Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Fiber reinforced hybrid phenolic foam
(USC Thesis Other)
Fiber reinforced hybrid phenolic foam
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
FIBER REINFORCED HYBRID PHENOLIC FOAM
by
Amit Desai
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(MATERIALS SCIENCE)
December 2008
Copyright 2008 Amit Desai
ii
Acknowledgments
First person I would like to give my gratitude is Dr. S.R. Nutt for being a
great mentor, friend, advisor and a teacher. For all I have achieved I would like to
thank Dr. Nutt for guiding me and providing me his invaluable advice as well as his
unshakable support through out my education at USC. He has helped me
significantly to develop professional skills and knowledge.
I would like to extend my appreciation to Dr. Edward Goo and Dr. Charles
Sammis for their agreement to serve on my PhD guidance committee and taking the
time and effort to evaluate my thesis.
Next, I would like to thank every person at the USC composite center where I
had the privilege of working with group of most wonderful colleagues. The
opportunity to collaborate my research with experts and scientists from all over the
world in particular Virginia Alonso and Maria Lujan Auad helped me increase my
knowledge in the fascinating world of materials science. In particular, Virginia
guided me immensely and was always prompt in coordinating efforts. I am also
indebted to Hongbin Shen as his expertise and suggestions were very useful for the
successful completion of this work. I am grateful to Mr. Warren Haby, the lab
manager at USC composite whose handyman expertise with got me through several
unforeseen challenges while conducting my experiments.
Last but not the least I would like to acknowledge my parents, my family in
the US and back home in India and my friends for their support and encouragement.
iii
Table of Contents
Acknowledgments ii
List of Tables vii
List of Figures viii
Abstract xi
Chapter 1: Introduction 1
1.1 Cellular materials 1
1.1.1 Synopsis 1
1.1.2 Applications 3
1.1.3 Plastic or polymer foam 4
1.1.3.1 Foam formation 7
1.1.3.2 Morphology 9
1.1.3.3 Foam mechanics 12
1.2 Phenolics and phenolic foam 15
1.2.1 Resole 15
1.2.2 Novalac 16
1.2.3 Phenolic foam 17
1.2.3.1 Novalac based phenolic foam 17
1.2.3.2 Resole based phenolic foam 19
1.3 Motivation 21
1.4 Toughening of phenolic foam 26
1.4.1 Physical and chemical modification 26
1.4.2 Fiber reinforcement 27
1.4.3 Hybrid concept 28
1.5 Scope of dissertation 30
Chapter 1 References 32
Chapter 2: Experimentation 35
2.1 Materials 35
2.1.1 Resins 35
2.1.2 Chopped fiber strand 35
2.1.3 Blowing agent 36
2.1.4 Curing agent 36
2.1.5 Surfactants 36
2.2 Synthesis 36
2.2.1 Plain phenolic foam 36
iv
2.2.2 Fiber reinforced phenolic foam 37
2.3 Material characterization tests 40
2.4 Design of experiments 40
Chapter 2 References 42
Chapter 3: Mechanical behavior of hybrid foams 43
3.1 Motivation 43
3.2 Experiments 46
3.2.1 Foam preparation 47
3.2.2 Mechanical tests 48
3.2.3 Morphology 48
3.3 Results and discussions 48
3.3.1 Compression test 68
3.3.1.1 Fiber type effects 49
3.3.1.2 Anisotropy 50
3.3.1.3 Fiber proportions 53
3.3.1.4 Compression stress strain relationship 53
3.3.1.5 Phenolics vs. Polyurethane 55
3.3.2 Shear test 56
3.3.2.1 Fiber type effects 56
3.3.2.2 Proportions 58
3.3.2.3 Shear stress strain relationships 59
3.3.2.4 Phenolics vs. Polyurethane 62
3.3.3 Theoretical modeling of foam properties 63
3.3.4 Foam structure 66
3.4 Conclusion 69
Chapter 3 References 71
Chapter 4: Modeling of fiber reinforced phenolic foams 74
4.1 Motivation 74
4.2 Experiment 77
4.2.1 Foam preparation 77
4.2.2 Statistical experimental design 77
4.2.3 Plan of experiments 82
4.2.4 Compression tests 84
4.2.5 Scanning electron microscopy 84
4.3 Results and discussions 85
4.3.1 Characterization of foams 85
4.3.1.1 Effect of fiber length on cell size 87
4.3.1.2 Effect of fiber weight fraction on cell size 88
4.3.1.3 Effect of blowing agent on cell size 90
4.3.2 Statistical model for glass fiber reinforced foam 90
4.3.2.1 Foams with 6mm, 9mm and 12 mm fibers 90
4.3.2.2 Foams with 3mm fibers 96
v
4.4 Conclusion 100
Chapter 4 References 102
Chapter 5: Modeling of hybrid foams 104
5.1 Motivation 104
5.2 Experiment 106
5.2.1 Foam preparation 106
5.2.2 Statistical experimental design 106
5.2.3 Plan of experiments 109
5.2.4 Compression tests 110
5.2.5 Shear tests 111
5.2.6 Scanning electron microscopy 111
5.3 Results and discussions 111
5.3.1 Statistical model for compressive properties 111
5.3.2 Statistical model for shear properties 115
5.3.3 Morphology of hybrid composite phenolic foams 119
5.4 Conclusion 122
Chapter 5 References 124
Chapter 6: Diffusivity and climatic simulations of hybrid foams 126
6.1 Motivation 126
6.2 Experiment 131
6.2.1 Materials and foam preparation 131
6.2.2 Moisture absorption and diffusivity model 131
6.2.3 Flammability test 133
6.2.4 Climatic simulation: Accelerated aging 134
6.2.5 Mechanical tests 135
6.2.6 Scanning electron microscopy 135
6.3 Results and discussions 136
6.3.1 Foam diffusivity and moisture absorption 136
6.3.2 Flammability 138
6.3.3 Accelerated aging 140
6.3.4 Mechanical performance 141
6.3.5 Scanning electron microscopy 143
6.4 Conclusion 146
Chapter 6 References 148
Chapter 7: Future studies 150
Chapter 7 References 153
Bibliography 154
vi
List of Tables
Table 1.1: Gibson-Ashby foam theory formulas 14
Table 1.2: Typical mechanical properties for novalac phenolic foam 19
(20
o
C)
Table 1.3: Flammability properties of various commercial foams 22
Table 1.4: Heat resistance of polymer foams 24
Table 2.1: Specifications of novalac resins 35
Table 3.1: Compressive properties of foams 50
Table 3.2: Shear properties of foam 57
Table 4.1: Matrix of two level factorial designs with one central point 79
(Experimental Design I)
Table 4.2: Matrix of two level factorial designs with one central point 80
(Experimental Design II)
Table 4.3: Foams with 6, 9 and 12 mm fibers 83
Table 4.4: Foams with 3mm fibers 84
Table 4.5: ANOVA for density in Experimental Design I 91
Table 4.6: ANOVA for density with significant effects 92
(Experimental Design I)
Table 4.7: ANOVA for Modulus (Experimental Design I) 93
Table 4.8: ANOVA for Modulus with significant effects 94
(Experimental Design I)
Table 4.9: ANOVA for strength (Experimental Design I) 95
Table 4.10: ANOVA for strength with significance effects 95
(Experimental Design I)
Table 4.11: ANOVA for density (Experimental design II) 96
vii
Table 4.12: ANOVA for density with significance effects 97
(Experimental Design II)
Table 4.13: ANOVA for modulus (Experimental design II) 98
Table 4.14: ANOVA for modulus with significance effects 99
(Experimental Design II)
Table 4.15: ANOVA for strength (Experimental design II) 99
Table 4.16: ANOVA for strength with significance effects 100
(Experimental Design II)
Table 5.1: Compressive property of hybrid foams with glass and aramid 110
fibers
Table 5.2: Shear property of hybrid foams with glass and aramid fibers 110
Table 5.3: Analysis of variance for compression modulus 112
Table 5.4: Analysis of variance for compressive strength 115
Table 5.5: Analysis of variance for shear modulus 116
Table 5.6: Analysis of variance for shear strength 118
Table 5.7: Average cell diameters for hybrid foams 122
Table 6.1: Heat loss distribution for heating a typical residence 126
Table 6.2: Heat loss distribution for cooling a typical residence 127
Table 6.3: Embodied energy of common insulation materials 128
Table 6.4: Moisture absorption test 132
Table 6.5: Flammability test results: UL 94 ratings assigned to materials 139
viii
List of Figures
Figure 1.1: Properties of cellular solids 2
Figure 1.2: Morphology of foam with different densities in a 2-d array 10
Figure 1.3: Formation of foam morphology in 2D 10
Figure 1.4: Polyhedral model of foam cell in 3D space 11
Figure 1.5: Gibson-Ashby models for a foam. (a) Open cell; 13
(b) Closed cell
Figure 1.6: Compressive and impact strength as function of rubber content 18
Figure 1.7: Burn test on foam roof panels (a) Phenolic foam during start 23
of experiment (b) Phenolic foam panel after 25 minutes under
flame (c) Expanded polystyrene after 12 minutes under flame
(d) End of Test
Figure 1.8: Fire on roof of Monte Carlo hotel Las Vegas caused due to 23
EPS foam
Figure 2.1: Hybrid mixer (a) and its working principle (b) 38
Figure 2.2: Comparison of mixing quality with different mixers 39
Figure 3.1: Typical compression stress-strain relationships of phenolic 54
foams: Loading direction is parallel to the foam rise direction
Figure 3.2: Typical compression stress-strain relationships of phenolic 55
Foams: Loading direction is parallel to the foam rise direction
Figure 3.3: Typical shear stress-strain relationships of phenolic foams. 60
Shear plane and loading direction are both parallel to the foam
rise direction.
Figure 3.4: Typical shear stress-strain relationships of phenolic foams. 61
Shear plane and loading direction are both parallel to the foam
rise direction
Figure 3.5: Comparison of Experimental results with various theoretical 65
models
ix
Figure 3.6: SEM images of glass fibers in matrix of “hybrid” foam 67
Figure 3.7: SEM images of aramid fibers in matrix of “hybrid” foam 68
Figure 4.1: Cube for 2 cube design 79
Figure 4.2: Cube for 2 square design 80
Figure 4.3: Variation of density with weight percentage of blowing agent 85
for unreinforced foam
Figure 4.4: Variation of (a) modulus and (b) compressive strength with 86
density of unreinforced foam
Figure 4.5: Effect of fiber length on cell size 87
Figure 4.6: Effect of fiber weight fraction on cell size 89
Figure 4.7: Response surface curve for 2
3
design 93
Figure 4.8: Response surface curves for 2
2
designs 98
Figure 5.1: 3 square schematic effect 108
Figure 5.2: (a) Contour map showing effects of fiber type on compressive 114
modulus.(b) 3-d plot showing effects of fiber type on
compressive modulus
Figure 5.3: (a) Contour map showing effects of fiber type on compressive 115
strength.(b) 3-d plot showing effects of fiber type on
compressive strength
Figure 5.4: (a) Contour map showing effects of fiber type on shear 117
modulus.(b) 3-d plot showing effects of fiber type on
shear modulus
Figure 5.5: (a) Contour map showing effects of fiber type on shear 119
strength.(b) 3-d plot showing effects of fiber type on
shear strength
Figure 5.6: SEM images of hybrid foam samples 120
Figure 6.1: UL 94 burn test 134
x
Figure 6.2: Plots for weight gain vs. square root of time 134
Figure 6.3: Evolution of moisture diffusivity for composite foams 138
a) 1 wt% aramid fibers b) Unreinforced foam c) 1wt % glass
fibers d) Hybrid foam – 1wt% aramid and 3wt% glass fibers
e) Hybrid foam – 3 wt% glass and 1wt% aramid
Figure 6.4: Flammability test results: Volume lost as a percentage of 139
initial volume
Figure 6.5: Accelerated aging results: Mass lost as a percentage of initial 140
mass
Figure 6.6: Climatic effects on compressive modulus 142
Figure 6.7: Climatic effects on compressive strength 143
Figure 6.8: SEM images showing cell lengths across climatic stress 145
Figure 6.9: SEM results: Cell length across climatic stress 146
Figure 7.1: Cross-section of steel stud cavity walls with phenolic 151
foam insulation between steel studs and concrete
Figure 7.1: Cross-section of steel stud cavity walls with phenolic 152
foam insulation between steel studs
xi
Abstract
Hybrid composites in recent times have been developed by using more than
one type of fiber reinforcement to bestow synergistic properties of the chosen filler
and matrix and also facilitating the design of materials with specific properties
matched to end use. However, the studies for hybrid foams have been very limited
because of problems related to fiber dispersion in matrix, non uniform mixing due to
presence of more than one filler and partially cured foams. An effective approach to
synthesize hybrid phenolic foam has been proposed and investigated here. Hybrid
composite phenolic foams were reinforced with chopped glass and aramid fibers in
varied proportions. On assessing mechanical properties in compression and shear
several interesting facts surfaced but overall hybrid phenolic foams exhibited a more
graceful failure, greater resistance to cracking and were significantly stiffer and
stronger than foams with only glass and aramid fibers. The optimum fiber ratio for
the reinforced hybrid phenolic foam system was found to be 1:1 ratio of glass to
aramid fibers. Also, the properties of hybrid foam were found to deviate from rule of
mixture (ROM) and thus the existing theories of fiber reinforcement fell short in
explaining their complex behavior.
In an attempt to describe and predict mechanical behavior of hybrid foams a
statistical design tool using analysis of variance technique was employed. The
utilization of a statistical model for predicting foam properties was found to be an
appropriate tool that affords a global perspective of the influence of process variables
such as fiber weight fraction, fiber length etc. on foam properties (elastic modulus
xii
and strength). Similar approach could be extended to study other fiber composite
foam systems such as polyurethane, epoxy etc. and doing so will reduce the number
of experimental iterations needed to optimize foam properties and identify critical
process variables.
Diffusivity, accelerated aging and flammability of hybrid foams were evaluated
and the results indicate that hybrid foam surpassed several commercial foams and
thus could fulfill the current needs for an insulation material which is low cost, has
excellent fire properties and retains compressive stiffness even after aging.
1
Chapter 1. Introduction
1.1 Cellular materials
1.1.1 Synopsis
Cellular solids are assembly of cells with solid edges or faces, packed
together so that they fill space. They have interconnected network of solid struts or
plates which form the edges and faces of cells. Natural cellular materials have been
used by man for centuries: wood, cork, sponge and coral are few examples of this
category. The use of cellular materials accounts for over thousand of years where
during Roman times cork was used for bungs in wine bottles. Cellular materials can
be broadly divided into two main categories: natural (bones, wood, sponge etc) and
synthetic (ceramic, metal and polymer foams).
The single most important feature of a cellular solid is its relative density, that
is the density of the cellular material ( ρ*), divided by that of the solid from which the
cell walls are made, ( ρ
s
). Fig. 1.1 shows range of properties of cellular materials:
Compressive strength, density, thermal conductivity and Young’s modulus [1]. The
bar with dotted shading shows the range of the property spanned by conventional
solids and the solid bar shows the extension of this range made possible by foaming.
This enormous extension of properties creates applications for foams which cannot
easily be filled by fully dense solids, and offers potential for engineering ingenuity.
The heterogeneous structure of cellular solids makes the job tougher for materials
scientist who tries to understand it. Primarily, cellular solids involve at least one
2
more phase than a solid. Secondly morphology is greatly complicated at the added
interfaces. They may be either closed cell where every gas bubble in foam is
encapsulated by solid matrix or open cells where the cells are interconnected with
each other.
Figure 1.1 Properties of cellular solids
Also, anisotropy may exist with respect to preferred direction such as the foaming
direction of open-molded polymer foam. Other effects such as additives (fillers and
3
fibers), matrix solid properties, composition, and relative density can add further
complexity in analyzing the system of cellular material. Most of these complex
factors are ultimately related to the performance of cellular solids and thus cannot be
avoided.
1.1.2 Applications
The four diagrams of Fig 1.1 relate directly to four major applications of
cellular materials: thermal insulation, packaging, structural use and buoyancy [1].
Primarily filled by air or other gas, cellular materials exhibit extremely low thermal
conductivity and thus are a popular choice for insulation applications. The
applications vary from as simple as disposable coffee cups to complex applications
such as insulation materials of booster rocket for the space shuttles. Because of their
low cost, fire resistance and light weight properties foam materials are extensively
used for insulation applications in the areas of construction and transportation.
Wood, cancellous bone and coral are all natural cellular solids and are
extensively used for structural application to support large static and cyclic loads.
Natural cellular materials have been used by man for a long time for structural
applications. The most common example is sandwich structures comprising of strong
thin face sheets “sandwiched” around a thick, low-density cellular material as the
core to support the skins. Balsa wood, foams and aluminum or paper-resin based
honeycombs are popular choice of core materials for sandwich structures. These
sandwich structures have excellent strength and stiffness and are used for
applications in skis, space vehicles, racing yachts etc where weight is critical.
4
Sandwich panels are found in nature, too: the skull is made up of two layers of
dense, compact bone separated by a lightweight core of spongy bone.
The third major application of cellular solid is in packaging. The strength of
foams can be adjusted over a wide range by controlling its relative density. These
foams can undergo large deformation at almost constant stress and thus they absorb a
large amount of energy without generating high stress. This property is extremely
useful for protecting fragile objects from impact damage during transportation.
Because of their closed cells, polymer foams are extensively used for
floatation structures and as floatation in boats. Foams are much more damage
tolerant than floatation bags or chamber and because of their closed cells they retain
buoyancy even when extensively damaged; do not rust or corrode and are unaffected
by extended immersion in water. Cellular materials have been used as the core of a
sandwich structure forming the deck and hull of the boat, providing structural
rigidity as well as buoyancy.
Overall, cellular materials possess many unique physical and mechanical
properties that enable a wide range of application in devices, structures and other day
to day common consumer products.
1.1.3 Plastic or polymer foams
Almost any solid can be foamed. Plastic or polymer foams also referred to as
expanded plastics, generally consist of minimum of two phases, a solid polymer
matrix and a gaseous phase derived from a blowing agent [2]. The solid phase is
continuous and load bearing, whereas the gas phase can be either discrete, as in
5
closed cell foams or continuous as in fully open cell foams. Other solid phases may
be present in form of fillers, either fibrous or other-shaped fillers which may be of
inorganic origin such as glass, ceramic or metallic or they may be polymeric in
nature.
Polymer foams can be further categorized in different ways depending on the
materials and properties involved. Foams may be flexible or rigid, depending on
whether their glass transition temperature is below or above room temperature [2].
Thermosetting foams are formed through an irrereversible chemical reaction(s) and
can not be further melted or dissolved. However thermoplastic foam is made from
molten plastic and it remains fusible and soluble. Plastic foams can be produced in a
variety of densities [3]: low density (<100 kg/m
3
), medium (100-400 kg/m
3
) and high
density (>400 kg/m
3
).
Below are some of the most common ways to synthesize polymer foams.
These foams can be manufactured by mechanical, chemical or physical means [2]:
• Mechanical whipping of gases (frothing) into a polymer system (melt,
solution or suspension) which hardens either by catalytic or heat or both, thus
entrapping the gas bubbles in the polymer matrix.
• Thermal decomposition of blowing agents generating either nitrogen or
carbon dioxide or both by application of heat, or as the result of the
exothermic reaction during polymerization.
• Expansion of dissolved gas in a polymer mass on reduction of pressure in the
system.
6
• Volatilization of gases produced by the exothermic heat of reaction during
polymerization such as occurs in the reaction of isocyanate with water to
form carbon dioxide.
• Volatilization of low boiling-point liquids, such as fluoro- or chloro-alkanes
or pure hydrocarbons, within the polymer mass as the result of the
exothermic heat of reaction, or by the application of heat.
• Expansion of gas filled beads by application of heat or expansion of these
beads in a polymer mass by the heat of reaction, e.g., expansion of
polystyrene beads in a epoxy resin system.
• Incorporation of hollow microspheres into a polymer mass. The microspheres
may consist of either hollow glass or hollow plastic beads.
Foams can be produced by several techniques as discussed here:
• Extrusion of foams using expandable beads or pellets.
• Compression molding of foams.
• Spraying of foams.
• Reaction injection molding (RIM), usually by impingement.
• Continuous slab-stock production by pouring or impingement, using multi-
component foam machines.
• Injection molding of expandable beads or pellets.
• Foaming-in-place by pouring from a dual-or multi-component head.
• Lamination of foams.
• Rotational casting of foams.
7
• Production of foam composites.
• Precipitation foam process where a polymer phase is formed by
polymerization or participation from a liquid which is latter allowed to
escape.
• Frothing of foams, either by introduction of air or a low boiling-point volatile
solvent.
1.1.3.1 Foam formation
The preparation of polymeric foams is a highly complicated process involving
formation of gas bubbles in a liquid system, followed by the growth and stabilization
of these bubbles as the viscosity of liquid polymer increases, resulting ultimately in
the solidification of the cellular resin matrix. The formation of internally blown
foams takes place in several stages. In the first stage the blowing agent generates a
gas in solution in the liquid phase until the gas reaches a saturation limit in solution,
and becomes supersaturated. The gas finally comes out of solution in the form of
bubbles. The formation of bubbles represents a nucleation process since a new phase
is formed. The presence of a second phase may act as a nucleating agent [4].
In order to disperse a given volume of gas in a unit volume of liquid, the free
energy ( ΔF) of the system should increase by an amount of energy as a function of
surface tension ( γ) and total interfacial area (A)
ΔF = γA 1.1
When the surface tension of the liquid is lowered, either by heat or by the addition of
a surfactant, the free-energy increase associated with the dispersion of the gas will be
8
reduced and will aid the development of fine cells which correspond to a large value
of A.
The gas pressure in the spherical bubble is greater than the pressure in the
surrounding liquid by a difference Δp, as shown in the equation below as a function
of radius R of the bubble:
Δp = 2γ / R 1.2
Hence, the gas pressure in a small bubble is greater than that in a large bubble.
For two bubbles of radii R1 and R2, the difference in pressure Δp
2
, is given by the
equation:
Δp = 2γ [ (1/R1) – (1/R2)] 1.3
Thus, in a liquid system, a diffusion of gas takes place the small bubbles from small
bubbles into large bubbles, resulting in disappearance of small bubbles, while the
large bubbles grow in size with time. It is also obvious that low values of γ, by
adding surface-tension depressant reduces the pressure differences between bubbles
of different sizes and hence lead to better bubble stability. Thus choice of surfactant
is a very critical part for making polymer foams.
Another factor which affects the bubble stability is temperature, since an
increase in temperature reduces both surface tension and viscosity, which results in
thinning of cell membrane and may promote cell rupture.
Another factor which is critical for cell stability is the drainage of the liquid in
the bubble walls which is due to gravity and capillary action. This can be retarded by
an increase in viscosity at the film surface. The proper balance between viscosity and
9
gas evolution at the surface can be controlled by suitable choice and concentration of
catalyst and surfactant, presence of nucleating agent, choice of chemical blowing
agent depending whether the foam is thermosetting or thermoplastic.
1.1.3.2 Morphology
Cellular polymer result from the nucleation, growth and expansion of gas
bubbles in a melt or reaction liquid system. Fig 1-2 represents idealized structures at
different stages of the expansion process. Large differences in the resultant foam cell
morphology are expected along with varied foam densities, if gas bubbles stabilize at
different stages as shown. For foams at low densities (cases (c) – (e) in Fig. 1-2),
foam cells have expanded to such an extent that they are touching each other and
their shape is polygonal. Fig. 1-3 provides another opportunity to study such
processes in two dimensions [3]. In real space, this shape distortion will produce a
polyhedron, as illustrated by Fig. 1-4. From the consideration of thermodynamic
stability, ideal space-filling is achieved by cells which have the shape of a β-
tetrakaidecahedron [5], which on average has 14 faces in each cell. 5.143 sides for
each face and an angle of 109
0
28'16'' between any two joint sides. In real polymer
foams, the cell shape is far from perfect, but statistical study of low density
polyurethane foam confirmed that the most probable geometries of foam cells were
similar to the β-tetrakaidecahedron [6].
In addition to cell shape, other morphological parameters have profound
effects on foam properties and thus have to be characterized individually [6]. The
10
prominent ones include cell size and distribution, anisotropy cell orientation,
proportion of closed cells and material distribution between cell struts and walls.
Figure 1.2 Morphology of foam with different densities in a 2-d array [3]
Figure 1.3 Formation of foam morphology in 2D [3]
11
Figure 1.4 Polyhedral model of foam cell in 3D space [1]
However, morphological characterization is a tedious task especially for
quantitative purposes. Several factors complicate measurements on foam cells. The
most difficult part, in general is to observe and determine the spatial configurations
of the cells without damaging them. As a result, the majority of reported
morphological data for foams are of cell size and distributions, measured from
sectional surfaces of foam specimens. The conventional approaches used to obtain
morphological information on foams materials and the associated problems have
been summarized by Hilyard and Cunningham [6].
12
1.1.3.3 Foam Mechanics
Several attempts have been made to model and analyze the mechanical
behavior of foams. Gent and Thomas [7] proposed one of the earliest theories to
relate the Young’s modulus E
f
of open-cell foam to the volume fraction of solid
material, Φ [7]. In this model, foam structure was idealized as a collection of thin
elastic threads of material that connects spherical joints which do not deform. It was
assumed that the motion of thread ends was determined entirely by the macroscopic
deformation and the total strain energy for the foam was obtained by summing the
individual contributions from randomly oriented threads, which deform only by
stretching. For low density foams where Φ is small, such a relationship is obtained,
E
f
/E = Φ/6, where E is the Young’s modulus of the solid connecting material.
Despite the popularity of this theory and its apparent success in correlating some
data, it is important to recognize that affine displacement is a kinematical condition
that does not ensure a balance of forces within the structure. Furthermore, when the
mechanics of the strut junction is neglected, mechanisms associated with the strut
bending are suppressed. These shortcomings severely limit the usefulness of this
approach.
In one of the other studies Ko [8] studied the relationship between network
geometry, strut deformation mechanisms, and stiffness of low-density open cell
foams, using closed-packed spheres as an idealized model. It was found that E
f
α Φ
2
for face centered cubic packing (FCC). The different results reflect a change in the
dominant deformation mechanisms: stretching leads to the former dependence,
13
whereas bending leads to the latter. However, Ko’s models involved junctions with
six or more struts, which are not found in real forms. Menges and Knipchild [9]
observed the bending behavior of foam struts through optical microscopy, and
closely studied the elastic response of a microstructural element- four identical struts
that meet at equal tetrahedral angles – under both stretching and bending forces.
Their simple analysis gave E
f
α Φ
2
, where Φ is small. Kraynik and Warren [10] based
on the same model detailed the mechanical considerations and derived the
relationships for general situations.
Some of the more popular work was carried out by Gibson and Ashby [11].
They accepted bending as the essential strut deformation mechanism and used
dimensional analysis to obtain E
f
α Φ
2
. Taking advantage of their intensive studies on
honeycomb structures, they proposed a simple cubic foam cell model and readily
applied the honeycomb analysis results to a foam structure. Most of their derivations
and analytical formulae are straightforward and simple in form, and therefore have
become the most popular choice for modern foam researchers. Here a brief summary
of the theory is given.
Figure 1.5 Gibson-Ashby models for a foam. (a) Open cell; (b) Closed cell [1]
14
Fig. 1-5 illustrates the Gibson-Ashby foam cell models, in which the only
difference between a closed cell and an open cell is the thin membrane on each face.
When the membrane is very thin, as in most low-density foams, the difference is
almost negligible. Table 1-1 lists the formulas and separate consideration is given to
each case. Many physical parameters such as density,
moduli etc are expressed in relative values, and this expression easily stressed the
strong dependence of foam properties on the corresponding properties of the solid
material.
Table 1-1 Gibson-Ashby foam theory formulas [1]
Notes: s refers to the solid whereas f refers to the foam. E-Young’s modulus; ρ-
density; Ø- fraction of solid material distributed in cell edges; ν-Poisson’s ration;
C,C’-proportionality constants; σ
y
-yield strength; K
IC
-mode I fracture toughness; l-
average cell size
15
1.2 Phenolics and phenolic foam
Phenolic are produced by polycondensation reaction between phenols and
formalin (40% aqueous solution of formaldehyde). The trifunctional phenol reacts
with the difunctional formaldehyde, resulting in three dimensional matrix when the
reaction is carried out beyond the gel point. After the gel point, further curing results
in thermoset. The ready to be used resin is processed to just before the gel point
during the polymerization process. Based on different reactant ration and catalyst,
the two types of phenolic resin result in very different physical and chemical
attributes as described in the next section.
1.2.1 Resole
Resole phenol-formaldehyde resin is a water soluble methylol (CH
2
OH)
bearing thermoplastic. The curing process to the final thermoset material can be
initiated by just heating the resole in a mold above its gel point. The resole resin
formed has reactive methylol and hydroxyl groups [12]. When heated, resoles form
larger molecules with methylene crosslinks without the use or addition of a curing
agent. The phenol-formaldehyde resin resinification reaction is a typical
polycondensation reaction since water is given off as a byproduct which may cause
annoying effects in phenolic products [13].
Typical resole resin comprises a mixture of monomers, dimmers, trimers and
small amounts of higher-degree oligomers. Due to this, the average molecular
weight of resole is small, typically a few hundred g/mol. Usually a commercial
resole resin is supplied in the form of concentrated aqueous solution, with the typical
16
viscosity range between 1000-6000 cP at room temperature and a solid content of
60-80 %. Because the average functionality of resole is greater than 2, there is no
need for an additional curing agent to solidify a resole phenolic. Resole in fact
experiences slow curing even at room temperature, so they are stored close to 0
o
C to
arrest this reactivity. While a resole can cure by itself over time at elevated
temperatures, an acidic catalyst is usually added to gain control of the curing speed.
The curing reaction is exothermic, and this attribute is useful when heat is needed to
evaporate the blowing agent for making foams [2].
1.2.2 Novalac
The polymerization of phenol-formaldehyde novalac resin is carried out in the
presence of an acid catalyst such as oxalic acid, sulfuric acid, hydrochloric acid,
formic acid and aromatic sulfuric acid. The gel point of the cure is deliberately
delayed by using a phenol-formaldehyde feedstock ratio of 1:0.8 in the first stage
(prepolymerization). Polymerization is carried out at 160
o
C by heating the mixture
for 2 to 4 hours at reflux. The novalac phenolic resins that are obtained have a shelf
life on infinite under normal storage conditions. More details are available in
literature [2].
The chemical structure of novalac phenolic is widely agreed to be linear,
consisting of hydroxyl benzenes bridged by methylene segments. Due to its
relatively high molecular weight, novalac normally is a solid with a melting point
around 110
o
C. The curing reaction for novalac is exothermic just like resole and this
17
attribute is useful in when heat is needed to evaporate the blowing agent for making
foams [2].
1.2.3 Phenolic foam
Phenolic foam can by synthesized by incorporating phenolic resin with
several additional chemicals. First, a cross linking agent is required for novalac or an
acid catalyst for resole phenolic. Second, a surfactant is necessary to stabilize the gas
bubbles and refining the cell shape and size. Third, a blowing agent supplies gas to
expand the foam and the concentration determines to a large extent the final density
of foam product. In addition, other additives such as performance modifiers or
processing agents may also be necessary. The details of chemistry involved for
resole and novalac resin are covered in the next few sections. Resole is a more
popular choice of resin for foam manufacturing because of its ease of processing.
1.2.3.1 Novalac based phenolic foam
Novalac based phenolic foams are more popular in the countries of the former
Soviet Union. Because of the brittle and fragile nature of novalac based foam,
acrylonitrile rubber is added to increase the impact strength [14]. The nitrile rubber is
a plasticizer and an integral part of the polymer structure (grafted copolymers).
Increasing the rubber content usually results in decrease in rigidity, thermal stability
and compressive strength. Fig. 1-6 illustrates compressive strength and impact
strengths as functions of the rubber content in phenolic foams.
In general, a typical formulation of expanded phenolic and phenolic-nitrile
foams [3] consist of a novalac phenolic resin, acrylonitrile rubber, HMTA, sulfur
18
(vulcanizer for rubber), and azodiisobutyronitrile rubber ( AIBN, blowing agent).
The typical process of making such foams is comprised of two steps: (a) preparation
of mixture and (b) charging the mixtures to molds with subsequent foaming and
hardening of the composition. The optimal processing temperature is reported below
150
o
C, and at higher temperature the foam properties deteriorate as thermal oxidative
degradation begins at 180-200
o
C.
Figure 1.6 Compressive and impact strength as function of rubber content [3]
Typical properties of novalac phenolic foam are listed in Table1-2. The
strength properties of the phenolic foam with density of 100 kg/m
3
are lower than
those of polystyrene and rigid PVC foam, but similar to polyurethane foam.
19
However, with respect to thermal stability, the phenolic foam exceeds all other
foams, retaining 50% of its strength after heating at 200
o
C for 5hr in atmospheric
oxygen. It was found in another study that this property of retention could be
improved up to the 350-400
o
C range if aluminum powder was added as the co-
blowing agent. However, the novalac based phenolic foams exhibited significant
shrinkage when subject to high temperature (above 80
o
C) in air for a short period of
time.
Table 1-2 Typical mechanical properties for novalac phenolic foams (20
o
C) [3]
1.2.3.2 Resole based phenolic foam
The viscosity range of the commercial phenolic resins that is suitable for
making foams is about 2,000-6,500 cP at room temperature. That are certain
requirements that phenolic resole must meet in order to be foamable and different
composition of resin may also be needed to achieve density varieties. In addition to
the varied P/F ratios, the method of resin preparation will influence the degree of
condensation and pH level, and this in turn will greatly affect the performance of
resin in foams.
20
Strong acids are generally employed to cure resole resin for foaming as heat-
curing alone requires long cure cycles. The catalyst works basically as an instigator
of the exothermic reaction. The heat generated by the condensation reaction is the
basis of foaming. Weak acids do not provide sufficient heat to form anything other
than dense foams. The most commonly used catalysts are hydrochloric, sulfuric,
benzenesulfonic, xylenesulfonic, toluenesulfonic, phenolsulfonic and phosphoric
acid. Many other inorganic and organic compounds and their combinations have
been reported in preparing phenolic foams [15], but they are mostly used to serve
special purposes.
Different kind of water-soluble nonhydrolyzable, acid-stable surfactants can be
used in phenolic foamable compositions. The nonionic types are the most commonly
employed agents. Particularly useful are siloxane-oxyalkylene copolymers [16] and
polyoxythylene sorbitan fatty acid esters [17]. The use of these agents has a
significant effect on the cured foam’s compressive strength, cell size and cell
uniformity. However, unlike some polymer foam systems, the use of surfactants with
phenolic resins does not appear to have a gross effect on the closed-cell content of
the resultant foams [17].
Phenolic foams of controlled densities are obtained by adding some volatile
liquids and solid blowing agents into the foam mixture. Polyhalogentated saturated
fluorocarbons with a boiling point between 30 and 100
o
C were preferred in early
phenolic foam formulations [18] and these foams had excellent fire retardant
properties and thermal insulation. However, due to the ozone-depletion effect of
21
chlorofluorocarbon (CFC) compounds, their use in modern industries is now largely
prohibited [2]. In modern days, volatile hydrocarbons such as pentane, hexane,
cyclohexane and heptane are widely accepted as substitutes. Meanwhile, chemical
blowing agents that release gas upon reaction with the acid catalyst are also used as
alternatives, and these include a number of carbonate or nitrite salts and a variety of
organic compounds.
There are several ways to prepare phenolic foams and broadly they can be
classified into spraying, continuous production and batch molding [15]. Each method
has to be matched with appropriate formulations that allow easy processing as well
as quality products. Several of the reported fabrication methods and chemical
compositions are employed for specific commercial applications of phenolic foams,
and many of these methods and formulations are patented.
1.3 Motivation
Phenolic foams have several significant and useful properties which makes
them an ideal candidate for several commercial and industrial applications. One of
the unique features of phenolic foam is its excellent non flammability. The
flammability properties of phenolic foam have been compared to other commercial
foams in Table 1-3. Phenolic foams also possess low smoke toxicity, no dripping
during combustion, low smoke density in addition to the fire resistance. Fig 1-7
shows a fire test performed on phenolic and expanded polystyrene foam (EPS) that
demonstrates excellent fire, smoke and toxicity (FST) properties of phenolic foams
(courtesy: Kingspan UK) compared to other polymer foams.
22
Table 1-3 Flammability properties of various commercial foams
The recent fire at the Monte Carlo Hotel in Las Vegas on January 2008 was
caused when the polystyrene foam used for architectural and decorative purposes
caught fire (Fig. 1.8). This fire costed hotel millions of dollars and several people
were treated for smoke inhalation. Several such accidents have been reported
worldwide caused because of foam catching fire and causing loss of lives and
property.
Phenolic foam is also very cost effective. A pound of foamable phenolic resin
costs around $1.50-$2.00 and are thus economically competitive. Thirdly, phenolic
foams are extremely stable over a wide range of temperature. They have shown no
change in performance even under -196
o
C and remains stable upto 200
o
C. Table 1-4
shows thermal stability of phenolic foam in comparison to other polymer foams.
Phenolic foams also exhibit excellent resistance to common industrial solvents
and liquids. The low thermal conductivity (k) of phenolic foam makes it an excellent
choice for insulation applications.
23
Figure 1.7: Burn test on foam roof panels (a) Phenolic foam during start of experiment
(b) Phenolic foam panel after 25 minutes under flame (c) Expanded polystyrene after
12 minutes under flame (d) End of Test
Figure 1.8: Fire on roof of Monte Carlo hotel Las Vegas caused due to EPS foam
24
Table 1-4 Heat resistance of polymer foams
Germans extensively used phenolic foam in aircrafts during World War II.
Because of its low cost phenolic foams has also become popular roof insulation
material. Phenolic foam in recent times has also been used as core materials in
sandwich structures for construction supplies. Fire safety is becoming an issue of
immense attention in the US especially in areas of aircrafts, building materials and
marine vessels, and conventional insulation materials such as EPS and polyurethane
foam (PU) will soon be replaced by highly fire resistant alternatives. Phenolic foam
because of its exceptional fire properties is a prime candidate for related applications.
However, phenolic foam has some shortcomings and that have severely
limited its structural applications. Phenolic foams are extremely friable, particularly
at lower density and this has led to serious problems such as debonding in sandwich
structures, possible damage during handling and dust pollution in the workspace.
Secondly, phenolic foams are corrosive due to the residues of acid catalyst.
However, significant progress has been made to improve these aspects of phenolic
foam that is discussed in further sections.
25
In general, the high friability of these foams arises from the brittle nature of
phenolic structure [2]. The high functionality and inherent structural rigidity of
phenolic resin produces a highly cross linked network with significant defects. The
water released during the condensation reaction results in high void content. These
voids constitute sites for crack nucleation and growth. Because of its extreme brittle
and friable nature, phenolic foam has not been considered for most structural
applications in which load bearing capacity is of primary concern. Some success has
been reported in making sandwich panels with phenolic foam core, severe problems
are encountered during production and service. Problem have been encountered
while bonding the foam to other materials, insufficient strength during bonding
core/skin surfaces, fragility during handling and machining of panels and dust
pollution in the work place area. The applications of phenolic foam have been
significantly limited due to these drawbacks.
Over the last few years, significant progress has been made to improve the
mechanical performance of phenolic foam by reinforcing them with chopped fibers
[20]. Several fold improvement in compression, shear, peel and tension properties
has been reported by using fiber reinforcements. This study was motivated by a need
to make phenolic foam stronger and more suitable for insulation applications and
also as a structural member in buildings by using the Hybrid concept i.e. reinforcing
foam with more than one type of fiber.
26
1.4 Toughening of Phenolic foam
1.4.1 Physical and Chemical modification
The toughening of phenolic foam has been classified into two main
categories: chemical modification and non-reactive approaches.
Several approaches have been performed to achieve chemical modification.
Flexible polymer segments are introduced into the phenolic backbone structure
during the preparation of resole resin. Some of the modifiers included are polyvinyl
alcohol [21], resorcinol [22], m-cresol [23], polyols [24] and isocyanates [25].
Sometimes a second polymer is added that is compatible with the resoles and capable
of being co-cured with the resole under the influence of the acid catalyst [26].
Improvements have been obtained by implementing this type of modification.
Physical strengthening involves adding high molecular weight polymers or
inert fillers. Polymer blends that have been used are polyesters [27], polyamides [28]
and styrene copolymers [29]. Also, ground fillers such as talc, mica, and cork flour
have shown to improve the texture and homogeneity of foams but increased density
of foams.
Efforts have been guided to reduce corrosiveness that is caused to due to acid
catalyst present in phenolic foam. Some metals and their oxides or salts, such as
CaO, zinc, aluminum, Na
2
CO
3
were added to neutralize foams. Phenolic foam has
also been exposed to ammonia for 24-36 hours to neutralize the acid residues in
foam [30].
27
1.4.2 Fiber reinforcement
Fiber reinforcement has been very popular with composite materials and
significant improvement of properties has been attained. Fiber reinforcement was
first applied to polyurethane foam, resulting in improved strength and stiffness [31].
After reaction injection molding (RIM) was successfully applied to the manufacture
of short fiber-reinforced composite materials, significant interest was drawn to
manufacture of short fiber-reinforced composite materials [32-34].
In one of the studies, non-woven glass fiber was impregnated with foamable
phenolic resin and then expanded upon heating, but this technique suffers from the
disadvantage that it is virtually impossible to impregnate all the spaces between the
fibers with the viscous resin, thus resulting in relatively low mechanical strength
[35].
Researchers have also used 3-D woven glass performs in an effort to obtain
isotropic reinforced foams [36] and, also foam core was stitched to face sheets with
continuous glass fiber to prevent delamination of sandwich structures [37]. The
manufacturing costs associated with such an approach are high and thus limiting its
physical applicability.
Almost seven fold improvement was reported in peel resistance of phenolic
foam reinforced with aramid fibers [38]. These foams reinforced with aramid fibers
were also found to exhibit better fracture toughness as compared to unreinforced
foams. Aramid fibers also reduced the mass loss of phenolic foam during the
friability test to five time lower level. Overall these aramid fiber reinforced phenolic
28
foams exhibited several fold improvement in friability and shear properties as
compared to unreinforced foam [39]. Phenolic foams reinforced with glass fibers
produced substantial improvement in the stiffness and strength as compared to
conventional counterpart. For loadings applied in the foam rise direction, the
Young’s modulus of glass fiber composite foam was found to be twice as much as of
the plain phenolic foam. A non-destructive technique (NDT) using Micro-CT when
applied to study foam architecture and fiber morphology revealed that glass fibers
tend to orient in the foam rise direction and thus explained the increase in stiffness of
composite phenolic foam reinforced with glass fibers [40].
1.4.3 Hybrid concept
Hybrid composites in recent times have been developed by using more than
one type, shape or size of reinforcement. These composites have been developed to
bestow synergistic properties of the chosen fillers and matrix. The study of these
hybrid composites is relatively new. The development of composites containing
more than one type of fiber reinforcement (hybrid composites) is motivated by the
ability to combine advantageous features of various fiber types, including improved
performance as well as reduced weight and cost. Understanding the mechanical
properties of hybrid composites is essential in order to optimize the design of new
hybrid materials [41].
Hybrid composites are gaining commercial significance for several reasons.
First, a wider spectrum of tailor-made physical and mechanical properties is possible,
thus facilitating the design of materials with specific properties matched to an end
29
use. Secondly, there are economic advantages replacing a more expensive
reinforcement or filler with cheaper materials. Thirdly, hybrids can achieve
synergistic effects and improvements in mechanical and functional properties [42].
The properties of hybrid composites depend on several factors, including the
interaction of fillers with the polymeric matrix, shape and size (aspect ratio) of
fillers, and orientation of fillers, to name a few. For example, hybridizing glass fiber
composites with carbon fiber is known to enhance fatigue performance and
environmental resistance compared to all glass composites [43]. The judicious
selection of banana and sisal fibers has been useful in developing value-added and
cost-effective hybrid composites having high tensile and flexural properties [44].
Although, success has been reported in developing hybrid composites, not
much progress has been achieved to develop hybrid foams reinforced with more than
one type of filler or reinforcement. The reason being addition of two or more fillers
or additives causes significant technical problems related to mixing, foam
morphology, incomplete foaming, increased viscosity and partial curing of resin.
Also for the hybrid foams developed so far, it was found that by adding more than
one filler, improvement was observed in one aspect or property of hybrid foam but
compromised on other properties significantly thus limiting the applicability of these
foams.
In one of the studies to develop hybrid foams, rubber particles obtained from
environmentally hazardous waste tires were incorporated in glass-microballon-
epoxy composites. Enhancement in ductility was achieved in hybrid foams without
30
significant loss in strength. However, compression modulus was found to decrease
significantly and thus limiting the application of these hybrid foams for several
applications [45].
In summary, hybrid foams have displayed substantial potential of improving
mechanical performances. However, due to technical difficulties, the reports of
hybrid foams are far from satisfactory in terms of practicality and economy.
Furthermore, very little effort and studies have been devoted to developing a
comprehensive knowledge of fiber reinforced hybrid phenolic foam.
1.5 Scope of Dissertation
The studies in this work were focused in scientific, academic and applied
interests.
The study was intended to explore and develop hybrid composite phenolic
foams reinforced with glass and aramid fibers. One of the aims of this study was
intended to explore and develop the potential of hybrid composite phenolic foam for
insulation applications in buildings and possibly as a structural member. The present
constitutes a search for and exploration on approaches to develop hybrid foams that
show significant improvement in properties compared to unreinforced counterpart
and other foams such as PU, EPS used in commercial applications.
Hybrid foam developed here shows significant promise for several purposes.
The hybrid foam retains toughness and stiffness, with significantly enhancement of
strength. Also, these properties can be tailored to specific requirements for particular
applications simply by varying the fiber proportions. These results indicate that
31
judicious selection of fiber reinforcements in optimal proportions can yield structural
foams that are strong, tough, fire-retardant, and cost competitive for a variety of
structural applications.
Moreover, introducing more than one type of fibers in hybrid foams revealed
unique mechanical properties in compression and shear. Several existing models
were evaluated to study the behavior of hybrid foams. However, none of these
models accurately fit the experimental data, indicating that the behavior of hybrid
composite foams involves mechanisms more complex than these models can
represent. So, the next part of this study was dedicated to develop a statistical model
to describe the effects of fillers to study the responses of modulus and compressive
strength. The statistical design approached employed in this study was found to
predict mechanical properties of hybrid foams, reduced the number of experimental
iterations, identified critical process variables and could be applied to study other
hybrid foam systems. Studying these hybrid phenolic foam systems advanced our
understanding of basic structure-property relationships for composite foams.
Finally, we performed a study to evaluate climatic simulation and diffusivity
of hybrid foams. We have developed a model to predict diffusion in hybrid foams
using fick’s second law of diffusion. The foams were accelerated aged in an
envoirmental chamber to study the impact of envoirment on foams over an extended
period of application. The data for hybrid foam was compared to conventional
polyurethane and EPS foams.
32
Chapter 1 References
[1] Gibson LJ, Ashby MF. Cellular solids: structure and properties. UK: Cambridge
University Press, 1997
[2] Landrock AH (ed.) Handbook of plastic foams. New Jersey: Noyes Publications,
1995
[3] Benning CJ. Plastic foams: Volume II: structure properties, and applications.
New York: John Wiley and sons 1969
[4] Saunder JH, Hansen RH. Plastic foams. Volume I. New York, 1972
[5] Williams RE. Space filling polyhedron: its relation to aggregates of bubbles,
plant cells and metal crystallites. Science 1964:161:276-7
[6] Hilyard NC Cunningham A (ed.) Low density cellular plastics: physical basis od
behavior. New York: Chapman & Hall 1994
[7] Gent AN, Thomas AG Mechanics of foamed elastic materials. Rubber Chem
Technology. 1963:36:597-610
[8] Ko WL. Deformation of foamed elastomers. J Cellular Plastics. 1965:1:45-50
[9] Menges G, Knipschild F. Estimation of mechanical properties for rigid
polyurethane foams. Polymer Engr Sci. 1975:15:623-627
[10] Warren WE, Kraynik AM. Foam mechanics: the linear elastic behavior of
polydisperse response of two-dimensional spatially periodic cellular materials.
Mechanics Mat 1987:6:27-37
[11] Gibson LJ, Ashby MF. The mechanics of three-dimensional cellular materials.
Proc Royal Soc London a 1982:382:43-59
[12] Goodman SH, Handbook of Thermoplastics. Second edition: 1998
[13] Lin-Gibson S. PhD thesis. Virginia Polytechnic Institute, 2001
[14] Moiseyev AA, Popov VA, Borodin MY. (eds.) Expanded plastics. New York:
Macmillian, 1963
[15] Knop A, Scheip W. Chemistry and application of phenolic resins. New York:
Springer-Verlag, 1979
33
[16] USP 3,271,331 1966
[17] USP 3,300,419 1967
[18] USP 3,389,094 1968
[19] Mao J, Chang J, Chen Y, Fang D. Chem Ind Engr 1998:15(3):38-43
[20] USP 6,841,584 2005
[21] USP 2,728,741 1955
[22] USP 2,582,228 1952
[23] British Pat. 586, 199 1947
[24] USP 2,772,246 1956
[25] USP 2,772,246 1956
[26] British Pat. 1,077,423 1965
[27] USP 2,895,173 1959
[28] USP 2,789,054 1957
[29] USP 2,678,298 1954
[30] WO 94-04604 1994
[31] USP 4,163,824 1979
[32] Morimoto K, Suzuki T, Yosomiya R. Polym Engr Sci. 1984:24(12):943; 1000
[33] Morimoto K, Suzuki T. Ind Engr Prod Res Dev 1984:23:81
[34] Yosomiya R, Morimoto K. Ind Engr Chem Prod Res Dev 1984:23:605
[35] JP 2001-228835
[36] USP 4,900,616 1990
[37] JP 03-086506 1991
34
[38] Shen H., Lavoie A.J. and Nutt S.R. Enhanced peel resistance of fiber reinforced
phenolic foams, Comp. A: Appl. Sci. Manufact. 2003: 34(10): 941-948
[39] Shen H. and Nutt S.R. Mechanical characterization of short fiber reinforced
phenolic foam, Comp. Part A. 2003:34(9): 899-906
[40] Shen H ., Nutt S.R. and Hull D. Direct observation and measurement of fiber
architecture in short fiber-polymer composite foam through micro-CT
imaging, Comp. Sci. and Tech. 2004:(13-14):2113-2120
[41] Chiang Martin Y.M., Wang X., Schultheisz, C.R. and He J. Prediction and
three-dimensional Monte-Carlo simulation for tensile properties of unidirectional
hybrid composites, Comp. Sci. and Tech. 2005:65 (11-12): 1719-1727
[42] Babu Prem, E.J., Savithri S. and Pillai U.T.S, Pai B.C. Micromechanical
modeling of hybrid composites. 2005: Polymer, 46 (18):7478-7484.
[43] Shan Y. and Liao K. Environmental fatigue behavior and life prediction of
unidirectional glass–carbon/epoxy hybrid composites. Int. J. Fatigue. 2005: (24):847-
859.
[44] Idicula M., Neelakantan N.R., Oommen Z., Joseph K. and Thomas S. A study of
the mechanical properties of randomly oriented short banana and sisal hybrid fiber
reinforced polyester composites, J of App. Poly. Science. 2005: 96 (5):1699-1709.
[45] Gupta N., Maharsia R., Jerro H.D. Enhancement of energy absorption
characteristics of hollow glass particle filled by rubber addition, Mate. Sci and Engg.
A. 2005: 395 (1, 2):233-240
35
Chapter 2. Experimentation
2.1 Materials
2.1.1 Resins
Resole phenolic. HRJ-14489, supplied by Schenectady International Inc.,
contains solid content of 82.6% free water 5.4%. At room temperature, the resin has
Brookfield viscosity of 5796 centi-Poise (cP).Normally it is stored in a refrigerator
with temperature preset to 40
0
F (4
0
C) for longer lifetime. Prior to use for making
foams, the cooled resin is placed under ambient condition for at least one hour.
Novalac phenolic. GP2074 (Batch#344) and GP2056 (Batch#264, epoxy
hardener), both in solid state, were provided by Georgia Pacific Resins, Inc. The
resin was used as received. Specifications for the two resins are listed in Table 2-1.
2.1.2 Chopped fiber strand
Fiber glass. Four length varieties: 3.2 mm, 6.4 mm, 9mm and 12 mm with
11μm in diameter. Supplied by Lauscha Fiber International and precoated with
compatible sizings.
36
Aramid fibers. Chopped Nomex® product of DuPont with length of 6.4 mm
(1/4 in.) and diameter: ~12 μm. Used as received.
2.1.3 Blowing agent
n-Pentane. 99% in purity with b.p.of 40
0
C. Aldrich Chemical Co.
Nonane, 99% in purity with b.p.of 151
0
C. Aldrich Chemical Co.
2.1.4 Curing agent
PSA (phenolsulfonic acid). Saturated aqueous solution. CRC-608, supplied by
Capital Resin Corp., Ohio
2.1.5 Surfactants
Pel-stab. Supplied by Peltron Corp., France.
Dabco. DC193, supplied by Air Products and Chemical Inc., Pennsylvania.
2.2 Synthesis
2.2.1 Plain phenolic foam
The process of preparing conventional phenolic foam on a large scale is
detailed in [30]. With appropriate choice of phenolic resin, acid catalyst, blowing
agent and surfactant, the key to success lies in the fast, efficient and thorough mixing
of all the components. For a lab-scale foam fabrication, a table drill press was
converted into a lab-scale mixer with adjustable rotation up to 2000rpm. The mixing
container was usually a 1000-mililiter polypropylene beaker, which was reusable.
Different molds were tried, but the most useful one was built of aluminum plates and
lined with plastic film. This mold was used and heated in a 65
0
C Fisher oven to
37
preserve the exothermic foaming. The mixing procedure consisted of3 steps. First,
phenolic resin and surfactants were weighed into the plastic beaker and mixed
thoroughly. Next, a pre-weighed blowing agent (pentane) was added to the mixture,
followed by mild stirring to avoid splash. When there was no more free liquid visible
in the creamy mixture, a pre-weighed amount of acid catalyst was poured in and
vigorous mixing was maintained for about 1 minute. The contents of the plastic
container were quickly maintained for about 1 minute. The contents of the plastic
container were quickly transferred to a preheated mold, and the mold was replaced in
the oven immediately. After about 1 hour, the foam was completely set, depending
on the size of casting. After two hours, the foam was safe to remove from the mold,
ready for neutralization. A paint pot (max. pressure 25psi) was used as the
neutralizing chamber, where the fresh foam slabs were submerged in an ammonia
atmosphere for more than 48 hours. The ammonia was supplied either through
external tank or by the vaporization of ammonium hydroxide solution under slight
heat. The neutralized phenolic foam had a yellowish color, and was subject to 2-3
days of further drying in a ventilated chamber before it was tested.
2.2.2 Fiber reinforced phenolic foam
Once a small amount of fiber was added to the resin mixture, the viscosity
increased tremendously, and the mixing became difficult. Moreover, the high
viscosity also led to large air bubbles introduced and entrapped during mixing, and
poor foam quality resulted. Therefore, different mixing tools and procedures were
necessary for obtaining fiber-reinforced phenolic foams.
38
Figure 2.1: Hybrid mixer (a) and its working principle (b)
The dual axis mixer (Keyence Hybrid Mixer HM-560) proved to be an
effective solution to the problem. The mixer involves two simultaneous rotation
modes at predetermined high speeds (see Fig.2.1), and the combined centrifugal
force drives the turbulent resin flow in the container, providing the shear force
needed for mixing and dispersing fibers. Unlike other mechanical stirrers, this hybrid
mixer introduces no air voids even in highly viscous fiber slurry. Fig 2.2 compares
the mixing result of the hybrid mixer with that from a conventional planetary mixer.
As a result o the high speed shear flow, the presence of fibers causes a significant
temperature rise during hybrid mixing. Thus, a cooling process was followed
immediately after dispersion of fibers. Accordingly, the general protocol for making
reinforced foam was developed.
39
Figure 2.2: Comparison of mixing quality with different mixers
Phenolic resin and surfactants were all weighed in a polyethylene container
suitable for use in the hybrid mixer. The container was then placed in the mixer and
subjected to 1 min. mixing plus 1min. degassing. After addition of the fiber,
container was subjected to additional 2 min. of mixing in the mixer. The container
was placed inside a refrigerator for 1-2 hours, allowing the contents to cool down to
below 10
0
C. Pentane and PSA were pre-weighed and poured into the cold container,
followed by another 2 min. of high speed mixing. The resultant mixture was
transferred into a pre-heated mold. The remaining procedures were the same as for
making plain phenolic foam.
40
2.3 Material characterization tests
Several characterization tests were carried out in this study. The details of each
test have been discussed in details in relevant section. The following tests were
performed using ASTM standards to evaluate various properties of foams:
• Mechanical testing : compression and shear
• Scanning Electron microscopy (SEM)
• Optical microscopy
• Climatic simulation - Diffusivity
• Accelerated aging
• Flammability test
2.4 Design of experiments (DOE)
To investigate the effect of each factor in a multi-variable system, a systematic
experiment-planning tool is necessary. The design of experiment (DOE) approach is
intended to serve this purpose. DOE is widely used in modern industries to study
complex systems in which multiple factors and the interactive effects between them
may be involved [1]. The primary benefit of DOE is minimum number of
experiments required to extract sufficient information about the system. In this study,
several obvious as well as subtle factors were identified that could affect, to varying
degrees, the mechanical performance of phenolic foam. These factors include the
formulation components, synthesis procedures and test conditions. A DOE plan was
41
first configured to screen out the unimportant factors, and then a second DOE plan
was launched to investigate the effect of each of the remaining ones. The details of
experiments and the software are covered in further chapters.
42
Chapter 2 References
[1] Porter SC, Verseput RP, Cunningham CR. Pharm Technol 1997: October: 1-7
43
Chapter 3. Mechanical behavior of hybrid foams
3.1 Motivation
Hybrid composites in recent times have been developed by using more than
one type, shape or size of reinforcement. These composites have been developed to
bestow synergistic properties of the chosen fillers and matrix. The study of these
hybrid composites is relatively new. The development of composites containing
more than one type of fiber reinforcement (hybrid composites) is motivated by the
ability to combine advantageous features of various fiber types, including improved
performance as well as reduced weight and cost. Understanding the mechanical
properties of hybrid composites is essential in order to optimize the design of new
hybrid materials [1].
Hybrid composites are gaining commercial significance for several reasons.
First, a wider spectrum of tailor-made physical and mechanical properties is possible,
thus facilitating the design of materials with specific properties matched to an end
use. Secondly, there are economic advantages replacing a more expensive
reinforcement or filler with cheaper materials. Thirdly, hybrids can achieve
synergistic effects and improvements in mechanical and functional properties [2].
The properties of hybrid composites depend on several factors, including the
interaction of fillers with the polymeric matrix, shape and size (aspect ratio) of
fillers, and orientation of fillers, to name a few. For example, hybridizing glass fiber
composites with carbon fiber is known to enhance fatigue performance and
44
environmental resistance compared to all glass composites [3]. The judicious
selection of banana and sisal fibers has been useful in developing value-added and
cost-effective hybrid composites having high tensile and flexural properties [4].
Because of the added complexity of hybrid composites, however, the
understanding of mechanical properties such as strength, modulus and fracture
toughness poses a challenge. As a crude estimate, a rule of hybrid mixtures (ROHM)
can be used to predict the properties for a hybrid system consisting of two
constituent composites [5].
P
H
= P
A
V
A
+ P
B
V
B
3.1
Where, P
H
is the property of the hybrid material, P
A
the corresponding property of
the first constituent composite, and P
B
the corresponding property of second
constituent composite. V
A
and V
B
are the relative hybrid volume fraction of the first
and second constituents respectively, and V
A
+ V
B
= 1. A positive or negative hybrid
effect is described as a positive or negative deviation of a certain mechanical
property from the ROHM [5].
Phenolic foam has been considered in the current research. Phenolic foam has
certain distinct advantages when compared with other polymeric foams. For
example, phenolics exhibit excellent fire resistance, including low flammability, low
peak heat release rate (PHRR), no dripping during combustion, low smoke density
and low toxicity [6]. In addition, phenolic foam is one of the less expensive polymer
foams commercially available. Phenolic foam is also thermally stable over a broad
45
temperature range, maintaining performance and stability from –196 up to 200
0
C.
The thermal conductivity is low, which has led to a broad range of applications as an
insulating material. Finally, phenolic foam is highly resistant to chemicals and
solvents [6]. However, structural applications of phenolic foam have been severely
limited because of the inherent brittleness and friability [7]. As a consequence,
phenolic foam is rarely used as a core material in sandwich structures.
In contrast, polyurethane (PU) foam is widely used for sandwich cores in
structural applications. PU foam can be moderately stiff and easy to process.
However, PU foam is highly flammable and produces toxic fumes during
combustion, a factor that precludes many practical uses. In contrast, phenolic foam
exhibits excellent fire resistance, including low flammability, but poor toughness and
friability characteristics. At present, there is a need for fire-retardant, non-toxic foam
for use in fire-critical sandwich structure applications. Furthermore, standards for
fire, smoke and toxicity (FST) properties are becoming increasingly stringent
worldwide, and limitations of conventional structural foams may preclude their
continued use.
Over the past few decades, there have been attempts to increase the toughness
of phenolic foams [6, 8]. Particularly, short fiber reinforcement of phenolic foam
was considered [9, 10]. Significant improvement in peel strength and toughness were
achieved by reinforcing phenolic foam with aramid fibers [7] (e.g., Nomex® and/or
Kevlar®), which have well-known affinity for phenolic. For example, Shen et al
found that the addition of only 3wt% short aramid fibers produced a six-fold increase
46
in peel strength, and addition of 5wt% fibers resulted in a seven-fold increase over
unreinforced foams. Furthermore, increasing the fiber length and the fiber loading
generally increased the toughness. Interestingly, aramid fibers were more effective
than glass fibers in enhancing the peel strength and abrasion resistance for equivalent
loadings and fiber length, while glass fiber additions enhanced foam strength,
stiffness and dimensional stability [9, 10]. Based on the above observations, we
hypothesized that the mixed use of glass and aramid fibers in hybrid phenolic foams
could produce an optimal combination of toughness and strength.
In this chapter, we report property enhancements observed in hybrid phenolic
foams reinforced with glass and aramid fibers blended in varied proportions. The
properties of the hybrid composite foams are assessed to determine the effects of
blending different fibers in different proportions. Properties are compared with
foams having only aramid or glass fibers. In addition, we compare reinforced
phenolic foams with polyurethane foams in order to assess the potential of the
material as a fire-retardant, non-toxic substitute in fire-critical structural applications.
The present chapter also includes attempts to fit the mechanical properties of hybrid
foam into some existing models of composite materials.
3.2 Experiment
3.2.1 Foam preparation
Phenolic foams were synthesized using a proprietary formulation [12] and a
patented technology [13]. The formulation was typically composed of phenolic
47
resole resin (solid content >80/100 parts) and appropriate amounts of pentane to
achieve desired foam densities. Polysulphonic acid (PSA) was used as a catalyst for
the reaction. When fiber reinforcements were introduced, the amount of
polysulphonic acid (PSA) catalyst was slightly increased to allow more time for
dispersing fibers. All foams were formulated to achieve a density between 190 to
250 kg/m
3 (
12 to 15 pcf).
The synthesis of reinforced phenolic foam sample was carried out by blending
chopped fibers with the phenolic resin using a high-speed dual axis mixer, as
described previously [13]. The glass fibers (Lauscha Fiber International) were 6.4
mm in length and 11 μm in diameter, and were treated with a silane coupling agent.
Aramid fibers (DuPont Nomex®) were 6.4 mm in length and ~12 μm in diameter.
After blending fibers and resin with around two minutes of mixing time, the mixture
was then poured into a mold and held at 80
0
C for one hour. The foams were further
neutralized overnight in a closed chamber with a source for ammonia.
3.2.2 Mechanical test
Test specimens were sectioned from foam slabs using a diamond blade band
saw. Special attention was given to the cutting direction with respect to the foam
rise direction, and the edges of foam blocks were avoided. Mechanical tests were
performed using a universal testing machine (Instron 1331) in accordance with
ASTM standards.
48
Compression testing was performed in accordance with ASTM D1021.
Compressive modulus was taken as the steepest initial slope of the stress-strain
curve, and strength was determined from the maximum load (in a range of strain
<10%). At least five replicates were tested for each specimen, and the results were
presented as the average value of all replicates.
Lap shear testing was performed in accordance with ASTM C273. Foam
specimens were bonded to stainless steel plates with a fast-cure epoxy adhesive. The
shear modulus was taken as the steepest slope of the stress-strain curve, and strength
as the peak stress value. At least five replicates were tested for each specimen, and
the results were presented as the average value of all replicates.
3.2.3 Morphology
Foam surfaces were examined by scanning electron microscopy (SEM,
Cambridge 360). Prior to sectioning, samples were submerged in liquid nitrogen to
avoid structural deformation and damage to the foam microstructure. Subsequently,
samples were sputter-coated with gold to impart electrical conductivity and reduce
charging artifacts. The operating voltage of the SEM was 10 kV.
3.3 Results and Discussion
3.3.1 Compression test
Like most plastic foams, phenolic foam exhibits a multi-stage deformation
response when subjected to compressive loading [15]. An initial steep rise in the
49
stress-strain curve is followed by a constant-stress plateau, during which cells
collapse by the bending and buckling of cell walls and edges. The effect of fiber
type, fiber proportion, and (loading direction) anisotropy are described in the
following subsections.
3.3.1.1 Fiber Type Effects
The fiber type had a distinct effect on the compression behavior of the hybrid
composite foams, as shown in Table 3-1. For compression properties measured
parallel to the foam-rise direction, glass fibers produced greater enhancements in
compressive properties than aramid fibers. For example, glass fiber-reinforced foams
showed an increase of almost 275% (compared with unreinforced foams), while the
increase in strength was 36% for aramid fiber foams. In contrast, aramid fiber
reinforcement produced no significant change in compression modulus, while glass
fiber reinforcement resulted in more than 130% increase. This phenomenon can be
attributed in part to the relatively high stiffness of glass fibers [17] compared with
aramid fibers [16], and to a higher degree of glass fiber orientation along the foaming
direction. Evidence of preferred orientation of glass fibers was reported by Shen et
al. [18], using micro CT (computerized tomography) imaging of composite foam.
50
Table 3-1 Compressive properties of foams
3.3.1.2 Anisotropy
The anisotropy of foam properties is apparent from the data in Table 3-1. For
example, in the direction normal to the foam rise direction, the glass fiber foam
showed an increase in modulus of 156% (versus a 275% increase for the foam rise
direction), and virtually the same strength in the two directions. However, for the
aramid fiber foam, the modulus in the foam rise direction was identical to the
unreinforced foam, while in the normal direction, the composite foam showed a 77%
increase. The strength increase was 36% and 47% for the foam rise and normal
directions. From these observations, it appears that the aramid fibers may not align
51
in the foam rise direction to the same extent that the glass fibers do, although this
assertion is speculative at present.
Gibson and Ashby [15] noted that most foam, especially those produced by an
open mold process, are anisotropic in the foaming and transverse directions. The
anisotropy may arise from two independent factors: foam structure and materials.
Using an elongated cubic foam cell mode, they derived a Young’s modulus
anisotropy ratio that depended on the structural anisotropy alone, as shown in
Equation 3.2.
E
║
/ E
⊥
= 2R
2
+ [1+ (1/R)
3
] 3.2
where E
║
is the Young’s modulus of the foam measured parallel to the foaming
direction, E
⊥
is the Young’s modulus perpendicular to the foaming direction, and R
is the shape anisotropy ratio, defined as the ratio of cell height (measured in the
foaming direction) to cell width (measured in the transverse direction). This
relationship is obtained for open-cell foams when the cell membranes are weak
relative to the cell edges, and thus their contribution to foam modulus can be
neglected [15].
Using the framework described above, the modulus anisotropy ratios of the
foams were calculated (Table 3-1). As the values show, the aramid fiber foams are
nearly isotropic, with E
║
/ E
⊥
approaching 1, whereas the glass fiber foams are
substantially anisotropic. For hybrid foams, the modulus anisotropy ratio is
approximately 1 (or slightly higher). The framework can be tested by analyzing PU
52
foams, which show a shape anisotropy ratio of approximately 1.2 [15]. Insertion of
this R value in Equation (1) yields a modulus ratio, E
║
/ E
⊥
of 1.62, which compares
with the measured modulus ratio of 1.54, as shown in Table 3-1. This indicates that
PU foam behaves like open-cell foam.
If relationship (2) is valid for plain phenolic foam, the shape anisotropy ratio R
should be ~1.25. However, for the fiber reinforced foams, the property anisotropy
should stem from material anisotropy as well as from a structural origin. The
presence of fibers may modify the process of cell formation during foaming, altering
the foam cell morphology from that of plain foam. Meanwhile, fibers in the foam
may acquire preferred orientations and non-uniform distributions, contributing to
property anisotropy. This can be observed for the hybrid foam data in Table 3-1,
where preferred fiber orientation of glass and aramid fibers in the foam leads to
anisotropy in the foam properties, and the modulus anisotropy ratio varies with
change in fiber proportion.
The variation in foam properties for axial and transverse loading directions
reflects the extent of fiber alignment, which depends on fiber type. Consequently, the
strength and modulus of the glass fiber phenolic foams are greater than those of the
aramid fiber counterpart. This result is also consistent for hybrid foams, where an
increase in modulus is observed in the parallel direction with increasing proportion
of glass fibers. However, when load is applied transverse to the foam rise direction,
the hybrid foams have higher modulus than the foams reinforced with either glass or
Nomex ® fibers.
53
3.3.1.3 Fiber Proportions
Blending glass and aramid fibers in varied proportions produced hybrid
composite foams with significant improvements in modulus and strength over
unreinforced foams, as shown in Table 3-1. The greatest improvements in modulus
and strength were observed when glass and aramid fibers were added in the ratio of
3:1 respectively. This ratio resulted in a 126 % increase in modulus and a three-fold
increase in strength for the hybrid foam (glass: aramid, 3:1) relative to the
unreinforced foam (for both axial and transverse directions). Fiber weight ratios of
1:1 and 1:3 (glass: aramid) yielded a nearly two-fold increase in modulus for both
types of hybrid foams, and increases in strength of 250% and 230%, respectively.
The results indicate significant increases in strength and modulus for the hybrid foam
relative not only to unreinforced foams, but also to composite foams reinforced with
only aramid fibers.
3.3.1.4 Compression Stress Strain Relationships
Gibson and Ashby [15] have described and analyzed the deformation behavior
of cellular materials under compressive loading. Phenolic foam exhibits multistage
deformation response when subjected to compressive loading. In Figure 3.1 and
Figure 3.2 initial part of the compression stress strain response is displayed (strain <
0.2). This is the portion of the deformation response that is most relevant for
engineering applications, and contains the key parameters of compressive modulus
54
and strength. The data from several compression test results for hybrid foams have
been summarized in Table 1.
Figure 3.1: Typical compression stress-strain relationships of phenolic foams. Loading
direction is parallel to the foam rise direction.
55
Figure 3.2: Typical compression stress-strain relationships of phenolic foams. Loading
direction is parallel to the foam rise direction.
3.3.1.5 Phenolic vs. Polyurethane (PU)
The data from Table 3-1 afford an opportunity for comparisons with PU
foams. Typically, PU foams are stiffer (higher modulus) than most phenolic foams
of equivalent density. The compressive modulus of hybrid phenolic foam (glass and
Nomex® in a ratio of 3:1) in the foam rise direction is only 20% less than PU foam
of the same density, while the compressive strength is comparable. Thus, the hybrid
composite phenolic foam appears to offer mechanical performance nearly
56
comparable to PU foams, and superior FST properties. Such foams would be useful
for sandwich structure applications with stringent FST requirements.
3.3.2 Shear test
By design, sandwich cores are intended to mechanically couple face sheets by
carrying shear loads [19]. Consequently, shear properties are among the most
important criteria governing the selection of core materials for sandwich structures.
To assess the potential of hybrid composite foams in such structural applications, the
shear properties were measured, and these are summarized in Table 3-2.
3.3.2.1 Fiber Type Effects
The hybrid reinforced foams exhibit marked increases in shear modulus and
shear strength, as shown in Table 3-2. The data show that glass fiber reinforcement
produces a greater increase in shear modulus than the aramid fiber counterpart at the
same fiber loading. For example, glass fiber composite foam shows a 78% increase
in shear modulus and 158% increase in shear strength relative to unreinforced foam.
In contrast, the aramid fiber foam shows only a 12% increase in modulus and a 14%
increase in strength relative to the unreinforced foam.
57
Table 3-2 Shear properties of foams
The shear data were measured only in the plane normal to the foam rise
direction, and are consistent with previous reports. For example, Shen et al. [20]
found that the shear resistance of fiber-reinforced phenolic foam was substantially
greater on the plane normal to the foam rise direction. However, the difference in
anisotropy between unreinforced and reinforced foam was small, indicating shear
properties of phenolic foam were insensitive to material anisotropy.
58
3.3.2.2 Proportions
The shear properties were measured for hybrid foam samples consisting of
aramid and glass fibers in weight ratios of 3:1, 1:1 and 1:3. In all three hybrid
foams, the shear strength and modulus increased substantially, as shown in Table 3-
2. Hybrid foams with equal proportions of glass and aramid fibers exhibited shear
properties superior to the other hybrid foam samples, which included a nearly three-
fold increase in shear modulus and a nearly five-fold increase in shear strength.
Hybrid foam samples reinforced with glass and aramid fibers in ratios of 1:3 and 3:1
showed increases in shear modulus of 67% and 96% respectively, and an almost
three-fold increase in shear strength.
For hybrid systems in general, both glass and aramid fibers enhanced the shear
properties of the foams. The 1:1 aramid and glass ratio hybrid foams exhibited the
highest shear modulus of all the foams tested. From our experimental observations,
we concluded that at the homologous fiber ratio, the hybrid foam system exhibited
superior shear modulus compared to other reinforced phenolic foams with varied
fiber proportions. Thus, the contribution of both fiber types is critical to the shear
performance of phenolic foams. However, when the optimum fiber ratio of 1:1 was
altered, the shear properties decreased by more than 50% compared to glass and
aramid fiber reinforced foams.
59
3.3.2.3 Shear Stress Strain Relationships
Figure 3.3 shows typical shear stress-shear strain curves for aramid fiber and
glass fiber phenolic foams. The unreinforced foam shows purely elastic behavior,
and the glass fiber foam shows similar behavior followed by a brittle failure,
although with higher strength and modulus. In contrast, the aramid fiber foam
shows a distinctly different failure behavior characterized by more graceful failure.
The foam continues to carry load well beyond the peak stress, showing a smooth
decline in stress that continues to large shear strains. This behavior is consistent with
the report by Shen et al. [20], and indicates substantially enhanced shear toughness.
The extensive energy absorption is attributed to the tenacity of the aramid fibers
embedded in the foam structure, which bridge the shear cracks and gradually pull out
of the foam matrix. The aramid fibers afford greater flexibility and chemical
compatibility than the glass fibers, thus accounting for the different behavior of the
two composite foams.
As shown in Figure 3.4, the hybrid composite foams exhibit similar stress-
strain behavior. This includes a smooth decline in stress after the peak stress,
extending to strains of 15-40%. As the weight proportion of aramid fiber is
increased, the shear modulus decreases, and the peak stress increases. After peak
stress, the stress decline becomes smoother and more gradual and the ultimate strain
increases significantly. The 3:1 hybrid foam (aramid-to-glass ratio) exhibits the
highest peak stress and the largest ultimate strain of the three hybrid foams. The
increase in ultimate strain and energy absorption is attributed to the tenacity of the
60
flexible aramid fibers, which resist pullout and result in extensive crack bridging and
more graceful failure, manifest as a gradual decline in stress with increasing strain.
Figure 3.3: Typical shear stress-strain relationships of phenolic foams. Shear plane
and loading direction are both parallel to the foam rise direction.
Viewed from the opposite perspective, as the glass fiber content in hybrid
foams is increased, the ultimate strain decreases. There also is an increasing
tendency towards brittle rupture, and a slight but measurable decrease in peak stress.
61
In the limit, i.e., in foams with all glass fibers, brittle rupture is observed, much like
unreinforced phenolic foam (Figure 3.3). However, the shear modulus also increases
with increasing glass content.
Figure 3.4: Typical shear stress-strain relationships of phenolic foams. Shear plane
and loading direction are both parallel to the foam rise direction.
Energy absorption is an important performance metric for cellular materials,
including metal and polymeric foams [15]. The energy absorbed during fracture
provides a useful means for comparing foam performance, especially with regard to
62
impact resistance and damage tolerance [22]. Values of strain energy density for the
composite foams were calculated from the areas under the stress-strain curves [23],
and are listed in Table 3-2. The strain energy density of aramid fiber foam is 150%
greater than that of unreinforced foam. In contrast, the glass fiber foams show a 55%
decrease in strain energy density, indicating diminished toughness. Significantly, as
aramid fibers are blended with glass fibers, the strain energy density values for the
resulting hybrid foams increase with increasing aramid fiber content, nullifying the
embrittling effect of the glass fibers. Hybrid foams with 3:1 aramid and glass fiber
exhibit a strain energy density nearly 2.4× greater than plain phenolic foam. The
gradual and continuous decline in stress beyond peak stress may translate into
improved damage tolerance and more graceful failure of the hybrid foams, both of
which are desirable for structural applications.
3.3.2.4 Phenolic vs. Polyurethane (PU)
Polyurethane foam is a widely used structural foam that is convenient to
manufacture and nearly isotropic in shear performance [24]. In the present study, we
selected PU foam as an additional benchmark for comparison with the hybrid
phenolic foams. The shear modulus of the 1:1 hybrid foam was nearly twice the
modulus of the benchmark PU foam of equivalent density. A 10% increase in shear
strength was also observed for the 1:1 hybrid foam. The surprising properties of the
foams are linked in part to the foam structure, described in Section 3.3.4.
63
3.3.3 Theoretical Modeling of foam properties
Several models have been proposed to describe the mechanical properties
(compression, tension etc.) of reinforced composite materials in terms of different
parameters [25, 26]. These models can be classified into two groups, one based on
the nature of the matrix, and the other based on the type of reinforcements.
According to the parallel and series models, Young’s modulus is calculated
according to the following equations [31].
Parallel model
M
c
= M
f
V
f
+ M
m
V
m
3.3
Series Model
M
c
= M
m
M
f
/ (
M
m
V
f
+ M
f
V
m
) 3.4
where, M
c,
M
m
and M
f
are the Young’s moduli of composite, matrix and fiber,
respectively, and V
f
and V
m
are the volume fractions of fibers and matrix
respectively.
The second model considered is the Hirsch model, which is a combination of
the parallel and series models [30]. According to this model, the Young’s modulus is
given by the following equation:
M
c
= x (M
m
V
m
+ M
f
V
f
)+ (1-x) M
f
M
m
/ (M
m
V
f
+ M
f
V
m
) 3.5
Where, M
refers to the modulus, V is the volume fraction, x is the stress transfer ratio
[31], and the subscripts c, f and m signify the composite, fiber and matrix,
respectively.
64
The Halpin-Tsai model has been used by several researchers to analyze
polymeric blends consisting of continuous and discontinuous phases [27]. This
model was also useful in determining the properties of composites that contained
discontinuous fibers oriented in the loading direction [28, 29]. According to the
Halpin-Tsai model [34], the stiffness estimate in the fiber direction, E
║
, is given by:
) 1 (
) 1 (
11
f
f
m
v
v
E
E
η
ξη
−
+
= 3.6
where,
η = (E
f
/E
m
-1) (E
f
/E
m
+ ξ)
-1
, ξ = 2 (l/d)
υ
f
is the volume fraction of fibers, and subscripts f and m refer to fiber and matrix
respectively.
The experimental results obtained from compression test data were plotted
with the results obtained by fitting values for the models discussed here. Figure 3.5
shows the volume fraction of fibers plotted as a function of the compression modulus
(parallel direction). Parallel, series and Hirsch models do not provide good fits to the
data, and fail to explain the behavior of hybrid foams. The assumption of uniform
stress or uniform strain is clearly an oversimplification in these models, because the
stress transfer mechanism of continuous fiber reinforced composite is different from
that of short fiber composites [31]. Shen et al. [18] reported that severe breakage of
glass fibers occurred during processing of reinforced foams, resulting in a wide
distribution of lengths. The Halpin-Tsai model was thus plotted for original fiber
length of 6mm and another arbitrary length of 3mm. The Halpin-Tsai model for
65
fiber length 3mm comes close to fitting the experimentally observed behavior of
hybrid foams. However, the model lacks the complexity needed to capture the actual
behavior, and thus does
not accurately predict the dependency of the compression modulus on the volume
fraction of fibers. This deficiency can be partly attributed to the complex foam
structure, described in the following section.
Figure 3.5: Comparison of Experimental results with various theoretical methods
To reasonably depict the mechanical behaviors of hybrid foams, more
advanced modeling techniques are required. The scaling law, for example, has
proven useful for modeling complex engineering systems where traditional methods
66
are typically tedious and time-consuming [33]. The scaling law (SLAW) differs from
classical dimensional analysis in that it selects the scaling law with the smallest
predictive error out of all the dimensionally correct models. The algorithm combines
a linear regression model of the experimental data with physical consideration of the
process, namely, that the units of the resulting models match the units of the
dependent variable. The output of the algorithm is a physically meaningful and
simple power law representing the process, and a set of dimensionless groups
ordered by relevance to the problem. Thus, application of SLAW may offer
advantages for the study of hybrid foams, as it affords one an opportunity to consider
aspects such as fiber length, fiber proportion and other parameters which directly
influence the mechanical performance of foam considered in the same model.
Present efforts are aimed at implementing the approach of scaling laws [33] to hybrid
foams for achieving a suitable model for the complex behavior of hybrid composite
foams.
3.3.4 Foam structure
Microscopic observations of the hybrid foams revealed interactions between
the different fiber types, as well as fiber-matrix interactions. Typical observations of
fracture surfaces from the 1:1 hybrid foam are shown in Figures 3.6 and 3.7. Figure
3.6 shows details of the interaction between glass fibers and the foam matrix. Fibers
appear to be intercellular (between cells), as opposed intracellular (spanning cells), a
probable consequence of fiber surface tension during foam expansion. The glass
67
fiber ends protrude from the foam matrix, which shows evidence of damage and
cracking in the vicinity. The fibers are largely bare of foam fragments, indicating
fiber pullout and failure at the interface. Cracks apparently initiate in the brittle foam
and deflect along fiber interfaces before propagating further.
Figure3.6: SEM images of glass fibers in matrix of “hybrid” foam
68
Figure3.7: SEM images of aramid fibers in matrix of “hybrid” foam
A distinctively different phenomenon was observed for aramid fibers in the
hybrid foam, as shown in Fig. 3.7. The micrograph shows a foam fragment adhering
to an aramid fiber extending from the fracture surface. Adhering to the aramid fiber
are numerous small fragments of phenolic foam, indicating unusually strong
cohesive strength that forces a combination of interface and matrix failure. Shen et
al. [7] reported a micro-peeling process for aramid fiber reinforced foam where tiny
69
fibrils had peeled off from an aramid fiber stem. This phenomenon undoubtedly
contributed to the observed property enhancements of hybrid foams. As the micro-
peeling process proceeds in the hybrid foam, the stress concentration is reduced, and
secondary cracks are induced that branch into adjoining regions, thereby enlarging
the fragmented or damaged volume. This crack branching, combined with the
flexibility of aramid fibers, results in extensive crack bridging in hybrid foams. As a
result, failure is more graceful, as evidenced by the stress-strain behavior shown
previously in Figure 3.4. Thus, despite the presence of glass fibers in hybrid foams,
the brittle quality of failure is significantly diminished, and the foam strength is 1.9×
greater than the plain glass reinforced counterpart at the same density.
3.4 Conclusion
The approach of hybrid fiber reinforcement of phenolic foams improves all
aspects of shear and compression properties. The optimum fiber ratio for the
reinforced hybrid phenolic foam system was the 1:1 ratio of glass to aramid fibers.
This homologous ratio produces a balance in the shear and compressive properties.
The hybrid foam retains toughness and stiffness, with significantly enhancement of
strength. These properties can be tailored to specific requirements for particular
applications simply by varying the fiber proportions. Consequently, the hybrid fiber
approach is well-suited to optimizing a broad range of foam properties.
A major issue motivating hybrid fiber reinforcement of phenolic foams
concerns the potential suitability of the foams for structural applications. The
70
evaluation of the hybrid composite phenolic foams has shown that the mechanical
performance is comparable to commercial PU foams of equivalent density. These
results indicate that judicious selection of fiber reinforcements in optimal proportions
can yield structural foams that are strong, tough, fire-retardant, and cost competitive
for a variety of structural applications.
Several existing models for predicting the behavior of short-fiber-reinforced
composites were evaluated to determine the applicability to hybrid foams. None of
these models accurately fit the experimental data, indicating that the behavior of
hybrid composite foams involves mechanisms more complex than these models can
represent. This deviation suggests that a better understanding is required to model the
mechanisms of fiber reinforcement in foams, and the dependence on fiber type, fiber
orientation, and critical length. Thus, considerable room for further optimization
remains. An empirical approach may produce incremental improvements, although a
predictive model for mechanical properties of composite foams is sorely lacking.
The next chapter discusses a statistical predictive model that was developed to
describe the compression properties of phenolic foam reinforced with glass fibers.
An analysis of variance (ANOVA) was applied to determine the behavior of
composite phenolic foam.
71
Chapter 3 References
[1] Chiang Martin, Y.M., Wang, X., Schultheisz, C.R. and He, J. (2005). Prediction
and three-dimensional Monte-Carlo simulation for tensile properties of
unidirectional hybrid composites, Comp. Sci. and Tech., 65 (11-12): 1719-1727.
[2] Babu Prem, E.J., Savithri, S. and Pillai, U.T.S, Pai B.C. (2005). Micromechanical
modeling of hybrid composites, Polymer, 46 (18):7478-7484.
[3] Shan, Y. and Liao, K. (2002). Environmental fatigue behavior and life prediction
of unidirectional glass–carbon/epoxy hybrid composites, Int. J. Fatigue, (24):847-
859.
[4] Idicula, M., Neelakantan, N.R., Oommen, Z., Joseph, K. and Thomas, S. (2005).
A study of the mechanical properties of randomly oriented short banana and sisal
hybrid fiber reinforced polyester composites, J of App. Poly. Science, 96 (5):1699-
1709.
[5] Shao-Yun, F., Yiu-Wing, M., Lauke, B. and Chee-Yoon, Y. (2002). Synergistic
effect on the fracture toughness of hybrid short glass fiber and short carbon fiber
reinforced polypropylene composites, Mater. Sci. Eng. (A), 323 (1-2): 326-335.
[6] Mao, J., Chang, J., Chen, Y. and Fang, D. (1998). Review of Phenolic foam,
Chem. Ind. Eng., 15(3): 38-43.
[7] Shen, H., Lavoie, A.J. and Nutt, S.R. (2003). Enhanced peel resistance of fiber
reinforced phenolic foams, Comp. A: Appl. Sci. Manufact., 34(10): 941-948.
[8] Knop, A. and Scheip, W. (1979). Chemistry and application of phenolic resins.
New York; Springer
[9] Takanori, K. (1993). JP 05-286069.
[10] Manabu, H. (1993). JP 05-318506.
[11] Halpin, J.C. and Kardos, J.L. (1972). Moduli of crystalline polymers employing
composite theory, J. Appl. Phys., 43(5):2235-2241.
[12] Cohen, MI. (1994). WO 94/04604.
[13] Nutt, S.R. and Shen, H. (2005). US Patent 6,841,584.
72
[14] Boeing Part Specification (BPS) D124. (1991), Air, Panels and Plenum
Insulation, Rigid Plastic Foam, for low pressure Systems, Boeing.
[15] Gibson, L.J. and Ashby, M.F. (1997). Cellular solids: structure and properties,
Cambridge, UK.
[16] Technical guide of Nomex® Brand Fiber (2002). Technical information,
DuPont.
[17] Lauscha fiber International (2002). Technical data sheet.
[18] Shen, H., Nutt, S.R. and Hull, D. (2004). Direct observation and measurement
of fiber architecture in short fiber-polymer composite foam through micro-CT
imaging, Comp. Sci. and Tech., 64 (13-14):2113-2120.
[19] Zenkert, D. (1995). An introduction to sandwich construction, London,
Chamelon.
[20] Shen, H. and Nutt, S.R. (2003). Mechanical characterization of short fiber
reinforced phenolic foam, Comp. Part A, 34(9): 899-906.
[21] Kaynak, C., Arikan, A. and Tincer, T. (2003). Flexibility improvement of short
glass fiber reinforced epoxy by using a liquid elastomer, Polymer, 44 (8): 2433-2439.
[22] Vaikhanski, L. and Nutt, S.R. (2003). Fiber-reinforced composite foam from
expandable PVC microspheres, Comp. Part A, 34 (12): 1245-1253.
[23] Broek, D. (1982). Elementary engineering fracture mechanics, Dordrecht:
Martinus Nijhoff: 469.
[24] FR-6700® Series foam product (2004). General Plastics manufacturing
company.
[25] Robinson, I.M. and Robinson, J.M. (1994). The influence of fiber aspect ratio
on the deformation of discontinuous fiber-reinforced composites, J. Mater. Sci., 29
(18):4663-4677.
[26] Sarasua, J.R. , Remiro, P.M. and Pouyet, J. (1995). The mechanical behavior of
PEEK short fiber composites, J. Mater. Sci., 30(13):3501-3508.
[27] Radesh Kumar, C., George, K.E. and Thomas, S. (1996). Morphology and
mechanical properties of thermoplastic elastomers from nylon-nitrile rubber blends,
J. Appl. Poly. Sci., 61 (13): 2383-2396.
73
[28] Halpin, J.C. and Tsai, S.W. (1967). Effects of Environmental Factors on
composite materials, Rep. No: AFML-TR-67-423, Air Force Materials Laboratory,
Wright Patterson Air Force Base, Ohio.
[29] Rosen, B.W. (1965), Fiber composite Materials (American Society for metals,
Metals Park, OH):58.
[30] Hirsch, T.J., J. Am. Con. Inst., (59):427-451.
[31] Kalaprasad, G., Joseph, K. and Thomas S. (1997), Theoretical modelling of
tensile properties of short sisal fiber-reinforced low-density polyethylene composites,
J. of Mat. Sci., 32(16): 4261-4267.
[32] Hyer, M.W. and White, S.R. (1998), Stress analysis of fiber- reinforced
composite materials, New York: McGraw-Hill.
[33] Mendez, P. and Ordonez, F. (2005), Scaling Laws from Statistical Data and
Dimensional Analysis, J. Appl. Mech., 72: 648 -657.
74
Chapter 4. Modeling of Fiber Reinforced Phenolic Foam
4.1 Motivation
Fiber reinforcements have been introduced in phenolic foams to address the
deficiencies in mechanical performance [1-3]. For example, Shen et al. [1] reported
significant improvements in peel strength, and compression and shear properties of
foam reinforced with glass and aramid fibers. In chapter 3 hybrid reinforcements
were used to tailor the properties of phenolic foam to specific requirements by
varying fiber proportions. In both cases, several-fold improvements in compression
and shear properties were demonstrated for foams by the judicious selection of
hybrid reinforcement compared to foams reinforced with only glass or aramid fibers.
Several approaches have been employed to model the properties of composite
foams, and these can be classified into three categories. In the first, the modeling is
based on micro-mechanical methods which require precise representation of the
internal structure of the material [4]. The second category involves a macro-
modeling approach in which deformation mechanism based on analysis of
constitutive equations derived from experimental data [5]. The third category
includes models derived from homogenization theories [6]. A good example of the
modeling based on foam microstructure is the well-known theory of Gibson and
Ashby, which offers the virtue of simplicity [7]. On the other hand, Saint-Michel et
al. modelled polyurethane foams [PU] using the Christensen and Lo equations [5], an
example of the macro-modeling approach. Both of these examples involve analysis
of conventional unreinforced foams.
75
The presence of filler in composite foams raises special challenges for
modeling mechanical behavior, and although there are a few examples in the
literature, success has been limited. In one case, Siegmann et al. [6] studied the low-
to medium-density foams filled with different fillers below 10 µm. They attempted
to model the linear behavior of filled foams using Kerner equations in two steps.
However, their analysis did not account for foam microstructure or for the influence
of filler size on foam properties. In chapter 3 conventional models (parallel, series
and Halpin-Tsai) were evaluated for fiber-reinforced composites to predict the
mechanical properties of hybrid foams. However, none of the models provided
accurate fits to the experimental data, suggesting the need for a mechanistic model
incorporating fiber length, loading, orientation, and foam density to predict the
complex behavior of hybrid foams.
Different analytical tools have been employed to predict the elastic behavior of
foams as a function of cell irregularity and relative density. For example, random
Voronoi models have been constructed to simulate the linear elastic behavior of open
cell foams [6]. Similarly, the dependence of elastic properties of the foams on the
relative density has been simulated using finite element analysis (FEA) [8, 9],
providing good accuracy and utility [8]. FEA analysis was combined with
experimental observations by Youssef et al. [10], when they obtained the
microstructure of polyurethane foam directly form X-ray tomographic data and
implemented a predictive finite element model of the mechanical behavior of foams.
76
An alternative approach to the mechanical modeling of foams involves the use
of statistical methods. Statistical approaches greatly reduce the number of
experimental iterations compared to other modeling approaches while
simultaneously yielding a wealth of information about multiple, interacting variables
which influence the system. Several recent works demonstrate the utility of a
statistical approach, known as analysis of variance (ANOVA) [11, 12]. For example,
Alonso et al [12] applied this approach to analyze the behavior of composite foams.
They developed a statistical model to describe the compressive properties of glass-
fiber reinforced epoxy foams and investigated the effects of simple variables such as
density, fiber weight fraction, and fiber length on the modulus and compressive
strength. A 2
3
central composite design was applied, which included three main
effects and three two-factor interactions. Such methods significantly reduce the
number of experimental iterations, combining increased efficiency with good
accuracy.
This chapter determines the relationship between composition, final
morphology, density and elastic properties of fiber reinforced composite phenolic
foam. A statistical model is utilized to describe the behavior of glass-fiber reinforced
phenolic foams when length, weight fraction, and density are varied within a
prescribed range.
77
4.2 Experiment
4.2.1 Foam preparation
Phenolic foam samples were produced using a proprietary formulation [13]
and patented technology [14], as described elsewhere [3]. Pentane was used as the
blowing agent for foaming and polysulphonic acid (PSA) was the acid catalyst. Short
glass fibers were used with chop lengths of 3, 6, 9, and 12 mm and an average
diameter of 11µm (Lauscha Fiber International). The fibers included a silane sizing.
Synthesis of reinforced phenolic foams was carried out by blending the resin
along with the surfactants, blowing agent, and catalyst in a high-speed, dual-axis
mixer (Keyence HM-101). After mixing, the glass fibers were incorporated into the
mixture, which was blended for an additional two minutes. The reinforced phenolic
foam was poured into a mold and held for 1 h at 80
0
C. The phenolic foam was then
cooled to room temperature before neutralizing overnight in a closed chamber of
ammonia. A total of fourteen different reinforced phenolic foam samples were
synthesized with different amounts of blowing agent, glass fiber weight fractions and
glass fiber lengths.
4.2.2 Statistical Experimental Design
Experimental design is widely used in the scientific and the engineering
communities for product improvement and optimization. Statistical approaches to
experimental design are useful for extracting the most meaningful conclusions from
78
experimental data. Only a brief discussion of statistical analysis of experimental
design is presented here, with a focus on factorial experiment, although more
detailed accounts appear elsewhere [15, 16].
The two aspects common to nearly all experimental investigations are the
design of experiment and the statistical analysis of the data, and these two aspects are
complimentary. Choice of experimental design involves consideration of sample size
(number of replicates), selection of suitable run order, and identification of
randomization restrictions, if any [15]. The key task for selecting an appropriate
model is identifying dependent and independent variables. One must then identify
factors or variables which cause the response and estimate the magnitude of the
response change. In this chapter, a 2
k
factorial design was implemented. The design
of experiments was divided into two sets, as explained below.
A 2
3
factorial design was chosen for the first part of the design of experiments
(Experimental design I). The design involved three factors (variables), each with two
levels of interest. High and low levels of the factors were designated as “+” and “-”,
respectively. Table 4-1 summarizes the design with all possible high/low
combinations from the range of three factors, and Fig. 4.1 represents graphically the
ranges of the three factors considered in the 2
3
design. Each one of the numbers from
the “cube” represents an iteration of the experimental design. Additionally, in the
first part of the design, one additional experiment is included which corresponds to
the midpoint of the ranges studied for the independent variables. The iteration of the
79
central point of the experimental design can be used as an estimate of the
experimental variation [16].
Table 4-1 Matrix of two level factorial designs with one central point
(Experiment Design I)
Figure 4.1: Cube for 2 cube design
80
A 2
2
factorial design was chosen for the second part of design of experiments
(Experimental design II), in which, there were two factors (variables), each with two
level of interest. Table 4-2 summarizes the design with all possible combinations
(high/low) from the range of the two factors. As with the first design, one additional
experiment is added which corresponds to the midpoint of the range studied for the
independent variables. Fig. 4.2 represents the range of the three factors considered in
the 2
2
design. The various dependent variables (or responses in the present case)
were density, fiber length, and fiber weight fraction, discussed in a later section.
Figure 4.2: Cube for 2 square design
Table 4-2 Matrix of two level factorial designs with one central point
(Experiment Design II)
81
The experimental designs were employed to fit a linear statistical model to the
response results, with factor levels coded as “low” or “high”. Statistical software was
used to perform analysis of variance (ANOVA) and multiple regression methods.
For the 2
3
design, this yields a statistical model that can be expressed as:
j i
j
ij
ii
i i
X X a X a a Z
∑ ∑∑
= ==
+ + =
3
1
3
1
3
1
0
4.1
where Z is the response considered, a
ij
are the regression coefficients, and X
ij
are the
effects studied. After the regression coefficients are determined, an estimation or
validation of the results is performed. Thus, each effect yields a P-value which
signifies the significance of that effect, either low or high. For example, a higher P-
value (>0.05%) indicates that an effect should be rejected in the variance analysis
because that effect lacks significance in the model. In addition, for determining the
fit of the model, the data must show a low scatter and an R
2
value (fraction of the
variance) near 100%. The responses can be plotted as a contour map or a response
surface (2D and/or 3D plot). Such graphs provide a visual depiction of the evolution
of several factors that influence the response studied, allowing one to optimize the
system conditions within the ranges considered.
For the 2
2
design, the statistical model is represented as:
j i
2
1 j
ij
2
1 i
2
1 i
i i 0
X X a X a a Z
∑ ∑∑
= ==
+ + = 4.2
82
4.2.3 Plan of Experiments
In the first part of the experimental work, nine phenolic foams were
formulated with different fiber lengths (6mm, 9mm and 12mm), fiber weight
fractions, and amount of blowing agent, resulting in composite foams with varied
density. In addition, five composite phenolic foams were formulated with similar
fiber lengths (3mm), but with different fiber weight fractions and blowing agent
content. Compression tests were performed on each of the specimens (parallel
direction only), and the data were analyzed using the Gibson and Ashby model [7].
Cell size distributions for the fiber-reinforced composite foams were obtained from
SEM images.
In the second part of this work, factorial design was employed to study the
effects of the main process variables on composite foam manufacture. For
experimental design I, a 2
3
central composite design was applied (9 runs: 2
3
+ 1
central point), which included three main effects and three two factor interactions.
Similarly, for experimental design II, a 2
2
central composite design was applied (5
runs: 2
2
+ I central point), which included two main effects and two two-factor
interactions. The variables studied included density ( ρ), fiber weight fraction (W),
and fiber length (L) for experimental design I, and for experimental II, the variables
included density ( ρ) and fiber weight fraction (W). The measured responses included
compression modulus (E) and the compressive strength ( σ
c
).
83
Tables 4-3 and 4-4 summarize the experimental conditions employed and the
results for experimental design I and II. Commercial software (Statgraphics Plus
®
)
was used to process the data. The model equations included first-order terms to
describe main effects, and second-order terms for interactions, yielding the
expression below for the response:
Z = a
0
+ a
ρ
ρ + a
L
L + a
W
W + α
ρL
ρL + a
ρW
ρW + a
WL
WL 4.3
The non-significant effects were discounted by applying analysis of variance, which
were included from the model regression, and to check the suitability of the model.
Table 4-3 Foams with 6, 9 and 12 mm fibers
84
Table 4-4 Foams with 3 mm fibers
4.2.4 Compression tests
Compression tests were performed in accordance with ASTM D1621.
Specimens, 30 mm square by 25.4 mm thick, were placed between steel platens, and
load was applied with a crosshead speed of 0.5 mm/min (0.02 in. /min). Compressive
modulus was determined from the steepest initial slope of the stress-strain curve, and
strength was determined from the maximum load (in a range of strain <10%). At
least five replicates were tested for each material, and the results were presented as
the average value of all replicates.
4.2.5 Scanning Electron Microscopy (SEM)
Fracture surfaces of foam samples were examined with a scanning electron
microscope (SEM, Cambridge 360) operated at 10 kV. The samples were submerged
in liquid nitrogen to avoid deformation and structural damage to the foams, and
85
samples were sputter coated with gold to impart electrical conductivity. SEM images
were analyzed with image analysis software.
4.3 Results and Discussions
4.3.1 Characterization of unreinforced and reinforced
Foams
The density of phenolic foam increased sharply as the amount of blowing
agent (alkane) is decreased in the formulation, as shown in Fig. 4.3. Thus, the
amount of blowing agent was critical to the density of phenolic foams.
Figure 4.3: Variation of density with weight percentage of blowing agent for
unreinforced foam
The compressive modulus (E) and compressive strength ( σ
c
) are plotted as functions
of foam density in Figs. 4.4a and 4.4b. The figures indicate that the foam strength
and modulus increase substantially with increasing foam density. The compressive
modulus and strength can be fitted to a simple power law, as follows:
n
EAρ = 4.4
86
m
c
B σ ρ = 4.5
where A and B are constants related to the physical properties of the foams, and n
and m are density exponents related to the structure and deformation mechanics of
the foams. Although equations (4) and (5) were developed for open-cell foams [1],
these expressions have been widely used to predict other properties for closed-cell
foams as well [12, 17], and they accurately describe the compressive strength and
modulus over the present density range. The corresponding value of coefficients in
Figs. 4.4a and 4.4b are 0.95 and 0.97, respectively, while the values obtained for “n”
and “m” are 1.5 and 1.9.
Figure 4.4: Variation of (a) modulus and (b) compressive strength with density of
unreinforced foam
87
4.3.1.1 Effect of fiber length on cell size
The effect of fiber length on cell diameter for composite phenolic foam (with 1
wt. % fiber and 2.5 wt. % blowing agent) is shown in Fig 4.5. The average cell
diameter for foams with 3 mm fibers is approximately 0.15 mm.
Figure 4.5: Effect of fiber length on cell size
88
However, for the foams with 6 and 12 mm fiber lengths the average cell is
reduced by an order of magnitude, to 0.012 and 0.014 mm, respectively.
We speculate the decrease in average cell size with increasing fiber length can
be attributed in part to heterogeneous nucleation of cells on the fibers. In addition,
significant fiber breakage occurs during processing, and the resulting fiber clustering
interferes with cell expansion locally and compromises foam expansion globally.
Longer fibers (6 and 12 mm) experience more extensive breakage compared to
shorter (3 mm) fibers, resulting in greater reductions in average cell diameter. The
phenomenon of fiber breakage is consistent with computer tomography (CT)
observations by Shen et al [18], who reported that glass fibers added to phenolic
foam were substantially reduced in length after mechanical mixing. In future work
CT imaging of foams will be studied to support the theory.
Similar observations were noted for foams with 5 wt% fiber and 2.5 wt%
blowing agent using fiber lengths of 3, 6, and 12 mm. Figure 6 compares these foams
with unreinforced foam at the same composition. The average cell diameter was 0.28
mm for unreinforced foams, while for composite foams with 3 mm, 6mm and 12 mm
fibers, the mean cell diameters dropped to 0.24 mm, 0.007 mm and 0.012 mm
respectively.
4.3.1.2 Effect of fiber weight fraction on cell size
The average cell diameters for foams with 1 wt. % fibers (Fig. 4.5) were 0.15
mm, 0.012 mm and 0.014 mm with fiber length of 3 mm, 6 mm and 12 mm,
89
respectively, whereas for foam samples with 5 wt. % (Fig. 4.6), cell diameters were
0.24 mm, 0.007 mm and 0.012 mm, respectively. Thus, altering the fiber weight
fraction (from 1 wt. % to 5 wt. %) had a neglible effect on the cell size distribution.
Figure 4.6: Effect of fiber weight fraction on cell size
These results are consistent with data for epoxy foams [13], where the size cell
distribution was not significantly affected by the weight fraction of fibers.
90
4.3.1.3 Effect of percentage of blowing agent on cell size
Table 4-3 summarizes the effect of blowing agent on cell size of composite
phenolic foams. Runs 1 and 2 produced a large increase in average cell size
(diameter). The increase in cell size was 650 % when the amount of blowing agent
was increased by only 1wt. %, while all other constraints (12 mm fiber length and of
5 wt% fibers) remained unchanged. However, the change in average cell diameter
with amount of blowing agent was negligible when fiber lengths were 6 mm and 3
mm, as shown in Tables 4-3 and 4-4. For example, comparing runs 3 and 5, at a fiber
length of 6 mm with 5 wt% fibers, there is no significant change in the average cell
size with an increase in the amount of blowing agent. Thus, the amount of blowing
agent affected cell diameter only when fiber lengths were long, while the cell size for
foams with shorter fibers (6 and 3 mm) was essentially unchanged.
4.3.2 Statistical model for glass fiber reinforced composite
Phenolic foams
The behavior of fiber reinforced foams was evaluated on the basis of measured
values of compressive modulus (E) and strength ( σ
c
), as summarized in Tables 4-3
and 4-4.
4.3.2.1 Foams with 6mm, 9mm and 12mm glass fibers
In the experimental design I, a factorial design was followed to study the
effects of key process variables on the manufacture of glass-fiber-reinforced
phenolic foams. A 2
3
central composite design was applied (9 runs), which included
91
three main effects and three two-factor interactions. The variables studied were
blowing agent (B), fiber weight fraction (W), and fiber length (L), and the respective
ranges were 2.5-3.5 wt%, 1–5 wt%, and 6–12 mm. The response measured for model
development was density ( ρ).
The influence of blowing agent (B), fiber length (L) and fiber weight fraction
(W) was studied through statistical design. The analysis of variance (ANOVA) for
density of the reinforced foams was carried out for a level of confidence of 95%, and
is shown in Table 4-5.
Table 4-5 ANOVA for density in Experimental Design I
In cases where the factor had a distribution F < 18.58% and P-values > 0.05,
the factor or factors were rejected because the effects presented a significance level
less than 95%. Thus, as shown in Table 4-5, the factors L, BL and WL were rejected
from the analysis of variance. The new ANOVA values for the density are shown in
Table 4-6. All factors display a confidence level of 95%. The corresponding revised
92
model describing the density for the composite foams (in the range studied), is given
by:
ρ (kg/m
3
) = 12.5615 + 21.3525 · B + 71.9488 · W -17.6675 · B W 4.6
Table 4-6 ANOVA for density with significant effects (Experimental Design I)
The response surface of the density as a function of fiber weight fraction and
amount of blowing agent is shown in Fig. 4.7a. Statistical results indicate that the
foam density is independent of fiber length.
In the second experimental part, variables studied included density ( ρ), fiber
weight fraction (W), and fiber length (L). The measured responses used for model
development were modulus (E) and compressive strength ( σ
c
). The analysis of
variance (ANOVA) for composite foam modulus was carried out for a level of
confidence of 95%, and is shown in Table 4-7.
93
Figure 4.7: Response surface curve for 2
3
design
Table 4-7 ANOVA for Modulus (Experimental Design I)
94
In this case the factors L, ρL and WL had a distribution F < 18.58% and P-
values > 0.05 and thus were rejected from the ANOVA (Table 4-8). Thus, the
corresponding revised model describing the modulus for the composite foams (in the
range studied), is given by:
E (MPa) = - 81.8488 + 1.16321 · ρ + 17.2642 · W – 0.185056 · ρW 4.7
Statistical results indicate that the composite foam modulus is independent of fiber
length, similar to results obtained in equation [19]. Fig. 4.7b shows the response
surface of the modulus as a function of fiber weight fraction and density.
. Table 4-8 ANOVA for modulus with significance effects (Experimental design I)
Finally an operation similar to the one for evaluating modulus was performed for
strength, and the factors L, ρL and WL were rejected from the analysis of variance.
This is shown in Table 4-9, and the new ANOVA values for compressive strength
are shown in Table 4-10. The model obtained for compressive strength is found to
be:
σ
c
(MPa) = - 2.08192 + 0.0316814 · ρ + 0.19398 · W – 0.00314429 · ρW 4.8
95
Table 4-9 ANOVA for strength (Experimental Design I)
Table 4-10 ANOVA for strength with significance effects
(Experimental Design I)
The results obtained from the statistical model support the assertion that the
compressive modulus and strength are independent of fiber length. This lack of
dependence is attributed in part to the fiber breakage that apparently occurs during
foam processing, as noted previously. Previous work on similar composite foams
showed that the fiber lengths were substantially reduced as a result of foam
processing [18].
The co-relations obtained above for density, compressive strength and
modulus leads to the next part of this work (Design II), in which a statistical
96
modeling was applied by varying the blowing agent and fiber proportion while
keeping fiber length constant.
4.3.2.2 Foams with 3mm glass fibers
A 2
2
central composite design was applied (5 runs), which included two main
effects and one two-factor interaction. The variables were blowing agent (B), and
fiber weight fraction (W), and the respective ranges were 2.5-3.5 wt%, and 1–5 wt%,
while the response measured for model development was density ( ρ). The influence
of the variables (B and W) was studied through statistical design. The analysis of
variance (ANOVA) for foam density, carried out for a level of confidence of 90%, is
shown in Table 4-11.
Table 4-11 ANOVA for density (Experimental design II)
In cases where the factor had a distribution F < 9.29% and P-values > 0.10, the
factor was rejected. Therefore, the factor BW was rejected from the ANOVA, as
shown in Table 4-11, and the new ANOVA values for density are shown in Table 4-
12. All factors display a confidence level of 90%. The corresponding revised model
obtained describing the composite foam density (in the range studied), is given by:
ρ (kg/m
3
) = 158.516 – 28.115 · B + 5.94625 · W 4.9
97
Fig. 4.8a represents the response surface of density as a function of fiber
weight fraction and amount of blowing agent. The 3-D curve shown in Fig. 4.8a
bears resemblance to the curve obtained earlier for the 2
3
design for density (Fig.
4.7a). In both these curves we observe a similar pattern where the density of the
foams increases with increase in weight percentage of fibers whereas with an
increase in amount of blowing agent the density of foams decreases.
Table 4-12 ANOVA for density with significant effects (Experimental Design II)
Next, a factorial design was followed to study the effects of key process
variables on foam manufacture. The variables studied were density ( ρ) and fiber
weight fraction (W), and the responses measured were modulus (E) and compressive
strength ( σ
c
).
Analysis similar to the one performed in Experimental design I was carried out
using statistical design to study the influence of density ( ρ) and fiber weight fraction
(W). Table 4-13 shows the analysis of variance (ANOVA) for modulus of reinforced
phenolic foam. The factors presenting a significant level less that 95% (modulus) and
90% (strength) were rejected and the revised models describing the modulus and
strength were:
E (MPa) = - 8.09045 – 0.189865 · W + 0.335407 · ρ 4.10
98
σ
c
(MPa) = - 0.169761 -0.0411862 · W + 0.00856372 · ρ 4.11
Figure 4.8 Response surface curves for 2
2
designs
Table 4-13 ANOVA for modulus (Experimental design II)
99
Table 4-14 ANOVA for modulus with significance effects (Experimental design II)
Table 4-15 ANOVA for strength (Experimental design II)
The response surface of modulus as a function density and weight fraction of
fiber is shown in Fig. 4.8b. The response surface resembles the surface of Fig. 4.7b
obtained for the 2
3
design of composite foam. Also, the response surface of
compressive strength as a function of density and weight fraction (Fig. 4.8c)
resembles the surface obtained for the 2
3
design of composite foam (Fig. 4.7c).
These figures help us study the variation of compressive strength and modulus as a
function of density and weight fraction of fibers. Also for the range of foam studied
the density and compression modulus is not affected much by the length of fibers
used. Tables 4.13, 4.14, 4.15 and 4.16 show corresponding data before and after lack
of fit for modulus and strength.
100
Table 4-16 ANOVA for strength with significance effects (Experimental design II)
Shen et al. [18] studied similar composite foams and observed a large variance
in the fiber length distribution (FLD) after foam processing, although the original
fiber length was constant (6.4 mm). The majority of fibers in the final foam samples
were approximately 3mm or less. The approach and experimental design can be used
as a predictive tool to reduce the number of experimental iterations required to obtain
foams with the specific strength and modulus values by simply varying the amount
of blowing agent and the weight fraction of fibers. Finally, the modeling approach
presented here can be applied to the design of more complex foams with hybrid
fibers [4] to obtain specific ranges of mechanical properties.
4.4 Conclusion
Composite foams with different fiber loadings, fiber lengths, and amount of
blowing agent were produced to obtain a broad range of densities, and the
compressive strength and modulus were measured. While a simple power-law
expression is sufficient to describe the mechanical properties of unreinforced foams,
composite foam systems required a more sophisticated description. For this purpose,
a statistical design approach was employed to analyze the mechanical properties of
101
composite foams. This tool was used to investigate the effects of simple process
factors (such as fiber loading, fiber length, etc.) on the mechanical properties of the
composite foams.
Analysis of the experimental data and statistical design led to the conclusion
that the compressive properties of the composite foams did not depend on the fiber
length, at least over the range of lengths employed, and depended instead on other
parameters, such as the amount of blowing agent and the fiber weight fraction.
Similar methods and approaches can be applied to other fiber-foam systems to
reduce the number of experimental iterations and understand the effects of process
variables on associated mechanical properties. For reinforced phenolic foams, it
would be useful to extend this method to the analysis of more complex systems, such
as “hybrid” foams reinforced with more than one type of fibers. In particular, the
statistical method employed here would be well-suited to investigating the effects of
multiple process parameters on foam mechanical properties. This leads us to the next
chapter where statistical approach has been applied to study mechanical properties of
hybrid foams.
102
Chapter 4 References
[1] Shen, H. and Nutt, S.R. (2003). Mechanical characterization of short fiber
reinforced phenolic foam, Comp. Part A, 34(9): 899-906.
[2] Desai, A., Auad, M.L., Shen, H. and Nutt, S.R. (2007). Mechanical Behaviors of
hybrid composite phenolic foam, Journal of cellular plastics (Article in Press).
[3] Shen, H., Lavoie, A.J. and Nutt, S.R. (2003). Enhanced peel resistance of fiber
reinforced phenolic foams, Comp. A: Appl. Sci. Manufact., 34(10): 941-948.
[4] Mao, J., Chang, J., Chen, Y. and Fang, D. (1998). Review of Phenolic foam,
Chem. Ind. Eng., 15(3): 38-43.
[5] Saint-Michel, F., Chazeau, L., Cavaille, J.Y., Chabert, E. (2006). Mechanical
properties of high density polyurethane foams: II Effect of filler size, Comp Sci. and
Tech., 66(15): 2709-2718.
[6] Siegmann, A., Kenig, S., Alperstein, D., and Narkis, M. (1983). Mechanical
behavior of reinforced polyurethane foams, Polymer Composites, 4(2): 113-119.
[7] Gibson, L.J. and Ashby, M.F. (1998). Cellular solids: structure and properties.
Oxford: Pergamon Press.
[8] Gan, Y.X., Chen, C., Shen, Y.P. (2005). Three-dimensional modeling of the
mechanical property of linearly elastic open cell foams, Int Journal of Solids and
Structures, 42 (26): 6628-6642.
[9] Zhu, H.X., Hodbell, J.R., Windle, A.H. (2004). Effect of cell irregularity on the
elastic properties of open-cell foams. Acta Mater, 48: 4893-900.
[10] Youssef, S., Maire, E., Gaertner, R. Finite Element modeling of the actual
structure of cellular materials determined by X-ray tomography. Acta Mater, 53(3):
719-730.
[11] Davim, J.P., Reis, P., Antonio, C.C. (2004). A study on milling of glass fiber
plastics manufactured by hand-lay up using statistical analysis (ANOVA), Compos
Structures, 64 (2):493-500.
[12] Alonso, M.V., Auad M.L., Nutt, S.R. (2006) Modeling the compressive
properties of glass fiber reinforced epoxy foam using the analysis of variance
approach. Comp Sci and Tech, 66(13): 2126-2134.
103
[13] Cohen, MI. (1994). WO 94/04604.
[14] Nutt, S.R. and Shen, H. (2005). US Patent 6,841,584.
[15] Montogomery, D.C. (1991). Design and analysis of experiments. Wiley and
Sons, Inc.: New York.
[16] Box G.E., Hunter W.G., Hunter J.S. (1978). Statistics for experimenters: an
introduction to design, data analysis and model building. John Wiley and Sons: New
York.
[17] Goods, S.H., Neuschwanger, C.L., Whinnery, L.L., Nix W.D. (1999).
Mechanical properties of a particle strengthened polyurethane foam, J Appl Polym
Sci, 74(11):2436-724.
[18] Shen, H., Nutt, S.R. and Hull, D. (2004). Direct observation and measurement
of fiber architecture in short fiber-polymer composite foam through micro-CT
imaging, Comp. Sci. and Tech., 64 (13-14):2113-2120.
[19] Bledzki, A.K., Zhang, W., Chate, A. (2001). Natural-fiber-reinforced
polyurethane microfoams, Comp Sci and Tech., 61(16): 2405-2411.
104
Chapter 5. Modeling of Hybrid Foams
5.1 Motivation
The combination of different types of fibers in a common matrix, known as
hybrid composite, has the potential of achieving balance between composite strength
and ductility [1]. Also hybrid composites can provide a wider spectrum of
mechanical properties and reduce the cost of composites by mixing reinforcements
such as glass fibers with more expensive high-performance fibers.
Several hybrid composites have been designed to improve abrasion resistance
and other mechanical properties, and to reduce coefficient of thermal expansion. In
chapter 3 phenolic foams were reinforced with glass and aramid fibers (Nomex® or
Kevlar), thus obtaining phenolic foams which were tough, stiff, and strong. Several-
fold improvement in compression and shear properties were demonstrated for foams
by judicious selection of fibers. Hybrid composites of glass fiber/wollastonite fiber
reinforced engineering thermoplastics (ETPs) were blended with various levels of
glass to wollastonite to obtain tailored composites showed high strength, good
dimensional stability, and warp resistance [2].
The mechanical properties of hybrid composites may or may not follow the
rule of mixtures (ROM) and this has attracted much attention [3]. The so-called
‘hybrid effect’ is defined in two ways: (a) any deviation from ROM [4] or (b) a
difference between the failure strain of the low elongation (LE) fiber in a hybrid
composite, and that of the LE fiber in a single-fiber type (SFT) composite [2,5].
105
Several approaches have been employed to model the properties of composite
foams. In chapter 4 an analysis of variance was applied to analyze and describe the
mechanical behavior of fiber-reinforced phenolic foam [6]. We studied material
variables such as fiber length, fiber weight fraction, etc., and developed a statistical
approach to predict effects of the variables on the elastic properties of composite
foams. In other work, researchers have employed analytical tools to predict elastic
behavior of foams as a function of cell irregularity and relative density. Siegmann et.
al. constructed Voronoi models to simulate linear elastic behavior of open cells [7].
They attempted to model linear behavior of filled foams using Kerner equations in
two steps.
For hybrid composites, complexity arises in modeling the mechanical
properties because of more than one type of fiber or filler in the matrix, and thus the
studies have been limited. Babu et al. studied the elastic modulus of hybrid
composites through micromechanical modeling [8]. They examined the effect of
filler concentrations, filler shape, and aspect ratio on the effective elastic properties
of polymer based hybrid composites. Desai et al. evaluated multiple composite
models, including parallel and series models, and the Halpin-Tsai, and Hirsch
models to fit experimental data [1]. As hybrid composite foams become more
widespread, there will be a need for a modeling approach which can reduce the
experimental iterations required to optimize properties and also serve as a predictive
tool for similar fiber foam systems.
106
In this chapter we determine the relationship between fiber compositions,
final morphology, and elastic properties of hybrid composite phenolic foams
reinforced with glass and aramid fibers. A statistical model was utilized to describe
the effects of glass and aramid fiber weight fractions, and the responses measured
were modulus and compressive strength. Foams were also examined to understand
the influence of hybrid foam morphology on the mechanical behavior.
5.2 Experiment
5.2.1 Foam preparation
The formulation used for foam production was composed of phenolic resol
resin (solid content >80/100 parts). Fiber-reinforced hybrid phenolic foams were
synthesized by blending chopped fibers with the phenolic resin in a high-speed dual-
axis mixer [9]. Aramid fibers (DuPont Nomex®) were 6.4 mm in length and ~12 μm
in diameter. The glass fibers (Lauscha Fiber International) were 6.4 mm in length
and 11 μm in diameter, and were treated with a silane coupling agent. Foams were
synthesized using a proprietary formulation [10] and a patented technology [9].
Polysulphonic acid (PSA) was used as a catalyst and pentane was used as the
blowing agent. A total of nine fiber-reinforced hybrid foams were synthesized.
Pentane and PSA was added in different amounts to obtain foams with a density of
~80 kg/m
3
.
107
5.2.2 Statistical Experimental Design
Design of experiments (DOE) has been used to : a) quantify the
interrelationship between variables , b) screen a large number of variables to identify
critical ones, c) develop new products, and d) optimize processes through control
schemes. The DOE method can reduce the number of experimental iterations
required to achieve necessary end results or properties. Analysis of variance
(ANOVA) is a collection of statistical models and associated procedures in which
the observed variance is partitioned into components due to different explanatory
variables. There are three conceptual classes of such models, including 1) the fixed
effect model that assumes data from normal populations which may differ only in
their means, 2) the random effects model, which assumes that the data describe a
hierarchy of different populations whose differences are constrained by the
hierarchy, and 3) the mixed effect model that describes situations where both fixed
and random effects are present [11].
In this study, we implemented 3
k
factorial design using analysis of variance to
study the influence of glass and aramid® fibers on the compression and shear
properties of hybrid foams.
A 3
2
factorial design was implemented and the design of experiments was
carried out to study the behavior of hybrid foams in compression and shear. The 3
2
factorial design featured 2 factors, each at three levels. The nine treatment
combinations for this type of design are shown in Fig. 5.1. A notation such as “20”
signifies that factor A is at its high level (2) and factor B is at its low level (0). The
108
effect of the factors on the response is measured by the change in the average
response in two or more combinations of levels. In this type of factorial design, there
are 9 combinations of treatments. Thus, there are eight degrees of freedom between
these combinations. The main A and B effects each have two degrees of freedom,
while the AB interaction has four degrees of freedom. If there are p* replicates, there
is a total of p* 3
2
-1 degrees of freedom, which corresponds to 3
2
(p*-1) degrees of
freedom for the errors.
Figure 5.1 3 square schematic effect
The combinations of treatments are written in standard order, which means
that one factor at a time is introduced successively, combining each level with the set
of the factors levels. The standard order for a 3
2
design is 00, 10, 01, 11, 21, 02, 12,
and 22 [12].
The experimental designs were employed to fit a linear statistical model to the
response results, with factor levels coded as “low” or “high”. Statistical software was
109
used to perform analysis of variance and multiple regression methods. After the
regression coefficients were determined, an estimation or validation of the results
was performed. Thus, each effect yielded a P-value which signified the significance
of that effect, either low or high. In addition, for determining the fit of the model,
low scatter in the data and an R
2
value (fraction of the variance) near 100% were
required. The responses were plotted as a contour map or a response surface (2D
and/or 3D plot).
5.2.3 Plan of Experiments
Nine samples of hybrid composite phenolic foam were formulated with glass
and aramid fibers. The proportion of blowing agent and PSA was measured for each
formulation to obtain hybrid foams with similar densities. Compression and shear
tests were performed on each of the specimens using the Gibson and Ashby
micromechanical co-relation for Young’s modulus [13]. SEM images were obtained
to determine the cell size distributions for the hybrid foams. The variables studied
were glass fiber weight fraction (G) and aramid fiber weight fraction (N). The
responses measured for developing the model were modulus (E) and compressive
strength (σ
c
). Table 5-1 and 5-2 summarize the experimental conditions employed
and the results of compression and shear testing of hybrid foams. Commercial
software (Statgraphics Plus
®
) was used to process the data.
110
Table 5-1 Compression property of hybrid foams with glass and aramid fibers.
Table 5-2 Shear property of hybrid foams with glass and aramid fibers.
5.2.4 Compression tests
The compression testing was performed using a universal testing
machine (INSTRON 5564 system) in accordance with ASTM D1621. Specimens
were cut with a diamond blade band saw and polished to a size of 30 mm × 30 mm ×
25.4 mm. The samples were compressed between two stainless steel platens using a
crosshead rate of 2.5 mm/min. From the stress-strain curves, the compressive
modulus values were determined from the steepest initial slope, and the strength was
111
calculated from the maximum load. Average values were determined from a
minimum of five replicates.
5.2.5 Shear tests
Lap shear testing was performed in accordance with ASTM C273. Foam
specimens were bonded to stainless steel plates with a fast-cure epoxy adhesive. The
shear modulus was taken as the steepest slope of the stress-strain curve, and strength
as the peak stress value. At least five replicates were tested for each specimen, and
the results were presented as the average value of all replicates.
5.2.6 Scanning electron microscopy (SEM)
A scanning electron microscope (SEM) was employed to observe the fracture
surfaces of foam specimens. Samples were carefully cut from the freshly peeled
surfaces using a razor blade. Gold sputtering onto the sample surface was used to
impart electrical conductivity. The operation voltage of the SEM was 15 kV. SEM
images were recorded in a high-resolution electronic format and processed later with
computer software.
5.3 Results and Discussions
5.3.1 Statistical model for compressive properties of hybrid
foams
A 3
2
screening design with 3 central points were followed to determine the
effects of the main variables on the production of glass and aramid fibers-reinforced
112
phenolic foams. The variables studied were glass fiber weight fraction (G) and
aramid fiber weight fraction (N). The responses measured for model development
were modulus (E) and compressive strength ( σ
c
). The influence of glass and aramid
fibers weight fraction (W) was studied through statistical design. The analysis of
variance (ANOVA) for compression modulus of reinforced phenolic foams is shown
in Table 5-3. This analysis was carried out for a level of confidence of 95%.
Table 5-3 Analysis of variance for compression modulus
In this case, all factors displayed a confidence level of 95%, and no factors
were rejected. The corresponding model obtained describing the modulus for the
composite foams (in the range studied), is given by Eq. (1):
E (MPa) = 25.39 – 6.745·G – 0.1·N + 2.955·G
2
+ 0.825·G·N 5.1
and the coefficient correlation R
2
= 99.09.
Figures 5.2a and 5.2b show similar plots for compression modulus with
different weight percentages of glass and aramid fibers, where Fig. 5.2a is a contour
map and Fig 2b is a 3-D plot. Glass fibers had a greater influence on compressive
113
modulus of hybrid foams than aramid fibers, and the values of modulus increased
with increasing proportion of glass fibers. However, when the glass fiber weight
fraction was less than 2%, the modulus remained almost constant with increasing
proportion of aramid fibers. For hybrid foams with glass fiber weight fractions
greater than 2%, an increase in modulus was observed. This is partly attributed to the
fact that glass fibers tended to orient in the foam rise direction [14], and glass fibers
are stiffer then aramid fibers.
A model was obtained describing the compression strength for hybrid
composite foam in the range studied. The equation describing compressive strength
of hybrid composite phenolic foam is given below:
σ (MPa) = 0.1374 + 0.0308·G + 0.065·N + 0.0098·G
2
- 0.022·G·N 5.2
All factors display a confidence level of 95%, and the coefficient correlation
R
2
= 98.21. The corresponding contour maps and 3-d plot are shown in Fig. 5.3a and
5.3b, respectively. Both aramid and glass fibers had similar influences on the
compressive strength of the hybrid foams. An increase in compressive strength was
observed at both extremes of contour plots. Hybrid foam specimens with higher
weight fractions of glass and aramid fibers exhibited higher strength. Implementing
this statistical approach with analysis of variance revealed the influence of fiber type
on the properties of foam. Table 5-4 shows the analysis of variance for strength.
114
(a)
(b)
Figure 5.2 (a) Contour map showing effects of fiber type on compressive modulus.
(b) 3-d plot showing effects of fiber type on compressive modulus.
115
Figure 5.3 (a) Contour map showing effects of fiber type on compressive strength.
(b) 3-d plot showing effects of fiber type on compressive strength.
Table 5-4 Analysis of variance for compressive strength
5.3.2 Statistical model for shear properties of hybrid foams
The analysis of variance (ANOVA) for shear modulus of reinforced phenolic
foams is shown in Table 5-5. This analysis was carried out for a level of confidence
of 95%. In cases where the factor had a distribution F < 18.58% and P-values > 0.05,
116
the factor or factors were rejected because the effects presented a significance level
less than 95%.
Table 5-5 Analysis of variance for shear modulus
As shown in Table 5-5, the factor GG is rejected from the analysis of variance.
The new analysis of variance values for the modulus are shown between parentheses
in Table 5-5. All factors display a confidence level of 95%.
The corresponding revised model obtained describing the shear modulus for
the composite foams (in the range studied), is given by:
E (MPa) = 13.9333 + 1.6·G + 1.4·N + 0.6·G·N 5.3
Coefficient correlation R
2
= 96.65.
The shear properties of hybrid foams were influenced by both glass and
aramid fibers. The contour and 3-d plot for shear modulus are shown in Fig. 5-4a and
5-4b, respectively. Glass and aramid fibers exhibited a similar influence on the shear
modulus of hybrid foams. The shear modulus increased with an increase in weight
fraction of glass and aramid fibers.
117
(a) (b)
Figure 5.4 (a) Contour map showing effects of fiber type on shear modulus.
(b) 3-d plot showing effects of fiber type on shear modulus
Table 5-6 shows analysis of variance values for the shear strength of hybrid
foams. The factor GG was rejected because it displayed a confidence level of less
than 95%. The new ANOVA values for the strength are shown in parentheses in
Table 5-6. The corresponding revised model obtained describing the shear strength
for hybrid composite foams (in the range studied) is given by:
σ (MPa) = 0.387222 – 0.0525·G – 0.0683333·N + 0.0725·G·N 5.4
Coefficient correlation R
2
= 99.88.
118
Table 5-6 Analysis of variance for shear strength
The contour and 3-d plots for shear strength are shown in Fig. 5-5a and 5-5b,
respectively. The shear strength of hybrid foams exhibited unusual behavior. Hybrid
foams formulated with 1 wt.% aramid fibers and different weight percentages of
glass fibers exhibited strength values that were nearly constant. However, almost
85% increase in shear strength was observed when weight fraction of aramid fibers
increased from 1 wt. % to 3 wt. %. Furthermore, for hybrid foams with 1 wt. %
glass fibers and different weight percentages of aramid fibers, value of strength was
almost constant.
119
(a) (b)
Figure 5.5 (a) Contour map showing effects of fiber type on shear strength.
(b) 3-d plot showing effects of fiber type on shear strength.
5.3.3 Morphology of hybrid composite phenolic foams
The unusual behavior of hybrid foams can be understood from the foam
morphology, which has been examined using scanning electron microscopy (SEM).
Hybrid foams with 3 wt. % and 1 wt. % of glass and aramid fiber respectively are
shown in Figure 5.6(5). Glass fiber ends protruded from the foam matrix, and
significant damage and cracking was evident in foam cells in the vicinity. Also,
glass fibers were randomly dispersed in the matrix, and extended from the fracture
surface, free of foam fragments, indicating fiber pullout. The cells were irregular in
size, and exhibited damage due to pullout of glass fibers from the foam matrix, as
shown in Fig. 5.6 (5).
120
Figure 5.6 SEM images of hybrid foam samples
The cell geometry and distribution for hybrid foams reinforced with 3 wt. % and 1
wt. % of aramid and glass fibers respectively is shown in Fig. 5.6 (6). Foam
fragments had a tendency to adhere to aramid fibers (not shown here) unlike glass
121
fibers, indicating a strong cohesive strength. The individual cells in these hybrid
foams with larger aramid fiber content showed uniform distribution and a
symmetrical cell size distribution throughout the foam microstructure, and each cell
was bounded by a well-defined cell wall. This phenomenon contributed to the
isotropic behavior of hybrid foams with larger weight percentages of aramid fibers as
found earlier in chapter 3.
Multiple factors affect the cell anisotropy in foams. Gibson and Ashby derived
a Young’s modulus anisotropy ratio that depended on the structural anisotropy alone,
as shown in Equation (5).
E
║
/ E
⊥
= 2R
2
/ [1+ (1/R)
3
] 5.5
where E
║
is the Young’s modulus of the foam measured parallel to the foaming
direction, E
⊥
is the Young’s modulus perpendicular to the foaming direction, and R
is the shape anisotropy ratio, defined as the ratio of cell height (measured in the
foaming direction) to cell width (measured in the transverse direction). This
relationship is obtained for open-cell foams when the cell membranes are weak
relative to the cell edges, and thus their contribution to foam modulus can be
neglected [13]. In chapter 3 we found the anisotropy ratio for hybrid foams
approximately to be 1 (or slightly higher).
Table 5-7 presents the average cell diameters for all the hybrid foam samples
studied here. No significant variations in the average cell diameter were detected,
although as discussed before, variations were observed in the patterns of cell
122
geometry based on the proportion of glass and aramid fibers. As a consequence of
varying fiber fractions hybrid foam samples exhibited isotropic or anisotropic
behavior.
Table 5-7 Average cell diameters for hybrid foams
5.4 Conclusion
Hybrid composite phenolic foams were produced with different weight
fractions of glass and aramid fibers. The compression and shear properties of these
hybrid foams were complex, and a statistical design tool was employed to describe
and predict the mechanical performance. The effect of fiber weight fraction of glass
and aramid fibers (controllable variables) on the mechanical properties was
investigated. The co-relations obtained for modulus and strength in this study can be
applied to predict mechanical properties of hybrid foams. Furthermore, these
relations can be used to tailor the properties to end use applications, thus
significantly reducing the number of experimental iterations needed.
123
The study of foam morphology revealed that the type and amount of fiber used
in the hybrid foams influenced the isotropic or anisotropic behavior of hybrid foams,
although the average cell size was not significantly affected. The statistical design
approach employed here can be applied to other hybrid foam systems (epoxy,
polyurethane etc.), and doing so will reduce the number of experimental iterations
needed to optimize foam properties and identify critical process variables.
This leads us to our next chapter, where climatic simulation and diffusivity of
hybrid foams have been studied to evaluate its potential for insulation and cladding
applications.
124
Chapter 5 References
[1] Desai, A., Auad, M.L., Shen, H. and Nutt, S.R. (2008). Mechanical Behaviors of
hybrid composite phenolic foam, Journal of cellular plastics, 44(1): 15-36.
[2] Jacobson, R., Caulfield, D., (2003). Hybrid Composites: Combining cellulose
fibers and woolastonite mineral fibers into a nylon 6 matrix, The seventh
international conference on woodfiber-plastics composites, May 19-20.
[3] Kretsis, G. (1987). A review of the tensile, compressive, flexural and shear
properties of hybrid fiber-reinforced plastic, Composites, 1:13-23.
[4] Marom, G., Fischer, S., Tuler, F.R. and Wagner, H.D. (1978). Hybrid effects in
composites: conditions for positive or negative effects versus rule of mixtures
behavior, Journal of Materials Science, 13: 1419-26
[5] Fukuda, H. (1983). An advanced theory of the strength of hybrid composites,
Journal of Materials Science, 19:947-82.
[6] Desai, A., Alonso, M.V., Nutt, S.R (2008). Modeling of fiber reinforced phenolic
foam, Journal of cellular plastics, Article in press.
[7] Siegmann, A., Kenig, S., Alperstein, D., and Narkis, M. (1983). Mechanical
behavior of reinforced polyurethane foams, Polymer Composites, 4(2): 113-119.
[8] Babu, Prem E.J., Savithri, S., Pillai, U.T.S., Pai, B.C. (2005). Micromechanical
modeling of hybrid composites, Polymer, 46:7478-7484.
[9] Nutt, S.R. and Shen, H. (2005). US Patent 6,841,584
[10] Cohen, M.I. (1994). WO 94/04604.
[11] Box G.E., Hunter W.G., Hunter J.S. (1978). Statistics for experimenters: an
introduction to design, data analysis and model building. John Wiley and Sons: New
York.
[12] Montogomery, J.S., (1991). Diseño y Análisis de Experimentos, Editora
Panamericana S.A., México
[13] Gibson, L.J. and Ashby, M.F. (1998). Cellular solids: structure and properties.
Oxford: Pergamon Press.
125
[14] Shen, H., Nutt, S.R. and Hull, D. (2004). Direct observation and measurement
of fiber architecture in short fiber-polymer composite foam through micro-CT
imaging, Comp. Sci. and Tech., 64 (13-14):2113-2120.
126
Chapter 6. Diffusivity and climatic simulation of hybrid
foams
6.1 Motivation
Energy conservation is an increasingly important aspect of modern building
designs. Among the factors that the building designer must consider, energy
conservation by use of thermal insulation is most effective [1]. Creating a low-cost
structural insulation material is also important for environmental reasons. As energy
costs rise, energy conservation in buildings is becoming increasingly important. Heat
is principally lost through air infiltration. A large reduction in heat loss, however,
can be realized by insulating attics and ceilings [2]. Table 6-1 and Table 6-2 compare
heat losses for an un-insulated and insulated house during heating and air-
conditioning periods. The tables show that insulation can save on energy costs by up
to 50%, with rooftops being the most effective locations to place insulation [3].
However, multiple factors must be considered when choosing an insulation material.
Table 6-1 Heat loss distribution for heating a typical residence
127
Table 6-2 Heat loss distribution for cooling a typical residence
The choice of insulation is among the most important factors affecting the
environmental impact of a building. Although insulation reduces building energy
consumption and provides other ongoing environmental benefits throughout a
building’s lifetime, insulation materials greatly differ in their environmental
advantages. As the United States Department of the Interior notes, insulation
manufacturing processes generate pollution as a result of fossil fuel combustion [4].
The simplest way to assess manufacturing impact is to compare the manufacturing
energy required, or embodied energy (2008). Table 6-3 reveals significant
differences in embodied energy costs of commonly used insulation materials [4]. The
table shows that expanded polystyrene (EPS) has an embodied value over 35 times
greater than cellulose. The data suggests that use of natural, locally available or
recycled material can drastically reduce embodied energy costs. Cellulose, mineral
wool, fiber glass, EPS and polyurethane foam (PU) are among the most commonly
used insulation materials.
128
Table 6-3 Embodied energy of common insulation materials [4]
Cellular plastics, because of the intrinsic low mass and low thermal conductivity, are
obvious choices for core materials in insulating panels for buildings. They are
relatively low-cost and include a vast variety of polymer foams with a range of
densities, elastic properties and strengths. The properties of foams depend on the
density of the foam, the foam material, and the cell structure. Popular choices of
commercial foams for the purpose of insulation are EPS and PU, although both have
their drawbacks. For example, embrittlement and dusting of PU foam insulation has
been reported in the living space of homes [5]. In addition, exposure of PU foam
samples to high temperature and humidity conditions reportedly causes deterioration
similar to that which can occur in service [6]. Secondly, PU and EPS foams are
flammable. As the standards for fire, smoke and toxicity (FST) properties become
increasingly stringent worldwide, and limitations of conventional structural foams
may preclude their continued use. However, as shown in Table 6-3, phenolics have
an embodied energy value of 799*10
5
J/kg which is 28% less than the value for EPS.
129
Assuming phenolic foam has an embodied energy comparable to dense phenolic, the
embodied energy of phenolic foam is expected to be considerably less than that of
EPS.
Hybrid composite phenolic foam reinforced with glass and aramid fiber is
considered in this study because of its outstanding properties as discussed in chapter
3 and chapter 5 [7]. Phenolic foam also has other distinct advantages relative to other
polymeric foams, including low flammability, low peak heat release rate (PHRR), no
dripping during combustion, low smoke density and low toxicity [8]. Phenolic foam
is also thermally stable over a broad temperature range, maintaining performance
and stability from –196 up to 200
0
C. The low thermal conductivity makes it an ideal
candidate as an insulating material. In addition, phenolic foam is one of the less
expensive polymer foams.
However, to use a foam or cellular material for insulation, several other factors
have to be taken under consideration, such as long-term durability and stability under
extreme environmental conditions, especially for service conditions involving
extreme heat and humidity. The thermal conductivity of the entrapped gas in the cell
represents an important contribution to overall foam thermal conductivity since
approximately 50% of the heat transfer through the foam occurs by conduction
through the gas phase [9]. A major problem in foam materials involves the decrease
in insulation value over time, and the lack of dimensional stability. Air diffuses into
the cells during the service life of foams and a high molecular weight blowing agent
diffuses out, which modifies the cell gas composition and ultimately causes a gradual
130
decrease in thermal resistivity of the foam. Thus, the diffusion behavior in foams is
important, since the mass transfer of gases in the air control the cell gas composition
for short to intermediate times, while the mass transfer of the blowing agent controls
long-term composition changes in the foam.
Efforts to develop predictive models for the effective diffusivity of foam have
met with limited success. Bart and Dul Cauze De Nazelle developed a discrete model
to predict the effective diffusivity of foam [10]. However, this and other discrete
models reported in literature are not realistic in describing the diffusion through
foams, mainly because of the basic assumption of a steady state in the cell walls that
is not necessarily valid.
In this study, we apply Fick’s second law of diffusion to predict the moisture
content and diffusivity of hybrid foams. A statistical approach is employed to
analyze the foam behavior and to predict moisture absorption. The flammability of
phenolic foam with different fibers is measured and compared to commercial EPS
foam. Accelerated aging of these hybrid foams is analyzed to determine if extended
exposure to intense heat and humidity weakens the material. Compressive properties
of composite foam are compared before and after accelerated aging to study
environmental stress and effects on foam properties and performance. The present
study also determines structure-property relations and the effects of fiber
reinforcement on foam morphology. Cell lengths before and after climatic simulation
are compared to study moisture absorption in composite foams.
131
6.2 Experiment
6.2.1 Materials and Foam preparation
Phenolic foams were synthesized using a proprietary formulation [11] and a
patented technology [12]. The formulation was typically composed of phenolic
resole resin (solid content >80/100 parts) and appropriate amounts of pentane to
achieve desired foam densities. Polysulphonic acid (PSA) was used as a catalyst for
the reaction. When fiber reinforcements were introduced, the amount of
polysulphonic acid (PSA) catalyst was slightly increased to allow more time for
dispersing fibers. All foams were formulated to achieve a density of 50 kg/m
3
(~3
pcf). The foams were synthesized at low density to compete with existing insulation
materials such a PU and EPS.
The details of manufacturing hybrid foam have been discussed in details in
chapter 3 and 5. Table 6-4 summarizes the different foam formulations used. For
comparison purposes hybrid foams reinforced with glass and aramid fibers (fiber
ratio 1:3 and 3:1) were produced, foam with only aramid fibers, foam with only glass
fibers and unreinforced foam.
6.2.2 Moisture absorption and diffusivity model
Foam plate samples (5 × 50 × 50 mm) were used for moisture absorption tests, in
accordance with ASTM D5229. The samples were dried in a vacuum oven at 60 ºC
132
until no weight change was observed. Dried specimens were placed in an
environmental chamber and maintained at 80 % relative humidity at room
temperature. The weight of these samples was monitored until specimens reached
equilibrium. Table 6- 4 summarizes the experimental conditions employed.
Table 6-4 Moisture absorption test
The mechanism of moisture diffusion can be described by Fick’s second law
using diffusivity constant (D), which normally represents non-steady state diffusion
in a polymer in three dimensions as follows [13]:
M
t
M
∞
= 1−
8
(2n + 1)
2
π
2
exp
−D π
2
4l
2
(2n +1)
2
t
⎡
⎣
⎢
⎢
⎤
⎦
⎥
⎥
0
∞
∑
6.1
Here M
t
is the weight gained at time t, and M
∞
is the maximum weight gained at the
equilibrium state. In the present work, one-dimensional diffusion through a sample
of thickness 2l was measured, and thus the relative moisture uptake was
133
approximated by the following expression (where only the terms for n = 0 appear)
[13]:
⎥
⎦
⎤
⎢
⎣
⎡
⋅
⋅
π ⋅ −
⋅
π
− =
∞
t
l 4
D
exp
8
1
M
M
2
2
2
t
6.2
An expression similar to Equation 6.1 was used to predict the moisture content,
given by:
⎥
⎥
⎦
⎤
⎢
⎢
⎣
⎡
⎟
⎟
⎠
⎞
⎜
⎜
⎝
⎛
⋅
⋅
⋅ − − =
∞
75 . 0
2
t
l 4
t D
3 . 7 exp 1
M
M
6.3
Fick’s second law is based on a linear relationship between the moisture gain
(M
t
/M
∞
) and time (t
1/2
). Consequently, the diffusivity coefficient can be determined
from the resultant slope of the following equation for small values of times (M
∞
/M
t
≤
0.5).
M
t
M
∞
= 4⋅
D⋅ t
π⋅4 ⋅ l
2
⎛
⎝
⎜
⎜
⎞
⎠
⎟
⎟
0.5
6.4
In this study, the non-steady state moisture absorption was considered, and a model
was utilized to predict diffusivity of composite foams.
6.2.3 Flammability test
Vertical burn tests were performed in accordance with UL 94 to measure
flammability [14]. Samples were loaded into the metal chamber, and a Bunsen
burner was lit and positioned beneath the sample (Figure 6.1). The flame was applied
to samples for two 10 second intervals separated by the time it took (if any) for the
combustion to cease after the first application. The length of time that the sample
134
retained flame for each application was measured. Three samples were tested per
specimen. In addition to combustion time, the volume of each sample was measured
before and after the test to determine the volume of material consumed.
Figure 6.1 UL 94 burn test
6.2.4 Climatic simulation: Accelerated aging
Accelerating aging was performed in accordance with ASTM D2126 [15].
Aged samples were subsequently prepared for compression and shear tests, and for
cell measurement.
Samples were first conditioned in an oven for 24 hours at 50ºC. Next, samples
were placed in a humidity oven. Controls were set at 40ºC and 85% relative humidity
(RH). Samples were conditioned for six weeks, then removed and weighed.
Variations of the standard aging were also performed. Compression samples were
prepared by sectioning to 1”x1”x1”. No preconditioning was performed on these
135
foam samples to extract moisture. Three samples were placed in the aging chamber
for each fiber type so that results of the post-conditioning tests could be averaged
after six weeks.
6.2.5 Mechanical tests
Test specimens were sectioned from foam slabs using a diamond blade band
saw. Special attention was given to the cutting direction with respect to the foam
rise direction, and the edges of foam blocks were avoided. Compression tests were
performed using a universal testing machine (INSTRON 5564 system) in accordance
with ASTM standards. Specimens, 30 mm square by 25.4 mm thick, were placed
between steel platens, and load was applied with a crosshead speed of 0.5 mm/min
(0.02 in. /min). Compressive modulus was determined from the steepest initial slope
of the stress-strain curve, and strength was determined from the maximum load (in a
range of strain <10%). At least five replicates were tested for each material, and the
results were presented as the average value of all replicates.
6.2.6 Scanning electron microscopy (SEM)
A scanning electron microscope (SEM) was employed to observe the fracture
surfaces of foam specimens. Samples were cut from the freshly peeled surfaces using
a razor blade. Gold sputtering onto the sample surface was used to impart electrical
conductivity. The operation voltage of the SEM was 15 kV. SEM images were
recorded in a high-resolution electronic format and processed later with computer
software.
136
6.3 Results and Discussions
6.3.1 Foam diffusivity and moisture absorption
The model described in section 6.2.2 was utilized to predict the saturated
humidity content of the composite foams. Fiber reinforced composite foams
exhibited lower moisture absorption rate compared to the unreinforced counterpart.
Figure 6.2a and 6.2b show plots of weight gain as a function of square root of time
for composite foam. Clearly, fiber reinforcement reduced the rate of moisture
absorption for composite foams. Hybrid foams (glass: aramid – 1:3) at the saturation
limit showed almost 100% less weight gain compared to unreinforced foam of
similar densities.
The effective diffusivity values predicted by Fick’s second law for phenolic
foams are displayed in Figure 6.3. A ~25 % reduction in effective diffusivity relative
to the unreinforced phenolic foam was predicted in hybrid foams with addition of
glass and aramid fibers. In general, the effective diffusivity decreased with
increasing fiber content because moisture absorption of reinforced composites
involves complex internal stress phenomena [16]. For example, the effective
diffusivity for glass-reinforced phenolic foam is influenced by cross linking and the
cell size reduction that results from fiber addition. Note that phenolic foam with 1
wt% aramid fiber exhibits a ~30 % increase in effective diffusivity with respect to
unreinforced phenolic foam because of the hydrophilic character of the fibers. The
137
results suggest that selection of fiber type is a critical factor that affects foam
diffusivity.
Figure 6.2 (a) (b) Plots for weight gain vs. square root of time
138
Figure 6.3 Evolution of moisture diffusivity for composite foams a) 1 wt% aramid
fibers b) Unreinforced foam c) 1wt % glass fibers d) Hybrid foam – 1wt% aramid
and 3wt% glass fibers e) Hybrid foam – 3 wt% glass and 1wt% aramid
6.3.2 Flammability
Figure 6.4 shows the flammability results for phenolic foam compared to
commercial EPS foam. Fiber reinforcement in phenolic foam had no effect on the
fire properties. However, phenolic foam overall showed significantly greater fire
resistance compared to EPS. EPS lost more than 70% volume after flame exposure,
while phenolic foams (both unreinforced and reinforced) exhibited less than 6%
volume loss during this test.
Table 6-5 also shows results obtained by using the classification system
defined by testing standard UL 94, in which V0 is a superior and V1 is inferior. For
foams identified with V0 rating the burning stopped within 10 second after two
applications of 10 seconds each of a flame to a test bar, whereas, burning stopped
within 60 second after two applications of 10 seconds each of a flame to a test bar for
139
foams with V1 rating. V0 rating for phenolic foam further highlights its superior
flammability properties.
Figure 6.4 Flammability test results: Volume lost as a percentage of initial volume
Table 6-5 Flammability test results: UL 94 ratings assigned to materials
140
6.3.3 Accelerated aging
Results of accelerated aging tests are also summarized in Figure 6-5. Mass loss
as a percentage of initial mass was calculated by subtracting the final mass of each
specimen after six weeks in the environmental chamber from the initial mass, then
dividing by the initial mass and multiplying by 100. The unreinforced phenolic foam
showed mass loss >10% over the six weeks at 40
0
C and 85% humidity.
Figure 6.5 Accelerated aging results: Mass lost as a percentage of initial mass
Fiber reinforcement significantly reduced mass loss in phenolic foams. For
example, almost 300% less mass lost was observed in hybrid foams (aramid: glass -
3:1) compared to unreinforced foam. EPS although lost 25% less mass than
unreinforced phenolic foam but lost almost 110% more mass over time in
141
comparison to hybrid foams. The results highlight hybrid foams long term durability
and its ability to withstand extreme environments. Also, foams with higher
percentages of aramid fibers exhibited greater dimensional stability compared to
other composite foams. These findings indicate that selection of fiber type and
proportion of fibers influences the aging process of composite foams. Thus, selection
of foam composition for insulation applications will be critical in service conditions
with high temperature and dry conditions.
6.3.4 Mechanical performance
Figures 6.6 and 6.7 show the compressive modulus and strength of foams
before and after six-week aging. The compressive modulus and strength of
unreinforced phenolic foam decreased by 35-40% after aging. EPS foams were
weaker than phenolic foams of similar density, and the modulus and strength of EPS
foam was almost 43.5% and 22% less than unreinforced phenolic foam after
undergoing the aging test. Hybrid foams again outperformed the other types of foam
studied here. For example, the hybrid foam with aramid: glass ratio of 1:3 was stiffer
than unreinforced foam, EPS foam and phenolic foams with either glass or aramid
fibers even after six weeks of aging, and the modulus and strength of this foam was
decreased by only 17% and 4% respectively. Although the mechanical performance
of foams at low densities is not of prime importance for insulation and cladding
applications, the preliminary results obtained for compressive properties indicate that
hybrid foams with varied proportions of glass and aramid fibers could be used as a
142
load bearing member in structural components of buildings. This is particularly true
for foams of medium to high density (160-275 kg/m
3
). The suitability of hybrid
foams structural application is presently speculative, and further studies are required
to validate the properties of hybrid foams for building materials.
Figure 6.6 Climatic effects on compressive modulus
143
Figure 6.7 Climatic effect on compressive strength
6.3.5 Scanning electron microscopy
Microscopic examination of foams revealed that fiber additions resulted in a
reduction in cell size. For example, for hybrid foams (aramid: glass – 1:3) the
average cell size was found to be almost 100% less compared to unreinforced foam.
These results are consistent with previous reports [7]. Glass fiber additions had a
similar effect. The glass fibers used for this study were treated with silane coupling
agent and that led to the formation of covalent bonds within the surface and the
development of cross-linked silane films [17]. Thus, the coupling agent bound the
organic material (phenolic foam) to the inorganic glass fibers and resulted in
extensive cross linking [17].
144
The reduction in cell size due to addition of fibers also influenced the water
absorption in composite foams. Hybrid phenolic foams exhibited lower water
absorption and diffusivity relative to unreinforced foams, and also compared to
foams reinforced with only glass or aramid fibers. The reduced water absorption in
the composite foams was obviously attributed to the cell size reduction associated
with fiber additions resulting in inhibited cell growth [7]. We speculate that
enhanced nucleation along with inhibited cell growth was responsible for reduced
cell size in hybrid foams.
Microscopic examination of foams before and after temperature humidity
exposure was performed. Figure 8 shows the SEM images of foams before and after
moisture absorption tests. The glass fiber ends protrude from the foam matrix, which
shows evident of cracking in foam vicinity. Whereas several small fragments of
phenolic foam are found adhering to aramid fibers in hybrid foams indicating a
strong cohesive strength that forces a combination of interface and matrix failure [7].
It is observed from Figure 8 that of all the foams studied here unreinforced phenolic
foams have the largest average cell size. The average cell size for each of these foam
samples was calculated before and after moisture absorption test and is reported in
Figure 9. From data in Figure 9 it is revealed that unreinforced phenolic foam have
average cell size almost 55% greater than hybrid foams. Also it is observed that the
average cell size for EPS foams is comparable to hybrid foams at same density. As
expected, no change was observed in cell size of foams after undergoing six weeks
of temperature humidity exposure in the envoirmental chamber.
145
Figure 6.8 SEM images showing cell lengths across climatic stress
146
Figure 6.9 SEM results: Cell length across climatic stress
6.4 Conclusion
The long-term durability and stability of hybrid composite foams was
evaluated and compared to commercial EPS foams which is a popular choice for
insulation purposes. Measurements were performed to determine moisture sorption
kinetics, flame resistance, and compressive strength.
A statistical approach along with Fick’s second law was applied to predict the
non-steady state diffusivity and water absorption of the hybrid composite foams. The
effective diffusivity decreased with increasing fiber content, and overall hybrid
foams exhibited the lowest diffusivity and moisture absorption of all foams studied.
Increased fiber loading, particularly glass fiber loading, resulted in increased cross-
linking due to formation of cross-linked silane films and thus reduced the effective
cell size. This finding is significant because increased water absorption in insulation
materials can drastically reduce conductivity and eventually its insulation and
147
strength over time. The accelerated aging test revealed that hybrid foams exhibited
dimensional stability superior to that of the other foams tested. Hybrid composite
phenolic foams lost significantly less mass in aging experiments compared to
unreinforced phenolic and commercial EPS foam.
Fiber reinforcement did not affect the flammability properties of phenolic
foam. As expected, all phenolic-based foams exhibited superior fire performance
compared to EPS. In particular, EPS showed significant volume reduction during
exposure to flame, a critical issue for service areas where weather conditions are hot
and arid.
The mechanical performance measured in this study, coupled with earlier
results [7] underscores the potential for applications of medium density hybrid foam
as a fire retardant, low-cost structural element for building structures, providing both
insulation and load-carrying capacity. Finite element modeling along with failure
analysis are in progress to validate behavior under applied loads and determine
suitability of hybrid foams for steel stud assembly in building applications.
We also plan to develop hybrid foams by reinforcing foams with natural fibers
such as jute, cellulose, bamboo etc. These fibers are available in abundance and bear
excellent properties.
148
Chapter 6 References
[1] Ibrahim, Said M.A. (1986). The thermal behavior of thermally insulated and
uninsulated buildings, Energy, 12(7): 615-622.
[2] Klempner, D and Sendijarevic, V. (2004). Polymeric foams and Foam
Technology, Hanser Publishers, Munich.
[3] Khemani, K. (1997). Polymeric foams, American Chemical Society, Washington,
D.C.
[4] http://www.afcee.brooks.af.mil/gree/case/accsfguide.pdf
[5] Brown, V.M., Crump, D.R. and Gardiner, D. (1988). Sources and concentrations
of formaldehyde in households dusts. In Indoor and Ambient Air Quality, ed. Perry,
R. & Kirk, P., Selper Ltd., London, p. 423.
[6] Brown, S.K. (1990). Poly. Deg. and Stab., 27(1) 121-43
[7] Desai, A., Auad, M.L., Shen, H. and Nutt, S.R. (2008). Mechanical Behaviors of
hybrid composite phenolic foam, Journal of cellular plastics, 44(1): 15-36.
[8] Mao, J., Chang, J., Chen, Y. and Fang, D. (1998). Review of Phenolic foam,
Chem. Ind. Eng., 15(3): 38-43.
[9] Alsoy, S. (1999). Modeling of Diffusion in closed cell polymeric foams, Journal
of cellular plastics, 35(3):247.
[10] Bart, G.C.J., Du Cauze De Nazelle, M.R. (1993), Journal of cellular plastics, 29
(1):29.
[11] Cohen, MI. (1994). WO 94/04604.
[12] Nutt, S.R. and Shen, H. (2005). US Patent 6,841,584.
[13] Droin-Josserand, A., Taverdet, J.L., Vergnaud, J.M. (1989). Modeling the
process of moisture absorption in three dimensions by wood samples of various
shapes: cubic, parallelepipedic, 23(1):259-271.
[14] UL plastic standards:http://www.ul.com/plastics/standards.html
[15] ASTM D2126-04
149
[16] Mehta, B.S., Dibenedetto, A.T., Kardos, J.L. (1977). Sorption and Diffusion of
water in glass-ribbon reinforced composites, J. Applied Polymer Science,
21(1):3111-3127
[17] Alonso, M.V., Auad, M.L., Nutt, S.R.(2006). Short fiber reinforced epoxy
foams, Composites part A: Applied Science and manufacturing, 37(11): 1952-1960
150
Chapter 7. Future studies
Future work will be devoted to validate incorporation of phenolic foam as a
structural member in the wall assembly of buildings. Studies will be performed to
evaluate the performance of phenolic foam as a potential material for insulation by
performing two experimental setups as discussed below:
We plan to assemble an 8ft x 8ft x 0.5in. Phenolic foam (unreinforced) wall by
stitching several smaller panels of size 2 ft x 2ft. The target density of the foams here
will be ~60-85 kg/m
3
. This panel will then be tested in the hot box using ASTM C
1363 procedure. The aim is to compare thermal behavior of walls insulated with
phenolic foams to the ones insulated with PU and Styrofoam® and observe the
difference in heat transfer through each wall assembly by calculating the heat
transfer coefficient. The heat transfer through walls will be calculated using methods
of Ibrahim [1] and Skujans et al. [2]. Figure 7.1 shows proposed application of
Phenolic foam as an insulation material in the wall assembly. Phenolic foam will be
introduced as an insulation material between the brick walls and steel stud assembly
as shown in Figure 7.1.
For assemblies where phenolic foam comes in contact with structural members
such as steel, corrosion of the steel member will be avoided using foams with lower
pH value. In this study foams were neutralized in an ammonia chamber for 24-48
hours to eliminate acidic residues. However, ammonia for obvious reason can cause
severe health hazards if inhaled in excessive amounts and limits its industrial scale-
up. We developed a new technique to neutralize foams by adding finely sized borates
151
during the last stage of foam processing. The foams developed using borate particles
were found to have a pH value greater than 7 that makes phenolic foam safe for
structural applications especially where foam has to be placed in vicinity to metals
and could cause their corrosion over time because of reaction between acidic reside
and the metal. This technique of adding borate particles is more practical especially
if foams were to be produced on an industrial scale. Future work will focus on a
more detailed study of this technique.
Figure 7.1 Cross-section of steel stud cavity walls with phenolic foam insulation
between steel studs and concrete.
We also plan to evaluate hybrid foam as a potential load bearing member for
buildings where the foam could be in-built between the steel studs as shown in
Figure 7.2. The steel studs are spaced 24 inches apart in conventional buildings. The
results from chapter 3 suggests that medium density hybrid foams (200-300 kg/m
3
) is
152
Figure 7.2 Cross-section of steel stud cavity walls with phenolic foam insulation
between steel studs
a very strong material, bears excellent mechanical properties and exhibits a more
graceful failure. Thus hybrid foams could not only serve as an insulation material but
also as a potential load bearing member in buildings. Taking advantage of these
properties of hybrid foams we predict that the stud spacing in the buildings (at
present 24 inches) could be increased and thus could lead to significant cost savings.
Finite element modeling along with failure analysis are in progress to validate
behavior under applied loads and determine suitability of hybrid foams for steel stud
assembly in building applications. We foresee phenolic foam as an ideal candidate
for insulation application in steel stud cavity walls.
153
Chapter 7 References
[1] Ibrahim, Said M.A. (1986). The thermal behavior of thermally insulated and
uninsulated buildings, Energy, 12(7): 615-622.
[2] Skujans, J, Vulans, A, Iljins, U and Aboltins, A (2007). Measurement of heat
transfer of multi-layered wall construction with foam gypsum, Applied Thermal
Engineering, 27:1219-1224
154
Bibliography
AFCEE homepage -http://www.afcee.brooks.af.mil/gree/case/accsfguide.pdf
Alonso, M.V., Auad M.L., Nutt, S.R. (2006) Modeling the compressive properties of
glass fiber reinforced epoxy foam using the analysis of variance approach. Comp Sci
and Tech, 66(13): 2126-2134.
Alonso, M.V., Auad, M.L., Nutt, S.R.(2006). Short fiber reinforced epoxy foams,
Composites part A: Applied Science and manufacturing, 37(11): 1952-1960.
Alsoy, S. (1999). Modeling of Diffusion in closed cell polymeric foams, Journal of
cellular plastics, 35(3):247.
ASTM D2126-04.
Babu Prem, E.J., Savithri S. and Pillai U.T.S, Pai B.C. Micromechanical modeling of
hybrid composites. 2005: Polymer, 46 (18):7478-7484.
Bart, G.C.J., Du Cauze De Nazelle, M.R. (1993), Journal of cellular plastics, 29
(1):29.
Benning CJ. Plastic foams: Volume II: structure properties, and applications. New
York: John Wiley and sons 1969.
Bledzki, A.K., Zhang, W., Chate, A. (2001). Natural-fiber-reinforced polyurethane
microfoams, Comp Sci and Tech., 61(16): 2405-2411.
Boeing Part Specification (BPS) D124. (1991), Air, Panels and Plenum Insulation,
Rigid Plastic Foam, for low pressure Systems, Boeing.
Box G.E., Hunter W.G., Hunter J.S. (1978). Statistics for experimenters: an
introduction to design, data analysis and model building. John Wiley and Sons: New
York.
British Pat. 586, 199 1947.
British Pat. 1,077,423 1965.
Broek, D. (1982). Elementary engineering fracture mechanics, Dordrecht: Martinus
Nijhoff: 469.
Brown, S.K. (1990). Poly. Deg. and Stab., 27(1) 121-43.
155
Brown, V.M., Crump, D.R. and Gardiner, D. (1988). Sources and concentrations of
formaldehyde in households dusts. In Indoor and Ambient Air Quality, ed. Perry, R.
& Kirk, P., Selper Ltd., London, p. 423.
Chiang Martin Y.M., Wang X., Schultheisz, C.R. and He J. Prediction and three-
dimensional Monte-Carlo simulation for tensile properties of unidirectional hybrid
composites, Comp. Sci. and Tech. 2005:65 (11-12): 1719-1727.
Cohen, MI. (1994). WO 94/04604.
Davim, J.P., Reis, P., Antonio, C.C. (2004). A study on milling of glass fiber plastics
manufactured by hand-lay up using statistical analysis (ANOVA), Compos
Structures, 64 (2):493-500.
Desai, A., Alonso, M.V., Nutt, S.R (2008). Modeling of fiber reinforced phenolic
foam, Journal of cellular plastics, Article in press.
Desai, A., Auad, M.L., Shen, H. and Nutt, S.R. (2008). Mechanical Behaviors of
hybrid composite phenolic foam, Journal of cellular plastics, 44(1): 15-36.
Droin-Josserand, A., Taverdet, J.L., Vergnaud, J.M. (1989). Modeling the process of
moisture absorption in three dimensions by wood samples of various shapes: cubic,
parallelepipedic, 23(1):259-271.
FR-6700® Series foam product (2004). General Plastics manufacturing company.
Fukuda, H. (1983). An advanced theory of the strength of hybrid composites, Journal
of Materials Science, 19:947-82.
Gan, Y.X., Chen, C., Shen, Y.P. (2005). Three-dimensional modeling of the
mechanical property of linearly elastic open cell foams, Int Journal of Solids and
Structures, 42 (26): 6628-6642.
Gent AN, Thomas AG Mechanics of foamed elastic materials. Rubber Chem
Technology. 1963:36:597-610
Gibson, L.J. and Ashby, M.F. (1997). Cellular solids: structure and properties,
Cambridge, UK.
Gibson, L.J. and Ashby, M.F. (1998). Cellular solids: structure and properties.
Oxford: Pergamon Press.
156
Gibson LJ, Ashby MF. The mechanics of three-dimensional cellular materials. Proc
Royal Soc London a 1982:382:43-59.
Goodman SH, Handbook of Thermoplastics. Second edition: 1998.
Goods, S.H., Neuschwanger, C.L., Whinnery, L.L., Nix W.D. (1999). Mechanical
properties of a particle strengthened polyurethane foam, J Appl Polym Sci,
74(11):2436-724.
Gupta N., Maharsia R., Jerro H.D. Enhancement of energy absorption characteristics
of hollow glass particle filled by rubber addition, Mate. Sci and Engg. A. 2005: 395
(1, 2):233-240.
Halpin, J.C. and Kardos, J.L. (1972). Moduli of crystalline polymers employing
composite theory, J. Appl. Phys., 43(5):2235-2241.
Halpin, J.C. and Tsai, S.W. (1967). Effects of Environmental Factors on composite
materials, Rep. No: AFML-TR-67-423, Air Force Materials Laboratory, Wright
Patterson Air Force Base, Ohio.
Hilyard NC Cunningham A (ed.) Low density cellular plastics: physical basis od
behavior. New York: Chapman & Hall 1994
Hirsch, T.J., J. Am. Con. Inst., (59):427-451.
Hyer, M.W. and White, S.R. (1998), Stress analysis of fiber- reinforced composite
materials, New York: McGraw-Hill.
Ibrahim, Said M.A. (1986). The thermal behavior of thermally insulated and
uninsulated buildings, Energy, 12(7): 615-622.
Idicula M., Neelakantan N.R., Oommen Z., Joseph K. and Thomas S. A study of the
mechanical properties of randomly oriented short banana and sisal hybrid fiber
reinforced polyester composites, J of App. Poly. Science. 2005: 96 (5):1699-1709.
Jacobson, R., Caulfield, D., (2003). Hybrid Composites: Combining cellulose fibers
and woolastonite mineral fibers into a nylon 6 matrix, The seventh international
conference on woodfiber-plastics composites, May 19-20.
JP 03-086506 1991
JP 2001-228835
157
Kalaprasad, G., Joseph, K. and Thomas S. (1997), Theoretical modelling of tensile
properties of short sisal fiber-reinforced low-density polyethylene composites, J. of
Mat. Sci., 32(16): 4261-4267.
Kaynak, C., Arikan, A. and Tincer, T. (2003). Flexibility improvement of short glass
fiber reinforced epoxy by using a liquid elastomer, Polymer, 44 (8): 2433-2439.
Khemani, K. (1997). Polymeric foams, American Chemical Society, Washington,
D.C.
Klempner, D and Sendijarevic, V. (2004). Polymeric foams and Foam Technology,
Hanser Publishers, Munich.
Knop, A. and Scheip, W. (1979). Chemistry and application of phenolic resins. New
York; Springer
Ko WL. Deformation of foamed elastomers. J Cellular Plastics. 1965:1:45-50
Kretsis, G. (1987). A review of the tensile, compressive, flexural and shear
properties of hybrid fiber-reinforced plastic, Composites, 1:13-23.
Landrock AH (ed.) Handbook of plastic foams. New Jersey: Noyes Publications,
1995
Lauscha fiber International (2002). Technical data sheet.
Lin-Gibson S. PhD thesis. Virginia Polytechnic Institute, 2001
Manabu, H. (1993). JP 05-318506.
Mao, J., Chang, J., Chen, Y. and Fang, D. (1998). Review of Phenolic foam, Chem.
Ind. Eng., 15(3): 38-43.
Marom, G., Fischer, S., Tuler, F.R. and Wagner, H.D. (1978). Hybrid effects in
composites: conditions for positive or negative effects versus rule of mixtures
behavior, Journal of Materials Science, 13: 1419-26
Mehta, B.S., Dibenedetto, A.T., Kardos, J.L. (1977). Sorption and Diffusion of water
in glass-ribbon reinforced composites, J. Applied Polymer Science, 21(1):3111-3127
Mendez, P. and Ordonez, F. (2005), Scaling Laws from Statistical Data and
Dimensional Analysis, J. Appl. Mech., 72: 648 -657.
158
Menges G, Knipschild F. Estimation of mechanical properties for rigid polyurethane
foams. Polymer Engr Sci. 1975:15:623-627
Moiseyev AA, Popov VA, Borodin MY. (eds.) Expanded plastics. New York:
Macmillian, 1963
Montogomery, D.C. (1991). Design and analysis of experiments. Wiley and Sons,
Inc.: New York.
Montogomery, J.S., (1991). Diseño y Análisis de Experimentos, Editora
Panamericana S.A., México
Morimoto K, Suzuki T. Ind Engr Prod Res Dev 1984:23:81
Morimoto K, Suzuki T, Yosomiya R. Polym Engr Sci. 1984:24(12):943; 1000
Nutt, S.R. and Shen, H. (2005). US Patent 6,841,584.
Porter SC, Verseput RP, Cunningham CR. Pharm Technol 1997: October: 1-7
Radesh Kumar, C., George, K.E. and Thomas, S. (1996). Morphology and
mechanical properties of thermoplastic elastomers from nylon-nitrile rubber blends,
J. Appl. Poly. Sci., 61 (13): 2383-2396.
Robinson, I.M. and Robinson, J.M. (1994). The influence of fiber aspect ratio on the
deformation of discontinuous fiber-reinforced composites, J. Mater. Sci., 29
(18):4663-4677.
Rosen, B.W. (1965), Fiber composite Materials (American Society for metals,
Metals Park, OH):58.
Saint-Michel, F., Chazeau, L., Cavaille, J.Y., Chabert, E. (2006). Mechanical
properties of high density polyurethane foams: II Effect of filler size, Comp Sci. and
Tech., 66(15): 2709-2718.
Sarasua, J.R. , Remiro, P.M. and Pouyet, J. (1995). The mechanical behavior of
PEEK short fiber composites, J. Mater. Sci., 30(13):3501-3508.
Saunder JH, Hansen RH. Plastic foams. Volume I. New York, 1972
Shan Y. and Liao K. Environmental fatigue behavior and life prediction of
unidirectional glass–carbon/epoxy hybrid composites. Int. J. Fatigue. 2005: (24):847-
859.
159
Shao-Yun, F., Yiu-Wing, M., Lauke, B. and Chee-Yoon, Y. (2002). Synergistic
effect on the fracture toughness of hybrid short glass fiber and short carbon fiber
reinforced polypropylene composites, Mater. Sci. Eng. (A), 323 (1-2): 326-335.
Shen, H., Lavoie, A.J. and Nutt, S.R. (2003). Enhanced peel resistance of fiber
reinforced phenolic foams, Comp. A: Appl. Sci. Manufact., 34(10): 941-948.
Shen, H. and Nutt, S.R. (2003). Mechanical characterization of short fiber reinforced
phenolic foam, Comp. Part A, 34(9): 899-906.
Shen, H., Nutt, S.R. and Hull, D. (2004). Direct observation and measurement of
fiber architecture in short fiber-polymer composite foam through micro-CT
imaging, Comp. Sci. and Tech., 64 (13-14):2113-2120.
Siegmann, A., Kenig, S., Alperstein, D., and Narkis, M. (1983). Mechanical behavior
of reinforced polyurethane foams, Polymer Composites, 4(2): 113-119.
Skujans, J, Vulans, A, Iljins, U and Aboltins, A (2007). Measurement of heat transfer
of multi-layered wall construction with foam gypsum, Applied Thermal Engineering,
27:1219-1224
Takanori, K. (1993). JP 05-286069.
Technical guide of Nomex® Brand Fiber (2002). Technical information, DuPont.
UL-plastics standards:http://www.ul.com/plastics/standards.html
USP 2,582,228 1952
USP 2,678,298 1954
USP 2,728,741 1955
USP 2,772,246 1956
USP 2,789,054 1957
USP 2,895,173 1959
USP 3,271,331 1966
USP 3,300,419 1967
USP 3,389,094 1968
160
USP 4,163,824 1979
USP 4,900,616 1990
USP 6,841,584 2005
Vaikhanski, L. and Nutt, S.R. (2003). Fiber-reinforced composite foam from
expandable PVC microspheres, Comp. Part A, 34 (12): 1245-1253.
Warren WE, Kraynik AM. Foam mechanics: the linear elastic behavior of
polydisperse response of two-dimensional spatially periodic cellular materials.
Mechanics Mat 1987:6:27-37
Williams RE. Space filling polyhedron: its relation to aggregates of bubbles, plant
cells and metal crystallites. Science 1964:161:276-7
WO 94-04604 1994
Yosomiya R, Morimoto K. Ind Engr Chem Prod Res Dev 1984:23:605
Youssef, S., Maire, E., Gaertner, R. Finite Element modeling of the actual structure
of cellular materials determined by X-ray tomography. Acta Mater, 53(3): 719-730
Zenkert, D. (1995). An introduction to sandwich construction, London, Chamelon.
Zhu, H.X., Hodbell, J.R., Windle, A.H. (2004). Effect of cell irregularity on the
elastic properties of open-cell foams. Acta Mater, 48: 4893-900.
Abstract (if available)
Abstract
Hybrid composites in recent times have been developed by using more than one type of fiber reinforcement to bestow synergistic properties of chosen filler and matrix and also facilitating the design of materials with specific properties matched to end use. However, the studies for hybrid foams have been very limited because of problems related to fiber dispersion in matrix, non-uniform mixing due to presence of more than one filler and partially cured foams. An effective approach to synthesize hybrid phenolic foam has been proposed and investigated here. Hybrid composite phenolic foams were reinforced with chopped glass and aramid fibers in varied proportions. On assessing mechanical properties in compression and shear hybrid foams exhibited a more graceful failure, greater resistance to cracking,and were significantly stiffer and stronger than foams with only glass and aramid fibers.The optimum ratio or the reinforced hybrid phenolic foam system was found to be 1:1 ratio of glass:aramid fibers. Also, the properties of hybrid foams were found to deviate from rule of mixture (ROM) and thus the existing theories of fiber reinforcement fell short in explaining their complex behavior.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Fiber-reinforced syntactic foams
PDF
Environmental effects on the hybrid glass fiber/carbon fiber composites
PDF
Metallic syntactic foams synthesis, characterization and mechnical properties
PDF
Multiwall carbon nanotubes reinforced epoxy nanocomposites
PDF
Fiber reinforced phenolic foam: climatic effects on mechanical properties and building applications in northern Thailand
PDF
Constitutive model of concrete confined by advanced fiber composite materials and applications in seismic retrofitting
PDF
Fatigue and fracture of pultruded composite rods developed for overhead conductor cables
PDF
Synthesis, characterization, and mechanical properties of nanoporous foams
PDF
Nanoclay-reinforced polyurethane
PDF
Ageing and mechanical failure of fiber reinforced polymers
PDF
Polymer flow for manufacturing fiber reinforced polymer composites
PDF
Chemical recycling of amine/epoxy composites at atmospheric pressure
PDF
Experimental study of acoustical characteristics of honeycomb sandwich structures
PDF
The potential applications of natural fiber reinforced bio-polymers (NFRbP's) in architecture
PDF
Deformation behavior and microstructural evolution of nanocrystalline aluminum alloys and composites
PDF
In situ process monitoring for modeling and defect mitigation in composites processing
PDF
Developing efficient methods for the manufacture and analysis of composites structures
PDF
Vacuum-bag-only processing of composites
PDF
Development of high frequency focused transducers for single beam acoustic tweezers
PDF
Sound transmission through acoustic metamaterials and prepreg processing science
Asset Metadata
Creator
Desai, Amit
(author)
Core Title
Fiber reinforced hybrid phenolic foam
School
Viterbi School of Engineering
Degree
Doctor of Philosophy
Degree Program
Materials Science
Publication Date
09/08/2008
Defense Date
07/31/2008
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
fibers,hybrid foams,insulation,OAI-PMH Harvest,phenolics
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Nutt, Steven R. (
committee chair
), Goo, Edward K. (
committee member
), Sammis, Edward (
committee member
)
Creator Email
aadesai@gmail.com,amitdesa@usc.edu
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-m1586
Unique identifier
UC1286675
Identifier
etd-Desai-2355 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-108228 (legacy record id),usctheses-m1586 (legacy record id)
Legacy Identifier
etd-Desai-2355.pdf
Dmrecord
108228
Document Type
Dissertation
Rights
Desai, Amit
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Repository Name
Libraries, University of Southern California
Repository Location
Los Angeles, California
Repository Email
cisadmin@lib.usc.edu
Tags
fibers
hybrid foams
phenolics