more difficult to enhance the permeability of a polymer for gas
separation applications without compromising polymer strength.17 Porous zeolite
molecular sieve and mesostructured silica particles have been shown to improve
the permeability polymer-based membranes.15' ”22'” For example, the addition
of 30% (w/w) MCM-41 to polysulfone boosts the oxygen permeability by 155%
without a loss in 02/N2 selectivity due to increased diffusivity through
nonselective voids.“5 However, the composite membranes became increasingly
brittle with increasing mesostructured silica loading. Similar improvements in
permeability at the expense of strength have been observed for amine-
functionalized polyimide membranes23 embedded with mesostructured silica.

We recently reported a large pore mesostructured silica, denoted MSU-
J,24 assembled from tetraethylorthosilicate as the silica source and the a, w-
diamine polypropylene oxide H2NCH(CH3)CH2[OCH2CH(CH3)]XNH2 (x = ~33,
trade named Jeffamine 02000) as the structure-directing agent. The latter amine,

which is commercially available under the trade name Jeffamine 02000, also has

49

been use as a cross-linking agent”‘27 for the formation of rubbery epoxy
polymers. In the present study we investigate the properties of as-made, amine-
intercalated MSU-J mesostructured silica for the synthesis of epoxy—silica
mesocomposites. For comparison purposes we also have used the surfactant-
free calcined version of MSU-J silica for epoxy composite formation. The three-

dimensional wormholeza' 29

pore network (average pore size 5.3 nm) and the high
surface area (~ 950 m2/g) of MSU-J silica is shown to substantially improve the
tensile properties of the polymer. Also, we report the unexpected enhancement in
the oxygen permeation properties for composite compositions derived from as-
made MSU-J mesostructures. The observed enhancement in oxygen

permeability may find use for the design of composite membranes based on

mesostructured forms of silica.

3.2 Experimental

Epoxy resin EPON 828, a diglycidyl ether of bisphenol (Resolution
Performance Products, MW ~377), was used for the preparation of epoxy - MSU-J
silica mesocomposites. The a, w diamine—functionalized polypropylene oxide
H2NCH(CH3)CH2[OCH2CH(CH3)]XNH2 (x = ~33) with an average molecular
weight of ~2000 was obtained from Huntsman Chemical Co. under the trade
name Jeffamine 02000. The amine was used both as an epoxy curing agent and
as a structure-directing agent for the synthesis of MSU-J silica. Sodium silicate
(27% SiOz, 14% NaOH solution) was purchased from Aldrich and used as a silica

precursor for the preparation of MSU-J mesostructured silica.

50

The Jeffamine 02000 porogen was mixed with an amount of aqueous HCI
solution equivalent to the hydroxide content of the sodium silicate solution. The
silica source was added to the porogen solution under vigorous stirring at
ambient temperature and the mixture was allowed to age at 25°C for 20 h. The
optimum molar composition for the formation of as-made MSU-J was 1.0 8102 :
0.83 NaOH : 0.125 Jeffamine 02000 : 0.83 HCI : 230 H20. The porogen-
intercalated as-made mesostructured product, denoted as-made MSU-J, was
recovered by filtration and dried in air at ambient temperature. A porogen—free
analog of the mesostructure, denoted calcined MSU-J, was obtained by
calcination of as-made MSU-J at 600°C for 4 h. The as—made and calcined forms
of the mesostructures were ground to a powder for future use.

A pre—determined amount of as-made or calcined MSU-J silica was added
to the epoxy resin and mixed at 50°C for 10 min. The amount of Jeffamine
02000 curing agent needed to achieve an overall NH : epoxide stoichiometry of
1 : 1 was then added to the mixture and mixed at 50 °C for another 10 min. For
composites prepared from the as-made mesostructure, the Jeffamine 02000
present in the pores of the mesostructure was counted as contributing to the
curing process. The resulting suspensions were out-gassed under vacuum and
transferred to an aluminum mold. Pre-curing of the nanocomposite was carried

out under nitrogen gas flow at 75°C for 3 h, followed by an additional 3 h cure at
125°C to complete the cross-linking.
In order to compare the optical transparency of an epoxy-as-made MSU-J

mesocomposite against a conventional epoxy - epoxy nanocomposite, an

51

octadecylammonium-modified montmorillonite (ODA-MMT, Nanocor Inc.) was
used as a reinforcing agent. Approximately 65% of the cation exchange sites of
the ODA—MMT clay are occupied by 00A cations and the remaining sites are
occupied by protons. A procedure analogous to that used to prepare the as-
made MSU-J-epoxy mesocomposites was used to prepare the clay composites.

X-ray diffraction (XRD) patterns were obtained on a Rigaku Rotaflex 200B
diffractometer equipped with Cu K, X—ray radiation and a curved crystal graphite
monochromator operating. Samples of the liquid epoxy-silica mixture were
prepared by pouring the suspension onto filter paper mounted on a glass slide
and blotting away the excess liquid. Cured nanocomposites were examined by
mounting a rectangular flat specimen into an aluminum X-ray sample holder.
XRD patterns with equivalent intensities were obtained for the upper and lower
surfaces of the cured mesocomposites containing 12 wt% silica, thus verifying
the uniform dispersion of mesostructured silica in the epoxy matrix.

Transmission electron microscopy (TEM) images were taken on a JEOL
JEM-100CX II microscope with a CeBe filament and an accelerating voltage of
120 kV. Sample grids of calcined mesoporous silicas were prepared via
sonication of a powdered sample in ethanol for 10 min and evaporating 2 drops
of the suspension onto a holey carbon film supported on a 3 mm, 300 mesh
coppergnd.

Tensile measurements on individually molded samples were performed at
ambient temperature according to ASTM standard 03039 using an SFM-20

United Testing System. The dog-bone shaped specimens used in the tensile

52

testing were 28 mm long in the narrow region, 3 mm thick, and 3 mm wide along
the center of the casting. At least four stress-strain data points with good
precision were obtained and the results averaged. For a typical instance, the
errors in modulus and strength for the composites containing 12%(w/w) as-made
MSU-J silica are t 1.2 and :l: 0.50, respectively, based on the 95% confidence
interval for the mean values. The original cross-sectional area at zero strain was
used to compute toughness.

Oxygen permeability data were recorded on a MOCON Oxtran 2/60
oxygen permeability instrument with a test gas containing 100% oxygen, nitrogen
(containing 2 vol% hydrogen) as a carrier gas. Film samples with a thickness
between 0.9 and 1.5 mm, area of 0.317 cm2, were cast from an aluminum disk
mold, and were loaded in the testing cells with adhesive aluminum foil masks.

Thermogravimetric analyses (TGA) were performed using a computer
controlled Shimadzu TGA-50 thermal analyzer. Typically, a 50 mg sample was

placed in a pyrex bucket and heated an ambient temperature to 800 °C at a rate

of 5 °C/min in a nitrogen flow of 50 mL/min.

Positronium Annihilation Lifetime Spectroscopy (PALS) was performed
with a standard fast timing spectrometer with a 0.30 ns time resolution. A 22Na
source that emits about 106 positrons/s is sealed between two thin sheets of Ni
foil and then sandwiched between two identical sheets (each sheet is about 2 cm
square) of the sample to be studied. Positrons are emitted from the source
simultaneously with a 1270 keV gamma ray (detected as the “start” signal) and

some 80% penetrate the sealing foil and stop in the target sample. About 30% of

53

the positrons stopping in the neat epoxy form positronium, the bound state of the
positron and electron, which tends to localize in free volume voids. Subsequent
annihilation of the positronium produces 511 keV radiation detected as the “stop”
signal and the time between each start and stop signal pair is recorded as a time
histogram. This lifetime spectrum is fitted for exponential lifetime components

wherein the lifetime and intensity are related to void size and porosity.

3.3 Results

A large pore mesostructured silica with a wormhole framework structure,
denoted MSU-J, was assembled from sodium silicate and Jeffamine 02000 as
the structure-directing porogen according to previously described methods.24 The
pore diameter (5.3 nm), pore volume (1.41 cm3/g) and a surface area (947 m2/g)
were determined by a N2 adsorption isotherm following calcination of the as-
made mesostructure at 600°C. It is presumed that the as-made mesostructure,
which contains 58% (w/w) Jeffamine 02000 porogen in the framework
mesopores, has similar textural properties.

Both the as-made and the calcined forms of the mesostmcture were used
in forming rubbery epoxy mesocomposites. The mesocomposites formed from
as-made MSU-J have the Jeffamine 02000 curing agent initially partitioned
between the silica mesophase and the prepolymer resin. For the
mesocomposites formed from the porogen-free calcined form of MSU-J,
however, all the curing agent is contained initially in the bulk pre-polymer prior to

the curing reaction.

 

Intensi

 

 

 

 

 

 

Figure 3.1 Small angle powder X-ray diffraction patterns: (a) as—made MSU-J
mesostructure containing 58% (wlw) intercalated Jeffamine 02000; (b) epoxy-
silica mesocomposite (12 wt% silica) formed from as-made MSU-J; (c)
mesostructured silica recovered through calcination at 600°C of a

mesocomposite containing 12% (wlw) silica.

55

Figure 3.1 compares the small angle XRD patterns for the as-made MSU-
J mesostructure (curve a) with the pattern of the mesostructure dispersed at the
12% (w/w) level in a cured epoxy composite (curve b). The presence of a low
angle reflection, corresponding to a d-spacing of 6.0 nm, is indicative of the
average pore—pore correlation distance for a wormhole framework. The
wormhole framework structure is verified by the TEM shown in Figure 3.23. The
loss of diffraction intensity upon dispersion of the mesostructure in the cured
composite is attributable to contrast matching between the silica walls and cured
polymer formed in the mesopores. Evidence supporting the contrast matching
effect is provided by the increase in intensity observed upon removal of the
polymer through calcination (curve c). Analogous XRD scattering behavior was
observed for the calcined MSU-J—epoxy composite system. These XRD data,
along with the TEM image of Figure 3.2b of the recovered mesostructure, clearly
show that the wormhole framework structure of MSU-J is retained in a epoxy

matrix upon cross-linking at 125 °C.

Tensile data were obtained from dog-bone-shaped specimens containing
1.0-12 wt% as-made and calcined MSU-J silica. A comparison of stress-strain
curves (more accurately load-displacement curves because the load is not
normalized with respect to cross sectional area of the specimens) as a function
of silica loadings for the rubbery epoxy mesocomposites is provided in Figure 3.3.
Both forms of the mesostructure provide composites with improved mechanical
performance. The modulus, strength, elongation and toughness of the epoxy

composites prepared from the as-made and calcined forms of the silica generally

56

 

Figure 3.2 TEM images of (a) calcined MSU-J silica and (b) the silica obtained
through calcination at 600°C a cured epoxy—silica mesocomposite (12 wt%

silica) formed from as-made MSU-J silica.

57

 

 

 

 

 

 

 

 

10
(a) 12%
3-
J 7%
§e
3.: 5%
2 4‘
...-0
"’ 2%
i 1%
24
Pfisfine
o‘rl'rrl'l'T'r‘
O 5101520253035
Strain(%)
12
(b) 12%
10-
A87
g 7%
§6~
g 5%
49
1%2%
2- . .
Pnsune
0

o ' é 71'0'1'5'2'0'2'5'30
Strain (%)

Figure 3.3 Comparison of stress-strain curves for pristine epoxy and MSU-

Jlepoxy mesocomposites containing different loadings of (a) as-made MSU-J

silica and (b) calcined MSU-J silica.

58

a
m
(D

 

 

 

 

 

$12 a 5'
E; ii
In I?
g 9
'u 8- -4 3
o a
5 ‘9.
2 3'
ca E
s 4- -2 u
I- 9.
0 I Tfi ' I o

 

 

6 ' i ' ll 6 8 jib '13
Silica Loading (wlw %)

 

 

 

 

 

 

 

 

 

4c 1000
354
" I -300
:2: 3°“. 9'
x _, (a
g 25 -600 g-
._ .
-9 20- 3
‘5 ' 400:"1
.5 15- ' a
g 10‘ 8 7
(D ‘ Va
. -200
5..
0 I ' I ' I ' I I I m 0

o 2 4 6 8 1'0 ' 1E

Silica Loading (wlw %)
Figure 3.4 Loading dependence of the (a) tensile modulus (b) tensile strength
(c) strain-at-break and (d) toughness for epoxy mesocomposites prepared

from MSU-J silica. The solid and open data points are for the composites

prepared from as-made and calcined MSU-J silica, respectively.

59

increase with increasing silica loadings (Figure 3.4). Table 3.1 summarizes the
values of tensile properties obtained at different silica loadings. Note that the
composites formed from the as-made silica have a higher elongation at-break, so
that the toughness of the composites increases more rapidly with increasing
silica loading in comparison to the composites formed from calcined MSU-J.

The most significant difference in the properties of the mesocomposites
formed from as-made and calcined MSU-J silica is observed in the oxygen
permeability profiles of the composites. Figure 3.5 presents the O2 permeability
of the rubbery epoxy composites as a function of as-made and calcined MSU-J
silica loadings. The oxygen permeability for the nanocomposites prepared from
the calcined mesostructure is nearly independent of the silica content of the
epoxy matrix over the entire loading range 0—12% (wlw) silica. The as-made
MSU-J system, in contrast, shows behavior similar to the calcined silica
composites only at low silica loadings (1-3 wt%). An abrupt and significant
increase in oxygen permeability occurs at 5 wt% loading. Relative to the pristine
polymer, a 5- to 6-fold increase in oxygen permeability is observed over the silica

loading range 7-12% (wlw). Oxygen permeability increases of this magnitude are

unprecedented among polymer-based nanocomposites containing dispersed
porous particulates.

PALS results for the neat epoxy and two samples of the as-made MSU-J
silica composites (2% and 10% w/w) are quite definitive. The neat epoxy displays
a very typical 2.7 ns positronium lifetime (with a relative intensity of 20%) that

corresponds to 0.7 nm diameter voids inherent to the material. The two spectra

60

from the MSU-J silica composites are almost identical except that the 10% w/w
film has a relative intensity of the 2.7 ns lifetime of only 18%. The 10% drop in
relative intensity corresponds to the 10% of positrons that stop in the silica phase
which does not have a 2.7 ns lifetime. There is no indication of any longer

positronium lifetimes as would be expected if the silica pores where vacant (5 nm

pores would correspond to lifetimes in the 50-80 ns range). Special long time-

range PALS runs with high sensitivity to such long lifetimes were performed and
an upper limit of 0.15% relative intensity could be set on any such components
(i.e. less than 1% of the 20% positronium intensity measured in the neat epoxy).
Heating in vacuum to 150 °C to drive off any absorbed water from environmental
exposure yielded identical spectra. The PALS data are completely consistent
with no voids in the as-made MSU-J silica mesocomposites, other than those
microvoids inherent to the cured epoxy. The silica pores are fully occupied by
cured polymer.

Another significant result derived from the epoxy—MSU-J mesocomposite
systems relative to natural epoxy clay nanocomposites is the optical
transparency of the mesocomposites. Figure 3.6 illustrates that the
mesocomposite containing 10% (w/w) as-made MSU-J silica is substantially
more transparent than an ODA-MMT composite containing the same loading of
montmorillonite clay. The mesocomposites made from calcined MSU-J have a
similar transparency. Thus, in comparison to clay nanoparticles, the refraction

index of the MSU-J silica more nearly matches that of the epoxy matrix. The

61

 

.3
0|

 

J

(a)

A
c.D

0|
1 .

(b)

 

 

 

02 Permeability (cc mil m"2 clay'1 I 104)

O

E'iré‘éii'o'iz
Silica Loadings (wlw%)

C

Figure 3.5 Dependence of oxygen permeability on silica loadings for
epoxy mesocomposites prepared from (a) as-made MSU-J silica and (b)

calcined MSU-J silica.

62

(C)

MIC HIGAN ST

I N l v I R Sl

~ --a--.m-n~

 

Figure 3.6 Optical image illustrating differences in the transparency of (a)
pristine epoxy polymer and epoxy composites prepared from (b) as—made
MSU-J silica at 10% (wlw) loading and (c) octadecylammonium-modified

montmorillonite (ODA-MMT) at 10% (w/w) silicate loading.

63

thermal stability of the mesocomposites formed from as-made MSU-J is
enhanced in comparison to the pristine polymer.
As illustrated in Figure 3.7, the temperature needed to achieve a weight

loss of 1.5% is increased by 50 °C relative to the pristine epoxy polymer. Along

with the delayed first degradation stage, the maximum degradation rate is shifted
from 365 °C for the pristine epoxy to 396 °C for the mesocomposite containing
12% (wlw) as-made MSU-J silica, as determined from the first derivative of the
TGA curves. Additionally, the charring processes, which starts at approximately

400 °C in both cases, is extended up to 800 °C for the as—made MSU—J

composite at a 12% (w/w) loading.

3.4 Discussion

The XRD and TEM results shown in Figures 3.1 and 3.2 clearly establish
retention of the mesophase phase structure upon dispersal of as—made and
calcined MSU-J silica in a cured epoxy matrix. Little or no settling of the
mesophase was observed during the polymer curing process, because the
intensity of the Bragg reflections from the upper and lower surfaces of the cured
composite specimens were equivalent over the entire range of silica loadings.

The substantial improvements in the modulus, strength and toughness
achieved for the MSU-J mesocomposites in comparison to the pristine polymer
(c.f. Table 3.1) most likely are a consequence of strong interfacial interactions
and adhesion between the epoxy matrix and silica mesophase. It is important to

note that favorable interfacial interactions are realized in these mesocomposites

64

 

c»
‘1’

Weight (%)
h

 

 

 

 

o ' 260 l 460 I 600 800
Temperature (°C)

Figure 3.7 Thermogravimetric analysis curves in air for (a) the epoxy
mesocomposite prepared from as-made MSU-J silica at a loading of 12%

(wlw) silica and (b) the pristine epoxy polymer.

65

 

without the need for organic modification of the silica surface. The high surface
area of the MSU-J silica (~950 m2/g) facilitates such interfacial interactions.
Similar improvements in tensile moduli, strengths and strain-at-breaks have been
observed for rubbery epoxy—organoclay nanocomposites.3°‘32 The 4- to 5.5-fold
increase in modulus at 12% (w/w) MSU-J loading is comparable to the 6-fold
benefit in modulus provided by organoclay nanoparticles at an equivalent loading.
Moreover, the benefit in tensile properties occurs with little or no sacrifice in
transparency (c.f. Figure 3.6) or thermal stability properties (c.f. Figure 3.7).

In the case of exfoliated clay nanocomposites, the anisotropic clay
platelets in the cured matrix can be partially aligned in the direction of the stress
and further enhance the tensile properties.30 This strain—induced alignment of
particles is less significant for the more isotropic MSU-J silica particles.
Differences in particle morphology and alignment under applied stress may be
the reason why the elongation-at-break is limited to 34% for the mesocomposites
formed from as-made MSU-J, whereas for nanocomposites formed from
exfoliated clay platelets the elongation-at—break is 50% at the same 12% (wlw)
loading.30 In any case, unlike conventional reinforced composites which sacrifice
elasticity and toughness in exchange for benefits in modulus and strength,3‘°"34
mesostructured silica and exfoliated clay are unique in providing substantially
improved tensile properties at relatively low particulate loadings.

We note that previous attempts to use three-dimensional mesostructured
silica as a polymer reinforcing agent have not demonstrated mechanical

improvements comparable to those achieved with exfoliated organoclays. For

66

 

instance, Nylon 66 composites containing 35% (w/w) of a mesostructured silica
denoted FSM12 exhibited only a two-fold increase in modulus. A mesostructured
organosilica13 and a silylated cubic MCM-4814 mesostructure have been used as
reinforcing agents for poly((3-trimethoxysilyl)propylmethacrylate) and poly(vinyl
acetate), respectively. The latter composites exhibited improvements in tensile
strength and modulus, but the elongation at break was sacrificed despite efforts
to achieve compatible particle-polymer interfaces through extensive organic
modification of the silica surface.

The oxygen permeabilities of the mesocomposites formed from calcined
MSU-J silica is independent of the particulate loading up to 12% (wlw) loading,
as expected for a uniform dispersion of more or less isotropic particles in the
polymer matrix (c.f. Figure 3.5, curve (b)). However, the composite formed from
as-made MSU-J silica shows an enhancement in oxygen permeability at a silica
loading 2 5% (wlw) silica (c.f. Figure 3.5, curve (a)). Both composite systems
show no evidence for the presence of free volume, as judged by PALS. That is,
the mesopores of the embedded silica in both mesocomposites systems are fully
occupied by cured polymer. Thus, the enhancement in oxygen permeability for
the mesocomposites made from as-made MSU-J silica cannot be attributed to
diffusion through vacant pores in the mesostructured silica.

The main difference between the composites formed from as-made and
calcined MSU-J silica lies in the initial distribution of the amine curing agent at
the initial stages of the polymerization process. For as-made MSU-J, the amine is

partitioned between the mesopores and the bulk resin, whereas for the calcined

67

silica, the curing agent initially resides exclusively in the resin. The latter
distribution of curing agent should lead to a uniform chain cross-linking
throughout the polymer matrix. Although protonated amines act as an acid
catalyst for epoxy-amine polymerization in the gallery regions of clays,35 the
silica-intercalated amine in as-made MSU-J is electrically neutral36 and acid
catalysis is not anticipated. However, the partitioning of the curing agent between
the silica phase and the bulk pre-polymer may affect the rates for diffusion of
resin and curing agent into and out of the pores and lead to a non-uniform chain
cross-linking, particularly in the silica - polymer interface region.”39 As illustrated
in Figure 3.8, a decrease in chain cross-linking near the silica interface would
provide a “corona” or an “interface” of more permeable polymer around the silica
particles and provide a percolation pathway for enhanced oxygen permeation.
This proposed decrease in chain cross-linking density and concomitant increase
in polymer flexibility in the interface region may also help explain why the
mesocomposites prepared from as—made MSU-J silica are notably tougher than
those prepared from the calcined form of the mesostructure (c.f. Table 3.1).
Moreover, 29Si MAS NMR spectroscopy (not shown) indicates the framework
cross-linking to be lower in as - made MSU-J (Q3/Q4 = 0.89) than in calcined
MSU-J silica (Q3/Q4 = 0.36). A lower degree of framework cross-linking for as-
made MSU-J may provide the silica reinforcing agent with more flexibility, thus
increasing toughness at he expense of modulus.

We note that zeolite-loaded polymer membranes19'21'22 have been

reported wherein the oxygen permeability increases via a percolation pathway at

68

(a)

-—- 02000
......... Epoxy resin

 

 

Y
Bulk polymer

Polymer Corona
or Interface

.A/MSU-J silica

    

Low Loading Higher Loading
Reduced cross-linking density

Figure 3.8 Schematic representations of (a) the proposed decrease in chain
cross-linking density around the as-made MSU-J silica particles and (b) the
proposed permeant gas pathways below and above ~ 5% (wlw) loading of as-

made MSU-J silica.

69

a loading of ~30% (w/w).22 For such membranes the zeolite pores are too small
(< 1 nm) to be filled by polymer and the permeant traverses the 0.30 cm3/g void
volume of the zeolite at the percolation threshold. The small pore sizes and
hydrophilic surface cause poor interaction with the polymer chains, resulting in

40'“ The deteriorated mechanical

inadequate zeolite particle dispersion.
properties of zeolite-epoxy composites41 substantiate zeolite de-wetting by the
polymer chains. Due to the poor zeolite particle dispersion, non-selective zeolite
voids cannot provide facilitated diffusion pathway for entire membrane in the
diffusion direction until the zeolite loading reaches a very high value (~30 wt%).
In the present work, the pore size (5.3 nm) and pore volume (1.41 cm3/g) of the
inorganic phase are much larger than a zeolite, and consequently, the pore are
filled with polymer. The high level of as-made MSU-J particle dispersion and the
formation of a permeable polymer corona around the particles provides for a

permeation pathway at a relatively low particle loading of 5% (wlw). Modeling

efforts will be needed to better estimate the corona size.

3.5 Summary

As-made and calcined forms of large pore (5.3 nm) mesostructured silica
with a wormhole framework structure, denoted MSU-J, have been used to form
rubbery epoxy mesocomposites containing 1.0—12 % (wlw) silica. The tensile
modulus, strength, toughness, and extension-at-break for the mesocomposites
formed from as-made and calcined forms of MSU-J silica are systematically

reinforced by up to 4.8, 5.7, 1.6, and 8.5 times, respectively, in comparison to the

70

pure epoxy polymer. The composites represent the first examples wherein the
reinforcement benefits provided by mesostructured silica particles are
comparable to those provided by exfoliated organoclay nanolayers at equivalent
loadings. Moreover, the reinforcement benefits are realized without the need for
organic modification of the silica surface, and the increases in tensile properties
occur with little or no sacrifice in optical transparency or thermal stability.

The oxygen permeability of the mesocomposites prepared form as-made
MSU-J silica increases dramatically at loadings 2 5.0% (wlw), whereas the
compositions made from the calcined form of the mesostructure show no
permeation dependence on silica loading. For instance, the oxygen permeability
of the mesocomposites containing 12 % (wlw) as-made MSU-J silica is six-fold
higher than that of the silica-free epoxy membrane. Positron annihilation lifetime
spectroscopy established the absence of free volume in the mesocomposites,
thus precluding the possibility of facile oxygen diffusion through the framework
pores of the silica. The increase in oxygen permeability is correlated with of the
partitioning of curing agent between the as-made mesostructure and the liquid
pre-polymer, which leads to coronas of permeable polymer with reduced chain
cross-linking in the vicinity of the silica particles. Mesocomposites made from
calcined forms of the mesostructured silica do not allow for curing agent
partitioning, and the oxygen permeability is not significantly influenced by the

silica loading.

71

Axum; v6 $0. 3995 m m>oEom B venom: mLBmLmQEB. m

 

 

 

 

ON mmm mdN Nvd 5:3. NV

o.m or

0N me wéw MNN Ndp N

9N mmN mdw 00;. hmd m

YN m

9N hm? Qmm woé N. —..m N mOszQEOOmeS—
QN m: QON wad mod —. 73w.)— “COED—NO
VF ONm Foo wdm F¢.m Q: NP

NF or

N? NNV ...Nm MON .vwd N.

mud 5mm Ndm NON VNd m

NN m

QN mm... QVN ov._. $0.0 N mmtmOQEOOOwOE
VN wa ...NN o F .—. V56 _. .3392 mUmEIm<
3 mom i F. K cod 8N o 8520c 2:95

6.» as); Acczv 3.33
A ot.->mc.~mE.___m._»oov cbzfiflm FIE we: 3% : Smcmbm wasnoE mcfimo. oEEmw
___nmoEm :9 x0 .mEoE. 39:: so... cam :05 26:9... 0:29... 8:5

 

.mozmanooOmoE

5-392388 new 55:8 583 mczmta ho mmquEa b=5moEma 596 new 3.39m .mEmfi 6:89... ...m 63¢...

72

3.6 References
(1) Novak, B. M. Adv. Mater. 1993, 5, 422.
(2) Giannelis, E. P. Adv. Mater. 1996, 8, 29.

(3) Kojima, Y.; Usuki, A.; Kawasumi, M.; Okada, A.; Fukushima, Y.; Kurauchi, T.;
Kamigaito, O. J. Mater. Res. 1993, 8, 1185.

(4) Yano, K.; Usuki, A.; Okada, A.; Kurauchi, T.; Kamigaito, O. J. Polym. Sci.
Polym. Chem. 1993, 31, 2493.

(5) LeBaron, P. C.; Wang, Z.; Pinnavaia, T. J. Appl. Clay Sci. 1999, 15, 11.

(6) Kresge, C. T.; Leonowicz, M. E.; Roth, W. J.; Vartuli, J. 0.; Beck, J. S. Nature
1992, 359, 710.

(7) Shi, J. L.; Hua, Z. L.; Zhang, L. X. J. Mater. Chem. 2004, 14, 795.

(8) Lu, Y. F.; Yang, Y.; Sellinger, A.; Lu, M. C.; Huang, J. M.; Fan, H. Y.; Haddad,
R.; Lopez, G.; Burns, A. R.; Sasaki, D. Y.; Shelnutt, J.; Brinker, C. J. Nature 2001,
411, 617.

(9) Nguyen, T. 0.; Wu, J. J.; Doan, V.; Schwartz, B. J.; Tolbert, S. H. Science
2000, 288, 652.

(10) Molenkamp, W. C.; Watanabe, M.; Miyata, H.; Tolbert, S. H. J. Am. Chem.
Soc. 2004, 126, 4476.

(11) Li, G. T.; Bhosale, S.; Wang, T. Y.; Zhang, Y.; Zhu, H. S.; Fuhrhop, K. H.
Angew. Chem. Int. Ed. 2003, 42, 3818.

(12) Kojima, Y.; Matsuoka, T.; Takahashi, H. J. Appl. Polym. Sci. 1999, 74, 3254.

(13) Ji, X. L.; Hampsey, J. E.; Hu, O. Y.; He, J. B.; Yang, 2. 2.; Lu, Y. F. Chem.
Mater. 2003, 15, 3656.

(14) He, J.; Shen, Y. B.; Yang, J.; Evans, 0. G.; Duan, X. Chem. Mater. 2003, 15,
3894.

(15) Reid, B. 0.; Ruiz-Trevino, A.; Musselman, l. H.; Balkus, K. J.; Ferraris, J. P.
Chem. Mater. 2001, 13, 2366.

(16) Pinnavaia, T. J.; Beall, G. W. Polymer-Clay Nanocomposites. ed.; John
Wiley and Sons: 2001.

73

(17) Maier, G. Angew. Chem. Int. Ed. 1998, 37, 2961.

(18) Kulprathipanja, S.; Neuzil, R. W.; Li, N. N. Separation of Fluids by Means of
Mixed Matrix Membranes. US Patent 4,740,219, 1988.

(19) Jia, M. 0.; Peinemann, K. V.; Behling, R. D. J. Membrane Sci. 1991, 57, 289.

(20) Duval, J. M.; Folkers, B.; Mulder, M. H. V.; Desgrandchamps, G.; Smolders,
C. A. J. Membrane Sci. 1993, 80, 189.

(21) Gur, T. M. J. Membrane Sci. 1994, 93, 283.
(22) Suer, M. G.; Bac, N.; Yilmaz, L. J. Membrane Sci. 1994, 91, 77.

(23) Balkus, K. J.; Cattanach, K.; Musselman, l. H.; Ferraris, J. P. Mater. Res.
Soc. Symp. Proc. 2003, 752, 91.

(24) Park, L; Wang, Z.; Pinnavaia, T. J. Chem. Mater. 2005, 17, 383.

(25) Triantafillidis, C. S.; LeBaron, P. C.; Pinnavaia, T. J. Chem. Mater. 2002, 14,
4088.

(26) Shan, L.; Verghese, K. N. E.; Robertson, C. G.; Reifsnider, K. L. J. Polym.
Sci. Pol. Phys. 1999, 37, 2815.

(27) Liang, W. J.; Kao, H. M.; Kuo, P. L. Macromol. Chem. Physic. 2004, 205,
600.

(28) Lee, J.; Yoon, 8.; Oh, S. M.; Shin, C. H.; Hyeon, T. Adv. Mater. 2000, 12,
359.

(29) Lee, J.; Han, S.; Hyeon, T. J. Mater. Chem. 2004, 14,478.

(30) Lan, T.; Pinnavaia, T. J. Chem. Mater. 1994, 6, 2216.

(31) Shi, H. 2.; Lan, T.; Pinnavaia, T. J. Chem. Mater. 1996, 8, 1584.

(32) Wang, Z.; Pinnavaia, T. J. Chem. Mater. 1998, 10, 1820.

(33) LeBaron, P. C.; Pinnavaia, T. J. Chem. Mater. 2001, 13, 3760.

(34) Ahmad, Z.; Sarwar, M. l.; Wang, 8.; Mark, J. E. Polymer 1997, 38,4523.

(35) Lan, T.; Kaviratna, P. 0.; Pinnavaia, T. J. Chem. Mater. 1995, 7, 2144.

74

(36) Only 4 mol% of NH2 groups of the 02000 molecules are protonated in the
aqueous solution before the silica source is added. It is caculated using pr (~4)
and the initial concentration (0.06 M) of the amine.

(37) Wind, J. 0.; Staudt—Bickel, C.; Paul, D. R.; Koros, W. J. Macromolecules
2003, 36, 1882.

(38) Chung, T. S.; Chng, M. L.; Pramoda, K. P.; Xiao, Y. C. Langmuir 2004, 20,
2966.

(39) Robinson, J. P.; Tarleton, E. S.; Ebert, K.; Millington, C. R.; Nijmeijer, A. Ind
Eng Chem Res 2005, 44, 3238.

(40) Zimmerman, C. M.; Singh, A.; Koros, W. J. J. Membrane Sci. 1997, 137, 145.

(41) Merkel, T. C.; Freeman, B. 0.; Spontak, R. J.; He, 2.; Pinnau, l.; Meakin, P.;
Hill, A. J. Chem. Mater. 2003, 15, 109.

75

CHAPTER 4

SILICA FOAM REINFORCED THERMOSET EPOXY MESOCOMPOSITES

4.1 Introduction

Thermoset epoxy polymers are used in a variety of applications, including
coatings, adhesives and electronics due to their high chemical resistance, good
adhesion, attractive optical properties and flexibility.“ 2 Layered silicate clays
have been widely investigated as reinforcing agents to improve mechanical,3' 4
thermal,“ optical,” 8 and barrier9 properties for thermoset epoxy polymers.
Complete dispersion (exfoliated) of individual clay nanolayers in the polymer
matrix generally provides greater phase homogeneity, resulting in a better
enhancement of properties in comparison to composites formed from intercalated
clay nanoparticles.8'10

To achieve complete exfoliation of clays11 organically modified clays
normally are used to import a hydrophobic environment to the clay layer surfaces
that are normally hydrophilic in nature. However, only intercalated
nanocomposites are obtained in many cases, despite the organic modification of
clay surfaces.2 Occasionally, delamination of clay layers is achieved under
severe processing conditions (high temperature12 or high pressure”) through the
use of auxiliary solvents14 or by processing under supercritical carbon dioxide”.

The absence of X-ray reflections in the low 20 range (> 10 nm layer separation)

are often interpreted as an evidence for exfoliated clay morphology.16 However,

76

disordered clay layers or aggregates of aligned clay tactoids separated on a
micron-scale by polymer voids are often observed in TEM images, even though
no XRD reflections are detected in the low 28 range of the XRD pattern of the
composite.16

Mesostructured silica materials with one- or multiple-dimensional pore
structures (e.g. MOM-41, SBA-15) have recently been used for the preparation of
polymer-based nanocomposites exhibiting improved performances in diode",
Li"-battery18, luminescence”, and conductivity20 applications. Relatively small

17, 18 or p0lymerizable structure-directing agents1921 have typically

monomers
been used to prepare (semi)conducting polymers for these applications. However,
only a few attempts have been made for improving the mechanical properties of
engineering polymers.”25 For instance, nylon 66 composites22 prepared under
high pressure (0.5 - 30 MPa) exhibited only two-fold increase in storage modulus
at a 35 wt% loading of FSM (MCM-41) silica. Silylated MOM-4123 and MGM-4824
organosilicas were used as reinforcing agents for poly((3-trimethoxysilyl)
propylmethacrylate and poly(vinyl acetate), respectively. In both cases tensile
modulus and strength increased but elongation-at-break was sacrificed, so that
no or a little improvement in toughness was observed. Recently polypropylene-
fllled MCM-41 composites prepared with aid of supercritical CO2 have been
reported.25 However, no clear evidence for polymer filling in the mesostructure

pore and only a few percents of reinforcement were observed in tensile

properties.

77

The filling of small mesopore silica by polymer chains sequences forcing
conditions such as high pressure”, silyation of the pore wallsZ3' 24, or supercritical
CO2 as a solvent”. For instance, hexagonal MCM-41 mesostructured silica (pore
size ~3 nm) with no organic modifier afforded brittle polyimide composites at high
silica loading (>10 wt%),26' 27 indicating poor interaction between the polymer
chain and the internal surface area of the silica pores. Large-pore
mesostructures are more likely to undergo pore filling by polymer chains without
using organic modifiers or complicated processing conditions. Moreover, large-
pore mesoporous silicas with high framework pore volume should provide
polymer nanocomposites with improved homogeneity due to the low intrinsic
density of the silica particles.

Recently, we reported epoxy mesocomposites reinforced by MSU-J silica
with a wormhole framework stmcture. Mechanical properties of the composites
were comparable to exfoliated clay-epoxy nanocomposites.28 The as-made and
calcined MSU-J silica particles (5.3 nm pore size) dispersed readily in the
thermoset epoxy polymer. The mechanical properties of the resulting
mesocomposites were reinforced at levels corresponding to 4.8-, 5.7-, 1.6-, and
8.5-fold increases in tensile modulus, strength, elongation and toughness,
respectively, compared to pristine epoxy polymer. These improvements in tensile
properties were realized despite the relatively isotropic particle morphology of the
silica particles. The pore size of MSU-J silica is comparable to the dam spacings

found for intercalated clays in many polymer-clay nanocomposite systems.

78

Mesocellular silica foams (MCFs) have been synthesized from TEOS in
the presence of a triblock copolymer surfactant (Pluronic P123) and 1,3,5-
trimethylbenzene (T MB) as a cosurfactant in aqueous acid.29 Nitrogen sorption,
XRD, and TEM analyses confirmed that the MCF products possess cellular pores
connected by windows. Subsequent studies” showed that increasing the content
of TMB leads to microemulsion droplets that form spherical pore structures.
Mesocellular foam structures exhibit very large cell sizes of 25—35 nm, window
sizes of 7—18 nm and high pore volumes up to 2.4 cm3/g. Another family of
mesocellular silica foam stmctures, denoted MSU-F, have been prepared from
cost-effective sodium silicate though electrically neutral assembly pathways.31
This synthetic approach to MSU-F silica foams generates open cell foam
structures with a relatively low cell to window size ratio (~1.5) after hydrothermal
treatment at 100 °C for >8 h unlike closed cell structures of MCFs formed at
similar post-synthetic conditions.”

In the present study a calcined mesocelluar silica foam (MSU-F) is used
as a reinforcing agent for a thermoset rubbery epoxy polymer. The MSU-F silica
foam exhibits very large pore size (cell size; 27 nm and window size; 15 nm)
comparable to clay layer separation of >10 nm in exfoliated clay-polymer
nanocomposites. The relatively high pore volume of the silica (2.2 cm3/g) also
provides a low intrinsic density of silica particles, promoting the formation of
homogeneous nanocomposites. The tensile modulus, strength, elongation, and
toughness of the resulting rubbery epoxy-MSU—F mesocomposites systematically

are enhanced up to 3.7, 6.8, 2.2, and 20.6 times at relatively low silica loading (9

79

wt%) compared to silica-free epoxy polymer. Although the modulus enhancement
is lower than the enhancement provided by epoxy-MSU-J mesocomposites, the
fractural properties of the MSU-F counterparts are more improved due to a more
uniform distribution of silica in the matrix. Moreover, no or little loss of optical
transparency and thermal stability was observed for the epoxy-MSU-F

mesocomposites.

4.2 Experimental

Epoxy resin EPON 828, a diglycidyl ether of bisphenol (Resolution
Performance Products), was used for the preparation of epoxy—MSU-F silica
mesocomposites. The a,w-diamine—terminated polypropylene oxide
H2NCH(CH3)CH2[OCH2CH(CH3)]XNH2 (x = ~33.1) with an average molecular
weight of ~2000 was obtained from Huntsman Chemical Co. under the trade
name Jeffamine 02000. For the synthesis of MSU-F silica foam, 3 triblock
copolymer (Pluronic P123, EO2oPO70EO2o, BASF) was used as a surfactant.
Sodium silicate (27% SiO2, 14% NaOH solution) and 1,3,5-trimethylbenzene
(T MB) were purchased from Aldrich and used as a silica precursor and a pore

expander, respectively, for the preparation of MSU-F mesostructured silica foam.

0
/ \
CH2—CH-CH2-O WOO-O ”CH2 CH- CH2O 000: 00 CH2- CH- O\CH2

CH3 CH3
EPON 828; II: 0(88 %) n1 (1o %) 2(2%). (4.1)

A large pore mesostructured silica foam, denoted MSU-F, was assembled

from sodium silicate and P123 as the structure-directing porogen according to

80

previously described methods.32 The P123 porogen was mixed with an amount of
aqueous acetic acid solution equivalent to the NaOH content of the silicate
source and stirred for 2 h. TMB was added to the porogen solution and the
mixture was stirred for an additional hour. The silica source was then added to
the porogen solution containing the pore expander (T MB) under vigorous stirring
at ambient temperature and the mixture was allowed to age at 25°C for 24 h. The
reaction mixture was then heated at 100°C for 24 h under static reaction
conditions. The porogen-intercalated as-made mesostructured product was
recovered by filtration and dried in air at ambient temperature. The molar
composition for the formation of as-made MSU-F was 1.0 SiO2 : 0.78 NaOH :
0.017 P123 : 0.83 CHaCOOH : 0.69 TMB : 230 H2O. A porogen—free analog of
the mesostructure, denoted calcined MSU-F, was obtained by calcination of as-
made MSU-F at 600°C for 4 h. The calcined form of the mesostructure was
ground to a powder for future use.

For the preparation of the rubbery epoxy polymer nanocomposites, a pre-
determined amount of calcined MSU-F silica foam was added to the epoxy resin
(EPON 828) and mixed at 50°C for 10 min. A stoichiometric amount of Jeffamine
D2000 curing agent was then added to the mixture and mixed at 50°C for
another 10 min. The resulting suspensions were out-gassed under vacuum and
transferred to an aluminum mold. Pre-curing of the mesocomposite was carried
out under nitrogen gas flow at 75°C for 3 h, followed by an additional 3 h cure at

125°C to complete the crosslinking.

81

N2 adsorption-desorption isotherms were obtained at 77 K on an ASAP
2010 volumetric adsorption analyzer. Calcined MSU-F silicas were outgassed at
150°C overnight. Pore size distributions were determined by applying the BdB-
FHH2 model33 to the adsorption and desorption isotherms.

Transmission electron microscopy (TEM) images were taken on a JEOL
JEM-100CX ll microscope with a CeBe filament and an accelerating voltage of
120 kV. Sample grids of calcined mesoporous silicas were prepared via
sonication of a powdered sample in ethanol for 10 min and evaporating 2 drops
of the suspension onto a holey carbon film supported on a 3 mm, 300 mesh
coppergnd.

Tensile measurements on individually molded samples were performed at
ambient temperature according to ASTM standard D3039 using an SFM-20
United Testing System. The dog-bone shaped specimens used in the tensile
testing were 28 mm long in the narrow region, 3 mm thick, and 3 mm wide along
the center of the casting. At least six stress-strain data points with good precision
were obtained and the results averaged. The errors in tensile data are based on
the 95 % confidence interval for the mean values.

Thermogravimetric analyses (TGA) were performed using a computer
controlled Shimadzu TGA-50 thermal analyzer. Typically, a 50 mg sample was
placed in a quartz bucket and heated an ambient temperature to 800°C at a rate
of 5 °C/min in a nitrogen flow of 50 mL/min.

Refractive indices of calcined MSU-F silica foam and pristine rubbery

epoxy sample were determined using liquid testing methods. Carbon

82

tetrachloride (n = 1.46, Aldrich, 99.8%), benzene (n = 1.50, Baker, 99.9%) and
benzyl alcohol (n = 1.54, Fisher Scientific Co.) were used as references. A 0.2 9
quantity of calcined MSU-F silica was added to 20 g of solvent. A pristine rubbery
epoxy sample (dog-bone sample used for tensile tests) was scratched with sand
paper to generate an opaque surface and then immersed in each solvent to

judge refractive index matching.

4.3 Results

Figure 4.1 presents the nitrogen adsorption-desorption isotherm and the
pore-size distribution for calcined MSU-F silica foam assembled at ambient
temperature for 24 h and then subjected to hydrothermal treatment at 100 °C for
additional 24 h. The isotherm curve and the relatively narrow hysteresis loop in
good agreement is consistent with a typical open cell foam structure.32 The pore
volume and BET surface area are 2.2 cm3/g, and 540 m2/g, respectively, and the
cell size (26.5 nm) and window size (14.9 nm) were judged using the BdB-FHH2

model.33

TEM images in Figure 4.2 verify the spherical pore structure and the cell
size matching the nitrogen sorption isotherm.

Tensile data were obtained from dog—bone-shaped specimens containing
1.0—9 wt% calcined MSU-F silicas. A comparison of stress-strain curves (more
accurately load-displacement curves because the load is not normalized with
respect to cross sectional area of the specimens) as a function of silica loadings
for the rubbery epoxy mesocomposites is provided in Figure 4.3. The

mesostructure provides composites with improved mechanical performance. The

modulus, strength, elongation and toughness of the epoxy composites prepared

83

 

 

1600
(a) 3 (b)

j

A
h
C
O
1

 

 

 

N2 volume adsorbed (cm 3/g, STP)
a
O
O

 

 

 

 

 

 

0 l ' I I I W I I j I
0 0.2 0.4 0.6 0.8 1 0 20 40 60 80 100
P/Po Cell diameter (nm)

Figure 4.1 (a) Nitrogen adsorption-desorption isotherm for calcined MSU-F1
silica foam prepared from postsynthesis hydrothermal treatment of the as-
synthesized MSU-F for 24 h at 100 °C. (b) Corresponding cell size (solid line)
and window size (dash line) distributions obtained from adsorption and
desorption branches, respectively. The distributions were determined by

applying the BdB-FHH2 model.”

84

 

.a

Figure 4.2 Transmission electron microscope (TEM) images of a calcined
MSU-F silica foam synthesized through postsynthesis treatment at 100 °C for
24 h. Images (a) and (b) are taken with high and low magnifications,

respectively.

85

 

 

 

 

9%
1o-
3. 7%
E
9, 6-
g 5%
a)
4-
3%
2- 1%
0%( ...... )
o '1'0'210'3'0'4'0'50
Strain(%)

Figure 4.3 Comparison of stress-strain curves for pristine epoxy ( """ ) and

MSU-F-epoxy mesocomposites containing different silica loadings.

86

 

 

 

 

 

 

 

 

     

 

 

    

 

 

12 6
10 - 5
g 5'
E a . 4 3.
g 25'
5 8
1: s - 3 10
o 3
E . 1g
2 :r
-- 4 1 2 A
In
I: . g
q; '0
I- - e.
2 - 1
o n l . l . l . I . o
0 2 4 6 8 10
Silica Loading (wt %)
60 2500
.l
50 "
/,___——0\{ .. 2000
A -1
:1 4o .. .. 1 g
3 21500 :-
- 8
g 30 (n
9 4 If.
g 1000 g
I: 2° . a
(D van
-I 500
10
o . . l . L . l . o
0 2 4 6 8 10

Silica Loading (wt %)

Figure 4.4 Loading dependence of the (a) tensile moduli, (b) tensile strengths,
(6) strain-at-break and (d) toughness for rubbery epoxy mesocomposites

prepared from calcined MSU-F silica.

87

from the calcined MSU-F silica generally increase with increasing silica loadings
(Figure 4.4). The values are summarized in Table 4.1 at different silica loadings.
All the tensile properties of epoxy-MSU-F mesocomposites generally increase
with increase in silica loadings. Although the elongation shows little dependence
on high silica loadings, the toughness is increased in proportion to the silica
loading, because toughness is a combined property of strength and elongation.
In addition to the remarkable improvement in the mechanical properties
another significant benefits derived from MSU-F silica foam is the retention of
optical transparency and thermal stability compared to pristine epoxy polymer.
Optical photographs of pristine epoxy polymer and epoxy-MSU-F
mesocomposite samples containing 5 and 9 wt% silica are shown in Figure 4.5
demonstrating the refraction index of the MSU-F silica nearly matches that of the
epoxy matrix. Liquid emersion testing methods were carried out to determine
refractive indices for the pristine epoxy polymer and calcined MSU-F silica foam.
The white color of the silica disappears in CCI4 solvent (n = 1.46) indicating the
refractive index of the silica foam is nearly 1.45 as predictable.” Figure 4.6
shows that MSU-F silica foam does not promote nor hinder thermal degradation
of the epoxy polymer. The temperatures at the maximum degradation rate, as
determined by the first derivatives of the corresponding TGA curves, are 362 °C

for both samples.

88

 

 

.mmZDO
0... 96:88:50 65 ho 0265th 6:... 05 :5: 6055.20: 925 mum: :ozmnmaov E:E_me Hm mQBSmaEE 6:... a
mom 8: Ba no.6 o. 5 a
9.: a? New 8.: N
8 609888
68 New id 56 m cmawELxcam:
48 e? m: 68 m
2: 6.8 go mad e
«on E e. K 8.0 8N o 6623. 656:6
.80: €53 36 Res; 586:... Res: 8382 case easem

3:636 REEL... 6665330... 83356

26:6...

26:2.

9.68. 8:6

 

62.6an80on 0-392168% 0:6 ..mEboa 66:0 626:: So 622.803 .mEcofi ncm 26:8. ...v 236...

89

 

Figure 4.5 Optical images illustrating differences in the transparency of (a)
pristine rubbery epoxy polymer and epoxy composites prepared from calcined

MSU-F silica at (b) 5 and(c) 10 wt%.

90

 

Weight (%)

 

 

 

 

0 200 400 600 800
Temperature (°C)

Figure 4.6 Thermogravimetric analysis curves in nitrogen for the silica-free
epoxy polymer (open circles) and the epoxy mesocomposites (solid circles)

prepared from calcined MSU-F silica at a loading of 9 wt%.

91

4.4 Discussion

A mesocellular silica foam (MSU-F) assembled under a near neutral pH
conditions is a good candidate as a polymer reinforcing agent due to its very
large pore size and pore volume. The large pore >10 nm of calcined MSU-F silica
is comparable to the average separation of exfoliated clay nanolayers in polymer-
clay nanocomposites. The improved mechanical properties for the resulting
mesocomposites are realized without need of organic modifiers. Calcined MSU-F
silica is readily dispersed in the polymer matrix and no observable particle
aggregates are found in the cured mesocomposites at 1-9 wt % silica loadings.
Previously reported epoxy—MSU-J mesocomposites exhibit nearly identical XRD
patterns between top and bottom of dog-bone specimens demonstrating little or
no settling of the mesophase during curing process.28 Positronium annihilation
lifetime spectroscopic (PALS) analysis substantiated the absence of voids in the
epoxy-MSU-J mesocomposites representing filled MSU-J silica mesopores with
epoxy polymer chains. Although the degree of dispersion of MSU-F silica
particles in a polymer matrix is not sensitive to the XRD technique owing to the
absence of Bragg reflections, the mesocelluar pores are certainly filled due to
large pore size. Retention of the silica foam structure in the epoxy polymer matrix
is confirmed by the nearly identical nitrogen sorption35 properties (not shown) of
the recovered MSU-F silica obtained by calcinating 9 wt% epoxy-MSU-F
mesocomposite sample. The isotherm of the recovered silica was nearly identical
to pristine MSU-F silica (c.f. Figure 4.1).

Good particle dispersion and the filling of the mesostructured pores with

92

polymer are required to maximize the interfacial interactions between the
polymer chain and the silica surface for reinforcing resulting mesocomposites.
Previously reported attempts to form nanocomposites using MCM-41 type silicas
with one—dimensional pore structure sacrificed elasticity and toughness. In this
respect, the composites behaved like conventional reinforced composites in
exchange for benefits in modulus and strength.” Calcined HMS (pore size 2.8
nm) and MOM—48 (pore size 3.2 nm) silicas with three-dimensional pore network
have been used as reinforcing agents for thermoset epoxy polymers.37 However
the unmodified silicas presented low level of dispersion in the polymer matrix.
Thus the dispersion of silica was not uniform composite materials. It is important
to note that large pore mesostructures are highly desired for high quality of
reinforced polymer nanocomposites without using organic modifier or auxiliary
processing steps to obtain cost- and labor-efficiency.

Significant improvements in mechanical properties have been obtained for
the epoxy-MSU-F mesocomposites (c.f. Figure 4.4 and of Table 4.1) compared
to silica-free rubbery epoxy polymer. The 3.7-fold increase in tensile modulus for
the mesocomposites at 9 wt% is comparable to that of MSU-J (~4-fold increase)
counterpart at the same loading. The higher surface area of MSU-J mesoporous
silica (~950 m2/g) facilitates interfacial interaction between the polymer and the
silica surface providing slightly higher modulus at the beginning stage in stress-
strain behavior. However, the fracture properties, tensile strength and elongation,
of the epoxy-MSU-F mesocomposites are superior to MSU-J composites.

Moreover the 6.8-fold increase in tensile strength of MSU-F composites is even

93

higher relative to exfoliated organo-montmorillonite nanocomposites (5.3-fold
increase)” at same silica loading (9 wt%). The highly improved tensile strength is
presumably provided by high homogeneity resulting from a high pore volume (2.2
cm3/g) of the mesocelluar structure. MSU-F silica with lower intrinsic density
compared to MSU-J particles, occupy a larger volume of the mesocomposites
leading to reduced distances between silica particles. It is also noteworthy that
the epoxy-MSU-F mesocomposites exhibit ~50% elongation at 5—9 wt% silica
loadings, despite the isotropic particle morphology. In comparison a more limited
elongation (~30% at 12 wt% silica loading) was obtained for epoxy-MSU-J
mesocomposites. The enhancement in strain-at-break for MSU-F composites is
comparable to 40—60 % elongation provided by anisotropic exfoliated nanolayers
aligning in the direction of applied stress.” Thus, the high pore volume and more
uniform dispersion of silica throughout the entire polymer matrix results in
improved fractural properties for epoxy-MSU-F mesocomposites.

The modulus of amorphous silica (~10 GPa) is much larger than the
modulus of the epoxy matrix (~3 MPa). The simple rule of mixture theory (eq. 4-
1) does not fit the observed moduli of the mesocomposites (see dash-dot line in
Figure 4.7). In many cases, a super-linear dependence of the Young’s modulus

EC is observed as a function of the particulate volume fraction vp.”' 4°

E =E (1-v)+Ev (4-1)
6 m p p p

, where Em and E,, are the moduli for the matrix and the particulate, respectively.

According to Elsheby’s prediction” the bulk modulus of a composite (Kc) and the

94

shear modulus of a composite (60) at a volume fraction of particulate are given

 

 

 

by:
K —K
K: Km ,A= P "' ,a=1+v"' (4-2)
l—Avp (KP—Km)a+Km 3(l—vm)
G —G _
Gm p ... fl=£4 5v," (4_3)

 

 

 

“ z l—va’ z (GP—Gm)fl+Gm’ 15 l-v
, where K... and K,, are the bulk modulus values for the matrix and the particulate,
respectively, and G... and 6,, are the shear modulus values of the matrix and the
particulate and v... is the Poisson ratio for the matrix. The bulk modulus is the
inverse of compressibility. Young’s modulus (E) for an isotropic, linearly elastic
composite containing spherical particulates is related to K, G, v through the
equations below, but we need Poisson ratios of the rubbery epoxy polymer and

amorphous silica to compute Km, K,,, G... and Gp.

 

 

 

9KG E K E _ 3K—2G

E = ,G = ’ V (4_3)41
(3K + G) 2(1+v)

3(1—2v)’ 'm
We know the approximate Poisson ratio for amorphous silica (~0.2)42 and a
rubber has a value near 0.5,41 the maximum value of Poisson’s ratio. As seen in
equations (4-2) and (4-3), K and G are independent of the Poisson ratio for the
particulate.

If the Poisson’s ratio of the polymer is taken to be 0.5, Km approaches
infinity and G... will be ~Em/3 (namely, Em ~3Gm). Accordingly, the difference in
predicted Young's modulus (EC) for the polymer at vm =0.5 (actually 0.49999) and
at 0.45 is 20 % (10 MPa vs. 8 MPa), where the Young’s modulus for the

mesoporous silica is ~10 GPa.43' 44 In Figure 4.7, the experimental Young’s

95

N
O

 

 

 

 

I

I

E
E1 |
E 1
CD
2
3 I
'5 O
O 10 ,1
E . [:1 'l'
.9
6 0 D
C C .-'
m . 0"—
I- l-""

I In

0 I l I r I

02'4'6 8'10‘12
Silica loading (wt%)
Figure 4.7 Experimental (symbols) and computed (lines) values of Young’s
modulus as a function of silica loadings. Solid circles: composites made from
calcined MSU-J and open squares: composites made from calcined MSU-F.
Dot line: Computed dependence taking only the silica wall volume as the
particulate volume fraction; dash line: assuming the particulate volume to be
the silica wall volume plus the pore volume of MSU-J; solid line: assuming the

particulate volume to be the silica wall volume plus the pore volume of MSU-F.

The prediction (dot-dash line) based on a simple rule of mixture is also plotted.

96

 

moduli for composites prepared from as-made MSU-J, calcined MSU-J and
calcined MSU-F are indicated with symbols. The dependence computed from
equations (4-2 - 4-4) assuming only the wall volume of the silica alone as the
particulate volume (dot line) provides a poor prediction of the experimental data.
A better correspondence to the data is obtained based on the filling factor of the
particulate to be the combined volume of the silica walls and the pore volume of
the mesostructure. Thus, it is clear that the entire volume of the mesostructured
silica particle (walls plus pores) contribute to the reinforcement. Moreover, the
Eshelby model predicts the silica with the lower structural density to provide more
reinforcement (see the predicted curves for MSU-F and MSU-J with foam and
wormhole structure, respectively). However, the denser MSU-J wormhole
structure is observed to provide more reinforcement than the foam structure. Also,
the experimental moduli are larger than the predicted values based on the
Eshelby model. This latter model assumes the reinforcing particles to be solid
rigid spheres. The mesostructured silicas used as reinforcement agents in the
present study provide a substantially larger surface area than expected on the
basis of solid spherical particles. Accordingly, the additional interfacial
interactions between the particulate and the polymer matrix should lead to
greater stiffness, in agreement with the higher moduli found for composites made
from MSU-J (950 m2/g) in comparison to MSU-F (540 mzlg). Moreover, the
Eshelby prediction values are higher than other advanced models (see examples
in Reference 40) indicating that the interfacial interactions must be taken into

account in these mesocomposites. However a molecular dynamic simulation

97

study for the dependence of Young’s modulus on particle size may be required
as smaller particulates generally provide an increased composite modulus
compared to larger particlulates.40

Liquid test results show the approximate refractive index of the rubbery
epoxy polymer is ~1.50, nearly matching that of benzene. Both calcined MSU-J
and MSU-F mesostructures possess identical refractive index (n ~ 1.45), as
confirmed by the disappearance of silica in CCI4 solvent. This result indicates that
the transparency of the silica particles is dependent not on the porosity but on
refractive index of the silica wall. A glassy epoxy polymer45 prepared form EPON
826 and Jeffamine 230 exhibits ~1.55 refractive index so that glassy epoxy
nanocomposites containing calined MSU-F silica lose their transparency at >5
wt% silica loading due to a mismatch of refractive indices between the polymer
and the silica.37 However, the epoxy-MSU-F mesocomposites show more
transparency compared to the epoxy-MSU-J mesocomposites at the same silica
loadings. Consequently, the large separation of the silica walls allows complete
pore filling by polymer and this leads to improved transparency for the epoxy-
MSU-F mesocomposites in comparison to MSU-J and montmorillonite
composites. The transparent composite samples obtained from exfoliated clay
nanolayers as approved to transparent clay analogues”, can be explained on the
basis of similar consideration. Moreover, the high transparency at 9 wt% silica
loading also substantiates the complete dispersion of MSU-F silica particles in

the entire epoxy matrix.

98

4.5 Summary

Mechanically improved thermoset epoxy mesocomposites have been
prepared from calcined silica foam, MSU-F which exhibits surface area of 540
m2/g, ,cell size of 26.5 nm (14.9 nm window size) and pore volume of 2.2 cm3/g.
Despite the absence of a surface organic modifier, the calcined MSU-F silica
foam is well dispersed in the epoxy matrix due to its very large pore diameter and
low intrinsic density. The cellular foam structure is retained in the polymer matrix
as judged by nitrogen sorption isotherm analysis of the residue silica obtained by
calcining the 9 wt% mesocomposite.

The tensile modulus, the tensile strength, the strain-at-break, and the
toughness for the resulting rubbery epoxy mesocomposites are systematically
enhanced up to 3.7, 6.8, 2.2, and 20.6 times, respectively, at relatively low silica
loading (9 wt%) relative to the silica-free rubbery epoxy polymer. The remarkable
mechanical enhancements are contributed by the large window and cell sizes of
the MSU-F silica foam which are comparable to the layer separation (>10 nm) of
the exfoliated clay layers in the polymer-epoxy nanocomposites. Moreover, the
mechanical improvements for the epoxy mesocomposites obtained from the
calcined MSU-F silica verify that mesostructured silica having a very large pore
size and a large pore volume is a good reinforcing agent for thermoset epoxy
polymers.

Although the modulus enhancement obtained with MSU-F foam particles
is less than the modulus increase observed for MSU-J wormhole particles, the

fractural properties of the MSU-F mesocomposites are substantially improved

99

due to homogeneous dispersion. The fracture properties, tensile strength and
elongation, of the epoxy-MSU-F mesocomposites are superior to MSU-J
composites. Moreover the 6.8-fold increase in tensile strength of MSU-F
composites is even higher than the tensile strength benefit obtained for to
exfoliated organo-montmorillonite nanocomposites (5.3-fold increase) at same
silica loading (9 wt%). The highly improved tensile strength is presumably
provided by the high homogeneity resulting from high pore volume of the
mesocelluar stmcture and the good dispersion of the foam particles. MSU-F
silica particles occupy a larger volume of the mesocomposites due to the lower
intrinsic density of the foam structure in comparison to the MSU-J wormhole
structure. The lower density of the foam structure leads to reduced distances
between silica particles and to better overall reinforcement of the polymer matrix.
It is also noteworthy that the epoxy-MSU-F mesocomposites exhibit ~50%
elongation at 5—9 wt% silica loadings despite the isotropic particle morphology
whereas limited elongation (~30% at 12 wt% silica loading) was obtained for the
epoxy-MSU-J mesocomposites.

The 3.7—fold increase in tensile modulus for the MSU-F mesocomposites
at 9 wt% loading is comparable to that of the MSU-J (~4-fold increase)
counterpart at the same loading. The relatively higher surface area of MSU-J
mesoporous silica (~950 m2/g) facilitates interfacial interaction between polymer
chains and the silica surface, providing a slightly higher modulus at the beginning
stage in stress-strain behavior. Theoretical modulus predictions established for

particulate-filled polymer composites were performed in order to investigate the

100

effects of the interfacial interactions and the pore volumes on the modulus
behavior. The experimental modulus values for both mesocomposites prepared
from each calcined MSU-J and MSU-F mesostructure are higher than the
theoretical values derived from a consideration of the volume fractions occupied
by silica walls only, as well as by the volume fraction occupied by the entire
mesostructured particle (silica walls + polymer-filled pores). The results
demonstrate that mesostructured silica particles function as reinforcement agents
by providing volume fractions beyond the volume fractions provided by non-
porous silica particulates at the same weight loading. Moreover, the large
surface area of the mesostructured particles increases the interfacial interactions
between the particulate and the polymer matrix and this leads to greater stiffness.
The importance of surface area in providing stiffness is verified by the higher
moduli found for composites made from MSU-J (950 m2/g) in comparison to

MSU-F (540 m2/g).

101

4.6 References

(1) Lange, J.; Nicolas, B.; Galy, J.; Gerard, J. F. Polymer 2002, 43, 5985.

(2) Becker, 0.; Simon, G. P. Adv. Polym. Sci. 2005, 179, 29.

(3) Lan, T.; Kaviratna, P. D.; Pinnavaia, T. J. Chem. Mater. 1995, 7, 2144.

(4) Messersmith, P. B.; Giannelis, E. P. Chem. Mater. 1994, 6, 1719.

(5) Kelly, R; Akelah, A.; Qutubuddin, S.; Moet, A. J. Mater. Sci. 1994, 29, 2274.
(6) Lee, D. C.; Jang, L. W. J. Appl. Polym. Sci. 1998, 68, 1997.

(7) Wang, Z.; Pinnavaia, T. J. Chem. Mater. 1998, 10, 1820.

(8) Brown, J. M.; Curliss, D.; Vaia, R. A. Chem. Mater. 2000, 12, 3376.

(9) Osman, M. A.; Mittal, V.; Morbidelli, M.; Suter, U. W. Macromolecules 2004,
37. 7250.

(10) Wang, Z.; Massam, J.; Pinnavaia, T. J. In Polymer-Clay Nanocomposites,
Edited by Pinnavaia, T. J. Beall, G. W., Wiley, New York, 2001.

(11) LeBaron, P. C.; Wang, Z.; Pinnavaia, T. J. Appl. Clay Sci. 1999, 15, 11.
(12) Liang, Z. M.; Yin, H.; Xu, H. H. Polymer 2003, 44, 1391.

(13) Vargas, E. S.; Lynch, M. W.; Ruda, M. J.; Chandrashekar, V. US Patent
6602966, 2003.

(14) Kotsilkova, R. J. Appl. Polym. Sci. 2005, 97, 2499.
(15) Li, c. Y.; Yang, J. J. Appl. Polym. Sci. 2005, 98, 22.

(16) Vaia, R. A. In Polymer-Clay Nanocomposites, Edited by Pinnavaia, T. J.
Beall, G. W., Wiley, New York, 2001.

(17) Park, J. H.; Kim, S.; Kim, Y. 0.; Park, 0. O. Nanotechnology 2005, 16, 1793.
(18) Xi, J. Y.; Tang, X. Z. Electrochim Acta 2005, 50, 5293.

(19) Nguyen, T. 0.; Wu, J. J.; Doan, V.; Schwartz, B. J.; Tolbert, S. H. Science
2000, 288, 652.

(20) Li, G. T.; Bhosale, S.; Wang, T. Y.; Zhang, Y.; Zhu, H. S.; Fuhrhop, K. H.
Angew. Chem. Int. Ed. 2003, 42, 3818.

102

(21) Lu, Y. F.; Yang, Y.; Sellinger, A.; Lu, M. C.; Huang, J. M.; Fan, H. Y.;
Haddad, R.; Lopez, G.; Burns, A. R.; Sasaki, D. Y.; Shelnutt, J.; Brinker, C. J.
Nature 2001, 411,617.

(22) Kojima, Y.; Matsuoka, T.; Takahashi, H. J. Appl. Polym. Sci. 1999, 74, 3254.

(23) Ji, X. L.; Hampsey, J. E.; Hu, O. Y.; He, J. B.; Yang, 2. 2.; Lu, Y. F. Chem.
Mater. 2003, 15,3656.

(24) He, J.; Shen, Y. B.; Yang, J.; Evans, D. G.; Duan, X. Chem. Mater. 2003, 15,
3894.

(25) Wang, M; Li, M. T.; Zhang, J. S. Mater. Lett. 2005, 59, 2685.

(26) Reid, B. D.; Ruiz-Trevino, A.; Musselman, l. H.; Balkus, K. J.; Ferraris, J. P.
Chem. Mater. 2001, 13, 2366.

(27) Balkus, K. J.; Cattanach, K.; Musselman, l. H.; Ferraris, J. P. Mater. Res.
Soc. Symp. Proc. 2003, 752, 91.

(28) Park, l.; Peng, H. G.; Gidley, D. W.; Xue, S.; Pinnavaia, T. J. Chem. Mater.
2005, ASAP.

(29) Schmidt-Winkel, P.; Lukens, W. W.; Zhao, D. Y.; Yang, P. D.; Chmelka, B.
F.; Stucky, G. D. J. Am. Chem. Soc. 1999, 121, 254.

(30) Lettow, J. 3.; Han, Y. J.; Schmidt-Winkel, P.; Yang, P. D.; Zhao, D. Y.;
Stucky, G. D.; Ying, J. Y. Langmuir 2000, 16, 8291.

(31) Kim, S. S.; Pauly, T. R.; Pinnavaia, T. J. Chem. Comm. 2000, 1661.
(32) Karkamkar, A.; Kim, S. S.; Pinnavaia, T. J. Chem. Mater. 2003, 15, 11.

(33) Lukens, W. W.; Schmidt-Winkel, P.; Zhao, D. Y.; Feng, J. L.; Stucky, G. D.
Langmuir1999, 15, 5403.

(34) Blanco, A.; Chomski, E.; Grabtchak, S.; lbisate, M.; John, 8.; Leonard, S.
W.; Lopez, 0.; Meseguer, F.; Miguez, H.; Mondia, J. P.; Ozin, G. A.; Toader, 0.;
van Driel, H. M. Nature 2000, 405, 437.

(35) Pore volume, BET surface area, cell size and window size of the recovered
MSU-F silica foam are 2.1 cm3/g, 509 m2/g, 26.9 nm, and 15.3 nm, respectively.

(36) Ahmad, Z.; Sarwar, M. l.; Wang, 8.; Mark, J. E. Polymer1997, 38,4523.

103

 

(37) Park, l.; Pinnavaia, T. J. Unpublished data.
(38) Lan, T.; Pinnavaia, T. J. Chem. Mater. 1994, 6, 2216.
(39) Richard, T. G. J. Composite Mater. 1975, 9, 108.

(40) Monette, L.; Anderson, M. P.; Wagner, H. D.; Mueller, R. R. J. Appl. Phys.
1994, 75, 1442.

(41) Handbook of Polymer Testing, edited by R. Brown, 1999, Marcel Dekker,
New York.

(42) Williford, R. E.; Addleman, R. 8.; Li, X. S.; Zemanian, T. S.; Bimbaum, J. C.;
Fryxell, G. E. J. Non-Cryst. Solids 2005, 351, 2217.

(43) Jung, S. 8.; Han, C. K.; Park, H. H. Appl. Surface Sci. 2005, 244, 47.

(44) Yamada, K.; Oku, Y.; Hata, N.; Seino, Y.; Negoro, C.; Kikkawa, T. J.
Electrochem. Soc. 2004, 151, F248.

(45) Triantafillidis, C. S.; LeBaron, P. C.; Pinnavaia, T. J. Chem. Mater. 2002, 14,
4088.

104

CHAPTER 5

MESOPOROUS SILICA STRUCTURES ASSEMBLED
FROM JEFFAMINE AND TOMAH3 SURFACTANTS
5.1 MSU‘I Wormhole Mesostructures

Electrically neutral supramolecular assembly pathways using primary
amine surfactants generally produce mesoporous materials with wormhole
framework structures as discussed in Chapter 1. Wormhole mesostructures
typically exhibit enhanced catalytic activity due to the three-dimensional pore
network1 and the improved textural porosityz relative to hexagonal
mesostructures prepared through strong electrostatic interactions.

Large-pore wormhole mesostructures (MSU-J) have been synthesized
from Jeffamine D2000 and D4000.3 The pore size and pore volume of MSU-J
silicas are in range of 5.3—14.3 nm and 1.4— 2.3 cm3/g, respectively, comparable
to those of SBA-15 silicas (see Chapter 2 for details). This appendix describes
other members of the MSU-J family of mesostructured silicas prepared frin
diverse Jeffamine and Surfonamine surfactants (see structures 5.1—5.4).

Three series of Jeffamine (M, D, and T series) surfactants are provided by
Huntsman Chemical. The prefix M, D, and T represent the presence of one
(mono-), two (di-), and three (tri-) amine groups, respectively, in the chemical
structures. The number designation after the prefix corresponds to the
approximate molecular weight. Basically, the Jeffamine surfactants are based on

a propylene oxide (P0) or mixed PO/EO backbone. The Surfonamines (see

105

CH3TO-CH2-CHHO-CHz-CHTNHZ JeffamineTM M2005: (5.1)
R

 

 

 

R' R, R’ = H or CH3
Molar ratio of PO/EO = 39/6
Mw = ~2000
CHziIO-CHz-fHJLNHz JeffamineTM T3000: (5.2)
x+y+z=50
: CH3; X M, = ~3000
CHd’O—CHz-CHW'NHZ
(LH Jeffaminem T5000:
- 3«Y x+y+z=81
CHZ‘FO-CHz-CHJ-NHZ Mw = ~5000
CH3‘ z
CH3—(CH2)11_13 O-CH2“CH NH2 Surfonamine“ ML300. (53)
gm M, = ~300
3 2
CgH19 O-CHz-(EH O—CHz-CfH-NHZ (5-4)
CH3 n CH3
SurfonamineTM MNPA1000:
n=135
MW = ~1000

structures 5.3 and 5.4) are monofunctional amines similar to the M-series but
they possess long-alkyl groups at the end of the hydrophobic tails. A
commercially available sodium silicate (Aldrich, 27% 8102, 14% NaOH) solution
was used as the silica precursor. Sodium silicate compared to another silica
source, such as TEOS, is cost-effective and suitable the use of the silica as a
reinforcement agent to enhance mechanical property of polymers. The
experimental procedures for the preparation of the MSU-J silicas in this appendix

are same as those of Chapter 2 except for the Surfactant/Si ratios (see Table

106

5.1). Numerous survey experiments indicated that the optimum Surfactant/Si

ratios provide intense XRD reflections, large surface areas and high yields.

Table 5.1 Physical properties for mesostructured wormhole silicas prepared from

sodium silicate and Jeffamine or Surfonamine surfactants.

 

 

Surf./Si . r Pore Pore
Surfactant molar Tgrgpfigbfie 33:11) (:3?) size volume
ratio 9 (nm) (cm3/g)
25 / 20 6.5/5.9 632 3.5 0.5
M2005 0.062 45/20 7.9/7.0 673 4.8 0.9
65 / 20 -1/ 7.1 885 6.1 1.4
25 / 20 6.4 / 5.9 936 5.0 1.3
T3000 0.083 45 / 20 6.6 / 6.4 964 6.0 1.5
65 / 20 6.2 / 7.2 881 6.7 1.8
25 / 20 J 1030 5.2 1.5
T5000 00” 45 / 20 3 882 8.2 2.0
ML300 0.25 25 / 20 4.6 / 4.2 770 2.8 1.1
MNPA1000 0.12 25/20 5.6/6.0 1012 5.1 1.4

 

1rBasaI spacings of as—made and calcined MSU-J silicas are separated by a
slash (as-made / calcined).
1 The XRD reflections are not resolved.

The XRD patterns of as-made and calcined MSU-J silicas prepared from
the sodium silicate and M2005, T3000, T5000, ML300 and MNPA1000 are
illustrated in Figures (5.1—5.4). Typical of wormhole mesostructures (e.g. HMS
and MSU-X), each sample exhibits a typical XRD pattern consisting of an intense
and relatively broad low angle reflection and a weaker, higher angle broad
shoulder due to lack of long-range order. Although a few MSU-J silicas do not

show well-resolved reflections in their XRD patterns (see Table 5.1), an intense

reflection characteristic of a regular pore spacing is in the 29 range 1—1.5°. The

107

 

(a)

 

 

 

 

 

 

E
C
3
E‘
9 o
1‘73 65 c
3
5‘ o
g 45 c
.9
E
25°C
'7 8

 

 

q (to)

   

O
O
‘P

 

Volume Adsorbed (cm3lg, srp)

 

   

r

 

 

N2
0

 

'012 031 '016 ‘ 018% 1
Relative Pressure (PIPo)

Pore Size (nm)

Figure 5.1 (a) XRD patterns, (b) N2 sorption isotherms, and (c) pore size
distributions for MSU-J silicas prepared from sodium silicate and M2005. The
dotted ("") and solid (—) XRD scans represent as-made and calcined MSU-J
silicas at different synthesis temperatures. lsotherms are measured for

calcined MSU-J silicas synthesized at 25 (o), 45 (1:1), and 65 (A) °C.

108

 

Intensity (arbitrary units)

 

 

_‘
N
0°

 

 

 

2

N Volume Adsorbed (cmalg, STP)
1%

 

 

 

 

 

1'0 1'
P/PO ' ' Pore size (nm)

Figure 5.2 (a) XRD patterns, (b) a TEM image (synthesized at 25 °C), (c) N2
sorption isotherms, and (d) pore size distributions for MSU-J silicas prepared
from sodium silicate and T3000. The synthesis temperatures are indicated as

25 (o), 45 (I), and 65 °C (A) (c and d).

109

)

.3
.h
D
O

 

 

 

 

 

 

 

    

 

 

O.
,_ . (a)
"1 1200
9 l
(a
E 100%
3 .
3 800-
.D
3 :1
g 600-
< .
w 400
E
2 -l
O
> 20m
N
z c -
0 012 014 016 018 1
0litflelative Pressure (PlPo)
' . it (b)
0.24
00.15-
1,
s
U 0.1-
0.05.
0 . . . , . . .
0 10 20 30 40

 

       

Pore Size (nm)

Figure 5.3 (a) N2 sorption isotherms and (0) pore size distributions for MSU-J

silicas prepared from sodium silicate and T5000. The synthesis temperatures

are indicated as 25 (o), and 45 (:1) °C.

110

 

     
  

(a)

MNPA1 000

 

Intensity (arbitrary units)

 

 

 

 

 

ML300 (C)

-'_ MNPA1000

 

 

N 2 Volume Adsorbed (cm3’g, STP)

 

  

 

 

 

 

20

   
 

s ‘ 10” ' 15
Pore slze(nm)

Figure 5.4 (a) XRD patterns (as-made; and calcined; —), (b) N2 sorption
isotherms, and (c) pore size distributions for MSU-J silicas prepared from
sodium silicate and Surfoamine surfactants (ML300; o and MNPA1000; E] in b

and c).

111

TEM image of the MSU-J silica prepared form sodium silicate and T3000 at
25 °C shows typical wormhole framework morphology with a regular pore size
(Figure 5.2b).

As shown in Chapter 2 the pore diameter of calcined MSU-J silicas
generally expands accompanied by increase of the pore volume with increasing
synthesis temperature over the range 25—65 °C. The temperature dependence of
MSU-J pore sizes is associated with a decrease in H-bonding at the silica-
surfactant interface with increasing temperature (details in Chapter 2).

The above results show that large-pore wormhole mesostructures have
been successfully synthesized through the supramolecular assembly of neutral
amine surfactants without the need for an auxiliary pore expander. Jeffamines
are well-known curing agents for thermoset epoxy polymers‘ and the diverse
Jeffamine chain lengths afford epoxy polymers with diverse mechanical
properties and glass transition temperatures. Therefore, the as-made and
calcined MSU-J wormhole mesostructures can be promising candidates as
reinforcement agents for the corresponding thermoset epoxy polymers. Moreover,
the mechanical properties of the epoxy polymers can be systematically modified

by the controlling pore size, pore volume and surface area of the MSU-J silicas.
5.2 MSU-G Type Lamellar Mesostructures (Tomah3 Surfactants)

A family of mesoporous molecular sieves (MSU-G) with vesiclelike

hierarchical structures has been prepared through a supramolecular assembly

112

pathway using electrically neutral gemini surfactants as the structure-directing
progens (composition 5.5) and a silica precursor, TEOS.5 The electrically neutral

CnH2n+1NH(CH2)2NH2 n = 10, 12, 14 (5.5)5
H-bonding assembly pathway based on the hydrolysis of TEOS in the presence
of the gemini surfactants form vesiclelike silica mesostructures analogous to a La
(bilayer lamellar phase) or L3 (wormhole or spongelike phase) intermediate. The
MSU-G vesicles exhibit unprecedented thermal (up to 1000 °C) and
hydrothermal (>150 hours at 100 °C) stabilities due to their very high degree of
framework (SiO4) crosslinking, as comfirrned by 2"Si MAS NMR spectroscopy
(CY/Q3 ~7). Moreover, redox-active Ti(|V) and acidic Al(lll) centers have been
incorporated into the MSU-G framework without altering the lamellar framework
or the vesicular hierarchical structure. The AI-substituted (2 mol%) MSU-G
derivatives are more efficient acid catalysts than Al-MCM-41 (2 mol%) for the
conversion of 2,4—di-tert-butylphenol and cinnamyl alcohol to a flavan.‘5 For
gemini derivatives with n = 12, 14 and m = 3, 4 (composition 5.6) hierarchical
vesicles have been prepared analogous to those assembled from the gemini
surfactants with m = 2.7 However, for n = 10, a new coiled slab structure (m = 3)
and an onion-like core-shell structure (m = 4) are formed.

CnH2n+1NH(CH2)mNH2 (c°n.m.o) n = 10, 12, 14 and m = 3, 4 (5.6)7

In this appendix, MSU-G type silica vesicles were synthesized from TEOS
in the presence of commercially available gemini surfactants, DA1214 and DA14.

The DA series of gemini surfactants provided by Tomah3 Chemical Inc. are

113

alkyloxypropyl-1 ,3-diaminopropanes. The number after DA represents the
number of carbon and oxygen atoms in tail group (compositions 5.7 and 5.8).

DA1214 CnH2n+1O(CH2)3NH(CH2)3NH2 n = 8 and 10 (5.7)

DA14 (CH3)2(CH2)30(CH2)3NH(CH2)3NH2 (5.8)
The chemical structures of the surfactants are provided by Tomah3 Inc. DA1214
is a mixture of octyl- (50 mol%) and decyl- derivatives. A distinct difference
between the DA gemini series and the former gemini surfactants (composition
5.6) used to prepare MSU-G silicas is the presence of an oxygen atom at the 7-
position from the secondary amino group.

The synthesis procedure for the preparation of vesicular MSU-G silicas is
similar to that of previously reported MSU-G mesostructures.5 In a typical
synthesis 0.8 g of DA1214 was dissolved in an ethanol/water solvent mixture (2
mL/18 mL) and 2.52 g of TEOS was then added to the surfactant solution. The
reaction mixture was stirred at ambient temperature for 30 min. and then heated
at 100 °C in an autoclave for 72 h under static conditions. The solid product was
filtered, washed with water, and dried in air. The surfactant was removed by
calcination at 600 °C for 4 h. The molar ratio of each reaction mixture was 1.0 Si :
0.25 Surfactant : 4.3 EtOH : 78 H2O. Partially protonated amine surfactants were
used by adding 1.0 mL of 1.0 M HCI solution (H*/Si = 0.083).

An as-made mesoporous material prepared from TEOS in the presence of
the neutral and protonated forms of DA1214 surfactant shows an intense low
angle reflection in the XRD pattern (not shown) similar to the diffraction feature of

the previously reported MSU-G vesicles. The TEM images (Figure 5.6 a, b) of

114

 

 

 

 

 

 

 

 

    
  
 
 
 
    

 

 

(a)
3.?
C
3
5‘
E \‘--_.. DA14
{E
3
:2“
fl)
C
3
E
"’\_ DA1214
0 2 4 6 ' 8' ' 10
28
A 500
g .(b)
(D
g. 400-
”E d
3
'3 300‘ DA1214
{2 .
8
ID 200'
(U
0
E
2 100-
O
>
zN

 

 

 

 

 

   

'012'014'016'018'1
P/Po
Figure 5.5 (a) XRD patterns (as-made; and calcined; -), (b) N2 sorption

O

Pore size (nm)

isotherms, and (0) pore size distributions for silica vesicles prepared from
TEOS and partially protonated Tomah3 DA surfactants (DA1214; o and DA14;

I in b and c).

115

 

.4 . t

Figure 5.6 (a) TEM images with low and (b) high magnification of MSU-G type

lamellar vesicles formed from TEOS and neutral DA1214 surfactant (as-
synthesized). (c) A TEM image of the calcined lamellar silica structure

assembled from partially protonated DA1214 surfactant.

116

the as-made mesostructure assembled under electrically neutral conditions
illustrate the vesicular morphology containing 5—20 silica layers in the vesicular
walls. Almost all of the particles (100—400 nm size) are broken vesicles. However,
the lamellar structure is collapsed by removing surfactant indicating lack of pillars
in the gallery region of the layers. In contrast, the partially protonated DA1214
surfactant formed by adding a small amount of hydrochloric acid (H"/Si molar
ratio = 0.0083 — 0.083) to the reaction mixture leads to lamellar silica
nanoparticles. Figure 5.5a (bottom) illustrates the well-ordered lamellar structures
and the XRD patterns of the as-made and calcined lamellar mesostructure
assembled from protonated DA1214 gemini surfactant (H‘ISi = 0.083). The
maximum in the HK pore size distribution curve occurs at 3.8 nm, corresponding
to capillary condensation of liquid nitrogen at a relative pressure ~0.3 (see Table
5.2). Although the silica assembled from the partially protonated gemini
surfactant possess a lamellar framework structure and textural properties similar
to those of MSU-G silica, the particle morphology cannot be described as being a
vesicle. However, the size of the lamellar silica particles is relatively more
uniform (100—200 nm) and each particle has a more regular number of silica
layers (~15).

No mesoporous materials with well-resolved XRD reflections were
synthesized from neutral DA14 gemini surfactant. From the XRD pattern (Figure
5.58, top) and N2 sorption properties (Figure 5.5b) of the calcined silica
assembled from protonated DA14 surfactant may be comparable to the lamellar

silica form by protonated DA1214 with same amount of acid in physical

117

properties. However, more detailed studies including TEM are required to
determine the framework structures of the silica prepared from the DA14 gemini

surfactant.

Table 5.2 Textural properties for lamellar silica vesicles assembled by TEOS in

the presence of Tomah gemini surfactants.

 

 

HCI/Si . a Pore Wall Pore
Surfactant molar 1183133)" g (3132?) size thickness” volume
'8th 5 1 / c 4649 (my) (nm) Jcm3/g)
. . - - - 0.39
DA1214 0.083 48/47 543 3.8 1.9 0.47
DA14 0.0 -° / -" 418 -°' - 0.98
0.083 -c / 5.0 479 3.8 2.2 0.65

 

5' Basal spacings of as-made and calcined MSU-J silicas are separated by a
slash (as-made / calcined).

b The framework wall thickness is determined by subtracting the HK pore size
from the d spacing of calcined sample.

C A low-intensity hump (one third of intensity of the as-made sample) in range of
1.5—2° 26.

d No observable peaks in HK pore size distributions.

9 No evident XRD reflections.

The protonated amine group possibly plays an important role in forming
silica bridges between silica layers. The tail of tomah3 surfactants contains an
ether (oxygen atom) group with a partial negative charge which causes a weaker
van der Waals interaction between the surfactant tails in comparison to the
previous gemini surfactants containing alkyl tail groups. The corresponding
increase in the surfactant chain volume results in an increase in the effective
packing parameter (9 value, Figure 1.3). Thus only silica lamellar layers in

absence of pillars in gallery region are generated. Once a small amount of acid

(H"/NH2 = 00166-0166) is added to the reaction mixture the larger effective

118

head group area of the partially protonated amino groups decreases the packing
parameter, to produce bridges between silica layers. However, lamellar silica
framework structures are not formed at higher concentrations of ammonium
groups (H+/NH2 >0.166), as confirmed by XRD and N2 sorption analyses (not
shown).

Mesostructured molecular sieves with a lamellar framework structure have
been synthesized from TEOS and commercially available gemini surfactants in
partially protonated form. A relatively low molar fraction of protonated amino
groups in the surfactants generates pillared lamellar silica frameworks. The
mesostructures are predicted to exhibit improved thermal and hydrothermal
stability and high catalytic performance as shown in previously in the published
literature5' 6. However, a more detailed analysis such as XRD, TEM and 29Si
NMR is needed to complete the characterization of the novel materials. Also, a
more detailed study of the reaction stoichiometry required for a more complete

description of the assembly reactions.

119

5.3 References
(1) Lee, J.; Yoon, 8.; Oh, S. M.; Shin, C. H.; Hyeon, T. Adv. Mater. 2000, 12, 359.

(2) Pauly, T. R.; Liu, Y.; Pinnavaia, T. J.; Billinge, S. J. L.; Rieker, T. P. J. Am.
Chem. Soc. 1999, 121, 8835.

(3) Park, L; Wang, Z.; Pinnavaia, T. J. Chem. Mater. 2005, 17, 383.

(4) Pinnavaia, T. J.; Lan, T. US Patent 5,801,216, 1998.

(5) Kim, S. S.; Zhang, W. Z.; Pinnavaia, T. J. Science 1998, 282, 1302.

(6) Kim, S. S.; Liu, Y.; Pinnavaia, T. J. Micropor. Mesopor. Mater. 2001, 44,489.

(7) Karkamkar, A. J.; Kim, S. S.; Mahanti, S. D.; Pinnavaia, T. J. Adv. Funct.
Mater. 2004, 14, 507.

120

APPENDIX

HIGH OXYGEN GAS BARRIER CLAY-EPOXY COMPOSITE FILMS

A.1 Introduction

Improving the gas barrier of polymers and plastics to small molecules (e.g.
02, H2, H2O, CO2 etc.) is an important need for a vast number of applications,
such as food and beverage packaging, liquid hydrogen storage and
transportation, moisture and air-proof coatings, etc. As mentioned in Chapter 1
the high aspect ratio of impermeable exfoliated clay nanolayers in a polymer
matrix presents an increased tortuous path length for permeant gas molecules.
For instance, PET (polyethylene terephthalate)-monmorillonite nanocomposites
exhibits up to 70% reduction in oxygen permeability compared to the pristine
polymer, depending on the degree of clay exfoliation, nanolayer orientation and
aspect ratio of the clay platelets.1

In this appendix, a new method is described for fabricating highly oxygen-
imperrneable polymer composites by impregnating smectite clay films with the
thermoset epoxy polymer. The glassy epoxy polymer described in this study can
have significant practical applications for the manufacturing of liquid hydrogen
and methane storage tanks and transfer pipes. The reduction in oxygen
permeability measured for the different epoxy—clay composite films were 2—3

orders of magnitude lower than for the pristine polymer. A recent study2 on the

121

permeability of epoxy-montmorillonite nanocomposites demonstrated a 70%
reduction of oxygen permeability coefficient at 5 vol% loading. The barrier
properties of the described epoxy—clay composite films were far better than those
for classical epoxy—clay nanocomposite samplesa, which were also prepared and

tested for comparison.

A.2 Sample Preparation

The epoxide resin used for the formation of the pristine glassy epoxy
polymer and the epoxy-clay composite films was EPON 826 (see structure 1.3)
and the curing agent was Jeffamine D230 (D230, structure 1.4). An industrially
purified Na+-montmorillonite (PGW, Nanocor Inc.) with a cation exchange
capacity (CEC) of 120 meq/100 g was used as the pristine clay for the
preparation of the Na+-montomorillonite clay films. A synthetic Li*-fluorohectorite
clay (FH, Corning Inc.) with a CEC similar to that of the PGW montmorillonite
was used as the pristine fluorohectorite clay for the preparation of the Li‘—
fluorohectorite clay films in this study.

For the preparation of the self-supported Na+—montomorillonite clay film
190 mL of an aqueous suspension containing 10.0 g of Na+-montmorillonite
(PGW) was blended at room temperature for 1-2 min to form a clay colloidal
suspension (with clay concentration of 5 wt%). The colloidal suspension was
then cast on a glass plate and was dried at room temperature and atmospheric
conditions for a period of 2 days. The formed clay film was then carefully peeled

of the glass plate. In this way a smooth, self-supported, continuous, semi-

122

transparent Na*-montmorillonite clay film with a uniform thickness of 60 pm was
produced. The thickness of the film could be varied depending on the initial
concentration of the aqueous clay colloidal suspension. For example, a 3 wt%
clay suspension resulted in a 35 pm thick film. Li+—fluorohectorite films were
prepared in the same fashion as the Na+—montmorillonite films.

Pristine glassy epoxy polymer disks were prepared by mixing the epoxy
monomer with the curing agent at 50°C for ~30 min, outgassing the liquid mixture
at room temperature and curing the mixture on an aluminum disk mold by using
an appropriate relief-coat at 75.0 for 3 h and at 125°C for an additional 3 h under
nitrogen. The formed epoxy disks were polished to an average thickness of ~0.2
mm (~8 mil).

For the preparation of the epoxy—clay composite films the self-supported
clay films were impregnated with a mixture of uncured EPON 826 and D230 and
were then cured to form flexible epoxy-clay composites in which the clay film
was encapsulated in the epoxy polymer in a sandwich-like fashion. The Na“—
montmorillonite film with dimensions approximately 8 cm x 8 cm was dipped in a
mixture of EPON 826 epoxy resin and D230 curing agent at room temperature
until it was well-wetted on both sides. The epoxy—D230 mixture was prepared by
mixing the liquid EPON 826 resin with the D230 at 50°C, for 30 min, under stirring.
Before dipping the clay film in the pre-polymer, the uncured epoxy-D230 mixture
was outgassed at room temperature for 30 min. The dipped clay film was then
suspended vertically at room temperature in air so that excess uncured mixture

was allowed to drop from the film. After 1 day of aging in air the epoxy resin was

123

partially cured. The epoxy-clay composite film was then heated in the oven at
60°C for 3 h and 125°C for 3 h in a nitrogen atmosphere so that the epoxy resin
was completely cured. The produced cured epoxy-clay composite film was
smooth, flexible, semi-transparent and uniformly covered on both sides with
cured epoxy polymer.

In order to estimate the tortuosity factor of the film the weight content of
the impregnated epoxy polymer in the Na+-montmorillonte clay film was
calculated by measuring the thickness and weight of the clay film, the pristine
epoxy disk, and the composite sample. The weight content of clay was converted
to a volume fraction using the bulk densities of the Na"-montmorillonite film (2.0
g/cm3) and the pristine epoxy polymer (1.18 g/cm3). The void volume of the Na+-
montmorillonite film prepared from a 3 wt% clay suspension was also estimated
using the bulk density of Na"-montmorillonite (2.6 g/cm3)4. The external epoxy
polymer coating of the composite film was excluded from the calculation of the
volume fraction.

X-ray diffraction (XRD) patterns were obtained on a Rigaku rotaflex 2008
diffractometer equipped with Cu K, X-ray radiation and a curved crystal graphite
monochromator, operating at 45 KV and 100 mA. The diffraction patterns were
collected between 1 and 12° at a scanning rate of 1°/min. Samples of the Na"-
montmorillonite powder and film were prepared by mounting the films on a glass
holder and a glass slide, respectively.

Oxygen permeability measurements were performed on a Mocon Ox-tran

2/60 oxygen permeability instrument with a 100% oxygen as test gas. The

124

specimens tested were the pristine epoxy polymer disks and the epoxy—clay

nanocomposite disks with an average thickness of ~0.2 mm (~8 mil), and the
epoxy—clay composite films with an average thickness ranging from ~0.10 to

~0.15 mm.

A.3 Results and Discussion

Different types of pristine clays have been cast from aqueous suspensions
into opaque (fluorohectorite) and semi-transparent (montmorillonite) films. The
clay film made from montmorillonite has a light brown color and is smooth,
continuous, flexible and semi-transparent. Self-supported films with different
lateral dimensions could be fabricated in this way. Optical images of the self-
supported Na*-montmorillonite film with a thickness of 35 pm are shown in Figure
A.1. The fluorohectorite films prepared from 5 wt% suspension by the same
method had a white color and a mean thickness of ~65 pm.

The XRD patterns of the Na+-montmorillonite powder and the form of the
clay film prepared from 3 wt% water suspension are shown in Figure 8.2.
Although base line corrections for each pattern are necessary to compare the
reflection intensities, the 001 peak of the film sample (Figure A.2 b) exhibits five
times higher intensity than that of the powder sample, indicating the nanolayers
are stacked in the film sample with high ordering along the c axis. In addition, the
intensity of the 004 peak for the film sample is 25 times higher than the
corresponding peak of the powder sample. Consequently the exfoliated individual

clay nanolayers are stacked with highly ordered perpendicular to the layer plane.

125

 

Figure A.1 Optical images of a self-supported Na+-montmorillonite clay film

with a thickness of 35 pm prepared from a 3 wt% aqueous suspension.

126

 

(001)

(004)
{105) (130)

 

 

I

I

Internsity (arb. units)
(1 10)

 

 

A

I

(b)
(a)

 

 

[Ah—fl

20

Figure A.2 X-ray diffraction patterns of (a) Na*—montmorillonite powder and
(b) the Na*—montmorillonite film prepared from a 3 wt% water suspension. The
intensity of the patterns from 12° to 80° was multiplied three-fold and shown

as an insert for clarity. Offsets are 500 counts for both sets of scans.

127

'r'f'r'
0 1O 20 30 40 50 60 70 80

 

A proposed restacked clay structure formed in films is illustrated in Figure A.3 (a).
From SEM images of the clay films (not shown) the clay films possess micro or
sub-micro voids which may generate the extremely high oxygen permeability (>
6000 cc-miI-m’z-day'1).

The barrier properties of polymer—clay nanocomposites have been initially
described based on a relatively simple tortuous path models, as is presented in
Chapter 1. In this model the silicate platelets obstruct the diffusion of gases
through the polymer matrix. The improvement in the barrier properties based on
this model is a function of the volume fraction and of the aspect ratio of the
platelets. Moreover, it is generally considered that a parallel orientation of the
clay platelets to the surface of the polymer—clay nanocomposite film is necessary
for improving the barrier properties of the polymers. This prerequisite is usually
difficult to achieve, especially in thermoset polymer—clay nanocomposites such
as epoxy nanocomposites, where the exfoliated individual clay nanolayers with
thickness of ~1 nm are randomly oriented upon curing of the polymer resin.

The dependence of the oxygen permeability of a classical epoxy—clay
nanocomposite on the clay content has been examined. The nanocomposites
were prepared from organically modified montmorillonite and fluorohectorite with
Jeffamine D2000 (see structure 1.7) and the same epoxy polymer.3 The
nanocomposite specimens had a thickness of ~0.2 mm (~8 mil) and the clay
nanolayers in the polymer matrix were intercalated or exfoliated depending on
the degree of onium ion exchange. The reduction in oxygen permeability is

achieved with the increase in the degree of nanolayer exfoliation but it was never

128

 

(a)
§—J K
\ —’~ 5

3"”

Epoxy resin
+
Jeffamine
D230

Cured
at 125 °C

”......0. ......
292 W999999 99“999999999 9
92292 92 9292 99 :92939 29 92929292929292929292 ,
9 9 9’929292"
“999092
”292929292 9 292929292 2
0
92929 292'. 1 ' 9 9 9 9 29 9 29292929 . 2“”
2929 29 29 92 9. 99 ‘ - 9 9.9.9 .9 9.9.9.9.9939
292:
”9999
2:222-99999 . .9 99....
..229 9 21' v 9 22929292929 9 9292929 .2 2: 292 29.2 2: 2 2: 2 22
M99999999; 9929 “992999 99 9u99n99
w9999999 v... 9999 “9929999
22.999 992-399999999399
22 929292912 292929 9 9 9 929
2.9 ..9 9. 292192992 292 "9 92 '929
.2929 9292929222929

0..
29292 9:‘2999 999 2‘ .
00v 000 0:“ 000 000'0"0 02:00
000..0:0 20.0.0::::: 00.: 2:20;::. 000.000.:202‘“

2:290.

2 20:92:. 999 .2 99:9 000 942
1222 92299 9292. 929:9 992.2:292 2:2 9";999 2299 29292929999999 222 929299
9999 929 929‘9'9":9 99 922999 92299999999
92 2999299999 2.;2 222999 222 2 2 292: 29292929292299 99292 29292 2299
29222999292929 22929992 9292 2292 29 .2922992929 22929299 22992929292 929292

 

Figure A.3 Proposed clay nanolayer structure of an oriented clay film in (a)
pristine and (b) epoxy-impregnated form. The mixture of epoxy resin and D230
represented by hatched pattern is impregnated in voids of the clay films and

cured at 125 °C to form high gas barrier composite films. Sodium ions in the

gallery regions of the clay are omitted for clarity.

129

higher than ~20°/o at 6 wt% clay loading. Higher clay loadings had a small
additional effect on the barrier properties of the nanocomposites. Although
greater reductions in gas permeability have been previously reported1' 68, they
refer to the reduction achieved for more permeable pristine polymers (such as
PET and rubbers) and/or to perrneants (e.g. water vapor) different from the
epoxy and oxygen system of interest in the present study.

In contrast to the barrier properties of classical epoxy—clay nanocomposite
samples, the oriented epoxy—clay composite films prepared in this work showed
a dramatic decrease in oxygen permeability, reaching 2—3 orders of magnitude
lower than the permeability of the pristine epoxy. This enormous increase of the
barrier properties of the produced epoxy composites was achieved by
impregnating the extremely permeable clay films (oxygen permeability 2 6 x 103
cc-miI-m’z-day“) with the permeable pristine epoxy polymer. The oxygen
permeability data for the epoxy-clay composite film specimens prepared from
different clay films are shown in Table A.1. In most cases, the reduction in
oxygen permeability is 2—3 orders of magnitude in comparison to the pristine
epoxy polymer and 3-4 orders of magnitude compared to the pristine clay films.
The reduction in permeability was not significantly affected by the thickness of
the composite clay film since a ~10 pm sodium montmorillonite composite clay
film resulted in a similar decrease in oxygen permeability as an epoxy—clay
composite film with a thickness of ~60 pm (see Table A.1). In general, the

differences in permeability among different composite film samples are difficult to

130

identify because the low oxygen transmission rates are subjected to relative high

standard deviation of the mean values.

Table A.1 Oxygen Permeability data of the epoxy—clay composite films.

 

Thickness Thickness of

 

x n
titty 0mm. ...
(mm) (mm) (cc-mII-m -day )
Pristine Epoxy - 0.20 98.9
Epoxy—(Na+-montmorillonite) 0.060 0.11 s 0.1
Epoxy—(Na+-montmorillonite) 0.035 0.10 0.97
Epoxy—(Li+-fluorohectorite) 0.065 0.14 1.2

 

On the bases of the density of a Na+-montmorillonite platelet and the bulk
density of a montmorillonite film, the void volume fraction in the Na+-
montmorillonite film is calculated to ~0.19. This is consistent with the volume
fraction of the impregnated epoxy polymer (~0.21) in the clay film which is used
for tortuosity factor estimation. lf clay disk plates of diameter L and width W are
dispersed in a parallel fashion in a polymer matrix, the widely used tortuosity
factor5 and the relative permeability are given in equation 1.6. Typical aspect
ratio for natural montmorillonite is 200 (200 nm in length and 1 nm in thickness).9
Thus, the tortuosity factor for the Na*-monmorillonite—epoxy composite film is 76
and its predicted permeability is 1.3 cc-mil/mz-day. This is slightly higher than the
oxygen permeability data shown in Table A.1. However, equation 1.6 is
applicable only in dilute (arp<<1) regime.10 The more developed version (equation

1.7)11 includes a quadratic term and provides a more reliable prediction (0.07

131

cc-mil/mz-day) of extremely low oxygen permeabilities consistent with the
observed values.

The high barrier properties of epoxy—clay composite films results from the
confinement of epoxy in the voids between stacked clay layers and the voids
formed between layer edges and folded layers (see Figure A.3). The
impregnation of the clay films by the liquid, uncured epoxy polymer precursors
occurs through diffusion into the voids of sub-micron dimensions that exist on the
surface of the clay films. Such cracks, which were observed by SEM images (not
shown here), allow the liquid epoxy resin to approach the inorganic clay

nanolayers in the center of the clay film.

A.4 Conclusion

A new concept in barrier technology is presented in this work wherein the
appropriate combination of two permeable materials resulting in a composite with
extremely low gas permeability. This unprecedented increase of barrier
properties can be achieved by impregnating an inorganic clay film with the
polymer precursors resulting in a cured composite that contain a central phase
consisting of highly oriented clay nanolayers. The high volume fraction of parallel
oriented clay nanolayers in impregnated epoxy polymer cause the permeability of
the polymer phase to be significantly lower than that of the bulk polymer, and to

contribute also to the extremely high barrier properties.

132

A.5 References

(1) Matayabas Jr, J. C.; Turner, S. R. In Polymer-Clay Nanocomposites, Edited
by Pinnavaia, T. J. Beall, G. W., Wiley, New York, 2001.

(2) Osman, M. A.; Mittal, V.; Morbidelli, M.; Suter, U. W. Macromolecules 2004,
37,7250.

(3) Triantafillidis, C. S.; LeBaron, P. C.; Pinnavaia, T. J. Chem. Mater. 2002, 14,
4088.

(4) Osman, M. A.; Mittal, V.; Lusti, H. R. Macromol. Rapid Comm. 2004, 25, 1145.
(5) Nielsen, L. E. J. Macromol. Sci. (Chem) 1967, A1, 929.

(6) Messersmith, P. B.; Giannelis, E. P. Chem. Mater. 1994, 6, 1719.

(7) Strawhecker, K. E.; Manias, E. Chem Mater 2000, 12, 2943.

(8) Kojima, Y.; Fukumori, K.; Usuki, A.; Okada, A.; Kurauchi, T. J Mater Sci Lett
1993, 12,889.

(9) Yano, K.; Usuki, A.; Okada, A.; Kurauchi, T.; Kamigaito, O. J. Polym. Sci. Part
A: Polym. Chem. 1993, 31 , 2493.

(10) Fredrickson, G. H.; Bicerano, J. J Chem Phys 1999, 110, 2181.

(11) Cussler, E. L.; Hughes, S. E.; Ward, W. J.; Aris, R. J. Membrane Sci. 1988,
38, 161.

133

II"LillilllfllalllllllljljllllelllzllliI