‘§§3.V

.2

. r.
x. . . ._ . ,
£3. a , 2.941;)...“ .,
.. «guns .5... flaw: .13 ad:
mumuhflfinw {a 33$th 2
I 3 . , J ‘ .11» I y;
we , . .. A? .4
. a '1. a a
,. . . sauna}? :4
, . mfim. 4
, #53»:

{11‘

o

I]
fix, i

In)

..
r .
.L‘ , (It: £1... tr ,
v I!
. 2 r ‘1»:
mafia,» a .3,
xfaamw ..

.354».

l I a
‘5‘] ' I- .
firsfivfismfiiwg

1‘ I59; fi‘coK—S V...
K... .2. Viahuwflf

. sxi .
I. T 0...! I: X,
.r..?&.uhnn9u..u..mmmn.e. .1.
again..ryihélxv

it.

511“ ..
3.1.. fat:
1: .6“de 0.....l

 

 

43's-

‘ k

1

If
m3

{757/40 37

 

fiBRARY , ‘
Michigan State
University

This is to certify that the
thesis entitled

ADHESION BETWEEN NEXTEL 312TM FIBERS AND
BLACKGLASTM SILICON OXYCARBIOE AND ITs EFFECT
ON COMPOSITE PROPERTIES

presented by

John Allen Helmuth

has been accepted towards fulfillment
of the requirements for the

Master’s degree in Materials Science

flii'L/IZ’LLéL/ 7/ ' 7/310?

Major Professor’s 3:961;er \-

May 26, 2003

MSU is an Affirmative Action/Equal Opportunity Institution

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

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 cJClRCJDateDuesz-p. 1 5

 

ADHESION BETWEEN NEXTEL 312TM FIBERS AND BLACKGLASTM SILICON
OXYCARBIDE AND ITS EFFECT ON COMPOSITE PROPERTIES

By

John Allen Helmuth

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Chemical Engineering and Materials Science

2003

ABSTRACT

ADHESION BETWEEN NEXTEL 312TM FIBERS AND BLACKGLASTM SILICON
OXYCARBIDE AND ITS EFFECT ON COMPOSITE PROPERTIES

By

John Allen Helmuth

Increasing demand for high strength, light weight composite structures for use in
gas turbine engine applications has fueled interest in ceramic matrix composites. CMC’s
typically exhibit high strength with low ductility and often require expensive
manufacturing processes, however, a polymer precursor to a ceramic matrix reinforced
with continuous ceramic fibers offers high strength coupled with ductile fracture modes.
Fabricated with conventional epoxy composite tooling, a Silicon Oxycarbide reinforced
with a Boron Nitride treated aluminoborosilicate fibers is capable of meeting 1300°F
service environments found in gas turbine engines. This continuously fiber reinforced
CMC lends its improved toughness to controlled fiber pull-out during failure which is
created through the Boron Nitride interphase. This work characterizes fiber-matrix
adhesion of untreated and BN treated Nextel 312TM fibers through fiber pull-out and
indentation testing. Contact angle and wettability measurements also are assessed.
Flexure and short beam shear testing of both unidirectional and woven composites were
used to verify adhesion observations. Environmental Electron Scanning Microscopy
(ESEM) performed in conjunction with composite processing provided insight to matrix
cracking and fiber-matrix differential expansion effects. This work served to quantify

performance gains associated with the BN treatment of Nextel 312TM composites.

ACKNOWLEDGNIENTS

I would like to express my thanks to Dr. L.T. Drzal for his guidance through my
graduate studies along with Mr. M.J. Rich and the staff and students of the Composite
Materials and Structures Center. Also, I would like to thank Dr. Richard Schalek for his
assistance with this project including numerous informative discussions. Additionally,
financial and technical support was provided by the Honeywell Corporation (formally
Allied Signal, Inc.), Pratt & Whitney Auto Air, Inc., and Jet Engineering, Inc. Finally, I
would like to thank my family for their patience and support throughout this extended

project, without whom, this would not have been possible.

iii

List of Tables

List of Figures

Table of Contents

Chapter 1: INTRODUCTION

Chapter 2: BACKGROUND

2.1 Matrix Materials

2.1.1

2.1.2

2.1.3

2.1.4

2.1.5

2.1.6

2.1.7

Glass and Glass-Ceramics

Mullite

Alumina

Zirconia

Silicon Carbide

Silicon Nitride

Carbon

2.2 Fiber Reinforcements

2.2.1

2.2.2

2.2.3

2.2.4

2.2.5

Glass

Alumina
Graphite\Carbon
Silicon Carbide

Silicon Nitride

2.3 Fabrication Methods

2.3.1

2.3.2

Hot Pressing

Hot Isostatic Pressing

iv

viii

ix

10

10

11

11

12

13

14

15

16

16

18

Chapter 3

2.4

2.5

2.3.3 Chemical Vapor Infiltration
2.3.4 Direct Metal Oxidation
2.3.5 Sol-Gel Method

2.3.6 Polymer Precursors

2.3.7 Silicon Oxycarbide

CMC Interphase

2.4.1 Mechanical Contribution (Frictional and
Residual Stresses)

2.4.2 Chemical Contribution

Failure Mode

2.5.1 Fiber-Matrix Debonding

2.5.2 Fiber Deformation and Fracture

2.5.3 Fiber Pullout

EXPERIMENTAL

3.1

3.2

3.3

BlackglasTM Polymer Precursor Reinforced with
BN Treated Nextel 312TM Fibers

3.1.1 BlaclcglasTM Matrix System
3.1.2 Nextel 312TM Fibers

3.1.3 Boron Nitride Treatment

Characterization of as Received, Desized, and BN
Treated Nextel 312TM Fibers

3.2.1 XPS Analysis
3.2.2 Auger Electron Spectroscopy
3.2.3 Stress-Strain Measurements

Composite Specimen Fabrication

19

2O

22

22

23

26

27

27

28

29

29

30

32

32

33

34

34

36

37

38

38

40

Chapter 4

3.4

3.5

3.6

3.3.1 Boron Nitride Fiber Treatment
3.3.2 Composite Prepregging

3.3.3 Polymer Cure

3.3.4 Pyrolysis

3.3.5 Reinfiltration Process
Adhesion Measurements

3.4.1 Interfacial Shear Strength
3.4.2 Fiber Pull-Out

3.4.3 Contact Angle and Wettability
Composite Performance Analysis
3.5.1 In-Situ ESEM Pyrolysis

3.5.2 Flexure Testing

3.5.3 Short Beam Shear Testing
3.5.4 Iosipescu In-Plane Shear Testing

Fracture Analysis

RESULTS AND DISCUSSION

4.1

4.2

4.3

Characterization of Fiber Samples
4.1.1 XPS Results

4.1.2 Auger Spectroscopy Results
4.1.3 Stress-Strain Measurements
Composite Specimen Fabrication
Adhesion Measurements

4.3.1 Interfacial Shear Strength

vi

41

41

42

42

t

46

46

47

47

50

50

50

51

52

52

52

57

57

58

62

62

Chapter 5

4.3.2 Fiber Pull-Out

4.3.3 Wettability

4.4 Composite Performance Analysis

4.4.1 In—Situ ESEM Pyrolysis

4.4.2 Flexure Strength

4.4.3 Short Beam Shear Strength

4.4.4 Iosipescu In-Plane Shear Strength
4.5 Fracture Analysis
4.6 Surface Chemistry of Fracture Surfaces
CONCLUSION

REFERENCES

vii

66

66

67

67

76

77

79

81

9O

96

97

Table 2.1
Table 3.1
Table 4.1
Table 4.2
Table 4.3
Table 4.4
Table 4.5
Table 4.6
Table 4.7
Table 4.8
Table 4.9
Table 4.10

Table 4.11

List of Tables

Properties of typical reinforcement and matrix materials.

Nextel 312TM properties.

Fiber surface chemistry (atomic%).

Sample weight gain through densification cycles.
Interfacial shear strength.

As-received flexure strength.

BN treated flexure strength.

BN treated short beam shear strength.
As-received short beam shear strength.
As-Received Iosipescu in-plane shear strength.
BN treated Iosipescu in-plane shear strength.
Fracture surface atomic concentrations.

Matrix atomic concentrations.

viii

34

53

61

63

76

77

78

78

79

8O

93

93

Figure 2.1
Figure 2.2
Figure 2.3
Figure 2.4

Figure 2.5

Figure 2.6

Figure 2.7

Figure 3.1
Figure 3.2
Figure 4.1
Figure 4.2
Figure 4.3
Figure 4.4
Figure 4.5
Figure 4.6

Figure 4.7

Figure 4.8

List of Figures

Schematic representation of hot pressing equipment.
Schematic of hot isostatic pressing equipment.
Schematic of chemical vapor infiltration method
Process description for directed metal oxide technique

Preparation of glasses, glass—ceramics and ceramics by
the sol-gel process.

Processing steps required to prepare the four types of
polymer derived ceramics.

Schematic representation of fiber-matrix interphase in a
composite material.

Schematic of interfacial testing system.

Pyrolysis process.

ESEM micrograph of as-received Nextel 312TM fiber surface.

ESEM micrograph of desized Nextel 312TM fiber surface.

ESEM micrograph of BN treated Nextel 312T“ fiber surface.
Typical weight gain curve.

IFSS micrograph of as-received sample.

IFSS micrograph of BN treated sample

ESEM in-situ pyrolysis. Top: As received fiber composite
at 650X and room temperature. Middle: Desized fiber

composite at 650X and room temperature. Bottom: BN
treated fiber composite at 1500X and 161°C.

ESEM in-situ pyrolysis. Top: As received fiber composite
at 1000X and 300°C. Middle: Desized fiber composite

at 1000X and 301°C. Bottom: BN treated fiber

composite at 2500X and 327°C.

ix

17

18

19

21

22

23

26

45

49

54

55

56

61

65

70

71

Figure 4.9

Figure 4.10

Figure 4.11

Figure 4.12

Figure 4.13

Figure 4.14

Figure 4.153
Figure 4.15b
Figure 4.16a
Figure 4.16b
Figure 4.17a
Figure 4.17b
Figure 4.18

Figure 4.19

Figure 4.20

ESEM in-situ pyrolysis. Top: As received fiber composite at
600X and 501°C. Middle: Desized fiber composite at 500x and
500°C. Bottom: BN treated fiber composite at 1500X and 499°C.

ESEM in-situ pyrolysis. Top: As received fibercomposite
at IOOOX and 599°C. Middle: Desized fiber composite

at IOOOX and 600°C. Bottom: BN treated fiber

composite at 1500X and 600°C.

ESEM in-situ pyrolysis. Top: As received fiber composite
at 600x and 900°C. Middle: Desized fiber composite

at 850x and 900°C. Bottom: BN treated fiber

composite at 1500X and 900°C.

ESEM in-situ pyrolysis. Top: As received fiber composite
at 3050X and 900°C. Middle: Desized fiber composite

at 1500X and 1000°C. Bottom: BN treated fiber
composite at 2500X and 901°C.

BN Treated and As-Received Flexure Load vs. Displacement.
Illustrating .05% strain to failure for as received and .5% for the
BN treated fiexure samples.

Mechanical testing summary.

Delaminated desized composite sample at low magnification.
Delaminated desized composite sample at high magnification.
As received low magnification fracture surface.

As received high magnification fracture surface.

BN treated low magnification fracture surface.

BN treated high magnification fracture surface.

Schematic diagram illustrating the possible locus of failure.
EDX spectra taken from a matrix region exposed during

fiber pull-out (4.19a) and an enlargement of the low energy
section of (4.19a) shown in (4. 19b).

(4.20a) EDX spectra taken from the fiber surface showing the

presence of boron and nitrogen and (4.20b) an enlargement of
the low energy section of the spectra shown in (4.20a).

74

77

8O

85

86

87

88

89

92

Chapter 1

INTRODUCTION

Recent cost reduction goals coupled with aggressive performance requirements have
hastened the development of structural composite materials utilized in hostile operating
environments. The aerospace and power generation industries demand increased
performance from their respective products through the incorporation of high performance
structural materials. Components fabricated from advanced materials must be capable of
meeting operating temperatures in the 1000°F to 2500°F regime, withstand supersonic air
velocities, and resist both thermal oxidative degradation and erosion. Gas turbine engine
and ground based power generation applications will benefit from the incorporation of these
materials as weight reductions and efficiencies gained from increased combustion
temperatures elevate specific power ratings. Component examples include exhaust tailcones,
interturbine seals, and thrust vectoring surfaces. Additionally, performance in automotive
applications is expected to increase as turbocharger rotors, valves and piston rings are
designed with advanced composite materials to reduce weight and enhance wear resistance.
Materials capable of operating in these environments have been researched for many years
and several candidates are under consideration.

Traditionally utilized high performance alloys have performed adequately in the
aforementioned operating environments but suffer from property degradation due to long
term elevated temperature exposure and hot gas corrosion. The weight associated with the

most commonly used alloys is typically three times that of conventional composite structures,

which are primarily polymer composites. Although advanced polymer matrix composites
have served industry well in replacing ferrous and nonferrous metallic structures they
typically are limited to exposure below 600°F, which warrants the continued development
of high temperature composite materials.

Composite materials capable of operating in the 1000°F to 2500°F have been
developed but reinforcement and matrix interactions have not been fully understood. Metal
matrix composites (MMC) have exhibited increased elevated temperature properties and
offer “tailorable” properties unlike conventional alloys. The MMC has been developed to
improve upon the performance of conventional high performance alloys by adding the
stiffness and weight reductions offered by platelet, particulate, and fibrous reinforcements.
This combination of materials demonstrates specific strength gains over traditional alloys,
but MMC are also susceptible to degradation at elevated temperatures. MMC structures
demonstrate a thermal expansion mismatch phenomenon resulting in fiber damage as stresses
between the matrix material and reinforcement material build during thermal cycling. While
numerous fiber and reinforcement treatments have been studied, the inherent coefficient of
thermal expansion mismatch phenomenon continues to hamper progress [1].

While many ceramic materials are stable at temperatures of 1000°F to 2500°F, they
are prone to uninhibited crack growth and do not possess the damage tolerance and load
carrying capability required of structural components. The addition of a reinforcement
constituent may modify the ceramic matrix to the extent that damage tolerance is increased,
as a result, ceramic matrix composites (CMC) potentially will replace many exotic alloys
now functioning from 600°F to 2500°F. The CMC future is largely dependent on the

complete understanding of structure-processing—property relationships and fiber-matrix

interactions, specifically, as they pertain to composite toughness.

While many materials may be selected for use in hostile environments based on their
mechanical performance, several economic concerns must be addressed prior to the final
application. Quite often materials, which are researched, have limited availability or require
exotic manufacturing processes. In the development phase, caution must be exercised to
insure advanced material candidates are selected from readily available sources within
acceptable industry cost structures. The aerospace industry has typically been utilized for
“high end” material research and development; however, the end product should be
accessible to automotive industries as well.

When cost is estimated, the figure is a result of many considerations. A
comprehensive review of costs such as raw material, tooling, fabrication, and by-product
disposal as well as final component weight reduction, durability and process ability must be
made. Traditional alloys such as titanium and nickel-based superalloys have relatively high
raw material costs and require expensive tooling schemes for end component fabrication.
High temperature MMC structures also utilize titanium and nickel—based superalloys for the
matrix constituent in addition to silicon carbide, tungsten or other fiber reinforcements to
obtain maximum performance. Fiber and matrix costs for the aforementioned systems are
costly and the required fabrication processes such as hot isostatic pressing can be limited by
the complexity of component geometry. Ceramic matrix systems, which offer fabrication
methods utilizing existing composite tooling and processing methods in conjunction with

low cost commercially available reinforcements, are the goal of more recent initiatives.

Chapter 2

BACKGROUND

Typically ceramic materials are thought to possess very low toughness and exhibit
susceptibility to sudden and catastrophic failure. Throughout ceramic development efforts to
increase toughness have been undertaken through the addition of various reinforcements.
Particulates, whiskers, chopped fibers, and more recently, continuous fiber reinforcements
have been employed to prevent rapid and uncontrollable crack propagation through structural
ceramic components.

The key to the CMC development is the “controlled” propagation of cracks and the
subsequent dispersion of crack energy within the laminate structure. A laminate structure,
which exhibits excellent toughness, will attain this attribute through graceful fiber debonding
and, under ultimate loading conditions, fiber pullout. This “strain” toughness is coupled with
reduced strength as the sliding or debonding nature of the system fails to efficiently transfer
the load carried by the brittle matrix to the high strength reinforcing fibers. In the event that
the debonding effect is eliminated, which may occur during composite processing or during
demanding therrnoxidative exposure, the load transfer may become so efficient that the
continuously reinforced CMC will perform much like a traditional monolithic ceramic.
Characterizing and controlling the interphase, which governs the ductile-brittle CMC, will
enable applications to expand. Along with interphase engineering, careful selection of
complimentary reinforcement and matrix materials must be made to insure satisfactory

performance from the final product. A material overview is provided in Table 2.1.

Table 2.1: Properties of typical reinforcement and matrix materials [2].

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Density Tensile Youngs’ Thermal Fiber
(g/cmz) Strength Modulus Expansion Diameter
(MPa) (GPa) (lO’°/°C) (um)
Fibers:
High Modulus Carbon 1.8 1900 530 radial 8.0 8.0
High Strength Carbon 1.8 2760 275 axial -
CVD SiC Monofilarnents 3.2 3450 415 4.8 140
Nicalon SiC Yarn 2.6 2060 220 3.1 10-15
Boron Monofilament 2.5 2750 400 4.7 100-200
Boron Nitride 1.9 1380 90 7.5 -
Boron Carbide 2.4 2275 90 4.5 -
Alumina 3.9 1400 385 8.5 20
Alumina-Borosilicates:
Nextel 312TM 2.5 1550 152 5.5 1 1
Nextel 440TM 3.1 2000 190 5.5 1 1
Whiskers: ’
SiC 3.2 21000 840 4.8 0.1-0.6
Alumina 3.9 21000 430 8.6 -
Boron Carbide 2.5 14000 480 4.5 -
Graphite 1.7 21000 700 2.0 -
Silicon Nitride 3.2 14000 380 2.8 -
Matrix Glasses
&
Ceramics:

Borosilicate Glass 2.3 100 60 3.5 -
Silica Glass 2.3 120 70 0.5-8.6 -
Mullite 2.8 185 145 5.3 -
Si3N4 3.2 410 310 2.8 -
SiC 3.2 310 440 4.8 -
Alumina (A1203) 3.6 250 330 8.5 -
Zirconia (ZrOz) 5.7 140 250 7.6 -

 

 

2.1 MATRIX MATERIALS:

The design and fabrication of high temperature structures may be customized for a
given environment through the use of many constituent materials. The demands placed on
advanced materials today cannot be met by utilizing a single material, but rather, require the
benefits achieved by combining materials to form composite structures. Current CMC
systems are fabricated from many commercially available constituents and recent
advancements in matrix and reinforcement materials have made it possible to target high
temperature hostile environments. Matrix materials of interest include, but are not limited to,
glasses, glass-ceramics, mullite, alumina, Zirconia, silicon carbide, and silicon nitride.
2.1.1 Glass and Glass-Ceramics:

Glass and glass-ceramic materials may be produced in large quantities and have
been utilized in flat glass applications, containers, tableware, fiberglass, and specialty glasses
requiring additional thermal stability. A wide variety of applications result from low cost of
manufacture and tailorable properties. The production of traditional materials such as soda-
lime-silica glass and borosilicate glass requires silica sand ($02), which has been sized from
40 to 140 mesh and cleaned to remove any refractory heavy metal impurities, to create the
glassy structure. Upon heating materials such as crushed limestone and soda ash through the
700°F to 1500°F range, CaO, MgO, and Na20 fluxes are generated with an accompanying
evolution of gases which serve to mix the melt. Increasing the melt temperature to 1800°F-
2200°F while additions of silica sand are made complete the basic melt required to produce
traditional glass products. Additionally, the melt may be supplemented with feldspatic sand,

which contains alumina (A1203) to reduce the thermal coefficient of expansion, increase

tensile strength, and improve resistance to chemical attack. In the production of fiberglass
and borosilicate specialty glasses used in laboratories, additions of borate rich materials such
as borax further increase thermal shock resistance while improving strength [3].

The traditional glass formulations have also been used in structural applications when
properly reinforced to enhance toughness. Glass matrix structures reinforced with silicon
carbide fibers have been fabricated and exhibit fracture toughness increases from 0.7 ksN in
to 27-ksi‘lin [4]. Carbon fibers have also been used to reinforce glass matrix composites,
however, the melt temperature required for the permeation of the fiber preform results in
oxidation of the reinforcement. Applications of glass matrix composites are somewhat
limited due to the matrix use temperature limit of 1300°F- 1400°F [5].

The further processing of traditional glasses to produce glass-ceramic matrix
compositions has been utilized to enhance thermal stability, strength, and toughness. Glass
exhibits a structure, which is noncrystalline and has short-range order. During glass
processing, great effort is expended to prevent crystallization because it creates a product,
which has poor optical properties in addition to reduced mechanical performance. The
controlled crystallization through the incorporation of nucleating agents such as Ti02
followed by heat treatment to produce a fine grained structure results in a glass-ceramic
structure. This matrix material is readily formed into complex shapes and contains near zero
porosity. The maximum glass-ceramic matrix service temperature depends on the
composition; however, the 1450°F to 2200°F range is typical. Commercially available glass-
ceramic systems include LiZO-Alzog-SiOZ and MgO-Aleg-SiOz, which possess near zero
thermal expansion coefficients with tensile strengths of 150—170 MPA. The glass-ceramic

materials offer improved properties to those of their amorphous glass counterparts and also

may be toughened substantially with reinforcements [6].

Fiber reinforced glass-ceramic composites have been fabricated using a melt process,
which begins with a powdered glass and preform. The glass-ceramic precursor is melted into
the preform and through subsequent heat treatment the fine-grained crystalline structure is
formed throughout the matrix. Silicon Carbide fibers have been utilized in this fabrication
process and composite structures with volume fractions approaching 45-50% have been
produced [7].

2.1.2 Mullite:

Mullite is regarded as a refractory having the composition (3Al203-28i02),
demonstrating exceptional resistance to thermal shock and low CTE of 5.5 x 106/°C.
Mullite is commonly created by the reaction sintering of A1203 and Si02 powders at
temperatures at or above 1600°C (2900°F). During processing, care is taken to obtain the
best possible properties by creating a fine-grained structure without Morphous boundary
phases. Mullite structures include monolithic and SiC reinforced composites which
demonstrate tensile properties of 200 MPA (29 ksi) and 425 MPA (61.5 ksi) respectively.
Mullite is also found in high refractory bricks where corrosion resistance is required [8, 9].
2.1.3 Alumina:

Alumina (A1203) may be utilized in a monolithic form or as the matrix constituent in
CMC structures when high temperature stability is required. A1203 is formed by the heating
of alumina hydrates producing 95-99% pure alumina at temperatures of 1200°C (2190°F).
Due to strong chemical bonding, the stable oc-Al203 structure demonstrates the greatest
hardness relative to other oxides and exceptional strength. Alumina monoliths possess

tensile strengths of 150-200 MPa at temperatures of 1000°C (1830°F). The processing of

alumina monoliths and composites is typically carried by the compaction of powders.
Applications of these structures range from piping linings and cyclones where wear
resistance and impact strength is valued to electrical substrates where thermal stability and
low electrical conductivity are required [10].

2.1.4 Zirconia:

Zirconia (Zr02) is manufactured from Zircon (Zr02-Si02), which is naturally
occurring in beach sands found primarily in Australia and South Africa. The Zircon may be
chlorinated with carbon to form zirconium and silicon tetrachlorides, which are later
separated through a distillation process. The Zr02 is distributed in a powder form, which has
been modified with Ca, Mg, or Y to create a stabilized structure consisting of precipitates of
tetragonal and or monoclinic phases dispersed in a cubic matrix. This partially stabilized
Zirconia (PSZ) exhibits excellent chemical resistance, high hardness and melting
temperatures approaching 2370°C (4300°F). Alumina modified PSZ (A1203 -Zr02) systems
demonstrate improved wear resistance and have been utilized as cutting tools. Whisker
reinforced tetragonal Zirconia polycrystalline composites have been fabricated by hot
pressing at 1450°C (2640°F), which further enhances toughness [11].

2.1.5 Silicon Carbide:

Silicon Carbide may be produced as a monolithic structure or processed to contain
particulate, whisker, or continuous fiber reinforcement. SiC possesses a structure comprised
of strong covalent bonding, which results in thermal and chemical stability. Silicon Carbide
has been utilized for electrical resistance heating elements, crucible materials and internal
furnace components. Gas turbine applications such as turbine vanes, shrouds and bushings

have been produced through hot pressing powders.

In terms of matrix applications found in advanced CMC’S, SiC is typically produced
through chemical vapor deposition processes. Reactant gases such as CH3SiCl3-H2 or SiH4-
CKHy are forced through a fibrous preform, which results in SiC deposition on the fiber
surface. The resultant SiC matrix material, when reinforced with SiC, Alumino-silicate
fibers or other oxide fibers, demonstrates high strengths approaching 400 MPa and fracture
toughness of 18-20 ksi‘l in (>20 Mpa‘lm). This matrix material possesses a melt temperature
of 2200°C (3992°F). Silicon Carbide matrix composites exhibit resistance to erosion,
corrosion and wear allowing applications to develop in hostile engine environments [12,13].
2.1.6 Silicon Nitride:

Silicon Nitride exhibits a variety of mechanical, thermal, and electrical properties
depending on the processing method utilized and subsequent structure, which is obtained.
Producing silicon nitride is achieved from an initial silicon powder, which is nitrided in
molecular nitrogen from temperatures between 1150°C (2100°F) to '1400°C (2550°F).
RBSN (Si3N4) is composed of a and B crystallographic forms and exhibits 12-30% porosity.

Complex shapes may be formed with little dimensional change and typical flexural strengths
of 22-50 ksi are possible [14]. Components may also be fabricated by hot pressing in
graphite dies coated with boron nitride. Hot pressed Si3N4 is processed at temperatures of
1650°C to 1850°C (3000°F to 3600°F) and pressures of 2200-4500 psi. Hot pressed
components are limited in complexity due to the uni-axial pressure application.
2.1.7 Carbon:

Carbon may be used as a matrix material and is most commonly reinforced
with 3-D carbon fiber preforms to form Carbon/Carbon composite structures. The carbon

preform is typically infiltrated with a phenolic resin and cured to form a rigid structure. A

10

subsequent pyrolysis in an inert atmosphere carborizes the phenolic and the resultant porous
carbon matrix requires chemical vapor infiltration (CVD) processes to fully densify the
composite structure. Typically, applications include brake disks for aircraft and high
performance automotive applications. The carbon matrix and fiber reinforcement is subject
to oxidation degradation over extended periods of high temperature exposure but component
weight reductions and improved frictional performance have made carbon a viable matrix

material for certain applications.

2.2 FIBER REINFORCEMENTS:

In order to increase the toughness of many monolithic glass or ceramic matrix
composites, a variety of reinforcements have been researched. Selection of the fibrous
constituent has been governed by factors such as comparable thermal expansion coefficients
to that of the matrix and thermal oxidative stability. Fibers such as glass, alumina, alumina-
borosilicates, carbon/graphite, silicon carbide, and silicon nitride are options for CMC
reinforcement.

2.2.1 Glass:

Glass fibers demonstrate chemical and thermal stability much like the bulk glass and
glass-ceramic matrix materials, which makes fiberglass reinforcements attractive for
aggressive environments. Bulk glass fiber production is most commonly carried out through
the direct melt method whereby molten glass is extruded through an orifice creating a fine
fiber typically 3 to 20 microns in diameter [15].

Varying the molten glass composition permits the fabrication of three basic forms of

commercial glass fiber. E-glass, which possesses a calcium aluminoborosilicate

ll

composition, is a general use grade, which has ambient temperature tensile strengths of 3445
MPa (500 ksi). This fiber is utilized in applications, which demand strength and electrical
resistivity. Applications, which require extremely high tensile loading, are best suited for
the S-glass magnesium aluminoborosilicate fiber, which exhibit 4585 MPa (665 ksi) ambient
tensile strength. C-glass is a soda-lime-borosilicate, which demonstrates 3310 MPa (490 ksi)
ambient tensile strength but is highly resistant to chemical attack particularly from acidic
compounds [16].

In terms of CMC structures, the glass fibers offer several options to assist in the
tailoring of a structure, however, the temperature limitations must be considered. The
aforementioned glass fibers have melt temperatures ranging from 1000-1450°C (1825-
2650°F) but the mechanical properties are reduced by 50% or more by the 538°C (1000°F)
mark, making glass an unacceptable reinforcement for CMC structures operating in the 1000-
2000°F range.

2.2.2 Alumina:

Alumina silicate fibers are thought of as glass fibers typically unless the purity
exceeds 99% a-Al203 and exhibits a polycrystalline structure. This level of purity and
structural order is what defines an alumina fiber, which is utilized most often in metal matrix
and ceramic matrix composite structures. The fiber is made by a continuous slurry spinning
process of alumina particles in an aqueous solution with organic polymer stabilizers. The
fiber is dried at low temperature and then fired to form a dense, high temperature stable or-
A1203 fiber. A silica coating is usually applied to the fiber to minimize the effects of surface
defects and thus improve the tensile strength, which is 14 to 19 GPa (2 to 2.8-106 psi) with a

tensile modulus of 390 GPa (57--10‘S psi). The alumina fiber has excellent elevated

12

temperature stability to 1000°C (1830°F) while maintaining approximately 100% of its
tensile strength and modulus [l7]. Alumina fibers exhibit the temperature capability for
most high temperature applications not requiring service above the 1000°C (1830°F) regime,
which makes the fiber a likely candidate for MMC and CMC structures.

2.2.3 Graphite/Carbon:

Carbon fiber reinforcement is used for many high strength lightweight structures.
The two fiber forms, carbon and graphite, are often mistakenly interchanged when composite
materials are discussed. The hexagonal base structure of carbon is highly structured and
exhibits high modulus and tensile strengths. Conversely, the bond strength of the hexagonal
planes is relatively weak and thus the transverse load carrying capability of the carbon
structure is low. If, during processing, sufficient pretreatment loading and heat treatment is
undertaken, a three-dimensional order will be obtained and this high performance fiber is
termed graphite. If, however, the fiber is found to have hexagonal planes twisted in relation
to each other and some degree of disorder exists the fiber is said to be a carbon fiber.
Commercially available material typically utilized in high performance composite structures
is the carbon fiber and it is derived from three basic raw materials.

The carbon fiber can be created from Rayon, Polyacrylonitrile (PAN), or Pitch based
precursors. In each of the three cases, the base material is pyrolized to create a carbon fiber.
Rayon has the poorest conversion rate and produces the lowest modulus end product while
the PAN and pitch based fibers can be processed to create very high modulus fibers in excess
of 725 GPa (110 x 106 psi). The carbon fiber may be heat treated in an inert atmosphere to
temperatures of 1000-3000°C (1830-5430°F) to form varying modulus properties with the

ultrahigh modulus fibers undergoing the highest thermal processing. At the conclusion of the

13

basic fiber production, a variety of treatments may be applied to tailor the carbon fiber
toward its end use.

The carbon fiber may be treated to improve its handling characteristics, adhesion to
various matrix materials and ability to resist thermal oxidation. Fiber tows are often coated
with an epoxy resin to prevent fiber damage as the tows or bundles are taken up on a spool or
woven into cloth. The fibers are also typically coated or sized with hydroxyl and amine
groups to improve laminate performance. Finally, carbon fibers have also been treated with
SiC, BN and other thermally stable compounds to protect the base fiber from oxidation.
Carbon fibers serve as an excellent composite reinforcement; however, their primary
limitation is rapid oxidation rates at temperatures above 700°F (370°C). Research has been
conducted to limit the oxdizing effects at elevated temperatures with various coatings,
however, due to the thermal mismatch between the coating and the fiber, cracking in the
protective barrier along with diffusion effects, the fibers are still susceptible to degradation
during extended service conditions. As a result the use of carbon fibers has been limited for
the environments targeted for CMC structures [18].

2.2.4 Silicon Carbide:

Silicon Carbide fibers are manufactured by two distinctly different methods, which
yield a fiber capable of reinforcing either MMC or CMC structures. The first method
produces bicomponent fibers that are formed by first taking a carbon core and subsequently
depositing, via chemical vapor deposition, the bulk SiC about the core to create a finished
fiber with an average diameter of 10-15 pm. The second method produces a monolithic
fiber by polymer pyrolysis and is commonly referred to as the N icalon SiC fiber. Both of the

carbide SiC fiber structures demonstrate advantages over other oxide fibers, in that, they

14

possess superior mechanical properties, especially higher modulus of elasticity and axial
compression in the case of large diameter fibers. The mechanical properties of the SiC fiber
are desirable for CMC reinforcement; however, the fiber is susceptible to thermal oxidation
above 1100°C (2010°F) due to grain microcrystallization. Despite the oxidation at these
temperatures, the SiC fibers show promise for CMC reinforcement as additional knowledge
is obtained with fiber sizings as they pertain to the oxidative stability [ 19].
2.2.5 Silicon Nitride:

Silicon Nitride fibers are currently under development but show great promise for
CMC reinforcement as bulk matrix Silicon Nitrides, particularly B-Sialon, has shown
excellent thermal oxidative stability and mechanical properties. The catalysts utilized for the
pyrolysis of polysilazanes has been developed and will be used to create Si3N4 fibers through
a spinning process which yield a precursor fiber that may be heat treated much the same as
PAN based carbon fibers. The service temperature of these fibers could reach 1500°C
(2730°F) with tensile strengths of 1000MPa(145 ksi) and tensile modulus of 3OOGPa (45-
106 psi). Ultimately, the final fiber properties will depend on the fiber mechanical
pretreatment and thermal processing used during processing, but the Silicon Nitride family of
matrix and fiber materials shows promise for superior CMC structural composites [17].

The aforementioned matrix materials and fiber reinforcements are the most widely
utilized in the fabrication of high performance composite structures and a summary of their
properties has been shown in Table 2.1. The key to a successful material selection for a
given fiber\matrix system is the careful evaluation of several criteria including service
condition, mechanical and thermal interactions between fiber and matrix along with the final

component geometry as it relates to the processing method. Several of the fiber\matrix

15

combinations may appear to serve well in an application but consideration to part geometry

and fabrication method must be undertaken.

2.3 FABRICATION METHODS:

The variety of materials available in the fabrication of high performance
composite components is also accompanied by a diverse assortment of processing
methods.

2.3.1 Hot Pressing:

Uniaxial hot pressing of ceramics is most probably the oldest technique of forming
components but also is the most limited with respect to part complexity which may be
achieved. Hot pressing of components requires construction of a mold with corresponding
inserts, which define the part shape and size. Pistons and punches, which transmit the
externally applied pressure to the die cavity, must be assembled. The appropriate powder,
having been previously sized, dried and binder prepared is loaded into the mold. The mold
assembly is placed in the fumace\press cavity and pistons are engaged with the pressure
source as shown in Figure 2.1. The furnace parameters derived for the particular material to
be processed must be programmed into the furnace controller. With the preparation of
material, tooling, and press equipment completed, the pressure and heat may be applied to
begin the cycle. After the consolidation process has been run to completion, the heat is
removed and the component is allowed to cool under pressure before removal from the press.
Examples of a typical parts is a flat plate, round billet or a cube with subsequent machining
or grinding required to finish the raw shapes into final usable products.

Boron Carbide (BC4) is a specific material traditionally hot pressed to form wear

l6

resistant items such as blasting nozzles and wear plates. The BC4 powder which is <10p.m is
typically mixed with sintering additives such as Al, Cu, or B203. The mixture is pressed at

2100°C (3800°F) for 30 minutes at 35 MPa (5 ksi). [20]

   
 

Water-cooled
hydraulic ram

  
    

Follow-up

Piston Water-cooled

retainer

Susceptor
insulating
block

 
  

    
   
 
 
     

   
    
  

   

"

V
S

N

7

R

   

/’

/ Susceptor
R\\\\\\\\ \\\\\‘

\\\

A

   

/
/

  

 

Cavity spacer J

Cavity column

I

g./////

k

Die

\
s

\\\.
s\\

#2
<’<s\\\\\ \\\\\\/ \~
“\\\ ////\\

Water-cooled base

A
//I

\\\\\\\\\\\&.‘/

    

//

\m‘mm‘t V '/

mm

blocks

   
  

§\\\\\\\\/\/\/\

    

'11,]

§ \\\\\\\\\\\\\\\\\\\\\‘ \\\\\\\\\\\\\‘
§§\ \\\\\§\\\\V\\\\\\\\ \\\}\\\I\\\

/\\‘

 

x\\\\\\\\\\\\‘

   

C

/
Susceptor support / insulating
blocks

Figure 2.1: Schematic of hot pressing equipment [20].

  

 

\ Water
cooled
induction

coil

Die body

. liner / sleeve

Insulation
" cavity

 

 

_:l

2.3.2 Hot Isostatic Pressing:

Hot isostatic pressing is an extension of uniaxial hot pressing. The pressure is
applied uniformly and in a direction normal to all part surfaces and results in high
interatomic bond strength along with no part deformation during pressure application.
Powders are compacted prior to the HIP process and encapsulated with ceramic, metal or
glass barriers. The purpose of the
encapsulation is twofold, first, it provides a
media in which the pressurizing gas may be

uniformly distributed and secondly, the barrier

 

 

prevents the pressurized gas from permeating
the work piece. The HIP process is capable of
processing many materials and is limited only

by the equipment capability with regard to size

 

 

 

and ability handle complex shapes. The

 

 

general press arrangement is shown in Figure
2.2.
Whisker reinforced CMC materials,

in particular, SiC reinforced Si3N4 can be

 

processed by the HIP process. Typically,

Figure 2.2: Schematic of hot isostatic

temperatures of 1950 C (3540 F) and press equipment [21],

pressures of 250 MPa (35 ksi) for several

hours are required to fully densify the material.

18

Components fabricated with this process were tested in three point bending and achieved
1000 MPa (150 ksi) flexure results. The HIP process may also be used to produce precision
pieces which are dimensionally very accurate as pressed and in the case of textile handling
guides fabricated from Si3N4, demonstrate exceptional wear properties. The HIP process is
traditionally utilized to produce smaller components and solid plates, round billets or cubes
which are subsequently processed to yield finished parts. The size limitation of typical HIP
furnaces creates problems when large structures are encountered but the performance
achieved by HIP produced components makes them an important option for the CMC
designer [21].
2.3.3 Chemical vapor infiltration (CVI):

Chemical vapor deposition has been used to create various coatings and fibrous

reinforcements for wear resistant, electronic, corrosion resistant and oxidation resistant

 

   

 

 

 

 

 

 

applications. Carbides, a“.-- - — -
. 7 %
nitrides oxides and borides 3' 4
9 HEATING ¢ HOT ZONE %
ELEMENT ¢ 4
have been processed utilizing Z awasr GAS Z
PERFORATED % é
. . LID / 2
the vapor deposrtron method. I. ;
>a//////lI///////Il/////ll ///////<?
The rocess includes heatin a / ' i ' ' i ': :5 ¢
P 8 g §¢
4 \ a
substrate fiber, component or ~nm
3:33: W
panel and depositing a solid \\\
via a gas phase to the substrate $933}? 13%?”
surface. CVD processes Raga“
include a furnace to enable the Figure 2.3: Schematic of chemical vapor

infiltration method [22].

19

substrate heating and a means of flowing reactive gases used in the deposition process to and
from the substrate. In terms of CMC structures the basic concept of CVD is used but the
target becomes a fibrous preform that requires densification with a suitable matrix material.
The ensuing densification is termed chemical vapor infiltration and is used to densify CMC
structures requiring high temperature performance. Figure 2.3 illustrates this process.

The CVI technique used to densify CMC performs is an isothermal/isobaric process
that utilizes low concentration reactant gases to complete densification. The process must be
carried out slowly as the preform may become sealed on the surfaces without a reasonable
diffusion path remaining to densify the center of the preform. Despite slow process
parameters, thick CMC sections will typically require a light surface machining as an
intermediate process step to that of the infiltration to open interconnected pores and restore
efficient diffusion paths. The process lends itself to large batch production but for larger,
high-density components the process time can be several months [22].

2.3.4 Directed Metal Oxidation:

The process of directed metal oxidation has been researched for nearly twenty years
with the primary goal of producing composite structures capable of performing in
applications such as turbine engine hot sections, piston engine combustion components and
other high temperature scenarios that require performance at temperatures up to 2000°C
(3630°F). The manufacture process for a high performance CMC begins with an inert barrier
placed at the surface of the fiber preform to prevent surface oxidation and the preform
assembly including the parent metal alloy are placed in the mold as shown in Figure 2.4. The
heating process takes the barrier film, preform, and parent metal to a condition that begins

oxidation of the parent metal. As the metal is oxidized, the matrix is created thus densifying

20

   
    

GROWTH
BARRIER

Figure 2.4: Process description for directed
metal oxidation process [24].
the CMC. The properties of the CMC are tailorable, in that, the preform architecture may be
engineered to meet specific requirements and the matrix properties may be modified as a
function of the heating process.

An example of the directed metal oxidation process is SiC reinforced A1203, which
has been completed to form heat exchangers, and related furnace components. The
aluminum parent metal is placed between the SiC preform layers and the matrix is grown to
complete the CMC. The matrix oxidation process has been shown to penetrate preform fiber
bundles to create a dense CMC when using untreated Nicalon SiC fibers and pretreated
fibers. The composites demonstrate 460 MPa (67 ksi) four—point flexure results at room
temperature with a 25% reduction in strength at 1400°C (2550°F). Fracture strength is 27.8
MPan (25.3 ksiN/in.) The performance of these composites has led to the widespread

manufacture of components but the tooling cost to mold at elevated temperatures has made

21

large-scale production a necessity to justify molds and related processing equipment. [24]

2.3.5 Sol-Gel Method:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

The sol-gel process is used . .
Solution / Sol
primarily to form thin F
coatings, films, fibers and r 1
microballoons. Sol-gel P°Wd°"fr°° POW“
techniques use hydrolyzable
alkoxides, which undergo a
Preforrn Monodispersed
chemical synthesis to form powers
|
high purity glass, glass- I 1
ceramic, and ceramic Amorphous Crystalline Uniform
compact
structures. The material may
. Gl
be amorphous or crystalline ass
which is determined with the
Glass-ceramic Polycrystalline Polycrystalline
initial solution or sols and the “mmc °°mmc

 

 

 

 

 

 

 

 

 

subsequent processing

Figure 2.5: Preparation of glasses,
glass-ceramics, and ceramics by the
sol-gel process [23].

conditions used [23]. Sol-gel

process is shown in Figure

2.5.

2.3.6 Polymer Precursors:
Polymer precursors have been used as means to capture the best of the low

temperature-processing world combined with high temperature ceramic component

production. The process involves the use of a preceramic polymer such as polysilastyrene,

22

vinylic polysilane, polysilazanes or silicon oxycarbides. Much like carbon\carbon composite

fabrication, these preceramic polymers are prepregged onto a unidirectional or woven

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Preoeramic polymer
Fiber Couposite l Coating Nonfugitive binder
I I I 7
r po r
fiber preform substrate powder
I I I ' I
Cure Mold to Cure
fiber shape polymer Mold
, I I I I
Multi cles
I ma Cure v—iiy Pyrolyze Cure
I I I I
Weave Ceramic
crow Pyrolyze mated ”£32.“
preform substrate
I I
Dense TI Fle— Dense
composite impregnate ""' ceramic

 

 

 

 

 

 

 

Figure 2.6: Processing steps required to prepare the four
types of polymer derived ceramics [25].

ceramic fiber materials and subsequently handled as standard epoxy prepreg materials. After
the initial cure and subsequent pyrolzation, the polymer transforms to an amorphous ceramic
material. Refer to Figure 2.6 for process flow diagram [25].

2.3.7 Silicon Oxycarbide:

The cure cycle for silicon oxycarbide precursor matrix material is typically low
enough to utilize tooling found in the manufacture of fiber reinforced epoxy composite

components. The preform or prepreg material is placed on the mold surface with fiber

23

orientation taken into consideration, vacuum bagged and cured under pressure in an
autoclave. Typical silicon oxycarbide polymer processing temperatures are 150°C (304°F)
and pressures range from 35-100 psi depending on part geometry and tool configuration.
After autoclave cure, the part is removed from the mold and placed in a furnace or kiln.

The pyrolysis process is completed with the component free standing during the
initial pyrolysis cycle or basic support surfaces may be provided via alumina blocks. The
pyrolysis is carried out in an inert atmosphere typically consisting of free flowing nitrogen.
A typical cycle utilizes a ramp rate of 2°F/minute from room temperature to 1700°F and a
soak at 1700°F for 1 hour. After the soak period, the ceramic structure is slow cooled to
room temperature where it is reinfiltrated to increase its density and reduce open porosity.

During the polymer-ceramic conversion the matrix material undergoes a substantial
density change from 1.25g/cm3 to 2.5g/cm3. This increase in density is accompanied with
matrix shrinkage and an increase in open porosity thus necessitating the polymer precursor
reinfiltration process. The reinfiltration process consists of numerous composite infusions
with low viscosity preceramic polymer material to fill porosity. The infiltrations are usually
performed with the part under vacuum to aid in the polymer infusion throughout the part
thickness. The polymer is cured and the composite is again pyrolyzed to convert the newly
introduced polymer to a ceramic. Typically, five or six infiltrations will be utilized to
achieve open porosity levels of 3-5% and a final density of 2.25-gr'n/cm3 [26].

The polymer precursor materials offer a unique advantage over many other CMC
matrix systems with respect to the fabrication methods required to produce finished goods.
Previously described CMC processing methods including hot pressing, chemical vapor

deposition and directed metal oxidation require the design, development and fabrication of

24

tooling capable of dimensional stability during processing temperatures ranging from 900°C
(1652°F) to 1950°C (3540°F). The therrnosetting polymer precursor technology offers initial
cure temperatures of 304°F (150°C) to cross-link the polymer. This temperature regime
allows traditional epoxy tooling to be used for complex CMC processing during the initial
fabrication step. The subsequent pyrolyzation process is typically completed “free-standing”
in a furnace that eliminates the need for costly tooling capable of CMC elevated temperature
processes. The flexibility to use existing polymer composite and aluminum tooling in the
fabrication of prototype and demonstration components has increased the speed at which in-
situ monitoring of finished components can be achieved. The demonstration components,
which include exhaust tailcones, thrust vectoring surfaces and exhaust gas mixers, have
provided valuable material design feedback with respect to oxidation, sonic fatigue, and
mechanical property degradation.

The aforementioned CMC materials typically suffer from the effects of long-term
oxidation resistance at elevated temperatures. The matrix, which has been reinforced with a
whisker or continuous fiber, may lose its strain to failure and damage tolerance capability as
oxidation degrades the reinforcements ability to toughen the structure. Controlling the fiber-
matrix interphase in an effort to prevent subsequent interactions while maintaining a viable
load path to arrest crack propagation is the most significant CMC issue currently. The
interphase stability issue will be addressed in this research through the survey of various
fibers pretreatment to provide an interphase capable of meeting the demands of current high

temperature applications.

25

2.4 CMC INTERPHASE

The interphase existing between fiber and matrix is best considered a three-
dimensional transition region. This interphase region, as shown in Figure 2.7, depicts a
diffuse area between the fiber and matrix, which effects overall composite properties [27].
The interaction of individual fiber and matrix bulk properties, fiber topography, morphology
and chemistry, along with fiber sizing will determine the composite performance. The
degree of adhesion between fiber and matrix as well as failure mode may be understood from

examining this three dimensional region.

 

 

     
 

  

Fifth-min“ Thermal. l
L and Mechanical
.1 Environments

 

 

 

 

 

Figure 2.7: Schematic representation of fiber-matrix interphase in a composite
material [27].

Adhesion between reinforcement and matrix materials is comprised of several
interrelated factors including mechanical, chemical and constituent bulk property
interactions. The degree of fiber-matrix adhesion found in a composite structure will

determine its strain capability. Understanding the contribution of each factor toward the total

26

bond strength will serve to better predict overall composite performance and allow an
estimation of failure modes-mechanisms [2].
2.4.1 Mechanical Contribution: (Frictional & Residual Stresses)

The mechanical contribution toward Overall fiber-matrix performance is comprised
primarily on fiber topography and residual stresses resulting from differential expansion
coefficients between constituents. The surface roughness of reinforcing fibers will result in
an increase of fiber pull—out force required to overcome frictional effects. Matrix materials,
which have a larger thermal expansion coefficient than that of the reinforcing fibers, will
yield a compressive force imparted to the fiber surface, which in turn will increase the
effective fiber pull-out force. At the conclusion of CMC processing, it is quite possible to
achieve large residual radial stresses between the fiber and matrix. These stresses may
induce fiber damage and or matrix cracking which may serve to undermine the interphase
region rendering it incapable of carrying crack energies efficiently through a given
component. Both frictional and CTE effects can be quantified via single fiber indentation
testing and the resulting load data may be used to obtain interfacial shear stresses [28-30].
2.4.2 Chemical Contribution:

Adhesion in CMC’s is also influenced by the chemical bond between fiber and
matrix. Strong interaction between fiber and matrix as a result of thermal oxidation is a
common problem as the interdiffusion of matrix and fiber materials can create a near
monolithic ceramic structure. This degree of bonding will adversely affect the composite
properties. If the interface bond is too strong, matrix cracks propagate normal to the fibers,
cutting through the fibers without deflection. This fracture behavior causes the composite to

have low fracture toughness as indicated by the flat fracture surface. To obtain the desired

27

fibrous, flaw-tolerant, failure requires a sufficiently low interfacial bond strength such that a
propagating crack causes fiber- matrix interfacial debonding and/or deflection of the matrix
cracks. Toughness and strength performance can be understood from studying fiber sizings

and surface modification effects [2,29].

2.5 FAILURE MODE:

CMC failure modes are of great interest to design engineers as these materials are
implemented in areas requiring structural load carrying capability. The failure must be
graceful and predicable unlike most monolithic ceramics. As composites are dynamically
loaded in service, energy due to impact may occur. This sudden energy input must be
uniformly distributed throughout the structure to avoid catastrophic failure. When crack
energy travels through the matrix, it will encounter the fiber reinforcement at which time a
controlled debonding between fiber and matrix is desirable. The controlled debonding will
serve to dissipate crack energy along the fiber axis throughout the structure. In the case of a
balanced isotropic lay-up, the energy will be distributed uniformly and thus minimize local
catastrophic failure.

The incorporation of treated fiber reinforcements targets the control of CMC failure
modes and enhances composite toughness. Toughness or strain capability is generally less
than 0 . 1% for monolithic structures while toughened systems may achieve rates in excess of
0.5% [31]. Failure mode research should consider the effects of chemical bonding, frictional
lpush-out forces and matrix cracking relationships to describe overall CMC failure

characteristics.

28

2.5.1 Fiber-Matrix Debonding:

The chemical bond between fiber and matrix must be sufficient to facilitate load
transfer but weak enough to yield slipping between the two constituents. In composites
consisting of a mullite matrix and Nextel 440TM fiber, hot pressing at temperatures of 1410°C
(2570°F) resulted in a fiber-matrix reaction which fused the two constituents together
resulting in a brittle fracture. The same Nextel 440TM fiber, when BN coated, did not react
with the matrix and allowed cracks to propagate along the fiber length indicating that the BN
treatment sufficiently prevented chemical bonding [2,29,32,33].

The work associated with fiber debonding from the surrounding matrix can be
quantified by considering the interfacial shear stress (1:), the fiber diameter (d), the debonded
length (1d) and the difference in strain (As) [34-38].

Per fiber basis:
Wdf = F 0 d
War = (T ' A) (A5 ' ld)

Wdf =T7'Cd12 '43?-

2.5.2 Fiber Deformation and Fracture:

Matrix cracking is best controlled and distributed when a controlled debonding
occurs. The work of deformation and fiber fracture can be expressed as the product of the
fiber tensile strength (0'), diameter and debonded length divided by the fiber tensile modulus
(E3). The critical transfer length is substituted for the debonded length in the instance that a

complete unbond does not exist [34-38].

29

Per fiber basis

Wff = (%)03 Ad

 

0'2”

w.=<}§> E: .121,
f

= 02712121,,

8E!

Wrr

 

2.5.3 Fiber Pull-Out
Fiber surfaces, which are sufficiently smooth, will contribute to low frictional effects
and consequently promote the fiber pull out needed to prevent catastrophic failure. The work

associated with pulling a fiber from the matrix is the product of the shear stress and the pull-

out length [34-38].

Per fiber basis
Wpf= (174.)1p
Wpf = wall:

The total work associated with the crack energy can be summarized as follows:

03712131
W, =mdl}9£+;—d—+mil:
2 3225,

The summation includes the contribution from fiber debonding, deformation-fracture and
pull-out. The common component found in the total work relationship is the interfacial shear
stress, which is non-linear and contributes to maximum energy adsorption at a specific
optimal point.

With the importance of interfacial shear strength on composite toughness highlighted,

a closer review of the fiber pretreatment and matrix interactions is necessary. Materials

30

development and processing of ceramic matrix composites that fail in a high strain (>0.5%),
damage tolerant manner are continually being sought. This active research area has lead to
the development of silicon oxycarbide/A1203 composites [39,40,43] and for the production of
complex-shaped components at low cost. Alunrinoborosilicateisilicon oxycarbide composites
have demonstrated the ability to retain enhanced fracture toughness and strength [41,42].
The objective of this work is to extend this research area by correlating the macromechanical
properties with the fiber- matrix micromechanical properties of Nextel 312TM
fiber/BlackglasTM composites. In particular, an understanding of how the boron nitride
coating affects the fiber/matrix adhesion, mechanical properties, and composite failure

behavior will be investigated.

31

Chapter 3

EXPERINIENTAL

The experimental work performed in this research program was concentrated toward
Nextel 312TM fibrous reinforcement composited in the Allied Signal 493 BlackglasTM
preceramic polymer matrix system. This preceramic CMC system offers ease of processing
and elevated service temperature benefits but as in many CMC systems requires fiber
pretreatment to enhance overall mechanical properties. The goal of this effort is to quantify
the stress-strain benefit gained from the use of nitrided Nextel 312TM fibers in the
BlackglasTM matrix as compared to either organically treated or bare Nextel 312TM fiber
reinforcements. Experimental efforts were divided into two sections consisting of
constituent characterization and finished composite property evaluation. Additionally, the
Blackglasm-Nextel 312TM composites were monitored in-situ during the polymeric
conversion to the ceramic in order to gain processing insight with respect to fiber matrix

interactions.

3.1 BLACKGLAS'” POLYMER PRECURSOR REINFORCED WITH BN
TREATED NEXTEL 312T“ FIBERS:

While a variety of CMC systems are commercially available, the polymer precursor
systems offer the greatest processing flexibility and low initial process temperatures and
pressures. The BlackglasTM polymer precursor system produced by Allied Signal and

reinforced with 3M Nextel 312TM woven fabrics shows promise in the ongoing research for a

32

toughened, continuously reinforced CMC. The Nextel 312TM fabric has been enhanced by
the addition of a BN rich surface treatment, which creates a stable interphase, capable of
facilitating load transfer and crack deflection at the matrix-fiber interface. The BN rich
region also provides a barrier to prevent fiber-matrix reactions that may reduce toughness as
elevated processing or service temperatures are encountered. The Allied Signal and 3M
CMC materials are commercially available and are able to be processed via commercially
available equipment.

3.1.1 Blaclrglas'rM Matrix System:

The matrix system developed by Allied Signal Inc. is a polymer precursor to a
ceramic. The BlackglasTM 493 polymer is a thermoset produced by the addition reaction of
proprietary siloxane monomers and the density of the cured polymer is 1.09 g/cm3. The
proprietary polymer precursor is cured and subsequent pyrolysis in an inert atmosphere at
900°C (1652°F) causes the conversion to the amorphous silicon oxycarbide SiCxOy with
evolution of Methane and Hydrogen as the byproducts of the phase conversion. This highly
refractory, silicon based glass contains 15-30% atomically distributed carbon, has a typical
composition of 24 wt% Carbon, 47 wt% Silicon, 28 wt% Oxygen, < 0.4 wt% Hydrogen, <
0.1 wt% Nitrogen (Si C2301“) and has been patented as Blackglasm by Allied Signal Inc.
The matrix, while in the pyrolysis cycle, demonstrates a char yield from the polymer state of
approximately 81 to 86 % by weight. The pyrolysis results in an amorphous silicon
oxycarbide glass retaining 43 to 45% of its initial volume and exhibiting open porosity
ranging from 20—30%. Upon the completion of the reinfiltration process, a fully densified
matrix is optimized at approximately 0.08 lbs./in3 (2.2 g/cm3). The matrix material is easily

reinforced with ceramic fibers to form CMC structures [44].

33

3.1.2 Nextel 312TM Fiber:

The 3M corporation

produces the Nextel 312TM ceramic
fibers. The alumina-silica ceramic
fiber has a composition of 62%
A1203, 24% SiO2,

14% B203.

Nextel 312TM fibers are available in

a wide variety of forms including

fabrics, tapes, sleevings and

cordage. Due to the variety of forms
available, N extel 3 12TM may be used
in number of

any design

environments. Typical physical
properties are summarized in the
Table 2.1. The fiber surface may be
modified via CVD or diffusion

methods to promote a suitable

interface for a number of ceramic

Table 3.1: Nextel 312TM properties.

 

Physical Properties

 

Composition

62% A1203, 24% SiO2,
14% B203

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Color White
Length Continuous
Fiber Density 2.7 gmlcc
Fiber Diameter (elliptical) Minor axis: 7—9 micron
Major axis: 10-12 micron
Surface Area <1mz/gm
Mechanical Properties
Tensile Strength 1725 MPa (250,000 psi)
Tensile Modulus 138 GPa (20 x 10‘ psi)
Elongation 1.20%
Thermal Properties
Continuous use Temperature 1204°C (2200°F)
Short Use Temperature 1371 °C (2500°F)
Lineal Shrinkage 2000°F 1.25%
(1093°C)
Melting Point 1800°C (3272°F)
Thermal Expansion Coefficient 3.0 x 1045 DIJU°C)
(ZS-500°C)
Specific Heat 0.25 BTU/lb/°F or
Callgrnl°C
Electrical Properties
Dielectric Properties 5.2 o 9.375 x 109 hertz

 

Optical Properties

 

 

Refractive Index

 

1.57

 

matrices. The Nextel 312TM reinforcement may be preimpregnated as any typical epoxy

prepreg material [45 ,46]. Table 3.1 summarizes various mechanical and physical properties

of Nextel 312““.

3.1.3 Boron Nitride Treatment:

The fiber pretreatment utilized to tailor the debonding process with Nextel 312TM

34

 

reinforcements is described within a patent created at the Boeing Company. The process
involves heating the Nextel 312TM fabric in an atmosphere containing ammonia, hydrogen
and nitrogen at a temperature of 2200-2600°F (1204—1427°C) for a period ranging from 5-90
minutes [47,48]. This process diffuses boron from the fiber bulk structure to the fiber
surface, reacts it with ammonia, and produces a thin, uniform boron nitride layer. The
coating transitions gradually from the BN surface to the parent boria (B203) found in the bulk
fiber thus creating a very stable interface for CMC fabrication [49].

While there are other methods of establishing the BN surface prior to compositing,
these methods usually involve exposing the fiber to a liquid boron oxide or boric acid, which
severely degrades the oxide based N extel 312TM fiber. Alternatively, the BN surface may be
created via CVD or sol-gel methods but maintaining a uniform coating thickness is difficult
especially in woven fabrics. In the CVD process, the reactant gasses create the BN coating
on the first hot surfaces they come in contact with and maintaining accessibility to all fiber
surfaces at constant temperature and with uniform gas composition is difficult. Likewise,
the sol-gel method is limited in the solution deposition uniformity. In any case, these
methods create a distinct coating unlike the previously described nitriding method developed
by the Boeing Company [50-56].

The nitriding method previously described under several patents was further
investigated with respect to process sensitivity for time, temperature and reactant gas
concentration. SIMS data indicates the NH3 reactant gas concentration, which was varied
from 5% to 20% had no effect on the coating depth or resultant BN concentration. The
nitriding process appears to be more sensitive to the time duration that was varied from 15,

30, 60, and 120 minutes. Under these process times and at 1 150°C (2102°F), the maximum

35

weight percent of boron present ranged from 10-22% at 25-50 angstroms. The boron
concentration tapered off in all cases to 5-8% at a depth of 100 angstroms from the fiber
surface. The 60-minute duration caused the highest concentration of boron at 23-wt%
located 35 Angstroms from the fiber surface. Using the 60-minute duration, the experiment
was run again varying process temperatures. Temperatures of 1150, 1225 and 1250°C
(2102°F, 2237°F, 2282°F) were used to understand their impact to BN formation. All peak
concentrations were within 21-23-wt% under this process profile but the higher temperatures
yielded a more broad depth profile with boron detected down to 5-8 wt% at 150 angstroms.
In summary, the 60-minute time at 1250°C (2282°F) yielded a 22-wt% boron concentration
at a depth of 70 angstroms. Appreciable (<8wt%) quantities of boron were noted at depths of

120 to 130 angstroms [57].

3.2 CHARACTERIZATION OF AS RECEIVED, DESIZED AND BN
TREATED NEXTEL 312'”M FIBERS:

Nextel 312TM fibers, as received from 3M, are treated with an organic sizing to
protect the fiber and facilitate handling through the spooling process. The fiber was
examined in this “as received” condition but also examined after a cleaning or stripping
process was conducted to expose the base fiber surface and finally an assessment was
conducted with a BN treatment present.

The as received fiber has a slight light gray appearance and is somewhat stiff as it is
removed from the spool. The fiber tows were used for the base fiber surface characterization
and the tows had approximately 3-5K filaments each. The desizing or cleaning process is

achieved by heating the sample Nextel 312TM in a ventilated furnace to 550°C (932°F) for 12

36

hours. After heat cleaning, the fibers appear bright white and will be more flexible and soft
to the touch. After the fiber desizing, one half of the desized fibers were BN treated via a
nitriding process previously described.

The three fiber tow samples were then analyzed to determine the surface composition
via x-ray photoelectron spectroscopy and auger spectroscopy. Additionally, the fibers were
mechanically assessed to determine if any stress-strain propertied were altered due to the
fiber tow processing.

3.2.1 XPS Analysis:

The fiber sample surface chemistry was assessed with x-ray photoelectron
spectroscopy (XPS). XPS is also referred to as Electron Spectroscopy for Chemical Analysis
(ESCA) but the technique will be referred to as XPS throughout this work. XPS analysis
begins with the sample bombardment with x-rays in a vacuum of 10'6 to 10’10 torr. The
impingent x-rays (hv) ionize the sample surface and the emergent core electron energy is
detected. The energy input less the core electron binding energy is the detected electron
energy as measured by the instrument detector. The relationship is shown below from a
physical and energetics viewpoint [58,59].

A + hv —) A” + e' (discreet energy-ESCA)
Where A = Atom or Molecule and A”3 = excited ion
E (kinetic) = E (photon) - B (binding)
Specifically in this work, the Perkin-Elmer model 5400 spectrometer was equipped
using a standard Mg Kora X-ray (1253.6ev) source operated at 300 watts, 15 kV and 20 mA.
The instrument used a 180° hemispherical energy analyzer operated in the fixed analyzer

transmission mode at a pass energy of 89.45 eV for survey spectra and 35.75 eV for

37

multiplex spectra, and a position sensitive detector. Spectra were collected using an analysis
area defined by a 3 x 10mm rectangle. Samples were mounted by clamping each end of a
two-centimeter length of fiber onto a sample holder using brass clips. All sample handling
was done with clean forceps and scissors. Spectra were collected using a 45° degree take-off
angle between the sample and the analyzer lens. As received, desized and BN treated fiber
tows were examined to determine the surface composition and obtain knowledge regarding
the chemistry present.

3.2.2 Auger Electron Spectroscopy:

Auger electron spectroscopy (ABS) was used to determine the fiber surface
composition uniformity and coating thickness if present. The AES technique is similar to the
XPS\ESCA technique in that the sample surface is ionized however it is achieved via an
electron beam rather than an x-ray source. The impingent electron beam causes vacancies in
an inner electron shell as an electron is removed. The vacancies are then filled by electrons
from outer shells, which cause x-ray photon emission or the emission of an Auger electron
[58,59]. The associated Auger electron energy is detected and used to determine the surface
composition within a very specific area. The AES instrument is capable of a spatial
resolution of approximately 200A. The electron beam lateral x-y resolution is approximately
.5 11 making it useful for analyzing specific features. The three samples were examined as
prepared as well as with a gold coat sputter due to the insulative nature of Nextel 312TM
fibers. The gold coat was selectively deposited over the fiber surface to prevent sample
charging and subsequent sample deterioration while permitting fiber analysis.

3.2.3 Stress-Strain Measurements:

Nextel 312TM fiber mechanical properties were assessed to determine what, if any,

38

effect the desizing and BN treatment processes may have had on the stress-strain capability.
Fibers may be degraded mechanically from the elevated temperature processing required to
remove organic sizings and elevated temperatures combined with an ammonia nitriding
atmosphere used to create the BN rich surface region. Mechanical tests were performed to
assess if any such degradation occurred with particular focus being placed on the BN treated
fibers, as the processing parameters are severe.

Single fiber tensile properties for; as received fibers, desized and BN treated fibers
have been prepared for tensile evaluation. The single fibers were mounted to stiff
rectangular paper frames until the sample could be loaded into the tensile testing machine.
Once loaded, the sample was pretensioned and the paper frame was severed allowing the
fiber to carry the uniaxial tensile load independently. The United Testing Systems tensile
machine was equipped with a micro load cell capable of resolving loads during tensile
failure. The samples are to be tested with a crosshead speed of 0.05 inches per minute. This
evaluation will determine the residual fiber strength due to processing when compared to the
as manufactured condition.

Polymer encapsulated fiber sample fabrication and tensile testing was planned to
enable an understanding of the critical fiber fracture length as it relates to interfacial shear
strength. The typical carbon fiber-polymer matrix approach includes bridging a single fiber
across a dog bone shaped mold and subsequently embedding the fiber in epoxy resin. The
sample would then be tested in tension under plane-polarized light to determine the fiber
critical fracture length, which is the length where insufficient load can be transferred to the
fiber and no further breaks occur. The critical length can be related to interfacial shear

strength via the following equation:

39

where (d) equals the fiber diameter, and IC is the critical fiber length [60].

3.3 COMPOSITE SAMPLE PREPARATION:

Unidirectional samples and woven fabric samples were prepared to accomplish the
mechanical property and in-situ pyrolysis evaluations. The sample preparation involves fiber
pretreatment, preimpregnation, polymer cure, pyrolysis and finally densification. Composite
samples were prepared with as received, desized and BN treated fibers.

3.3.1 Boron Nitride Treatment:

Fibers of Nextel 312TM (3M Corp.) aluminoborosilicate (64% A1203, 24% SiO2, 14%,
B203), with a tensile strength of 1725 MPa and a tensile modulus of 138 GPa, in a 5-harness
satin weave fabric woven with 900 denier tows were acquired. The fiber diameter was
approximately 12 um with a coefficient of thermal expansion of 3.0x106/°C. To protect the
fibers from damage and to decrease the amount of interfiber friction, the manufacturer coats
each fiber with an organic sizing. In this study, fibers coated with an organic sizing are
referred to as “as-received”. The as-received fibers were thermally treated to form 2 different
fiber surfaces. In the first fiber surface treatment, the fiber was heated from room temperature
to 315°C, held for 1 hour and slow cooled to remove the organic sizing. These fibers are
referred to as “desized” fibers. The second fiber-surface preparation technique consisted of
heating the as-received fibers in a high-temperature ammonia atmosphere. This treatment
formed a boron nitride rich surface layer ~200 nm thick and has been reported previously

[40]. These fibers are referred to as BN-treated fibers.

40

3.3.2 Composite Prepregging:

The pretreated fibers, either in tow form or 5-harness woven cloth, were prepregged
with a BlaclcglasTM preceramic silsesquioxane polymer 493 (Allied Signal, density of ~1.09
g/cm3). The fiber tow was prepregged on a Research Tool filament winder. Unidirectional
prepreg tape was allowed to have carrier solvents flash-off for improved tact and drape
characteristics. Resin to fiber ratio is typically 45% and 55% respectively. Each ply of the
tape material was aligned in a 0° orientation to form a true unidirectional composite
laminate. The samples were typically (20) plies thick with an average composite panel
thickness ofO. 100 inches. 2-D woven composite panels consisted of sixteen layers of fabric,
which were placed to create a 0/90-balanced lay-up. The solvent flash time and associated
resin to fiber content is similar to that of the unidirectional tape. The 5-harness woven fabric
panels averaged approximately 0.150 inches thick.

3.3.3 Polymer Cure:

The initial polymer curing of sample panels was performed in a laboratory press without
vacuum bagging. The laminate was placed between nonperforated Teflon® film to aid in
resin containment. Flat shim stock was utilized to create stops, which prevent over bleeding
of the laminate and facilitate consistent panel thickness. Panels were consolidated at 3040
psi while undergoing the cure cycle. The cure cycle ramps from room temperature to 140°F
(60°C) at 4°F/minute followed by a hold at 140°1= (60°C) for 1 hour. After the initial hold the
sample is ramped from 140°1= (60°C) to 304°1=(150°C) and held at 304°F (150°C) for lhour.
Following the final hold temperature, the panel is allowed to cool under pressure to room
temperature. All samples were deflashed and subsequently cut via a diamond saw into

specific coupons for tensile, flexure, short beam shear, Iosipescu shear and general-purpose

41

coupons. Individual samples were then weighed to determine their initial condition prior to
pyrolysis.
3.3.4 Pyrolysis:

Cured polymer samples were then pyrolyzed to form the ceramic matrix, which is
performed in flowing nitrogen to prevent oxidation. The kiln is purged with nitrogen and
also slightly over pressured to maintain an oxygen free environment. The samples are placed
free standing on flat alumina blocks without additional fixturing. The heating is from room
temperature to 1652°F at 2.25°F\min. After this ramp phase, the samples are allowed to
stabilize at 1652°F for 1 hour and are finally slow cooled to room temperature at which time
the Nitrogen gas flow is halted.

3.3.5 Reinfiltration Process:

Upon removal from the kiln, the samples require reinfiltration with polymer resin to
fill open porosity, which has resulted from the polymer to ceramic phase conversion. The
reinfiltration resin utilized is a lower viscosity version of the BlackglasTM 493 system.
Samples have a rough and grit like surface finish, which is lightly abraded to remove debris
and aid in opening pores, which the resin may then fill. The sample is then weighed and
placed in a vacuum bag containing reinfiltration resin. A vacuum is drawn on the bag
containing the samples thus drawing the resin into each sample uniformly. The resin is held
within the samples via this vacuum application while the press again consolidates the
composites under 30-40 psi. The preform temperature is raised from room temperature to
60°C at a rate of 2.2°C/min and held for 1 hour. After 1 hour, the temperature is raised to
150°C at 2.2°C/min and held for 1 hour. Following this heating segment, the preform was

press-cooled to room temperature under an applied load of 3040 psi.

42

Following the impregnation step, the preforms were pyrolyzed in flowing nitrogen by
heating the preform from room temperature to 900°C at l.25°C/min and holding for 1 hour.
The preform was then slowly cooled to room temperature. Further densification of the silicon
oxycarbide ceramic matrix composite occurred by applying 5 reimpregnation cycles. Each
reinfiltration cycle was achieved by infiltration and curing of the catalyst-added solution
using the procedure above followed by a 1 hour, 900°C pyrolysis. To assess the progress of
the densification, one panel was removed after each densification step and reserved for
characterization. The status of impregnation is designated by the symbol BG(n) where n
ranges from 0 to 5. For example, BG(O) represents a composite with no reimpregnation
whereas BG(S) corresponds to the sample obtained after 5 reimpregnations. The fiber volume
fraction, measured optically, was 55 to 60%.

Following the densification process, the open porosity was verified using alcohol and
calculated using Archimedes principle. The open porosity volume fraction has been
previously correlated to baseline sample volumes via Helium pycnometry. (41) Open
porosity fraction relative to true sample volume have also been correlated to mechanical test
results as a function of the number of infiltrations. From this data, open porosity
measurements along with the infiltration number provide a reasonable prediction of
mechanical properties. Typically samples processed through five infiltrations as further
cycling yields only marginal improvements in mechanical properties while adding

significantly to processing cost and time.

43

3.4 ADHESION MEASUREMENTS:
3.4.1 Interfacial Testing Systems:

Fiber/matrix interfacial shear strength (188) measurements were made using a
microindentation test. Refer to Figure 3.1 for schematic of equipment [62,63]. Samples
were prepared by mounting composite cross-sections in standard metallurgical mounts.
The samples were polished with SiC paper to a grit size of 2000 then wet polished with
0.5 pm alumina powder until smooth. The individual fiber indentations were made using
a 5 pm radius diamond indenter and recording the load versus displacement values. The
188 is calculated through the use of a finite-element analysis of the stressed area for each

fiber under examination. At least 5 fibers were measured for each type of composite

[28,60].

E aoasm sauce Ecstsezo oeeaofim ”3 seem

gaugirx

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.282 83> — —

 

 

 

 

45

3.4.2 Fiber Pull-Out Test:

Sample preparation and testing of Nextel 312TM fibers embedded in Allied Signal
BlackglasTM matrix micro-droplets is a useful means to determine the shear stresses along the
fiber-matrix interface as well as determine the frictional nature of the adhesion as the fiber
slides through the droplet. The micro-droplet method utilizes a single drop of matrix
material applied to the fiber. After the curing and pyrolzation process the droplet is secured
in a microvise designed to hold the droplet in a fixed location while allowing the fiber to be

drawn from it [45]. The basic relationship is as follows:

where 1.2 is the embedded fiber length, d is the fiber diameter and Cf is the uniaxial tensile
load imparted on the fiber. This evaluation ignores the varying shear stress distribution along
the fiber length which are considered more closely for polymer matrix composites but due to
the stiffness associated with the ceramic matrix the above relationShip will serve as the
beginning benchmark. Once the fiber bond is debonded, the net compressive force on the
fiber due to residual curing shrinkage and CTE mismatch will be realized through the

following relationship:
7 = flPo
where u is the coefficient of friction and p is the compressive forces acting normally on the
fiber surface [45].
3.4.3 Contact Angle and Wettability Measurements
Contact angle and wettability measurements provide insight with respect to the

natural affinity for the matrix material to adhere to the fiber. Typically the contact or

46

surface energy is described as follows:

75V = 731. + y“, c030
where ’st and 752 and *yLv are surface free energies of the solid-vapor, solid-liquid and liquid-
vapor interfaces respectively as conceptualized from a water droplet on a flat surface. In the
case of a fiber surface wetted by an uncured matrix system, when the angle 0 = 0°, the fiber
surface will be wet very easily [64]. Measurement of the contact angle between the three
Nextel 312TM fiber samples and the BlackglasTM polymer system should indicate to what

degree adhesion will develop between the fiber and matrix.

3.5 COMPOSITE PERFORMANCE ANALYSIS:
3.5.1 In-situ ESEM processing analysis:

An ElectroScan 2020 environmental scanning electron microscope (ESEM) was used
to monitor the polymer-ceramic composite pyrolization to the final CMC state. The cured
polymer sample measuring .2 x .2 inches by .196” thick was affixed to the sample crucible
with silver filled conductive adhesive. The ESEM was fitted with a hot stage capable of
1000°C and a Nitrogen supply was established. After loading the sample, a slight vacuum of
3-5 torr was established, an operating voltage of 20-25 kv initiated and the nitrogen
environment started. The sample was incrementally heated from room temperature to 900°C
while rnicrographs were taken at 100°C intervals and a continuous video was captured. The
in-situ pyrolysis procedure was designed to gain further understanding of the various

reactions that occur during processing through the polymer to ceramic conversion. Figure

47

3.2 below summarizes the basic pyrolysis process with associated gas evolution as heat is
applied.

During pyrolysis the effect of the mismatch in thermal coefficient between the
ceramic fiber and the polymeric matrix creates stresses, which can lead to rnicrocracking. As
previously stated, the fiber has a thermal expansion coefficient of 3.0 x 10‘6 Alll/°C while the
matrix has a varying coefficient as is progresses from a polymer to a ceramic at which time
the CTE is 1.9 x 10" Al/l/°C. In the final ceramic state and after cooling, a residual tensile
force would be expected based on expansion coefficient considerations alone, however, the
dynamic of matrix shrinkage must also be considered in the pyrolization process.

Along with the inherent fiber-matrix thermal expansion mismatch, the BlackglasTM
polymer itself has a radical volumetric shrinkage associated with the phase conversion.
During the pyrolysis, the shrinkage associated with the conversion from a polymer to an
amorphous carbon rich matrix has been known to create matrix cracking and fiber
debonding. Due to the magnitude of matrix shrinkage during processing, the effect of CTE
mismatch is secondary with respect to that of processing shrinkage. The BN treated, as-
received and desized fiber will be evaluated for their potential interaction with this matrix
phenomenon. The ESEM provides the unique possibility to observe the fiber interaction with

matrix shrinkage along with the expansion mismatch effects real time.

48

PYROLYSIS
REACTION

Methane Hydrogen
C2 and C3 Higher

Hydrocarbons

I We}?!
I
Heat E; Heat Heat Heat
; I
I II

Room 330 0C 450 °C 620 °C 900 °C
Temperature ‘

Preceramic ‘ SiCxOy
Ceramic

Polymer

 

 

     

Figure 3.2: Pyrolysis Process.

49

3.5.2 Flexure Testing:

The flexure strength was evaluated in 4-point bending (specimen dimensions: (0.730”
x 0.196” x 4”, span-to-depth = 16:1) using ASTM standard D790M, 1986. All shear
strengths were calculated using standard beam formulae. A minimum of eight specimens
was tested in this test configuration. The mean and standard deviation were then calculated.
The 4-pt. bend test was performed using a hydraulically driven MT S System 6 load frame
with a crosshead speed of 1.27 mm/minute.

3.5.3 Short Beam Shear Testing:

The interlaminar shear strength of each composite was determined by testing in short
beam shear (specimen dimensions: 0.250” x 0.150” x 2.5”, span to depth 4:1) configuration
(ASTM D2344 1984). A minimum of 8 specimens was tested in this test configuration. The
mean and standard deviation were then calculated. The interlaminar short beam shear tests
were-performed at Auto-Air Composites, Inc. in Lansing, Michigan on an Instron Corp.
Series IX Automated Materials Testing System. The crosshead speed was 1.27 mm/minute.
3.5.4 Iosipescu Testing:

The in-plane (interlaminar) shear strength (specimen dimensions: 0.75” x 0.196”
x 3.5”) was determined using the Iosipescu shear test (ASTM D5379/D5379M 1993)
with the warp direction in—line with the load direction. A minimum of 8 specimens was
tested in this test configuration. The mean and standard deviation were then calculated.
The in-plane Iosipescu shear test was performed using a hydraulically driven MT S

System 6 load frame with a crosshead speed of 1.27 mm/minute.

50

3.6 FRACTU RE ANALYSIS:

An ElectroScan 2020 environmental scanning electron microscope (ESEM) and a
JEOL J SM-6400V SEM were used to examine the surface-of failed 4—pt. bend specimens.
The JEOL system was equipped with a Noran Vantage System for energy dispersive x-ray
spectroscopy (EDX). The EDX system is equipped with a Moxtek window having a lower
limit of detection of beryllium with a Noran Extreme detector with a resolution of 126 eV.
The fracture analysis of BN treated, as-received and desized flexure samples will be
conducted to determine the loci of failure and the corresponding mode of failure. The nature
of the failure, whether it is adhesive between either the matrix and fiber or treatment and
fiber, cohesive within the matrix, treatment or fiber will provide insight into the ability of the

CMC to adsorb energy efficiently.

51

Chapter 4

RESULTS & DISCUSSION

4.1 Characterization of Fiber Samples:
4.1.1. XPS Results:

X-ray Photoelectron Spectroscopy (XPS) results (with an error of :05 at.%) are
summarized in the table 4.1. The as-received fibers show large carbon, oxygen, and nitrogen
concentrations consistent with an organic sizing. Minor amounts of boron, aluminum, and
silicon are present with trace amounts of chlorine and sulfur. Since the XPS probing depth is
~6 nm, the presence of boron, aluminum and silicon (fiber constituents) suggest that the
sizing may not have uniformly covered the fiber. The XPS results also indicate that desizin g
the as-received fibers removes much of the nitrogen and carbon, and exposes the oxygen,
aluminum, boron and silicon fiber constituents. As expected, the surface chemistry of the
desized fibers more closely reflects the chemistry of the aluminoborosilicate fibers. The BN-
treated fiber surface composition has 22% C, 21.4% 0, 23.4% N, and 21.0% B. Smaller
amounts of aluminum and silicon are also present. The results indicate that the boron coating
treatment produced a fiber surface rich in boron and nitrogen. The high carbon content
probably results from adventitious carbon. No attempts at measuring the BN thickness were

made, but the coating thickness is believed to be on the order of 100 to 300 nm [40].

52

Table 4.1: Fiber Surface Chemistry (atomic%).

 

 

Element As Received Desized Boron Nitride Treated
Carbon 63.9 27.0 22.0
Oxygen 21.0 45.8 21.4
Nitrogen 10.7 - 23.4
Boron 1.2 5.2 21.0
Aluminum 1.6 18.7 8.1
Silicon 1.0 3.3 4.2
Sulfur 0.3 - -
Chlorine 0.3 - -

 

 

Nextel 312TM fiber surface morphologies for the as-received, desized and BN-treated
samples are shown in figures 4.1, 4.2, and 4.3 respectively. The figures show that each fiber
has a slightly different morphology. The organic sizing on the as-received fiber surface
appears nonuniform with an uneven thickness. The underlying fiber microstructure is slightly
evident. In contrast, the desized fiber shows a uniform microstructure (though individual
grains cannot be discerned). The ESEM examination of the BN-treated fiber surface

indicated a uniform coverage of coating material.

53

 

Figure 4.1: ESEM micrograph of as-received Nextel 3121'M fiber surface.

54

 

 

Figure 4.2: ESEM micrograph of desized Nextel 312TM fiber surface.

55

 

 

Figure 4.3: ESEM micrograph of BN-treated Nextel 312TM fiber surface.

56

4.1.2 Auger Spectroscopy Results:

Auger Electron Spectroscopy (AES) was used to characterize the single fiber samples
surface chemistry and obtain a depth profile of the constituents present. During the analysis
of the fibers, it was noted that the insulative capacity of an aluminoborosilicate fiber created
a sample charging condition resulting in sample damage. The fibers were gold sputter coated
to disperse the beam energy and minimize the associated beam damage incurred by the
sample. The efforts to protect the fiber were unsuccessful and the depth profiling was not
possible with the surface modified or desized Nextel 312TM samples.

4.1.3 Stress-Strain Measurements:

The three fiber samples used in the experimental work were reviewed for their
possible degradation due to surface treatment processes. As-received samples were not
exposed to pretreatment heating, which preserved their parent tensile strength of 1725 MPa.
The desized samples were thermally cleaned at a temperature of 315°C to remove any
organic sizing compounds. This low temperature exposure is below any published
degradation levels for the Nextel 312TM fiber system thereby maintaining the fiber tensile
strength of 1725 MPa. The N extel 312TM single fiber samples, which were Nitride treated in
an ammonia atmosphere at approximately 1150°C to form a BN rich surface composition,
were investigated further. The nitriding process required to form the BN layer has been
described previously via a reference to the Boeing patented process. Allied Signal has
modified this baseline generic process to facilitate BlackglasTM composite performance and
the exact parameters of the Nitriding process are proprietary [57]. After the proprietary
nitriding process, Allied Signal performed testing of BN treated single fibers in uniaxial

tension. The data suggests that the initial fiber tensile strength of 1725 MPa was reduced 7-

57

10% to 1595 MPa. Tensile property reduction of this nature is believed to be a result of
Boria leaching from the bulk fiber structure.

The BlackglasTM polymer precursor used in this experiment proved extremely brittle
after the initial cure and exhibited extensive cracking. Single fiber fragmentation coupons
were, therefore, omitted from the testing because the BlackglasTM ceramic matrix was too
brittle for sample preparation and would otherwise not yield the birefringent patterns or
photoelastic response necessary to obtain critical fiber lengths required to compute the
interfacial shear strength. Polymer state BlackglasTM encapsulation would reveal little about
the interfacial shear stress in the final CMC system and therefore was not considered either.
Single fiber testing with the ceramic matrix would yield important data for determining the
interfacial shear strength however the microcracking which occurs during processing makes
it impossible to process single fibers in with a matrix rich dog bone geometry.

4.2 Composite Sample Processing Results:

Several reports have studied the pyrolytic conversion of polysilsequioxanes to silicon
oxycarbides [65]. These studies indicate that pyrolyzing a polysiloxane polymer produces
silicon oxycarbide with a high ceramic yield (70-80 wt%) containing a network of
microcracks and porosity due to volume shrinkage (55%). The amount of microcracking and
porosity increases when fabricating a silicon oxycarbide\aluminoborosilicate fiber composite
due to fiber constraints and the dissimilar shrinkage between the matrix and fibers [66,67].
Thus several infiltration cycles are required to fill the cracks and densify the matrix of the
composite.

The processing data in Figure 4.4 shows the % weight change —relative to the

composite after one pyrolization cycle, but with no impregnation of resin (BG(0))—— as a

58

function of infiltration cycle for as-received and BN—treated fiber composites (no statistical
difference between the two types of composites was measured). The green composite
consists of 58-wt% fiber (p = 2.7 gm/ml) and 42-wt% resin (p = 1.1 gm/ml). The first
pyrolysis cycle going from the cured state to the BG(0) is the most important step since this
determines not only the number of subsequent infiltration/densification and pyrolysis steps to
follow, but also the microstructure and properties of the final composite. On the first
pyrolysis step, the green compact experiences a 7.1% weight loss and an increase in density.
The weight loss results when chemical reactions produce gases (hydrogen and hydrocarbon)
that evolve from the polymer [68]. Fortunately, much of the porosity present after BG(O) is
interconnected and will be filled with resin on subsequent infiltrations.

0n subsequent pyrolysis cycles, the composites gain weight. The most dramatic
weight gain occurs after the second pyrolization (BG(1)) where a 15.6% gain is observed.
Less dramatic increases are observed for subsequent pyrolization/infiltration cycles. The total
weight gain in the fifth impregnation was 33.3%. After five infiltrations the apparent (or
open) porosity (the extent of the closed porosity is unknown) did not change suggesting that
the maximum densification, using the current procedure, had been reached. As Rangarajan et
a1. points out, this method of silicon oxycarbide matrix composite fabrication leads to closed
porosity in the form of micro- and nano-sized pores [67].

Conversion of the pre-ceramic polymer (without fibers) to silicon oxycarbide was
also monitored. The volume decreased to 55% of the initial volume. The typical composition
of the silicon oxycarbide glass, as determined by XPS, was 24 wt% carbon, 47 wt% silicon,
28 wt% oxygen, <0.4 wt% hydrogen and <01 wt% nitrogen. This composition results in a

matrix composition of SiC 12101.06. The matrix char yield after pyrolysis was 81 to 86% and

59

the density increased from 1.10 gm/ml (polymer) to 2.11 gm/ml (ceramic). The surface
chemistry of the ceramic matrix material was measured and consisted of 17.6 at.% carbon,
38.4 at.% oxygen, and 44 at.% silicon. No evidence of nitrogen, which is present as a gas
during the pyrolization process, was detected in the matrix material.

Attempts at fabricating composites with desized fibers were unsuccessful due to
spontaneous delamination of the fiber and matrix plies after each pyrolization cycle. The
exact cause of the delamination is not known; however, a strong fiber-matrix chemical
reaction combined with resin shrinkage and differences in matrix and fiber coefficient of

thermal expansion (CTE) can lead to high residual stresses causing delamination [69].

160-

 

 

 

150-
1- - *
4° +Welght
130i ()
120
110
100l I I l I U I
geeseee
h (5 (5 (5 0 (5 (5
(5 m In 1n n: In to

Figure 4.4: Typical weight gain curve through
densification process.

Table 4.2: Sample Weight Gain Through Densification Cycles.

 

 

 

Process Stage Weight (g) Cumulative Percent Delta Thickness (in. )
- weight gain% ‘
Green 121.91 **** **** .1490
BG(O) 113.21 100 -7.13 .1480
BG(l) 130.93 1 15.64 15.64 .1424
BG(2) 140.79 124.35 7.53 .1412
BG(3) 146.34 129.25 3.94 .1405
BG(4) 149.28 131.85 2.01 . 1405
BG(S) 150.98 133.35 1.14 .1405

 

61

4.3 ADHESION MEASUREMENTS:

4.3.1 Interfacial Testing System:

The microindentation technique test measures the fiber-matrix interfacial shear
strength (IF SS) of an individual fiber. The microindentation test consists of loading the
cross-section of an individual fiber oriented normal to the polished surface with a conical
indenter. The load is incrementally applied until a crack (debonding) develops at the
fiber/matrix interface. Reference Figure 4.5 and Figure 4.6 for typical debonding indications
about fiber circumference. The load at debonding is assumed to represent the shear strength
of the interface. The debonding load is influenced not only by the fiber/matrix bond strength
but also by the fiber and matrix properties (such as moduli and Poisson’s ratios) and fiber
packing. An analytical closed form solution for the stress distribution around the indented
fiber is not available so the calculation of an average shear stress relies on a finite element

calculation [70]. The data reduction scheme uses the following formula:

IFSS = Aé 0.875696 3"- —0.018626ln d" —0.026496
df Ef df

Where fg is the load at debonding, Gm is the shear modulus of the matrix, Eris the tensile

 

modulus of the fiber, do is the distance between the nearest fiber and the loaded fiber, dris the
diameter of the fiber, and A is a conversion factor. For the silicon oxycarbide matrix, Gm =
73.75 MPa, and v = 0.28. The fiber tensile modulus is 138 GPa, with a tensile strength of 1.7
GPa.

The IF SS results are for the as-received and BN-treated fiber composites are reported
in Table 4.3. For the as-received fiber composite, the IFSS was found to be 8.9 i 2.1 GPa.

This stress level exceeds the fracture strength of the fiber and the fracture stress of the silicon

62

oxycarbide. Tested fibers that exhibited cracking were discarded and not used in the
calculation of the IFSS average. The anomalously high value of IFSS results because the
FEM model does not include residual stresses caused by the radial shrinkage stress of the
matrix around the fiber or the longitudinal compression of the fiber caused by matrix
shrinkage. These residual stresses in the as-received fiber composite could be extremely high.
The effect of fiber spacing can also have important consequences on the debond strength
(less for high modulus materials). This certainly will affect the outcome somewhat, but not
enough to alter the conclusion: the IFSS of the as-received fiber composite is very high.
The data in Table 4.3 for the IFSS of the BN-treated fiber composite is 860 :1: 34
MPa. Experimentally, the tested fibers debonded gently and clearly showed interfacial crack
formation. The standard deviation is 4.0% of the mean value. This level of uncertainty could
result from several factors including: the testing of fibers from different regions of the
composite (differences in the thermal residual stresses of the innermost and outermost plies

could affect debond strength) [71].

Table 4.3: Interfacial Shear Strength Results.

 

 

 

 

 

 

 

 

Sample Strength Strength
(MPa) (ksi)
BN treated 860 :l: 34 125 i 4.9
As received 8860 :1: 2090 1289 :I: 305
Desized Samples delaminated during processing.

 

63

 

Figure 4.5: IFSS micrograph of as-received sample.

 

Figure 4.6: IFSS micrograph of BN treated sample.

65

4.3.2 Fiber Pull-Out Test:

The fiber pull-out test was not completed due to the sample fabrication difficulty
encountered with this matrix-fiber combination. A single fiber could not be cast in the
matrix material due to matrix cracking and specimen instability. The cured polymer exhibits
visible macrocracking and the pyrolyzed polymer features 55% volumetric loss during
processing to 900°C (1652°F). In order to complete the pull out test, the matrix must be
uniformly fixtured while the fiber is placed in uniaxial tension, which could not be achieved
with the 493 Blackglas'm ceramic matrix system.

4.3.3 Wettability

The BlackglasTM polymer resin system 493A was utilized in the fabrication of all
composite panels. The resin system has been designed for ease of wet laminating practices
and RTM (resin transfer molding) process, in that the viscosity is extremely low. The
viscosity measures 5 centipoise, which enables the resin to permeate the fiber preform or
fabric lay up as the application dictates. During fabrication evaluations, the resin was easily
infused into the fabric with no noticeable air bubbles generated as the fabric was saturated.
Typically, air bubbles are generated when a high viscosity system is forced into the
reinforcement as entrapped air attempts to exit the layup. The spontaneous wet out of the
fabric was a clear indication that the wettability as observed was very good and that surface

tension issues have little effect on the resin to adhere to the fiber system.

66

4.4 COMPOSITE PERFORMANCE ANALYSIS:
4.4.1 In-Situ ESEM Processing Analysis:

The polymer to ceramic conversion was observed real time via an ElectroScan 2020
environmental scanning electron microscope equipped with a hot stage and nitrogen
environment. The instrument was operated at 20-25 kv while evacuated at 3-5 torr. 2-D
composite panels which were previously polymer cured were sectioned into .2 inch square by
.196 inch thick test pieces. Individual cycles were run for the as-receives-desized and BN
treated samples. Samples were heated from room temperature to 900°C in order to reproduce
the conversion process and witness the contribution of the three different fiber surfaces.
Micrographs, Figure 4.7 through Figure 4.12, were taken at RT-161°C, 301-327°C, 500°C,
600°C and 900°C respectively. The temperatures were selected due the process milestones
as shown in Figure 3.2 that illustrates the various byproducts throughout the conversion
process with corresponding temperatures.

The initial micrographs that were taken at RT-161°C serve as a baseline image where
no apparent changes in the structure were perceivable. At this initial stage there is evidence
of fiber debonding and matrix cracking in all three samples. The polymer had, during the
initial cure and subsequent cooling, experienced cracking due to shrinkage inherent to the
curing process. Neat polymer resin samples had displayed evidence of cracking after initial
polymer cure, which indicates the initial cracks witnessed at RT-161°C were not created
solely from fiber-matrix thermal expansion mismatch.

Micrographs taken at 301-327°C indicate some enhanced signs of cracking in both
the matrix and about the fiber circumference. This temperature range is that which begins

the evolution of methane C2 + C3 gases and also where volumetric shrinkage of the matrix

67

occurs. The as-received and desized samples appear to have a higher frequency of matrix
(fiber to fiber) cracks while the BN treated sample reveals more pronounced fiber debonding.

Increasing the temperature to 500°C begins the release of higher hydrocarbons and
further matrix shrinking. In both the as-received and desized samples further widening of the
matrix cracks is evident which could be attributed to adhesion of the fiber to the matrix
which drives the volumetric shrinkage forces within the sample to relieve themselves via
macrocracking within the matrix. The BN treated samples, however, have a different
appearance as the fibers appear to unbond more discretely leaving the matrix more in tact in
comparison the other samples. The BN treated sample appears to have fibers debonding and
pushing out of the sample surface which again could be a result of the coupon attempting to
relieve stresses built in the matrix due to the volumetric shrinkage.

Samples elevated to 600°C are beginning to evolve hydrogen gases and continue to
crack both within the matrix and around individual fibers. The as-received samples show
signs of fiber debonding as the fibers appear to be moving in and out of the surface. The
desized sample reveals more widened matrix cracking and a lesser degree of fiber unbonding.

The BN sample continues to displace fibers along their axis as slippage and debonding
occurs.

The final pyrolysis temperature of 900°C continues with hydrogen evolution and
completes the matrix shrinkage of 55% and associated density increase from 1.09 g/cm3 to
2.2 g/cm3. The adhesion between the Nextel 312TM fiber and the BlackglasTM matrix appears
to be generally in tact while the as received sample demonstrates smaller matrix cracks and
evidence of more localized unbonding around individual fibers. This more incremental

mechanism to relieve internal processing stresses, may be the reason the as-received samples

68

remain intact while the desized sample were completely delaminated to the state where
further testing was not possible. The as-received samples were coated with an organic
coating which most probably decomposed to form a thin layer of carbon as the samples were
processed. The carbon boundary layer may have, in an inert atmosphere such as Nitrogen,
provided the necessary decoupling path that the composite relieved process stresses more
gracefully and did not delaminate, as did the desized sample. The BN treated sample appears
to unbond consistently throughout the processing stages as was evidenced by the large in and
out of plane displacements made by individual fibers. It would seem then that a mechanism
to relieve internal stresses must be provided to prevent the composite from delaminating and

preventing subsequent infiltration processing.

69

 

Figure 4.7: ESEM in-situ pyrolysis. Top: As Received fiber composite at 650X
and room temperature. Middle: Desized fiber composite at 650X and Room
temperature. Bottom: BN Treated fiber composite at 1500X and 161°C.

70

 

Figure 4.8: ESEM in-situ pyrolysis. Top: As Received fiber composite at 1000X
and 300°C. Middle: Desized fiber composite at IOOOX and 301°C. Bottom:
BN Treated fiber composite at 2500X and 327°C.

71

 

Figure 4.9: ESEM in-situ pyrolysis. Top: As Received fiber composite at 600X
and 501°C. Middle: Desized fiber composite at 500x and 500°C. Bottom: BN
Treated fiber composite at 1500X and 499°C.

72

 

Figure 4.10: ESEM in-situ pyrolysis. Top: As Received fiber composite at
1000X and 599°C. Middle: Desized fiber composite at IOOOX and 600°C.
Bottom: BN Treated fiber composite at 1500X and 600°C.

73

 

Figure 4.11: ESEM in-situ pyrolysis. Top: As Received fiber composite at 600X
and 900°C. Middle: Desized fiber composite at 850X and 900°C. Bottom: BN
Treated fiber composite at 1500X and 900°C.

74

 

Figure 4.12: ESEM in-situ pyrolysis. Top: As Received fiber composite at
3050X and 900°C. Middle: Desized fiber composite at 1500X and 1000°C.
Bottom: BN Treated fiber composite at 2500X and 901°C.

75

4.4.2 Flexure Data:

The flexural strength is equal to the maximum stress in the outer surface at the moment
of break. As Table 4.4 and Table 4.5 indicate, the flexural strength of the BN-coated
composite was 178.2 MPa versus 25.3 MPa for the as-received composite. In addition, the
strain-to- failure was less than 0.05% for the as-received fiber composite and 0.5% for the
BN—coated fiber composite. Increases of ~700% in flexural strength and ~1000% in the
strain- to-failure suggests that the BN-coated fiber composites are stronger and tougher in
flexure than the as-received fiber composites. Load-displacement data shown in figure 4.13
illustrates relative strain capabilities. The strength of the BN-coated fiber composites agree
with the flexural strengths reported by Zhao et al. [72]. At ultimate strength, both of the
composites failed on the tensile stress side. The low values of strength observed for the as-
received composites are close to that reported for the matrix material alone and suggests that
the BlackglasTM matrix and not fiber/matrix interface debonding is playing the dominant role
in determining the ultimate strength. I

Table 4.4: As-Received Flexure Strength

 

 

 

 

 

 

 

 

 

 

 

Sample ID. Maximum Stress (psi) Maximum Stress (MPa)
ARFlex 1 3049 21
ARFlex 2 3794 26
ARFlex 3 3889 27
ARFlex 4 4315 30
ARFlex 5 4148 29
ARFlex 6 3898 27
ARFlex 7 3264 23
ARFlex 8 3144 22

Mean: 3688 25
SD: 476 3

 

 

 

 

 

76

Table 4.5: BN Treated Flexure Strength

 

 

 

 

 

 

 

 

 

 

 

 

Sample ID. Maximum Stress (psi) Maximum Stress (MPa)
BNFlex 1 Not Tested Not Tested
BNFlex 2 24966 - 172
BNFlex 3 26633 184
BNFlex 4 25999 179
BNFlex 5 25202 174
BNFlex 6 24479 169
BNFlex 7 27137 187
BNFlex 3 27064 187
BNFlex 9 25969 179

Mean: 25935 179
S.D.: 992 7

 

 

 

 

 

WAS-Received [IBM-Treated

250

 

 

 

 

 

;? /
/

0 500 1000 1500 2000
Displacement (Inches 1110")

Figure 4.13: BN Treated and As-Received Flexure Load vs. Displacement.
Illustrating .05 % strain to failure for as received and .5% for the BN
treated flexure samples.

8

 

 

 

 

 

 

4.4.3 Short Beam Data:
The data in Table 4.6 and Table 4.7 also indicates that BN-treated fibers show a 600%

increase in the interlaminar shear strength as determined using the short beam shear test.

77

Low values of interlaminar shear suggest that the matrix properties are playing the major

role in determining the shear strength of the composite.

Table 4.6: BN Treated Short Beam Shear Strength

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Sample ID. Interlaminar Shear (psi) Interlaminar Shear (MPa)
BNSBS 1 4658 32
BNSBS 2 4711 32
BNSBS 3 4756 33
BNSBS 4 4777 33
BNSBS 5 4607 32
BNSBS 6 4569 32
BNSBS 7 4328 30
BNSBS g 4731 33
BNSBS 9 4799 33
BNSBS 10 4685 32
Mean: 4662 32
S.D.: 138 1

 

Table 4.7: As-Received Short Beam Shear Strength

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Sample ID. Interlaminar Shear (psi) Interlaminar Shear (MPa)
ARSBS 1 752 5
ARSBS 2 770 5
ARSBS 3 816 6
ARSBS 4 773 5
ARSBS 5 708 5
ARSBS 6 740 5
ARSBS 7 698 5
ARSBS s 700 5

Mean: 745 5
S.D.: 42 0

 

78

 

 

4.4.4 Iosipescu In-Plane Shear Strength:

The data in Table 4.8 and Table 4.9 shows the as-received fiber composite has in-
plane shear strength of 9 MPa, while the BN-treated composite has shear strength of 102
MPa. This represents a ~1000% increase in the interlaminar shear strength when the
composite is fabricated using BN-treated fibers. In addition to differences in the shear
strength, the as-received fiber composite failed when cracks developed and propagated in
straight-lines perpendicular to the loading direction and between the notches. On the other
hand, the BN-treated fiber composite failed with cracks developing and propagating between
the notches along a line approximately 45° to the loading direction. Macroscopically, the
BN-treated fiber composites had some fiber pull-out, while the as-received fiber composites
had smooth, flat surfaces.

Table 4.8: As-Received Iosipescu In-Plane Shear Strength

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Sample ID. Maximum Shear Stress (psi) Maximum Shear Stress (MPa)
AR 1 2552 18
AR 2 250 2
AR 3 333 2
AR 4 2274 16
AR 5 259 2
AR 6 1940 13
AR 7 445 3
AR 3 2124 15
AR 9 2911 20
AR 10 280 2
Mean: 1337 9
S.D.: 1110
79

t .—

 

Table 4.9: BN Treated Iosipescu In-Plane Shear Strength

 

 

 

 

 

 

 

 

 

 

 

 

 

Sample ID. Maximum Shear Stress (psi) Maximum Shear Stress (MPa)
BN 1 14382 99
BN 2 15762 . 109
BN 3 11307 73
BN 4 15615 108
BN 5 12681 87
BN 6 14359 99
BN 7 17204 119
BN 8 16366 113
BN 9 16305 112

EN 10 13819 95
Mean: 14780 102
S.D.: 1829 13

 

 

 

 

 

As-Fteceived
Fiber Composite

BN-Treated
Fiber Composite

Strength
(MPa)

 

4-pt. Short Iosipescu
Flexure Beam In-plane

Shear Shear
Figure 4.14: Mechanical Testing Summary

80

Both types of observed failure modes are generally acceptable; however, the large
standard deviation of the as-received fiber composites indicates a large amount of scatter in
the individual tests. Of the ten tests, five of the test values were at or below 3 MPa, while the
other five test values were above 13 MPa. If the 5 test values at 3 MPa and below are
disregarded, the average shear strength is 16.4 :1: 2.7 MPa. This is still significantly below the
value of 102 MPa for the BN—treated fiber composite, but with a much smaller standard
deviation. The reason(s) for the large amount of scatter in the as-received fiber composites is
unknown, though one possible contribution to the observed scatter in the shear strength could
be the nonuniformity of the organic sizing. If during repeated infiltrations the sizing is
consumed, regions of the as-received fibers will behave similar to desized fibers. The desized
regions would generate large stress concentrations and cause the composite to fail at a lower
stress level. Another possibility is the presence of small, unobservable, composite
manufacturing flaws or small cracks introduced in when machining the notches. Regardless
of the origin, the differences in shear strength between the as-received fiber composite and
the BN-treated fiber composite is substantial.

The summary data provided in Figure 4.14 illustrates the significant improvement
gained when the Boron Nitride system is utilized. The BN treatment also provided a

more predictable failure as demonstrated through the lower standard deviations noted.

4.5 FRACTURE ANALYSIS:
The fracture properties of ceramic matrix composites are governed by matrix cracking
followed by interaction of cracks with the fibers and interfaces. Within the composite, energy

dissipative processes such as fiber-matrix debonding, crack bridging, fiber pull-out, multiple

81

crack formation and crack deflection can be observed [73]. To determine the fiber-matrix
damage and fracture process for each composite type, specimens failed in 4-point flexural
tests were examined. In all cases, during mechanical testing, cracking sounds were heard near
the failure load, but no visual indication of failure was observed until catastrophic failure.

The ESEM micrographs in Figure 4.153 and Figure 4.15b show the interlarrrinar
fracture surfaces of desized fiber composites. Figure 4.15a is a low magnification view
showing matrix cracking caused by resin shrinkage. The large cracks seen in the matrix are
indicative of a weak matrix, a consequence of the dissimilar shrinkage of fiber and matrix
during pyrolysis and cool down from temperature. A higher magnification in Figure 4.15b
shows how the crack propagated near the fiber surface. This is consistent with XPS results
that indicated delamination cracks propagated through the matrix and that the locus of failure
was cohesive in the matrix.

The micrographs shown in Figure 4.16a and Figure 4.16b illustrate of the failure
behavior of as-received fiber composites. Figure 4.16a shows a macroscopically rough
surface indicating some crack deflection as it passed through the composite. A higher
magnification view shown in Figure 4.16b indicates a microscopically smooth surface with
no evidence of hackles or scallops or fiber pull-out. In particular, the crack path shows a
smooth transition across the fiber-matrix interface. These observations indicate strong fiber-
matlix adhesion. This also explains why the composite properties of flexural and shear
strength are dominated by the mechanical properties of the matrix. These results are
consistent with the XPS examination of the fracture surface chemistry that indicates exposed

fiber surfaces. These results suggest that the locus of failure alternates between the fiber-

82

matrix interface and cohesive failure of the matrix. The high magnification view also shows
the presence of small voids and microcracks at some of the fiber- matrix interfaces.
ESEM micrographs of BN-treated fiber composite fracture specimens are shown in
Figure 4.17a and Figure 4.17b. The fracture surfaces seen here are significantly different than
either the desized or the as-received fiber composites. Figure 4.17a shows that
macroscopically the fracture surfaces are extremely rough with a substantial amount of fiber
pull-out. In addition, fiber fracture continues at various planes throughout the composite,
resulting in a brush- like rough surface of broken fibers and/or bundles. At a higher
magnification fiber-matrix separation is observed and resin fracture occurs between fibers. In
this case, as Figure 4.17b shows, fiber pull-out results in a cupping surface. The presence of a
large amount of matrix debris suggests a large amount of energy dissipation during crack
propagation through the composite. These failure surface observations correlate well with the
mechanical properties. Crack propagation by debonding the fiber- matrix interface is typical
throughout the crack-propagation path for the higher strength sample. This type of cracking
consumes more energy and makes other energy dissipating mechanisms, such as fiber
pullout, fiber bridging, and crack deflections, possible. The amount and extent of fiber/matrix
debonding give an indication of how well the energy-dissipating mechanisms perform. The
extensive debonding along the crack propagation path can be attributed to the relatively
homogeneous distribution of the fibers and matrix, which makes extensive fiber pullout and

fiber bridging possible.

83

 

Figure 4.15a: Delaminated desized composite sample at low magrification.

84

 

Figure 4.15 b: Delaminated desized composite sample at high magnification.

85

 

Figure 4.16a: As received low magnification fi'acture surface.

86

 

Figure 4.16b: As received high magnification fracture surface.

87

 

Figure 4.17a: BN treated low magnification fracture surface.

88

 

Figure 4.17b: BN treated high magnification fiacture surface.

89

4.6 SURFACE CHEMISTRY OF FRACTURED SPECIMENS:

Controlling the fiber-matrix interface behavior is a key factor in developing
appropriate mechanical properties of composites. Determination of the locus of failure at the
fiber- matrix interface provides direct information on hOw the fiber coating affects the load
transfer from the matrix to the fiber. As the diagram in Figure 4.18 shows, for the case of
silicon oxycarbide composited with BN treated aluminosilicate fibers, failure can occur (1)
cohesively in the matrix, (2) at the matrix—coating interface, (3) at the coating- fiber interface,
(4) cohesively in the matrix, or (5) cohesively in the BN coating. To help determine the locus
of failure, XPS analysis of as-received, desized and BN-coated composite specimens
fractured in 4-point bend tests (or delaminated) was performed and the results are reported in
Table 4.10. Since the sampled area is approximately 2 m2, the XPS data are averaged over
a large spatial area and accumulated from fiber surfaces, fiber ends, and matrix material. The
large sampling area makes a definitive identification of the locus of failure difficult, but does
provide useful information that can help guide the determination.

Data for the spontaneously delaminated desized fiber composites indicates the
presence of carbon, oxygen and silicon. The lack of an aluminum signal suggests complete
coverage of the fibers by matrix material and cohesive failure of the matrix near the fiber
surfaces. The as-received fiber composite shows large amounts of carbon, oxygen and
silicon, with small amounts of nitrogen, boron and aluminum. In contrast to the desized fiber
composites, small amounts of aluminum, nitrogen and boron are seen in the as-received fiber
composites. This suggests that at least some exposed fiber surface is being detected. Though
the locus of failure for the as-received composite cannot be made with certainty, differences

between the desized and as-received composites suggest differing loci of failure.

The composition of the BN-treated fiber composite listed in Table 4.10 indicates a
large amount of boron and nitrogen in a 1:1 correspondence. In addition, a small oxygen
concentration was measured. These results suggest that the BN coating covers the surface of
the fiber and excludes the aluminum in the fiber from being probed by the x-ray. Since XPS
analysis of 4 pt. flexure specimens is suggestive, but inconclusive in determining if the fiber-
matrix locus of failure occurred at the fiber-BN interface or the BN— matrix interface, EDX
analysis of the fracture surface was performed.

EDX was used to analyze the BN-coated fiber surfaces exposed during the fracture
process and the matrix material (fiber channels) exposed during fiber pull-out. (Note that
because of the inherent surface roughness of the fracture surfaces extreme care was taken to
acquire accurate EDX spectra of the convex fiber and concave channel surfaces.) Figures
4.19a and 41% show an EDX spectrum taken from a region of the fracture surface where
fiber pull-out (channel) occurred. No evidence of boron (0.185 keV) is seen in the spectra,
though nitrogen (0.392) is clearly evident. The nitrogen signal could result from nitrogen
atoms incorporated into the matrix during composite fabrication (though nitrogen was not
detected in the XPS analysis of the matrix, the sampling depth of XPS is much less (2 pm
compared to 6 nm) than EDX) or could be remnants from the BN coating. An EDX spectrum
taken from the fiber surface is shown in Figures 4.20a and 4.20b. The EDX spectrum shows
boron and nitrogen on the fiber surface. The EDX results, combined with the XPS results,
indicate the locus of failure of BN-coated fiber composites occurs primarily at the BN-
matlix interface. The possibility of failure within the BN coating cannot be completely
eliminated; however, the data suggests that the majority of the coating remains on the fiber

surface. This result indicates that the thin BN coating forms a stronger bond with the fiber

91

than the matrix, and has enough BN-matrix bond strength to provide load nansfer, yet weak

enough to allow fiber pullout during crack propagation.

Shear Stresses

.
.u.
sieve:
(«H‘ Muuuuuu
.-.n‘;.:.».o.-.5.-.p.p.;.s
. n u
.2

a
o ..»u
a

v v
‘11

1‘

.u. .~4~4<a~-e«
s P. u \

.~4

»

a'a
u'tvurt

    

Propagation

MAmNI—i

BN Coating

Figure 4.18: Schematic diagram illustrating the possible locus of failure.

92

Table 4.10: Fracture Surface Atomic Concentrations (%)

 

 

 

 

 

 

 

 

 

 

Element Boron Nitride As Received Desized
Treated

Carbon 58.1 52.4 52.8
Oxygen 18.0 28.4 31.3
Nitrogen 8.1 0.8 -
Boron 7.9 0.8 -
Aluminum 0.6 3.4 -
Silicon 7.2 14.2 16.0
Sulfur - - -
Chlorine - - -

 

 

 

 

Table 4.11: Matrix Atomic Concentrations (%)

 

 

 

 

 

Element BlackglasTM Matrix
Carbon 17.6
Oxygen 38.4
Silicon 44.0

 

 

 

93

 

 

 

 

 

0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80

Figure 4.1%

 

        

 

 

0.10 0.30 0.30 0.40 0.50 0.60
Figure 4.191:

Figure 4.19: EDX spectra taken from the matrix region exposed during fiber pull-out
(4.19a) and an enlargement of the low energy section of (4. 19a) shown in (4.1%).

94

    

B C N
0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80

 

Figure 4.20a
O

     

         

0.10 0.20 0.30 0.40 0.50 0.60

Figure 4.20b

Figure 4.20: (4.20a) EDX spectra taken from the fiber surface showing the presence of
boron and nitrogen and 4.20b and enlargement of the low energy spectra shown in 4.20a.

95

Chapter 5

Conclusions

This work investigates the effects of a boron nitride coating on the mechanical
properties of an aluminoborosilicate fiber/silicon oxycarbide composite. The mechanical,
chemical, and fracture properties of composites fabricated with BN treated, desized and as-
received fibers are compared. Large internal stresses formed during the
infiltration/pyrolization of the desized fiber/silicon oxycarbide composite caused the
composite to spontaneously delaminate. Fractographic examinations of flexure test
specimens indicate that the BN treated fiber/silicon oxycarbide matrix composite has a
brush-like surface with extensive fiber pull-out. In contrast, the as-received fiber/silicon
oXycarbide composite demonstrated cracks that propagate with little deflection resulting in a
smooth surface. Energy dispersive x-ray spectroscopy of the failed surfaces indicated that the
presence of the BN coating which caused the locus of failure to occur at the BN-treated
fiber/silicon oxycarbide matrix interface. Consequently, the mechanical properties of the BN
treated fiber/silicon oxycarbide composite increased by approximately 85% over the as-
received fiber/silicon oxycarbide composite. In addition, the BN-treated fiber/silicon
oxycarbide composites showed enhanced strain-to-failure. The high mechanical strengths
combined with a damage tolerant failure behavior suggest fiber-matrix bond strength strong
enough to impart strength to the ceramic, but weak enough to provide energy absorption and

crack deflection.

96

References

97

 

10.

11.

12.

13.

14.

15.

N. Chandra, C.R. Ananth, and H. Garmestani, "Micromechanical Modeling of
Process-Induced Residual Stresses in Ti-24Al-11Nb/SCS-6 Composite," Journal of
Composites Technology & Research, JCI‘RER, Vol. 16, No.1, January 1994, pp.37-
46.

Lowden, R.A., "Characterization of Fiber/Matrix Interactions in Ceramic
Composites"; Dissertation Abstracts International.

F. Woolley, “Melting/Fining”, Engineered Materials Handbook, Vol. 4, ASM, 1991
pp. 386-392.

G. Ziegler, W. Huttner, “Engineering Properties of Carbon-Carbon and Ceramic-
Matrix Composites”, Engineered Materials Handbook, Vol. 4, ASM, 1991 p. 841

A.E. McHale, “Engineering Properties of Glass-Ceramics”, Engineered Materials
Handbook, Vol.4, ASM, 1991 p. 875

A.E. McHale, “Engineering Properties of Glass-Ceramics”, Engineered Materials
Handbook, Vol. 4, ASM, 1991 pp. 870-871

A.E. McHale, “Engineering Properties of Glass-Ceramics”, Engineered Materials
Handbook, Vol. 4, ASM, 1991 pp. 874-875

A.I. Kingon, R.F. Davis, M.M. Thackeray, “Engineering Properties of
Multicomponent and Multiphase Oxides”, Engineered Materials Handbook, Vol. 4,
ASM, 1991 p. 763

Kingery, W.D., Bowen, H.K., Uhlmann, “Introduction to Ceramics, Second Edition”,
John Wiley & Sons, 1976 p.305

M. Miyayarna, K. Koumoto, H. Yanagida, “Engineering Properties of Single
Oxides”, Engineered Materials Handbook, Vol. 4, ASM, 1991 p. 752.

J. D. Sibold, “Wear Applications”, Engineered Materials Handbook, Vol. 4, ASM,
1991 p. 973.

P.T.B. Shaffer, “Engineering Properties of Carbides”, Engineered Materials
Handbook, Vol. 4, ASM, 1991 pp. 806-807.

Kingery, W.D., Bowen, H.K., Uhlmann, “Introduction to Ceramics, Second Edition”,
John Wiley & Sons, 1976 pp.79-81

S. Hampshire, “Engineering Properties of Nitride”, Engineered Materials Handbook,
Vol. 4, ASM, 1991 pp. 812-813.

F. V. Tooley, “Fiberglass”, Engineered Materials Handbook, Vol.4, ASM, 1991 pp.

98

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

402-407.

P.F. Aubourg, C. Crall, J. Hadley, R.D. Kaverman, D.M. Miller, “Glass Fibers”,
Engineered Materials Handbook, Vol.4, ASM, 1991 pp. 1027-1031.

D.D. Johnson, H.G. Sowman, “Ceramic Fibers” Engineered Materials Handbook,
Vol. 1, ASM, 1991 p. 60.

NW. Hansen, “Carbon Fibers”, Engineered Materials Handbook, Vol. 1, ASM, 1991
p. 111-113.

T.Schoenberg, “Boron and Silicon Carbide Fibers”, Engineered Materials Handbook,
Vol. 1, ASM, 1991 p. 592-593.

P.T.B. Shaffer, “Engineering Properties of Carbides”, Engineered Materials
Handbook, Vol.4, ASM, 1991 pp. 804-805.

H.T. Larker, “Hot Isostatic Pressing”, Engineered Materials Handbook, Vol. 4,
ASM, 1991 pp. 195-198.

D.P. Stinton, T.M. Besmann, R.A. Lowden, B.W. Sheldon, “Vapor Deposition”,
Engineered Materials Handbook, Vol. 4, ASM, 1991 pp. 215-220.

L.C.Klein, “Sol-Gel Process”, Engineered Materials Handbook, Vol. 4, ASM, 1991
pp. 209-213.

A.W. Urquhart, “Directed Metal Oxidation”, Engineered Materials Handbook, Vol.
4, ASM, 1991 pp. 232-235.

D.R.Petrak, “Polymer-Derived Ceramics”, Engineered Materials Handbook, Vol. 4,
ASM, 1991 pp. 223-226.

Tobin, Leung, Gonzcy “LC3 Summary”, Allied Signal, Northrup-Grummon, 1998.
Tauseef Chaudry, MS. Thesis 1995, Interphase Diagram

D.B. Marshall, "An Indentation Method for Measuring Matrix-Fiber Frictional
Stresses in Ceramic Composites," J. Am. Ceram. Soc. 67 [12] C-259-60 (1984).

DB. Marshall and A.G. Evans, "Failure Mechanisms in Ceramic Fiber/Ceramic
Matrix Composites," J. Am. Cerarn. Soc. 68 [5] 225-31 (1985).

K.T. Faber and Daniel R. Mumm, Department of Materials Science and Engineering,

Robert R. McCormick School of Engineering and Applied Science, Northwestern
University, Evanston, IL 60208.; "Interface Design in Brittle-Brittle Systems," 1994.

99

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

45.

46.

47.

W. Toreki, Polyrn. News, 16, 6 (1991)

Rajendra U. Valdya, Joseph Fernando, K.K. Chawla and M.K. Ferber, "Effect of
Fiber Coating on the Mechanical Properties of a Nextel480m-Fiber-Reinforced
Glass Matrix Composite." Materials Science and Engineering A150 (2) February 29,
1992, PP. 161-169.

Singh, R.N.; Brun, M.K.; "Effect of Boron Nitride Coating on Fiber-Matrix
Interactions," Advanced Ceramics Materials 3, (3), May 1988, 235-237.

Wells, H.K., Beaumont, P.W.R.: J. Mater. Sci. 17, 397 (1982)

Harris, B. Beaumont, P.W.R., deFerran, E.: J. Mater. Sci. 6, 238 (1971)

Harris, B., Morley, J ., Phillips, D.C.: J. Mater. Sci. 10, 2050 (1975)

Outwater, J.C., Murphy, M.C.: Proc. 24th Ann. Tech. Conf. SP1, Paper ll-C (1969)

Mullin, J .V.: Analysis of Test Methods for High Modulus Fibers and Composites,
349, ASTM STP 521, Philadelphia, Pa (1973)

TM. Besmann, B.W. Sheldon, R.A. Lowden, and DP. Stinton, Science, 253, 1104
(1991)

S. Campbell, and ST. Gonczy, Cer. Eng. Sci. Pro., 15/4, 327 (1994)
R.Y. Leung and M. Glazier, Cer. Eng. Sci. Pro., 15/7-8, 209 (1995)

W. Zhao, P.K. Liaw, R. Belardinelli, D.C. Joy, C.R. Brooks, and C]. McHargue,
Met. Mat. Trans. A, 31A, 911 (2000)

P.K. Khandelwal, and W.D. Wildman, Am. Cer. Soc. Bull., 65/2, 326 (1986)

Allied Signal, BlackglasTM Technology and Applications, An Advanced Ceramic
Matrix Material for High Temperature Composites l994-Product Bulletin.

Holtz, Allan R., and Grether, Michael F.; "High Temperature Properties of Three
NextelTM Ceramic Fibers." 32nd International Sampe Symposium and Exhibition at
Anaheim Convention Center, April 6-9, 1987.

Pipes, Byron R.; Johnson, DD. and Kurst, K.; "Nextel 312TM Ceramic Fiber
Polymeric Composites." 3-M Technical Bulletin.

Simpson, RH. and Veremnieks, J ., "Barrier Coated Ceramic Fiber and Coating

100

 

48.

49.

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

Method"; US. Patent #4,605,588, August 12, 1986.

Rice, R.W.; "BN Coating of Ceramic Fibers for Ceramic Fiber Composites"; US.
Patent #4,642,271, February 10, 1987.

Simpson, RH. and Verzemnieks, J .;"Boron Nitride Coated Ceramic Fiber and
Coating Method"; US. Patent #4,948,662, August 12, 1986.

Partlow, Deborah, P.; "A Cerarrric-Ceramic Composite with Low Dielectric Constant
and Nonbrittle Failure"; Advanced Ceramic Materials, 3, (6), 553-56, 1988.

Campbell, S. and Gonczy, S.; "In-situ Formation of Boron Nitride Interface Coatings
on Nextel 312TM Continuous Fiber 1: Nitriding Process and BlackglasTM Ceramic
Matrix Composite Properties".

Campbell, S. and Gonczy, S.; "In-situ Formation of Boron Nitride Interfaces on
Nextel 312TM Continuous Ceramic Fiber 11: Oxidation of BlackglasTM Ceramic
Matrix Composites".

Fernando, J .A.; Chawla, K.K.; Ferber, M.K.; Coffey, D.; "Effect of Boron Nitride
Coating on Tensile Strength of Nextel 480TM Fiber"; Materials Science and
Engineering, A154, (1992), 103-108.

Singh, R.N.; Brun, M.K.; "Effect of Boron Nitride Coating on Fiber-Matrix
Interactions"; CeramicEng. Sci. Proc. 8, (7-8) P.P. 636-643 (1987).

Lane, Jay, E.; Pebler, Alfred, R.; "Failure Characteristics of Low Dielectric Constant
Ceramic Composites Reinforced with BN-Coated Fibers"; Ceramic Eng. Sci. Proc.
10, (9-10) P.P. 1213-1222, (1989).

Hazlebeck, D.A.; Glatter, I.K.; Streckert, H.H.; "Novel Sol-Gel Coating Techniques
for Ceramic Tows: In-situ Curing vs. Reaction Bonding." Ceramics Eng. Sci. Proc.
12, (7-8), P.P. 1075-1085 (1991).

SS. Campbell and ST Gonczy, Allied Signal Inc. PO. Box 5016, Des Plaines, IL
60017-5016, and Edgar Leone, Allied Signal, 101 Columbia Rd, Morristown, NJ
07962, and M.J. McNallan, 2095 ERF, University of Illinois at Chicago, 842 W.
Taylor St., Chicago, IL 60607, "The Effect of Processing Parameters on the Surface
Nitridation of Nextel 312TM Ceramic."

Hercules, David M., Hercules, Shirley H., University of Pittsburgh, Pittsburgh, PA
15260, “Analytical Chemistry of Surfaces, Part I: General Aspects”

Hercules, David M., Hercules, Shirley H., University of Pittsburgh, Pittsburgh, PA
15260, “Analytical Chemistry of Surfaces, Part H: Electron Spectroscopy”

101

60.

61.

62.

63.

65.

66.

67.

68.

69.

70.

71

72.

73.

Drzal, LT. and Herrera-Franco, Pedro, J .; "Composite Fiber-Matrix Bond Tests",
Engineered Materials Handbook, Vol. 1, ASM, 1991, pp. 391-403

A.G. Tobin, R.M. Leek, and Tom Donnellan, Northrop Grumman Research &
Developement Center, Bethpage, NY 11714, "Process-Performance Relationships in
3-D BlackglasTM Ceramic Composites," Presented at the American Ceramic Society
Meeting January 7-12, 1995, Cocoa Beach FL.

E.J.H. Chen and J .C. Young, Comp. Sci. and Technol., 42, 189 (1991)

D.H. Grande, J .F. Mandel], and K.C.C. Hong, J. Mat. Sci., 23, 311 (1998)

Hull, Derek, Cambridge University Press, 1987, “An Introduction to Composite
Materials”, pp.38-40.

J. Annamalai, W.N. Gill, and A. Tobin, Cer. Eng. Sci. Pro., July-August, 225 (1995)

M. Allahverdi, W.R. Cannon, and 8.0 Danforth, J. Am. Ceramic Soc., 83/12, 2929
(2000)

S. Rangarajan, R. Belardinelli, and P. Aswath, J. Mat. Sci., 34, 515 (1999)

S.H. Yu, R.E. Riman, S.C. Danforth and R.Y. Leung, J. Amer. Cer. Soc., 78/7, 1818
(1995) '

T. Koufopoulos and RS. Theocaris, J. Comp. Mat., 3, 308 (1969)

1.0. Outwater and M.C. Murphy, in Proceeding of the 24th Annual Technical
Conference, The Society of the Plastics Industry (1969)

E.J.H. Chen and RB. Croman, Comp. Sci. and Technol., 48(1-4), 173 (1993)

W. Zhao, P.K. Liaw,R. Belardinelli, D.C. Joy, C.R. Brooks, and CJ. McHargue,
Met. and Mat. Trans. A, 31A, 911 (2000)

El. Hurwitz, P. Heimann, S.C. Farmer, and D.M. Hembree, J. Mat. Sci., 28, 6622
(1993)cv

102

    
     

   

IIIIIII‘I‘I

I

llll‘ll

2487

  

I

III
1293 0

IIIIIIIII

3