; a?” m

5:9...

. h. .. i...
“Wartmu

1r... a." :.

flak?

 

mmflmw
5...»...2.

.f
«(E s t...
#flwfiéfi $1.

Luau
3!}... W?

5.94....
a

I‘... aha.

61 I. {1,315 ‘
.6 . r.
I (I . A

 

.
(Magnum...

1....
. .5

‘.
.2;

THESIS

2 illlillilllilllllllllllllllllllHUI]Illlilllllllllllllllllll

x 31293 1 17313
(‘11?) (>80

LIBRARY
Michigan State
University

 

 

 

 

This is to certify that the

dissertation entitled

COLLOIDAL PROCESSING OF Sij/A1203 CERAMIC MATRIX CCMPOSITES

presented by

Mingli Zhang

has been accepted towards fulfillment
of the requirements for

DOCTOR OF PHILOSOPHY degree in MATERIALS SCIENCE
AND MECHANICS

(fly/1M} (Q; 30

 

ajor profc sor

Dateflflfl’m/gt’xi/Jj/f/ (ii)

MS U is an Affirmative A ction/Eq ua! Opportunity Institution 0-12771

 

 

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

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1/90 mm"

COLLOIDAL PROCESSING OF Sij/A1203 CERAMIC MATRIX COMPOSITES
By

Mingli Zhang

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

DOCTOR OF PHILOSOPHY

Department of Materials Science and Mechanics

1998

ABSTRACT
COLLOIDAL PROCESSING OF Sij/A1203 CERAMIC MATRIX COMPOSITES
By

Mingli Zhang

By manipulating the interparticle forces of ceramic colloidal systems, suspensions of '
difl‘erent dispersing abilities were obtained. Changing the concentration of ionic species,
pH, polyelectrolyte and the solution altered the net interparticle force. The zeta potentials
of single component suspensions of varying concentration of ionic species,
polyelectrolyte addition, pH and solution (electrolyte, ethanol, or a mixture of electrolyte
and ethanol) were measured. The stability ratio, W, which measures the efi‘ectiveness of
the potential barrier in preventing the particles from coagulation, and the interaction
between the components were discussed. SiC whisker reinforced A1203 suspensions
were slip cast at difi‘erent processing conditions (SiC whisker volume fraction, solids
loading, ball milling, sintering aids, polyelectrolyte and pH). Green Specimens were
cold-isostatically pressed and pressureless sintered in a flowing nitrogen atmosphere.

Homogeneous Sij/Alzog green bodies with densities of2.55eo.07 g/ern3 (~65%
theoretical density) were obtained. Bulk densities of 3.80i0.06 g/ern3 (96% theoretical
density), 3.79:0.06 g/ern3 (97% theoretical density), and 3.40eo.07 g/etrr3 (89%
theoretical density) were obtained at 1600°C for composite samples containing 5, 10 and
20 vol% SiC whiskers, respectively. Bulk densities of the 10 vol% Sij/A1203
composites were 3.79i0.06 g/cm3 at pH 11 and 3.66:0.07 g/cm3 at pH 4, respectively.

pH 11 was determined to be the Optimum processing pH for SiCalA1203 composites with

sintering aids (2 wt% Y203 and 0.5 wt% MgO) and 2.5 vol% polyelectrolyte (Ammonia
salt of a polymeric carboxylic acid). The final microstructure revealed homogeneous and
near fully densified composites. Compositions of the composites were characterized
using Energy Dispersive X—ray Spectrum (EDX).

The efi‘ect of the aspect ratio of SiC whisker, the Y203 content and the choice of
polyelectrolyte were examined. The whisker aspect ratio reduction had a more dramatic
efi'ect on achieving dense composites in 20 vol% SiC whisker when compared to 5 or 10
vol% SiC whisker composites. High densities with 10 vol% SiC whisker were attainable
without aspect ratio reduction. Sintered densities increased with increasing Y203 content.
For 10 vol% Sicwl A1203, the densities remained almost constant at Y203 additions
greater than 0.5 wt%. 2.5 vol% polyelectrolyte stabilized the suspension and resulted in
high composite densities. However, further addition of polyelectrolyte, past the
adsorption saturation limit, served to leave excess polyelectrolyte in suspension. This
excess polyelectrolyte lowered the composite density due to depletion flocculation.
Sintered densities of 10 vol% Sij/Ale3 composites slip cast fi'om a mixture of ethanol
and water were only 2.9mm; g/cm3 (75% theoretical density).

The fracture toughness of A1203 was remarkably improved with increasing whisker
content. The fracture toughness of 20 vol% Sij/Al203 composite was twice the fracture

toughness of the lmreinforced A1203 matrix.

To my mother and my daughter,

whose love and support is the source of my strength

iv

ACKNOWLEDGMENTS

I wish to express my sincere appreciation to Dr. M. J. Crimp, my advisor, for her
guidance throughout my doctoral studies and reviews of my dissertation. I would also
like to thank Dr. E. D. Case for his guidance in hardness and fracture toughness
measurements and sample preparation, his generosity for letting me use his Indentation
Tester and his encouragement. I would also like to thank Dr. H. Eick for his corrections
on my dissertation and his suggestion to have a committee meeting to discuss my
dissertation and to provide an opportlmity for me to get help from all the committee
members. I would also like to thank Dr. P. Kwon for being my committee member. I
appreciate all the committee members, for their precious time on reviewing my
dissertation, attending the comprehensive and oral examinations.

I also wish to thank the following group of people: D. Oppermann for providing
help in CIP and furnace equipment, B. Wilson for his help in BSA and EDX
measurements, Michelle for her proofreading the last part of my dissertation, and all the

help and friendship from other fellow graduate students.

TABLE OF CONTENTS

LIST OF TABLES ................................................................................. ix
LIST OF FIGURES ................................................................................ x
INTRODUCTION .................................................................................. 1
CHAPTER 1 LITERATURE REVIEW ........................................................ 6
1.1 Theory of Colloidal Stability .............................................................. 6
1.1.1 Electrostatic Repulsion .............................................................. 6

1.1.2 VanDerWaals Attraction ......................................................... 10

1.1.3 Stability Ratio ........................................................................ 11

1.1.4 Refinements of DLVO Theory ................................................... 13

1.2 Processing Techniques of Ceramic Matrix Composites ............................ 16
CHAPTER 2 EXPERIMENTAL PROCEDURES ........................................... 25
2.1 Materials .................................................................................... 25
2.2 Zeta Potential Measurement .............................................................. 25

2.3 Viscosity Measm'ement ................................................................... 28
2.4 Sedimentation Experiment ................................................................ 28

2.5 Stability Ratio Calculation ................................................................. 28
2.6 Dispersion and Slip Casting .............................................................. 29

2.7 Sintering ..................................................................................... 31

2.8 Microstructure and Composition ......................................................... 31

2.9 Density Measurement ..................................................................... 32
2.10 Hardness and Fracture Toughness Measurement .................................... 32

CHAPTER 3 RESULTS AND DISCUSSION34
3.1 Summary and Overview ofResearch Approaches.....................................34

3.2 Colloidal Stability and Suspension Parameters of Sij/Al203 in Aqueous and
Nonaqueous System .......................................................................... 37

3.2.1 Zeta Potential ........................................................................... 37

3.2.2 StabrlrtyRano43

3.2.3 Viscosity ............................................................................... 50
3.2.4 Sedimentation Bulk Density ......................................................... 50
3.2.5 Microstructure .......................................................................... 53
3.2.6 Conclusions ........................................................................... 53

3.3 Colloidal Processing of Sij/Al203 CompOSltesS7

3.3.1 Monolithic Alumina ................................................................ 57
3.3.2 Processing of Sij/Al203 Composites in Aqueous System ................... 59

3.3.3 Processing of SiCdAl203Composites in Nonaqueous System. . . . . . ....65
3.3.4 Discussion and New Approaches71

3.4 Densification of SiCJAl203 Composites through Improved Processing

Approaches .................................................................................. 72
3.4.1 Colloidal Processing Combined with Mechanical Methods.. ..................... 72
3.4.2 Sintering Aids ......................................................................... 78
3.4.3 Polyelectrolyte ........................................................................ 82
3.4.4 Discussion .............................................................................. 92

vii

3.5 Improved Properties of Sij/Al203 Composites ....................................... 93

3.5.1 Sintered Density ..................................................................... 93
3.5.2 Microstructure and EDX Analysis ................................................ 98
3.5.3 Hardness and Fracture Toughnessllo
CHAPTER 4 CONCLUSIONS ................................................................ 125
REFERENCES .................................................................................. 128

viii

LIST OF TABLES

Table 1 Composite processing and mechanical properties .................................. 20

Table 2 Major characteristics of Tokai TWS-100 SiC whisker, Sumitomo AKP-30
and AKP-SO alumina powder and Y203 powder [44]27

' Table 3 Stable pH range in various media. .. ..................................................... 49
Table 4 Compatibility of stable pH range for difi‘erent components in aqueous media
- . . ........................................................................................................................ 83

Table 5 Compatibility of stable pH range for difi‘erent components

in 99% ethanol and 1% electrolyte ................................................................. .83
Table 6 Comparison of polyelectrolyte addition to iep and C-potential for A1203 ......... 86
Table 7 Comparison of iso-electric point of A1203 between this research and

published data [26,30] ..................................................................................... 86
Table 8 EDX composition of 10 and 20 vol% Sij/Al203 composites ......................... 98

Table 9 Difi'erent grain sizes for the 5 and 20 vol% SiCw loading.........................105

LIST OF FIGURES

Figure 1 A schematic diagram of an agglomerate in an array of particles [4] . ..2

Figure 2 Defects originating from agglomerates caused by
a) agglomerate sintering faster than the matrix and

b) matrix sintering faster than the agglomerate [5] ................................. 3
Figure 3 Incompatible shrinkage between the matrix and dispersed whiskers [4]

............................................................................................................................ 4
Figure 4 Model of charged spherical particles in a liquid indicating

potential distribution <b(r), where r is distance [12] ................................. 8
Figure 5 Schematic of the interaction energy as a fimction of separation

between two particles in suspension [11] ............................................ 12
Figure 6 Model for treating the efi‘ect of adsorbed polymer on zeta-potential.

The model allows for the possibility of a shift In the shear plane as well as

a modified 1011 distribution In the difi‘use layer [25]... ... .....15
Figure 7 Plot of C versus pH showing the location of the iSO-electric mpoint (i. e. p. ...)

and the regions ofstability [29]... ... .........18
Figure 8 The properties of ceramic product [40] ............................................... 24
Figure 9 Relationship between difierent parts of the ceramic process [43]......... .......24
Figure 10 A flow chart of experimental procedures ........................................... 26

Figurell The processing approach of high performance ceramic composites.
Improved performance is achieved through controlled processing approaches.

*Density, Fracture Toughness, Hardness ........................................... 36
Figure 12 Zeta potential of A1203 versus pH at difi‘erent electrolyte strengths. . . ...38
Figure 13 Zeta potential of Sij versus pH at different electrolyte strengths. . . . . . .39

Figure 14 Zeta potential versus pH for Sij at varying ratios of ethanol to electrolyte
under 0.001N KN03 and 0.5 vol% solids content ................................ 41

Figure 15 Zeta potential versus pH for A1203 at varying ratios of ethanol to electrolyte
under 0.001N KNO3 and 0.5 vol% solids content ................................ 42

Figure 16 Stability ratio versus pH for SiC,./Al203 in 0.001N KN03 and
ethanol:electrolyte (0:1) .............................................................. 44

Figure 17 Stability ratio versus pH for Sij/Al203 in 0.001N 10103 and
ethanol:electrolyte (1 :4) .............................................................. 45

Figure 18 Stability ratio versus pH for Sij/Al203 in 0.001N KNO3 and
ethanol:electrolyte (1:1) ........................................................... ' ...46

Figure 19 Stability ratio versus pH for Sij/A1203 in 0.001N KNO3 and
ethanol:electrolyte (4:1) ............................................................... 47

Figure 20 Stability ratio versus pH for SiWM203 in 0.001N KN03 and
ethanol:electrolyte (99: 1) ............................................................. 48

Figure 21 Viscosity of 10 vol% SiC.../Al203 with 40 wt% solids content at varying ratios
of ethanol to electrolyte ............................................................... 51

Figure 22 Sedimentation bulk density versus pH for A1203 and Sij in electrolyte and
ethanol, respectively (0.001N KNO3) .............................................. 52

Figure 23 SEM photomicrographs of 10 vol% Sin/A1203 green body in a) electrolyte,
b) ethanol:electrolyte (1 :1), and c) ethanol:electrolyte (99:1) (0.001N KN03,
non-ball-milled, pH 4, no polyelectrolyte and sintering aids)....... . .. . .54 ,

Figure 24 SEM photomicrographs of 10 vol% SiWM203 partially sintered samples:
a) electrolyte, b) ethanol:electrolyte (1 :1) (0.001N KN03, non-ball-milled,
pH 4, no polyelectrolyte and sintering aids, sintered at 1650°C for 1h,

nitrogen atmosphere) ..................................................................................... 56
Figure 25 Colloidal processing approaches ..................................................... 58

Figure 26 SEM photomicrographs of green microstructures of 10 vol% Sij/Al203
composite processed from difl‘erent pHs: a) pH 2, b) pH 3, 0) pH 4, d) pH 8,
e) pH 10, and f) pH 11 (0.001N KN03, non-ball-milled, no polyelectrolyte
and sintering aids) ........................................................................................ .60

Figure 27 SEM photomicrographs of sintered microstructures: a) 5 vol% SictJAl203
and b) 10 vol% Sij/Al203. Sintered at 1650°C for 1h, nitrogen atmosphere
(0.001N KNO3, non-ball-milled, pH 4, no polyelectrolyte and sintering aids)
........................................................................................................................ 63

Figure 28 Green density versus whisker loading (0.001N KN03, non-ball-milled, pH 4,
no polyelectrolyte and sintering aids) ............................................................ 64

Figure 29 SEM photomicrographs showing the green microstructures of
10 vol% Sij/Al203 composites: a) 100% electrolyte,
b) ethanol:electrolyte (1:4), c)ethanol:e1ectrolyte (1:1),
(1) ethanol:electrolyte (4: 1) and e) ethanol:electrolyte (99:1)
(0.001N 10103, non-ball-milled, pH 4, no polyelectrolyte and sintering aids)
........................................................................................................................ 66

Figure 30 SEM photomicrographs of sintered microstructures of 10 vol% Sij/Al203.
a) electrolyte and b) ethanol:electrolyte (1:1) (0.001N KNO3, non-ball-milled,
pH 4, no polyelectrolyte and sintering aids, sintered at 1650°C for 1h, N2)
...................................................................................................................... 69

Figure 31 Green density variation of 10 vol% SiWAl2O3 at difierent ratios of ethanol to
electrolyte (0.001N 10103, non-ball-milled, pH 4, no polyelectrolyte and
sintering aids) ................................................................................................. 70

Figure 32 Green density of 5 and 10 vol% SiC.,/Al203 as a fimction of ball milling time
(0.001N KNO3, ball-milled, pH 4, no polyelectrolyte and sintering aids)
....................................................................................................................... 73

Figure 33 Sintered density of 5 and 10 vol% Sij/A1203 as a function of ball milling
hour (0.001N 10103, ball-milled, pH 4, no polyelectrolyte and sintering aids,
sintered at 1600°C for 1h, nitrogen atmosphere) ............................................ 74

Figure 34 Green density Of 5 and 10 vol% Sij/Al203 as a function of solids loading
(0.001N 10103, non-ball-milled, pH 4, no polyelectrolyte and sintering aids)
........................................................................................................................ 76

Figure 35 Sintered density of 5 and 10 vol% SiWM2O3 as a flmction of solids loading
(0.001N KNO3, non-ball-milled, pH 4, no polyelectrolyte and sintering aids,
Sintered at 1600°C for 1h, nitrogen atmosphere) ........................................... 77

Figure 36 The A1203-Y2O3-Si02 phase diagram [63] 79

Figure 37 Green density of 5 and 10 vol% Sij/Al203 as a function of Y203 content
(0.001N KNO3, ball-milled, pH 11, 2.5 vol% polyelectrolyte) ...................... 80

Figure 38 Sintered density of 5 and 10 vol% Sij/A1203 as a flmction of Y203 content
(0.001N 10103, ball-milled, pH 11, 2.5 vol% polyelectrolyte, sintered at
1600°C for 1h, nitrogen atmosphere) ............................................................. 81

Figure 39 Zeta potential for A1203 versus pH at difl‘erent polyelectrolyte concentrations
(0.001N 10103) .............................................................................................. 85

xii

Figure 40 Green density of 10 and 20 vol% Sij/Al203 as a function of polyelectrolyte
concentration (0.001N 10103, ball-milled, pH 11, 0.5 wt% MgO and
2 wt% Y203) .................................................................................................. 88

Figure 41 Sintered density of 10 and 20 vol% Sij/Al203 as a function of polyelectrolyte
concentration (0.001N 10103, ball-milled, pH 11, 0.5 wt% MgO and
2 wt% Y203, sintered at 1600°C for 1h, nitrogen atmosphere) ..................... 89

Figure 42 Green density of 10 and 20 vol% Sij/Al203 as a flmction of ball milling hour
(0.001N 10103, ball-milled, pH 11, 2.5 vol% polyelectrolyte, 0.5 wt% MgO
and 2 wt% Y203) ........................................................................................... 90

Figure 43 Sintered density of 10 and 20 vol% SiCo/Al203 as a function of ball milling
hom' (0.001N 10103, ball-milled, pH 11, 2.5 vol% polyelectrolyte,
0.5 wt% MgO and 2 wt% Y203, sintered at 1600°C for 1h, nitrogen
atmosphere) ...................................................................................................... 91

Figure 44 Green and sintered density of 10 vol% Sij/Al203 as a function of pH
(0.001N 10103, ball-milled, 2.5 vol% polyelectrolyte, 0.5 wt% MgO and

2 wt% Y203, sintered at 1600°C for 1h, nitrogen atmosphere).
........................................................................................................................ 94

Figure 45 Comparison of sintered density of Sij/Al203 composites between
this research and other published results [3,54,73] . . . ...95

Figure 46 The sinterd density versus the green density (5-20 vol% Sij/Al203,
0.001N 10103, ball-milled, pH 11, 2.5 vol% polyelectrolyte, 0.5 wt% MgO
and 2 wt% Y203, sintered at 1600°C for 1h, nitrogen atmosphere) ............. 97

Figure 47 EDX composition spectra: a) 10 vol% Sij/A12O3, b) 20 vol% SijlAl203
(0.001N 10103, ball-milled, pH 11, 2.5 vol% polyelectrolyte, 0.5 wt% MgO
and 2 wt% Y203, sintered at 1600°C for 1h, nitrogen atmosphere) .............. 99

Figure 48 SEM photomicrograph of the fractln'e surface of 5 vol% Sij/Al203
(0.001N KNO3, non-ball-milled, pH 4, no polyelectrolyte and sintering aids,
sintered at 1600°C for 1h, nitrogen atmosphere) .......................................... 101

Figure 49 SEM photomicrograph of fi'acture surface of 10 vol% SiC../Al203
(0.001N KNO3, ball-milled, pH 11, 2.5 vol% polyelectrolyte, 0.5 wt% MgO
and 2 wt% Y203, sintered at 1600°C for 2h, nitrogen atmosphere) ............ 102

Figure 50 SEM photomicrograph of fiacture surface of 20 vol% Sij/Al203
(0.001N KNO3, ball-milled, pH 11, 2.5 vol% polyelectrolyte,
0.5 wt% MgO and 2 wt% Y203, sintered at 1600°C for 4h, nitrogen flow)
..................................................................................................................... 103

xiii

Figure 51 SEM photomicrograph of fracture surface of 20 vol% Sij/Al203
(0.001N 10903, non-ball-milled, pH 11, 2.5 vol% polyelectrolyte,
0.5 wt% MgO and 2 wt% Y203, sintered at 1600°C for 4h, nitrogen
atmosphere) .................................................................................................. 1 04

Figure 52 SEM photomicrograph of the fracture surface of 5 vol% Sij/A1203
(0.001N KNO3, non-ball—milled, pH 4, no polyelectrolyte and sintering aids,
sintered at 1600°C for 1h, nitrogen atmosphere) ......................................... 106

Figure 53 SEM photomicrograph of the fi'acture surface of 20 vol% Sij/Al203
(0.001N KNO3, ball-milled, pH 11, 2.5 vol% polyelectrolyte, 0.5 wt% MgO
and 2 wt% Y203, sintered at 1600°C for 4h, nitrogen atmosphere) ............ 107

Figure 54 SEM photomicrograph of the fracture surface of 20 vol% SiCa/Al203
(0.001N 10103, ball-milled, pH 11, 2.5 vol% polyelectrolyte,
0.5 wt% MgO and 2 wt% Y203, sintered at 1600°C for 4h, nitrogen flow)

..................................................................................................................... 108
Figure 55 SEM photomicrograph of the fiacture surface of 20 vol% Sij/Al203

(0.001N 10903, non-ball-milled, pH 11, 2.5 vol% polyelectrolyte,

0.5 wt% MgO and 2 wt% Y203, sintered at 1600°C for 4h, nitrogen

atmosphere) .................................................................................................. 109

Figure 56 Krc of composites versus whisker loading showing the comparison between
this research and other published results [3, 74] .......................................... 111

Figure 57 Hardness of composites versus whisker loading (0.001N KNO3, ball-milled,
pH 11, 2.5 vol% polyelectrolyte, 0.5 wt% MgO and 2 wt% Y203, sintered at
1600°C for 1h, nitrogen atmosphere) ........................................................... 112

Figure 58 Comparison of Krc of 20 vol% SiCw Composites with ball milling and without
ball milling (0.001N 10903, pH 11, 2.5 vol% polyelectrolyte, 0.5 wt% MgO
and 2 wt% Y203, sintered at 1600°C for 1h, nitrogen atmosphere) ............ 113

Figure 59 Comparison of Hardness of 10 and 20 vol% Sij composites with ball milling
and without ball milling (0.001N 10903, pH 11, 2.5 vol% polyelectrolyte,
0.5 wt% MgO and 2 wt% Y203, sintered at 1600°C for 1h, nitrogen

atmosphere) ................................................................................................... 1 14

Figure 60 SEM photomicrographs of the fiacture surface of 10 vol% Sij/Al203
showing (a) long whisker protruding length and (b) deep whisker pullout holes
(0.001N KNO3, ball-milled, pH 11, 2.5 vol% polyelectrolyte, 0.5 wt% MgO
and 2 wt% Y203, sintered at 1600°C for 1h, nitrogen atmosphere) ............ 116

xiv

Figure 61 Hardness of 10 vol% SiC../Al203 as a flmction of pH (0.001N 10903,
ball-milled, no polyelectrolyte and sintering aids at pH 4 and
2.5 vol% polyelectrolyte and 2 wt% Y203 at pH 11, sintered at 1600°C for 1h,
nitrogen atmosphere) .................................................................................... l 17

Figure 62 Hardness of 10 vol% Sij/Al203 as a function of Y203 content
(0.001N 10903, ball-milled, pH 11, 2.5 vol% polyelectrolyte, sintered at
1600°C for 1h, nitrogen atmosphere) ........................................................... 118

Figure 63 Hardness of 10 vol% SiCa/A1203 as a function of density (0.001N KNO3,
ball-milled, pH 11, 2.5 vol% polyelectrolyte, 0.5 wt% MgO and 2 wt% Y203,
sintered at 1600°C for 1h, nitrogen atmosphere) ......................................... 119

Figure 64 Hardness of 10 vol% SiWAl2O3 as a flmction of ball milling time
(0.001N 10103, ball-milled, pH 4, no polyelectrolyte and sintering aids,
sintered at 1600°C for 1h, nitrogen atmosphere) .......................................... 120

Figure 65 Hardness of 10 vol% SiCo/Al203 as a flmction of polyelectrolyte
concentration (0.001N 10903, ball-milled, pH 11, 0.5 wt% MgO
and 2 wt% Y203, sintered at 1600°C for 1h, nitrogen atmosphere) ............. 122

Figure 66 Hardness of 10 vol% Sij/Al203 as a function of solids loading
(0.001N 10903, non-ball-milled, pH 4, no polyelectrolyte and sintering aids,
sintered at 1600°C for 1h, nitrogen atmosphere) .......................................... 123

Figure 67 Krc and hardness of sintered composites versus the green density of their green
bodies, respectively (5-20 vol% SiC..,/A1203, 0.001N 10903, ball-milled,
le 1, 2.5 vol% polyelectrolyte, 0.5 wt% MgO and 2 wt% Y203, sintered at
1600°C for 1h, nitrogen atmosphere) .......................................................... 124

INTRODUCTION

The brittleness of alumina ceramics limits their application as structural materials.
Reinforcement of alumina with SiC whiskers can improve the fracture toughness [1-3].
The most common consolidation method for ceramic powders is via dry pressing. Since
dry powders are naturally agglomerated, these agglomerates produce large, crack-like
voids during densification, which can be a major strength degrading flaw. The
agglomerates pack together during consolidation to produce compacts with difi'erentials
in packing density leading to poor densification.

Figure 1 is a schematic showing an agglomerate in an array of particles [4]. The
difl'erential shrinkage during sintering results in defects originating from agglomerates
(circumferential cracks and radial cracks in Figure 2). For whisker or fiber reinforced
composites, the incompatible sintering between the matrix and dispersed whiskers or
fibers is more serious (pore, crack and whisker bending in Figure 3). Since it is
considerably difficult to obtain homogeneous matrix packing between whiskers, there
will commonly be inhomogeneous shrinkage between whiskers resulting in bending
(exaggerated) of one or both whiskers [4].

Often combined with dry processing, dry milling is used to help mix and add various
organics such as binders. The dry milling process introduces contaminates, large inclu-
sions and whisker damage. The elimination of polymer binders is not only time consum-
ing (days) but can also produce disruptive phenomena. Conventional powder processing
also causes poorly distributed second phases and particle packing that encourages

abnormal grain growth. Furthermore, the resultant components that have been received

O O O O O O O

O O O O O O O

O O O O O O O

O O . O O 'PO.O 0
£093

0 O O O O

O O O O O O 0

Figure 1 A schematic diagram of an agglomerate in an array of particles [4].

matrix

agglomerate

Figure 2 Defects originating from agglomerates caused by
a) agglomerate sintering faster than the matrix and
b) matrix sintering faster than the agglomerate [5].

 

 

A B
A
\\
‘3‘ '5 \\a\§
i i it
.i i ,
i it 3.; If”
\ z":
\ °\ 10:

Figure 3 Incompatible shrinkage between the matrix and dispersed whiskers [4].

hot-pressed have anisotropic properties due to the partial planar whisker orientation
developed during the hot pressing [6].

The purpose of this research was to determine the optimum conditions for particle/
whisker codispersion, consolidation and sintering densification of optimized green
bodies. By colloidal processing, stable, binder free, homogeneous suspensions and green
bodies at various processing parameters (SiC whisker volume fraction, solids loading,
ball milling time, sintering aid content, polyelectrolyte content and p11) are produced.
These homogeneous green bodies are pressurelessly sintered at relatively low sintering
temperatures and short sintering times. The final microstructure and composition are
characterized. The final ceramic matrix composites exhibit improved mechanical

properties.

CHAPTER 1 LITERATURE REVIEW

1.1 Theory of Colloidal Stability
DLVO theory (developed by Deryagin and Landau [7] and, independently, by
Verwey and Overbeek [8,9]) is accepted as the basic theory of colloidal stability [10].
According to the DLVO theory, the interaction between colloidal particles is a

superposition of electrostatic repulsion and van der Waals attraction.

1.1.1 Electrostatic Repulsion

The sm'face charge required for stability can be the result of one of the following
four processes [10].

(a) Dissociation of surface groups. This is the charge-determining mechanism for
several oxides. Silica, for example, can acquire a negative charge by dissociation of sur-
face silanol groups.

(b) Adsorption of potential-determining ions. An example is adsorption of Ag+ and
1' ions on AgI, rendering the AgI particles positive or negative, respectively.

(c) Adsorption of ionized surfactants. Carbon black dispersions, for example, can
become negatively charged by adsorption of anionic surfactants.

(d) Isomorphic substitution. This mechanism is common in clay minerals, e.g., iso-
morphic substitution in the octahedral A1203 layers and the resultant charge-balancing by
exchangeable counterions.

Charged particles attract oppositely charged ions (i.e., counter-ions) around the

particles. The charged surface and the surrounding counter-ions form the electric

double layer. When the particle moves in the liquid, some ions “shear ofl” [11], and
the resultant potential between the sm'face and the shear layer plane is the zeta potential.
The zeta potential is identified with the value of the double-layer potential at the outer
edge of the Stem layer of fluid that remains attached to the particle during the relative
motion of the particles and liquid during electrophoresis (Figure 4). The region
extending outwards fi'om the Stern layer is termed the difiuse region, in which electro-
static forces are opposed by those of thermal agitation, and the potential decays approxi-
mately exponentially. In Figure 4, 02 represents the electrokinetic (zeta) potential and
(but, the Nemst (double-layer) potential, which is the potential at the surface of the
particle relative to a remote point in the liquid. The higher the zeta potential, the thicker
the double layer, and the more stable are the particles against coagulation.

1n the DLVO theory, the distribution in a difl’use double layer is assumed to obey the

Poisson-Boltzmann distribution [13,14]

2.
raven—.2222, carpi—1%) (1)

where ‘1’ is the electrostatic potential, 8,, and e, are the dielectric constants of water and

solvent, respectively, Z, is the valence of the ion i, or is the local concentration of ion 1, T
is the absolute temperature, e is elementary charge, and k is Boltzmann’s constant. Two
bOImdary conditions plus a reference point for the surface potential are required to solve

equation (1).

 

 

shear

diituse / surface
\ region — /
\ +
\ \ ~ — _.. I /
Stern
layer

Figure 4 Model of charged spherical particles in a liquid indicating
potential distribution <D(r), where r is distance [12].

Applying the Debye-Huckel approximation to equation (1) gives:
‘1’(x) = ‘1", exp(-Kx) (2)

where III: is the Debye screening length or the characteristic double—layer thickness and

is given by

s a kT “2
= ' ° 3

 

where ‘1’o is the surface potential, x the distance from surface, and co is the bulk ionic
concentration. The exponential decay of the potential requires low-potential conditions;
i.e., it is valid only sufibiently far from the surface that the potential approaches zero.
This solution is not valid near the particle surface where the potential is high.

Hogg, Healy and Fuerstenau [15] developed a theory for nonidentical particles. In
the HHF theory [16], the two difi‘erent particles are treated as two plates having a

constant potential, separated by a distance 2d with the bormdary conditions

‘1’ = We, asan and ‘11 = To, asx—)2d. (4)

The solution gives

‘1'0’ - it, cosh(2xd)
sinh(2xd)

 

‘1’ = w, cosh(loc) +[ ]sinhotx) . (5)

When two particles approach each other, double layer overlap occurs. Two cases
can be distinguished, and may be denoted as the constant potential and the constant
charge interactions [10].

In the first case, it is assumed that during the encounter of two colloidal particles the

surface potential remains constant, and as a result of double layer overlap, the surface
charge should decrease. In the second case it is assumed that during the encormter the
surface charge remains constant, in which case double layer overlap leads to an increase
in the surface potential ‘1", [10].

The free energy of the double layer increases [10] upon interaction. Work must be
performed to bring the particles closer together. The repulsive energy VR(d) is the work
required to bring the particle surfaces fi'om infinity to a distance d. The fi'ee energy
should be calculated as a function of d. If F(oo) represents the free energy at infinite
separation, VR follows as [10]

Vlt(d)= 2[1=(d)-F(<=<>)]- (6)

For two identical spherical particles at small potentials

 

V, .wjo’n(r..~<R-2«>) <7)

where a is the particle radius, and R is the distance between the centers of the particles.

For two nonidentical spherical particles, the repulsive energy is [17,18]

2‘1’ ‘1’ ..
Va = s,£,7t[fl—](‘Ps,2 +93% 0' 0’ hp +exP( 1a)] + 1n(1-°XP(-ZIOC))]' (8)

 

 

a, +0] T22 +910): 1-exp(-tcc)

1.1.2 Van Der Waals Attraction
The attractive forces existing between colloidal particles are called London-van der
Waals forces [19]. The attractive potential energy between two colloidal particles is large

enough to compete with double layer repulsion. The attractive energy is [11]

V. =-A.fl‘_’z_ . (9)
6x a, +aj

10

where A is the Hamaker constant, x is the separation distance of the two particles, at and
aj are the particle radii of particles i and j, respectively. The Hamaker constant can be

calculated fiom dielectric constant by Bleier’s method [20]

 

A, =113.7 (‘9' '1): (10)
(a, +1)3’2(e,. + 2)”2 ' *
The total interaction energy is
v1=vR+vA, (11)

The attractive force, VA, dominates at small separations. A primary minimum in the
interaction energy curve results at particle/particle contact. As particle separation
increases, double-layer repulsion dominates. The total interaction energy curve shows a
peak. This is the energy barrier of particle coagulation [11]. A secondary shallow
minimum develops at large distances, the attractive energy again dominates. Figure 5 is a
schematic of the interaction energy as a function of separation between two particles in
suspension [1 1].

If particles approach close to the primary minimum, they form hard agglomerates.

The secondary minimum also can result in coagulation, which forms soft agglomerates

[11].

1.1.3 Stability Ratio
The stability ratio, W, measures the efl‘ectiveness of the potential barrier in prevent-
ing the particles from coagulating:

_ Number of collisions between particles
Number of collisions that result in coagulation

 

(12)

ll

 

 

 

. ‘ ‘ l
', Double-Layer i
/R¢P“l3|°fl (Va) ' p _ I

/Enasy Burier (Ea)

 

 

 
 
  

Total Energy, VT (V1- - VR + VA)

 

D .

Interaction Energy

 

Minimum

Primary Minimum

 

 

 

Figure 5 Schematic of the interaction energy as a function of separation

between two particles in suspension [1 1].

12

For two nonidentical particles of radii a, and a,- and a distance of separation x, the Wij (sta-

bility ratio for stability between particle i and j) is [21]:

"° ( V dx
Wij=(a,.+aj) I exalt-TX?) (13)

 

a,+aj

where VT=VA+VR.

Hogg, Healy and Fuerstenau [22] developed a variable W1- (total stability ratio) to

describe the total stability of the system:

_ n2 (l—n)2 2n(1—n) 4
W”[W..+ We i We ] (14)

where n is the proportion of the type 1 particles, and (l-n) is the proportion of the type 2

particles.

1.1.4 Refinements of DLVO Theory

The general equation describing the interparticle forces consists of attractive and
repulsive terms [23]:
V, = VA (van der Waals) + VR (electrostatic) + V3 (steric) + Vs (salvation) . (15)

Surfactants with high molecular weight promote steric stabilization. The solvation
forces are short range hydration and hydrophobic forces. The solvation and steric forces
can be attractive or repulsive in nature. Polyelectrolytes are added to suspensions to
promote steric stabilization. In this case, the particle surface potential, ‘15, is replaced by
the zeta potential C [25]. Figure 6 shows the shift in the shear plane after absorbing
polyelectrolyte on the particle surface. A parameter [3 measures how the presence of the

Polyelectrolyte affects the free energy of the ions. The additional free energy is

13

incorporated into the Boltzmann equation.

 

The relative C—potential (Z) is expressed as Z = if , where q, is the Q-potential in
the presence of polyelectrolyte and Co is the C—potentialowithout the presence of a
polyelectrolyte.

Using the appropriate boundary conditions, the relative C-potential is

calculated as [24]:

 

(fl
Lorna) . (16)

In most cases, it was found that Z>1[24]. The experiments showed [24] that polymer
adsorption can lead to large increases in the absolute value of the C—potential of biological
(red blood cell), organic (polystyrene latex) and inorganic (Si02) surfaces. Brooks [25]
shows that even if the adsorption of the polymer leaves the surface charge and the shear
plane unafl‘ected, an increase in [CI is predicted due to the change in the ion distribution in
the difl’use double layer. The change is produced chiefly by the influence of the excluded
volume of the polymer on the free energy of the double-layer ions. The overall efl‘ect, if
the surface charge is kept constant, is to increase the extension of the double layer and
hence to increase the surface potential of the underlying particle (see Figure 6, which
incorporates the possibility that the shear plane may be shifted away fi-om the particle

surface by the adsorbed polymer without causing a decrease in ICI [24]).

14

Rigid Free
6'0

I
{am-pl sin-o
|

  
     

9! ('1 to presence at
«norm layer

4,”) 'llflbm of
«sored low

  

 

Figure 6 Model for treating the effect of adsorbed polymer on zeta-potential.
The model allows for the possibility of a shifi in the shear plane as
well as a modified ion distribution in the diffuse layer [24].

1.2 Processing Techniques of Ceramic Matrix Composites

Colloidal processing involves the manipulation of interparticle forces. The only
possible way of avoiding coagulation is by developing repulsive forces between particles,
either V,(electrostatic) or V,(steric). This can be done by pH control and changing the
electrolyte or its concentration in the first case and by using macromolecules (polyelec-
trolytes) which may be adsorbed onto the particle surface in the second. Colloidal
methods are used to break apart weakly bonded agglomerates, eliminate strongly bonded
agglomerates, eliminate inorganic and organic inclusions greater than a given size,
homogeneously mix two or more powders, and consolidate powders to high bulk
densities.

In the electrostatic approach, ions or charged molecules are attracted to, or
dissociated fi'om the particle surfaces to produce a system of similarly charged particles.
When the repulsive double-layer electrostatic forces between the particles are greater
than the attractive van der Waals forces, the particles repel each other to produce a
dispersed system. The net interparticle force (in aqueous solutions) can be altered by
changing the type and concentration of the ionic species, the pH and by addition of
specific adsorbing organic or inorganic compounds or polyelectrolytes [26]. When the
particle charge approaches zero, the particles flocculate. The zeta potential provides a
convenient experimental measure of such forces. ESA (Electrokinetic Sonic Amplitude)
is the pressure amplitude generated by the colloid particle per electric field strength.
O’Brien’s theory [27,28] for electro-acoustic effects in a dilute suspension of spherical
particles is used to calculate the suspension zeta potential fiom the measured ESA.

By understanding the zeta potential to pH relationship, the suspension properties of

16

individual ceramic suspensions may be manipulated to give stable, multi-component
composite suspensions. Figure 7 is an example of how electrophoresis data can be used
to indicate suspension dispersion characteristics for Single component materials [29].

The graph shows how the zeta potential varies with pH and also illustrates the regions in

which the suspension undergoes flocculation (lmstable, I; I <2va) and dispersion (stable,

I§|>25mv). The state of agglomeration of the primary particles in the liquid will depend

on the stability of the dispersion against coagulation.

The critical coagulation concentration is the concentration of the indifl'erent electro-
lyte at which the potential energy barrier opposing coagulation is zero. Therefore,
electrolyte concentrations less than the critical coagulation concentration should be used.

With the steric approach, bifunctional macromolecules attach themselves to the
particles. The macromolecular additives are usually completely soluble in the dispersing
fluid, but are designed with certain fimctional groups to bind them to the particles. The
particles then repel one another once their separation distance is less than the radius of
gyration of the macromolecule.

Polyelectrolytes are multi-charged macromolecules that may absorb on or anchor to
surfaces of ceramic particles in a polar (often aqueous) medium. Observations suggest
that certain polyelectrolytes can be most eflicient in dispersing a wide range of ceramic
powders. It is known that they can produce an electrostatic double layer surface charge,
thereby opposing the attractive van der Waals forces in addition to ofi‘ering the possibility
of steric interaction. J. Cesarano 111 and 1. A. Aksay [30] concluded that once enough
polyelectrolyte has been adsorbed, there is a stabilization effect due solely to the

polyelectrolyte, even in the absence of appreciable repulsion due to electrostatics.

17

STABLE FLOCCUL ATED 5 TA 8 LE

i e.p.

C(mV)

 

I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
l
I
I
I
I

Figure 7 Plot of Q versus pH showing the location of the iso-electric point (i.e.p.)
and the regions of stability [29].

18

The interparticle forces produced by polyelectrolytes may be controlled by altering pH,
ionic strength, temperature, molecular weight and concentration.

Most processing of whisker composites has been made by hot pressing [1-2, 31-34],
see Table 1. Skeletal networks of SiC whiskers prevent the development of full
densification. The methods of using high pressure have achieved high densities [32].
Progress has also been made in the area of pressureless sintering. Tiegs and Becher [3]
(Table 1) have shown that A1203 and 16 vol% SIC“, can be sintered to 95% of theoretical
density at l700—1850°C, using 0.5 wt% MgO and 2 wt% Y203 (for liquid phase Sintering)
to give reasonable mechanical properties. However, higher loadings of whiskers limited
achievable denSities in pressureless sintering.

The mechanical properties of A1203-SiC whisker composites by pressureless
sintering were not as good as those for hot pressing specimens because the densification
is inhibited as a result of whisker interference with particle rearrangement and composite
shrinkage.

Slurries were milled to reduce the aspect ratio of SiC whisker. The aspect ratio was
reduced to improve the density by improving the particle/whisker packing for increased
green density and enhanced the ability of the whiskers to rearrange themselves during
sintering. Increasing the amormt of liquid phase present during sintering also improves
densification by aiding whisker rearrangement. However, as the amormt of liquid phase
reaches about 30%, no further improvement in densification is observed because of
decomposition of the liquid phases [3].

Porter, Lange and Chokshi [32] fabricated A1203/Sij composites by slip casting

and pressureless sintering. Pressureless sintering was discontinued because slip cast

19

Tablel Composite processing and mechanical properties.

 

 

 

 

 

 

 
 
  

 

 

   
 

 

 

 

 

 

 

Materials Fabrication Green IFinal density Fracture Flexure Reference
methods and density (% (% of toughness strength
conditions of theoretical Krc (MPa)
theoretical density) 3.11110
density) M )
203+ 1850- '“ 95% 9 00 echer and
0- 19000c, ei
0V°1% 0-60Pa, 1-2]
ij acuum.
203+10 P 140MPa 7-59% 10vol%Sij 7 330 iegs,
20vol% 420MPa 91-96% echer and
iCw inter 1700- 20vol%Sij ' ard
M80 1850°C Ar 66-89% 3.35]
Y203
203 31 .2MPa F" 97-100% 6.8-9.5 350-650 omeny,
30vol% 1800- aughn and
ij 1950°C, Ar erber [31]
E203... ssure 59-65% 99.5% 3.6-4.6 391-652 E2“,
-15vol% tration ge and
iCw 24MPa okshi
1500°C [32]
ili-1203+ IHP 22MPa 6 >99% 3.8-10 500-620 Smith,
0-30vol°/o 1827°C, Singh and
Sij Ar Sgitergood
A1203+ CIP 250MPa 40-42% 1680°C: "' l" u and
0:30vol% intenng of A1203: 93%
5‘“ wder at 1800°C: 35]
o 10-20vol%
1800 c, Sij: 98%
° 30vol%Sij
[86%

 

 

 

 

 

 

20

 

Table 1 (cont’d).

 

 

 

 

 

 

 

 

 

Materials Fabrication Green [Final density Fracture Flexlne Reference
methods and density (% (% of toughness strength
conditions of theoretical Krc (MPa)
theoretical density) a-mm
may) (MP )
A1203+ lip casting, 67% 5vol%Sij: * * Sacks,
Eamon/r trifugal 92% and
ij ' g, 15vol%Sij ojas [37]
intering at [88%
1600°C, N2 30vol%Sij
%
A1203+ 1P, sinter F" 97% 5.5 414 ACMC
9.1vol% [34]
lsij
03+ inter-i- 100% 6.7 541 ACMC
9.3vol% [34]
.i_C.!
#203 65.8% 99% "' F“ Cesarano
1111 and
Aksay
[30]

 

 

 

 

 

 

 

 

 

Note: HP: hot press, CP: cold press, 1?: isopress, CIP: cold isostatic press,
HIP: hot isostatic press, ACMC: Advanced Composite Materials Corporation
* No data available.

21

components with a 15 wt% whisker loading sintered at 1550°C in vacuum for 30 min
with no applied pressure, failed to densify.

. Smith [3 3] at Argonne National Laboratory reported that well mixed and nearly
fully dense Sij/Al203 composites were fabricated by wet blending the constituents and
uniaxially hot-pressing the resulting powder.

Hu and Rahaman [36] at the University of Missouri made A1203/SiC whisker
composites by free sintering of coated powder. SiC whiskers were coated with a thick
cladding of fine-grained A1203 powder by controlled heterogeneous precipitation in a
concentrated suspension of whiskers. After calcination, a loose powder consisting of SiC
whiskers coated with A1203 was obtained. The sinterability of the composites formed
from the coated 'whiskers was significantly higher than that for similar composites formed
from mechanically mixed powders. Increasing the whisker aspect ratio produced a
decrease in the sinterability of the composites formed from the coated powders. Doping
with 250ppm Mg to control the grain growth ofthe A1203 matrix led to an increase in the
sinterability of the composites.

Sacks, Lee and Rojas [37] at the University of Florida produced homogeneous
A1203/Sij composites by suspension processing. By optimizing the conditions for
particle/whisker codispersion, castable suspensions could be prepared at total solids
concentration of 50 vol%. Green bodies with a high relative density (~66% TD) were
obtained with SiC whisker contents in the range of 5 to 30 vol%. Although densification
was severely inhibited by the SiC whiskers, significantly higher sintered densities were
Obtained by suspension processing as compared to dry processing.

Cesarano III and Aksay [30] studied the processing of highly concentrated

22

.- I '
[03'

11nd:

adsc

for:

hot

M

3h

(65-70 vol%) aqueous OI-Al203 suspensions stabilized with polyelectrolytes. By
understanding the chemistries of the polyelectrolyte and particle surface, polyelectrolyte
adsorption behavior and polyelectrolyte rheological efl‘ects, the viscosity and rheology
for all solids levels may be controlled. Their results Show that processing from highly
concentration polyelectrolyte-stabilized suspensions leads to higher consolidated
densities (100% Of theoretical density) and these benefits carry over in the form of
reduced Sintering temperatures. :-
Cesarano 111 and Aksay’s studies [30] show the potential for colloidal processing
techniques. The sintering temperature for composites decreased by achieving

homogeneous microstructure in consolidated states compared to conventional processes. 1

 

Difl‘erent ways of fabricating the A1203/Sij composites have been tried. Besides the
above-mentioned methods, microwave sintering, fieeze-drying, sol-gel [38], and
whiskers coated by amorphous silica [39] have been used to improve processing.

Figure 8 indicates that the properties of ceramic products are dependent on product
microstructure and ceramic processing [40]. Advances [41,42] from two extremes of the
fabricating process (powder, right at the beginning, and sintering, the last operation) have

demonstrated the central importance of the green compact [43] (Figure 9).

23

 

[erratum cowosmou

 

 

~ A

[cerium Pnocessm

utmosmicrune ]/

 

 

 

 

 

(INTRINSIC I

 

: "09531155

 

 

 

Figure 8 The properties of ceramic product [40].

 

 

 

0mm

 

 

POWDER A
' PROGSS

 

FORMING

 

 

 

\...../'

CONMCT

I

HICROSTRWTURAL .
EVOLUTOI ON
SINTERING

 

 

 

 

 

 

Figure 9 Relationship between different parts of the ceramic process [43].

24

CHAPTER 2 EXPERIMENTAL PROCEDURES

A flow chart of experimental procedures is summarized in Figure 10.

2.1 Materials
The materials used in this study were a-Al203 powder (Sumitomo, AKP-30 and

AKP-50) and B-SiC whisker (T okai, TWS-100). Major characteristics of Tokai TWS-
100 SiC whisker, Sumitomo AKP-30 and AKP-SO alumina powder and Y203 powder as
supplied by the manufacturers [44] are listed in Table 2. For SiC whiskers, the length
and diameter are 5-15 pm and 0.3-0.6pm, respectively, giving an aspect ratio of 10 to 40.
The AKP-30 and AKP-SO A1203 powder diameters are 0.3-0.5um and 0.1-0.3 um,
respectively. Y203 powder (Rhone-Poulenc Basic Chemicals Company) and MgO

powder (Rhone-Poulenc Basic Chemicals Company) are used as sintering aids. The

pinity of MgO as labeled by manufacturer is 99.9 wt%.

2.2 Zeta Potential Measurement
Matec Applied Sciences BSA-8000 System was used to measure the ESA potential
of particle suspensions where the solids loading of each suspension was 0.5 vol%. Each
type of powder was suspended in 220ml deionized water of varying electrolyte
concentration (0.001-0.1N KNO3), polyelectrolyte (0.03-0.5 vol% ammonium salt of
polymeric carboxylic acid), or varying ethanol/electrolyte ratios. The pH was varied for
this system by titration of 1N 111903 and 1N KOH. Suspensions were dispersed using an

ultrasonic probe.

25

 

Materials and their colloidal characteristics
(ESA measurement. stability ratio calculation,
viscosity measurement and sedimentation experiment)

 

l

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Particle Ivvhisker codispersion
(control electrolyte, polyelectrolyte, solid loading)
1
pH Is adjusted under constant stirring
l
L Ultrasonic agitation j
1
Bali mll
(reducing whisker aspect ratio and breaking agglomerates)
l
I Ultrasonic agitation ]
l
I Slip cast ]
J
I Cold isostatiealiy press I

 

1

Green body microstructure(SEM)
Green density measurement

l
I Sinter I

l

Composite characterization
SEWEDX(microstructure and composition)
Density measurement

l

[Hardness and Fracture Toughness measurement]

 

 

 

 

 

 

 

 

 

 

Figure 10 A flow chart of experimental procedures.

26

 

 

Table 2 Major characteristics of Tokai TWS-100 SiC whisker, Sumitomo AKP-30,
AKP-SO alumina powder and Y203 powder [44]

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Grade TWS-100 AKP-30 AKP-SO Y203 I
crystal Form B a a —
Diameter (um) 0.3-0.6 0.3-0.5 0.1-0.3 1-3
Length (pm) 5-15 - - -
Aspect Ratio 10-40 - - -
Density (g/cm’) 3.20 3.98 3.98 5.01
Loose Bulk Density (g/cm3) 0.06-0.12 0.7-1.0 oou -
Pack Bulk Density (g/cm3) - 1.1-1.5 0.9-1.3 -
Specific Surface Area (xiii/g) 2-4 6-10 9-15 -
Purity (wt%) 99 >99.99 >99.995 99.99
Particulate Content (wt%) <1 - - -
Coeficient of Thermal Expansion 5.0 - - -
from RT to 1400°C (x10’6/°C)
Implnities(wt%)
Si02 0.5 - - -
Si - 0.004 0.0025 <0.005
Na - 0.001 0.001 -
Ca 0.05 - - <0.005
Mg 0.02 0.001 0.001 <0.005
Fe 0.05 0.002 0.002 <0.005
Cr 0.05 - - -
Co 0.05 - - -
A1 0.08 - - <0.005
Cu - 0.001 0.001 -

 

 

27

2.3 Viscosity Measmement
Viscosity measurements of particle suspensions were made as a function of pH
using a Brookfield Digital Viscometer, Model DV-II. The #3 spindle, 100mm, and
0.001N K1903 electrolyte solution were used for all the viscosity measurements. The
accuracy of the measurements is within 5%. Measurements were then repeated for

suspensions of varying ethanol/electrolyte ratios.

2.4 Sedimentation Experiment
Sedimentation densities for pure A1203 and pure Sij were determined for 5-10
vol% solids content suspensions in 0.001N 10903 electrolyte. The suspensions were
ultrasonically dispersed and the pH adjusted by addition of either acid (HNO3) or base
(KOH). The test tubes were then covered and the suspension allowed to settle until the
sedimentation height reached stable value. The sedimentation volumes were calculated

from the sediment height and used to determine the sedimentation densities.

2.5 Stability Ratio Calculation
From the zeta potential data as a function of pH, the stability ratios for homo- and
W011. WSic-Sic (811d Wmosmm) 811d wSiC-A1203r ICSPCCfiVCIYr were 931'
culated using Stable Suspension (c)[l6,45]. The relative volume fraction of components
was 0.5, the electrolyte concentration was 0.001N 10103, and the Hamaker constant was

calculated from the dielectric constant using a method developed by Bleier [20].

28

2.6 Dispersion and Slip Casting

From ESA results and stability ratio calculations, diflerent pH regions were chosen
to process the suspensions. Trial experiments demonstrated that the unstable slurry had
the consistency of ice cream and failed to cast. Suspensions prepared at pH 4 are stable
because the stability ratio (W) is greater than 1025 for all the possible interactions of
A1203/A1203, A1203/Sij and Sij/Sij. pH 11 was also chosen for processing com-
posites with sintering aids because Y203 dissolves at pH<7. Therefore, pH 4 and pH 11
were used for processing composites.

Individual suspensions were slowly mixed using a magnetic stirrer. After the solids
content adjustment to 35-50 vol%, the composite suspension was homogenized by ultra-
sonic agitation (Branson Sonifier 250). The 120W output and 20kI-Iz ultrasonic waves
were applied to suspensions for 20 minutes. An ultrasonic probe was embedded directly
in the slmry. To prevent overheating, the suspension containers were cooled in a cold
water bath.

A1203, sintering aids (Y 203 and MgO) and SiC whisker were dispersed and mixed in
electrolyte or a mixture of ethanol and electrolyte. Some suspensions had varying
concentrations of a polyelectrolyte (0.03-0.5 vol% ammonium salt of polymeric
carboxylic acid). The pH was adjusted by additions of 111903 or KOH.

By using a combination of pH adjustment, electrolyte, polyelectrolyte, and
ultrasonic agitation, A1203 particles, SiC whiskers and sintering aids (Y 203 and MgO)
were codispersed at a relatively high overall solids concentration (50 vol%) while
maintaining the relatively low viscosities desired for casting.

Slurries were ball-milled using A1203 grinding cylinders, which were 12mm in

29

plan ll
moved

mdcmi

‘39 d

diameter and 15mm in height. The grinding cylinders and Slurries were filled in rubber-
lined polyethylene jars. After 25-50 hours, the ball milled Slurries were ultrasonicated for
20 minutes before slip casting. Slips were poured into porous plaster molds where the
liquid was absorbed by capillary action into the mold leaving a layer of ceramic against
the mold wall.

Porous molds were made from No. 1 Plaster powder fiom the American Gypsmn
Company. Plaster and water were mixed in a ratio of 100g plaster/75ml water. The
mixed plaster slm'ry was poured into a plastic weighing boat and allowed to thicken for
25-30 minutes. A negative mold was inserted into the thickened plaster slurry. The
plaster mold was air-dried for 24 hours. After the plaster mold dried, the plastic boat was
removed and the negative mold was pulled out. Plaster molds were then put in clamps
and cut into two halves by a hand saw. The split molds facilitate easier release of the
ceramic green specimen. Before slip casting, the two halves of the plaster mold were
held in place by elastic bands.

During slip casting, slurry was periodically added until the desired wall thickness
formed. Solid casting was applied by leaving suficient slurry in the mold to get solid
green body. The green body was left in the mold for 36 hours. After further air drying,
samples were ground on 400 grit abrasive paper to smooth the surfaces. Samples were
heated to 100°C and held at this temperature for two hours to remove any residual
moisture. The weight and the dimensions of the green bodies were then measured by
electronic balance and calipers, respectively. The green density was obtained fi-om the
weight and dimensions of the samples. Five specimens were measured for each green

density data point. The green compacts were subsequently vacuurn-sealed into double

30

layered soft rubber bags and cold isostatically pressed at 310MPa (Iso-Spectrurn, Inc.).

2.7 Sintering
A high temperatme graphite-element furnace (Model 1000-4560, Thermal Technol-

ogy Inc.) under flowing nitrogen at 1600-1700°C was used for pressureless sintering the
composites. The samples were covered with A1203 powder in a closed graphite crucible.
The temperature was increased at 10°C/minute from room temperature to 1000°C, then
5°C/minute and 2°C/minute from 1000 to 1450°C and 1450°C to the final sintering
temperature, respectively. The samples were then held at the final sintering temperature

for 30 minutes. Samples were cooled inside the furnace at a cooling rate of 20°C/minute

toroomtemperature.

2.8 Microstructure and Composition

Green compacts and sintered composites were characterized using Scanning
Electron Microscopy and Energy dispersive X-ray spectroscopy to examine their
structure and composition. Fracture surfaces of green compacts and sintered composites
were gold coated for three minutes and examined in the Hitachi S-2500C Scanning
Electron Microscope (SEM) at an accelerating voltage of 15-20kV and in an
Environmental Scanning Electron MicrOSCOpe (ESEM) at an accelerating voltage of 10-
15kV. Energy dispersive X-ray spectroscopy analysis was performed using an X-ray
detector in the ESEM to measure the composition of the samples. An elemental dot-

mapping technique was used to distinguish the whisker composition (at%) distribution.

31

can a:

rabid

2.9 Density Measurement

Bulk densities and apparent porosities were measured by the Archimedes
displacement method using deionized water as the immersion medium. Relative densities
were determined by dividing the bulk density by the theoretical densities. Because there
were no data for the glassy phase, the theoretical density were calculated fi'om each
component’s density and its wt% by a superposition method.

Samples were boiled in deionized water for one hour to release any air bubbles
present at the sample surface and open pores. Three specimens were measured for each

sintered density data point.

2.10 Hardness and Fracture Toughness Measmement

Vickers hardness and fiacture toughness determinations used the indentation tech-
nique of polished specimens at a load of 5kg. Different loads were tried to find the most
appropriate load. 1kg was not enough to make a penny-like crack. 5kg turned out to be
the appropriate load. Sintered samples were cut using an Accutom-S high speed
diamond wheel cutting machine, ground using SiC grinding papers (400 and 600 grit) and
then polished using diamond pastes (10, 6, and 1 pm). Indentation measurements were
made on a Vickers Hardness Tester. Five impressions were made on each specimen and
the lengths of the diagonals and the radial cracks produced by indentation were measured
using the built-in optical microscope.

The Vickers hardness was calculated using the applied load and the length of the

32

halfdiagonal of the indentation with the following formula [46, 47]:

H=o.47P/a2 (17)
where H is Vickers hardness, P is the applied load and a is the length of the half-diagonal
indentation.

A dimensional analysis of indentation fi'acture has shown that the indentation crack

length, c, should be related to the impression radius a by [48]:

K, (<1>/H)aé = F, (c/ a)F2 (v, ,u,(Ry /a)) (18)
where Kc is fracture toughness, H is hardness, (I) the constraint factor (about 3), v Pois-
son’s ratio, u the friction coemcient between the indenter and the material, R3' the plastic
zone radius, and F1 and F2 indicated that c/a depends strongly on Kc (1)/Ham, but is
approximately independent of F2 [48]. From Evans and Charles’s simplified results
[48,49] based on theoretical assumptions and material characteristics Singh et a1. [50]
assessed fracture toughness of whisker reinforced Si3N4 composites by simplifying
equation (1 8) using the following relation:

2 3
I — .—
— E 5 c 2

where c is one-half the median crack length and E the elastic modulus. Equation (19)
was chosen to calculate fracture toughness in this study because this equation is
specifically derived fiom the whisker reinforced composites. The value of E was taken

fi'om the reference [51].

33

Inc its
:lclial pic
.21 ch:
3ng Al;

Figure

CHAPTER 3 RESULTS AND DISCUSSION

3.1 Summary and Overview of Research Approaches

The research of this dissertation has two goals. The first is an in depth study of the
colloidal properties and suspension parameters of the SiCa/Al203 system. Based on the
colloidal chemistry, fundamental theories, and recent developments, the colloidal system
of SiW A1203 is fully examined and discussed. The second is to obtain high
performance ceramic matrix composites through improved processing approaches.

Figure 11 illustrates a processing approach to produce high performance ceramic
composites. In a step by step method, it was established how the processing must be con-
trolled. There are four aspects to the concept of colloidal processing and improved pro-
cessing approaches: 1) colloidal processing was used to control surface chemistry and
suspension stability after a systematic investigation of colloidal interactions of the
SiCa/A1203 system; 2) ball milling served to control the aspect ratio of the SiC whiskers
and to break down agglomerates; 3) sintering aids were added to relax the back stresses
imposed by the SiC whisker networks to enhance the densification of these composites;
and 4) a polyelectrolyte was used to control the surface chemistry and to stabilize the
mixed SijlAl2O3 suspension and sintering aids, Y203 and MgO. By manipulating these
four processing steps it was possible to tailor-make composite microstructure and
alleviate whisker network efi‘ects on sintering to achieve high performance which means
high density, fracture toughness and hardness.

Details of the mechanisms of non-aqueous suspensions or aqueous suspensions with

SPCCial addition of polyelectrolyte are less known than those for general aqueous

34

amnion
tiresome
Tie second
tabbed l
is lowert
iii-in me
it. dead
finality l
reticular l
Nii‘electrc

P013010:

suspensions. Consequently, the first effort was to explore A1203 and Sij in mixtures of
electrolyte and ethanol and the aqueous system with the addition of a polyelectrolyte.
The second step has been the development of Sij/Al203 where colloidal processing is
combined with ball milling to form uniform deagglomerated green bodies. Ball milling
also lowered the aspect ratio of the SiC whiskers, which serves to decrease the number of
whisker networks. To eliminate or minimize the impact of defects or problems related to
Sij densification, sintering aids were added to the system. Sintering aids create gain
bormdary liquid phases and relax the back stress from the SiC whisker networks. Of
particular importance in forming ceramic matrix composites (CMCS) involves
polyelectrolyte additives that are able to reverse the zeta potential or shift the iso-electric
point of ceramic powders. Polyelectrolyte creates a steric barrier that restabilizes the
suspension even after the addition of Y203, disrupting stable interactions in the Sij/
A1203 suspension system. Four difl‘erent methods were being explored and were com-
bined in this research. This processing research combines the use of colloidal surface
chemistry, mechanical methods, sintering aids and polyelectrolyte additives to achieve
control over compaction, consolidation, forming and densification of CMCs. SiC
whisker reinforced A1203 suspensions were slip cast using numerous processing
conditions (SiC whisker volume fiaction, solids loading, ball milling, sintering aids,
polyelectrolyte and pH). Green specimens were cold isostatically pressed and

pressureless sintered in a flowing nitrogen atmosphere.

35

 

Goal of this research

 

colloidal processing
+mechanical method
A +sintering aids
+polye1ectrolyte
-> control surface chemistry
of the mixed suspensions
and sintering aids

 

 

 

 

 

colloidal processing

+ mechanical method

+ sintering aids
+control densification

Performance‘

 

 

 

 

colloidal processing
+ mechanical method

_. control aspect ratio
break agglomerates

 

 

 

 

colloidal processing

_>control surface chemistry
and suspension stability

 

 

 

 

 

 

Loose Powdirs .
. & Whiskers Improved Processrng ___>

Figurell The processing approach of high performance ceramic composites.
Improved performance is achieved through controlled processing
approaches. *Density, Fracture Toughness, Hardness

36

:rc

“ii S"

\]l
x.

has 111“

V

3.2 Colloidal Stability and Suspension Parameters of Sij/Al203 in Aqueous
and Nonaqueous Systems

By manipulating the interparticle forces of ceramic colloidal systems, suspensions of
difiemnt dispersing abilities were obtained. The net interparticle force was altered by
changing the concentration of the ionic species, pH, and the solution. The zeta potentials
of single component suspensions of varying concentration of ionic species, pH and solu-
tion (electrolyte, ethanol, or a ratio of electrolyte and ethanol) were measured. The
stability ratio W, which measures the effectiveness of the potential barrier in preventing
the particles fi'om coagulation, and the interaction between the components were

discussed.

3.2.1 Zeta Potential

Electrolyte was added to a suspension to control the particle double layer. The
critical coagulation concentration (c.c.c.) is the concentration of the electrolyte at which
the suspension is stable. Ifthe electrolyte concentration is higher than the critical
coagulation concentration, the particle’s double layer is compressed, causing rapid
coagulation. By using the sedimentation technique, the critical coagulation concentration
of A1203 occln's at 0.001N 10103. Figures 12 and 13 Show the zeta potential of A1203
and SiCw versus pH for various electrolyte strengths, respectively.

Figures 12 and 13 showed that the zeta potential for both A1203 and Sij is geater
at 0.001N 10103 than at other electrolyte concentrations. Therefore, the 0.001N KNO3
was selected for all the suspensions in this research.

Next the zeta potential for A1203 and Sij in a mixture of electrolyte and ethanol

was measured. The zeta potentials of SiC versus pH, for varying ratios of

37

 

 
 
  

 

 

 

 

 

 

+0.001N KN03
—D—0.01N KNO,

3. +0.1N KNO‘
.5.
3
“
C
8
O
o.
g .10 ~
"‘ .20 .

.30 J

40 TI T 1 TI

3 7 9 11
pH

Figure 12 Zeta potential of A1203 versus pH at different electrolyte strengths.

38

Fig

 

+ 0.001 N KNO3
—{3— 0.01 N KNO3

+0.1N KNO3

 

 

 

 

zeta porential (mV)

 

 

 

Figure 13 Zeta potential of Sij versus pH at different electrolyte strengths.

39

.sinemm

+3: 05'he 2:12
am: PM
mitt: 5135136
3:. 9H plots.

:13 zen Pc
is: 11 figure
725.13 100% 1
153219301 15 3‘

lie absoll
It" If. draw
than (D) 0?

files the :

1 r'
l
.\
K . c-
\ ea

ethanol/electrolyte, are plotted in Figure 14. SiC, which has been shown to have an
acidic surface [52]. Thus, in 100% electrolyte, the iso-electric point is in the acidic
region of the zeta potential plot (pH=5.2iO.5). After increasing the ethanol content, the
iso-electric point moves towards higher pH values (pH=7). In 99% ethanol and 1%
electrolyte suspensions, there are no more iso-electric point showing in the zeta potential
versus pH plots, which show positive zeta potential values for all pHs tested.

The zeta potential versus pH for A1203, at varying ratios of ethanol/electrolyte, is
shown in Figure 15. A1203 has been shown to have a basic surface characteristic [52].
Thus, in 100% electrolyte, the iso-electric point is in the basic pH region (pH=8.8i0.5).
As ethanol is added, the iso-electric point decreases to a more acidic pH (pH=7 .3).

The absolute value of the zeta-potential for Sij and A1203, as shown in Figures 14
and 15, decreases as the ratio of ethanol to electrolyte increases. Since the dielectric
constant (D) of ethanol is approximately 1/3 that of water, as the ethanol content
increases, the suspension dielectric constant decreases assuming a rule of mixtm-es

calculation. Therefore, the thickness of the double layer (UK) will decrease with

increasing ethanol content. The dependence of (UK) on dielectric constant can be seen in

l
1 DH 3
I; - [872712252 I (20)
Equation (20).

4O

.>E.U

 

   
 
  

+ 100%elecrmlyte

-D- ethnnol:electrolyte(l:4)
+ ethanol:electrolyte-(1:1)
-)(- ethanol:elecu'olyte(4:l)
-)K- ethnol:electrolyte(99:l)

 

 

 

C(mV)

.20 .4

«30-

40‘

 

 

60 ...............

Figure 14 Zeta potential versus pH for Sij at varying ratios of ethanol to electrolyte
under 0.001N KN03 and 0.5 vol% solids content.

41

JNCI \\A

 

   
 
  

 

 

 

 

60 +lm°loelemlyte

50 -D—eth|ol:elemlyu(lz4)
; 40 +e¢hmlzeumlyuam
3:; 30 —)(-ethnol:electnlyu(4:l)
U 20 +ethnolze|emlyu09m

10

o ........................

-10

-20

.30I I l l I l I l I

3 4 5 6 7 8 9 10 11

Figure 15 Zeta potential versus pH for A1203 at varying ratios of ethanol to electrolyte
under 0.001N KNO3 and 0.5 vol% solids content.

42

a .

DLVO theory predicts that the particle double layers provide a repulsive force against
flocculation. Therefore, the zeta potential decreases due to compression of the double
layer. From zeta potential measurements, SiCw and A1203 suspensions become unstable

as the ratio of ethanol to electrolyte increases, in agreement with DLVO predictions.

3.22 Stability Ratio

Figures 16 to 20 give the stability ratios as a fimction of pH under constant
electrolyte (0.001N 19103) concentration and varying ethanol to electrolyte ratios. Table
3 shows the stable pH ranges, as indicated by logW>lO [23], for each of the
ethanol/electrolyte ratios. From Figures 16-20 and Table 3, it is seen that the stability
against homocoagulation between Sij/Sij shifis fi-om pH 3 to 5 and from pH 6 to 11,
in pure electrolyte, to pH 3 to 8 for 99% ethanol and 1% electrolyte. Stability against
homocoagulation between A1203/A1203 shifts from pH 3 to 5 and from pH 10 to 11 for
100% electrolyte to pH 3 to 5 for 99% ethanol and 1% electrolyte. Stability against
heterocoagulation between Sij/A1203 is in regions of pH 3 to 6 and pH 9 to 11 for
aqueous suspensions and from pH 3 to S for 99% ethanol and 1% electrolyte suspensions.

Analysis shows that the stability generally decreases as ethanol content increases.

At pH>8 range the stability decreases to zero as ethanol content increases to 99%. The
stable range for interactions between Sij and A1203, and A1203 and A1203, at 99%
ethanol and 1% electrolyte, remains at pH<5. However, the stable range for interactions
between SiCw and Sij expands to pH 8 in the 99% ethanol and 1% electrolyte
suspensions. Therefore, at pH<5, interactions for SiCW to Sij, SiC“, to A1203, and A120;

10 A1203 are stable in 99% ethanol and 1% electrolyte.

43

 

Ifll

350 1

 

300 ...... +SijISij

+ Sij/Alz o3
+Al203IA1203

2501

 

 

 

log 1., W

150 .
1001

50‘

 

 

Figure 16 Stability ratio versus pH for Sij/A1203 in 0.001N KNO3 and
ethanol:electrolyte (0:1).

 

 

 

 

 

 

350 l
300 1» WW -o-sac,/sscw
6 , +SiC Al 0
250« \p 9 ‘ “l 2 3
‘~. 0 O +A1203IA1203
; zoo « 9 o '
3 \ o
a Q 9 ,'
2 150 “ x ‘
fi‘A \ : 9
100 - . W‘ .' /
C ‘ A ’ O
. . .l
50 O * A. .‘
.... A o A‘
0 1% f v v fl h
3 4 5 6 7 8 9 10 11
pH

Figure 17 Stability ratio versus pH for Sij/A1203 in 0.001N KN03 and
ethanol:electrolyte (1:4).

45

 

+SiC\,/Sicw

-O- SijIAl 2O3

 

 

 

 

 

 

Figure 18 Stability ratio versus pH for SijlA1203 in 0.001N KN03 and
ethanol:electrolyte (1:1).

46

350

 

 

 

 

300
+ SiC ISiC
W W
250 + SijIAl 20 3
+ A’2 03 W 2° 3
3 200
2
3 150

100

50*

 

 

 

Figure 19 Stability ratio versus pH for Sij/A1203 in 0.001N KN03 and
ethanol:electrolyte (4:1).

47

 

+ Sicw lSiC w
300 1 + Sic:w IA! 203

 

+A1203/A12 o3

 

 

 

'0910 W
§

 

 

Figure 20 Stability ratio versus pH for Sij/A1203 in 0.001N KN03 and
ethanol:electrolyte (99:1).

48

5:201 an:
fifteen Si
2:33 am

3r; and

The stable range for interactions between Sij/A1203, and A1203/A1203 in 99%
ethanol and 1% electrolyte become smaller. However, the stable range for interactions
between SiC“, and SiCW in 99% ethanol and 1% electrolyte expands to pH 8. DLVO
theory accounts for the total energy of interaction between the repulsive interaction
energy and van der Waal’s attractive energy as:

VT=VA+VR . (11)

As stated previously, the repulsive energy, VR, decreases as the ethanol content is F
increased due to the compression of the double layer. It is believed that the decrease in
V3 is substantially great and results in an overall decrease in V1. Thus, suspensions

generally become increasingly tmstable as the ethanol content is increased. Bleier [20]

 

I - t."‘ LGL'A'D.

predicted the stability of SiC in ethanol and water respectively by calculation of V A, and
has shown that VA is higher in ethanol compared to VA in water. It is believed that the
increase in V A for SiC in ethanol results in the stability improvement for the interactions

between SiCw and Sij in 99% ethanol and 1% electrolyte at pH<8.

Table 3 Stable pH range in various media

 

 

 

 

 

 

 

Sij/Sij Sij/A1203 A1203/A1203

100% electrolyte 3-5, 6-11 3-6, 9-11 B-7, lO—ll
ethanol:electrolyte 3-5, 6—1 1 3-5, 10—1 1 -6, 11
(1:4)
ethanol:electrolyte 3-4, 7-11 3-6 3-7

1:1)
ethanol:electrolyte 3-4, 9-11 3-6 3-7
(4:1)
Ethanolzelectrolyte 3-8 3-5 3-5

99:1)

 

 

 

 

 

49

3.2.3 Viscosity
Figure 21 shows the viscosity of 10 vol% SiWA1203 suspensions with 40 wt%
solids loading at varying ratios of ethanol/electrolyte. For 100% electrolyte, the
suspension viscosity is low at pHs<8 and pHs>11 and high from pH 9 to 10. When the
ethanol content increases, the viscosity increases dramatically at high pHs. But at pHs
<4, even in 99% ethanol and 1% electrolyte, the 10 vol% Sig/.4120, suspensions have
low viscosities. These measurements verify the stability predictions, where high stability

coincides with lower viscosities and low stability coincides with higher viscosities.

3.2.4 Sedimentation Bulk Density

Figure 22 shows plots of sedimentation bulk density as a function of pH for A1203
and Sij in 100% electrolyte and a mixture of 99% ethanol and 1% electrolyte
suspensions. A comparison of Sij and A1203 sedimentation densities in 100%
electrolyte and in a mixture of 99% ethanol and 1% electrolyte shows that the
sedimentation density in the electrolyte at high pHs is higher than that in a mixture of
99% ethanol and 1% electrolyte. The stable suspensions at low pHs from stability
prediction have high sedimentation densities and tmstable suspensions at high pHs have
low sedimentation densities. The low sedimentation densities are believed to be due to

the loose packing of agglomerates.

50

 

 

+ 100%eleetmlyte

-C}— ethanol:electrolyte“ :9)
+ ethnolzeleetrolytead)
-)(-— ethanol:eleemlytemm

 

 

 

 

n(10'3N.sIm2)

 

 

 

Figure 21 Viscosity of 10 vol% SiC,.,/A1203 with 40 wt% solids content at varying ratios
of ethanol to electrolyte.

51

 

 

 

 

 

1'4 1 + A12 03 ,100%e1ectmlyte
1.2 4
-0— A12 03 ,ethnokeleetrolytemm
r 1 ..
E . + SiC w .100'loelectrolyte
0 0.8 '
\
a l
”g 0.6 I . -x— SiC w ,emnomtecmryumzr)
O . .
CL 0.4 - ""I l "
0.2 V — -~—.—.——— —-
o . ,

 

2 3 4 5 6 7 8 9 1011

Figure 22 Sedimentation bulk density versus pH for A1203 and Sij in electrolyte and
ethanol, respectively (0.001N KNO3).

52

1'4.

 

3.2.5 Microstructure

The green body microstructures (Figure 23) show the SiCw to be evenly distributed
and tmagglomerated in electrolyte, a mixture of ethanol and electrolyte (1 :1), and a
mixture of 99% ethanol and 1% electrolyte. There is no obvious heterocoagulation
between Sij and A1203 and only slight homocoagulation between A1203 particles at
99% ethanol and 1% electrolyte, consistent with the stability predictions.

In Figure 24, the microstructures of partially sintered (1650-1700°C, 2h) 10 vol%
Sij/A1203 samples show that Sij is well distributed within the partially sintered

A1203 matrix. The SiC“, and A1203 are uniformly distributed throughout suspensions in

 

100% electrolyte and ethanol:electrolyte (1 :1). a ‘

3.2.6 Conclusions

As the ethanol content increases, the absolute values of the zeta potential for both
SiCw and A1203 decrease. The iso-electric point likewise shifts for Sij to more basic
pHs and for A1203 to more acidic pHs. This leads to a decrease in stability for both
materials, as verified by sedimentation experiments. A comparison of Sin, and A1203
sedimentation densities in 100% electrolyte and in a mixture of 99% ethanol and 1%
electrolyte shows that the sedimentation density in the electrolyte at high pHs is higher
than that in a mixture of 99% ethanol and 1% electrolyte. The stable suspensions at low
pHs from stability prediction have high sedimentation densities and unstable suspensions
at high pHs have low sedimentation densities.

Ethanol and a mixture of ethanol to electrolyte reduce the dielectric constant of both

53

   

Figure 23 SEM photomicrographs of 10 vol% Sij/A1203 green body in a) electrolyte,
b) ethanol:electrolyte (1 :1), and c)ethanol:e1ectrolyte (99:1) (0.001N KN03,
non-ball-milled, pH 4, no polyelectrolyte and sintering aids).

54

 

c) ’ )
2 t d
3 (con
e
Figur

   

b)

Figure 24 SEM photomicrographs of 10 vol% Sij/A1203 partially sintered samples:
a) electrolyte, b) ethanol:electrolyte (1:1) (0.001N KN03, non-ball-milled,
pH 4, no polyelectrolyte and sintering aids, sintered at 1650°C for 1h,
nitrogen atmosphere).

56

suspensions. The zeta potential decreases due to compression of the double layer. The
stability ratio shows that the stability generally decreases as the ethanol content increases.
' But at pH<8 the interactions between SiC/SiC are more stable as the ethanol content
increases. Green body microstructure shows the Sij and A1203 to be well dispersed
both in electrolyte and in ethanol:electrolyte (1:1). However, there is a slight homo-
coagulation of A1203 in ethanol:electrolyte (1:1). The partially sintered microstructure

again shows good dispersion of Sij.

3.3 Colloidal Processing of Sij/A1203 Composites
Figure 25 showed colloidal processing approaches in the beginning of this research.
Successful processing in practice required trial and error to determine the proper
conditions. The theoretical results were used as a guide in the searching of the proper

processing.

3.3.1 Monolithic Alumina
It is recognized that inhomogeneities (i.e., local variations of chemical composition,

grain size or density) within the green compact can limit the extent of densification
attained. It is important to eliminate inhomogeneities from monolithic powder compacts
in the form of packing difi‘erences, grain size difi‘erences or composition difl'erences [53].
Colloidal processing was used to achieve homogeneous green bodies. A green density of
64% of the theoretical density was obtained by slip casting. Colloidal processing yielded
higher densities at lower sintering temperatures and shorter times. A 99% of the theoreti-

cal density was obtained at 1400°C (one hour) for AKP-30 alumina powder with 0.3-

0.5um particle size at pH 4 and 0.001N KNO3 electrolyte.

57

 

 

limit imposed by whisker’s large aspect ratio

p—--—-—--———-—--———-_—-—d

 

colloidal processing
Sij/A1203 composites
in a mixture of ethanol and electrolyte
.> prevent SiC whisker
to form large networks

 

 

 

Densification

 

colloidal processing
SijlA1203 composites
in electrolyte
+5ij networks inhibit densification
low density, low hardness, better ch

 

 

 

 

colloidal processing
monolithic alumina

-> good density, hardness
low fracture toughness

 

 

 

 

 

 

Loose Powders .
& Whiskers Improved Processrng ___>

Figure 25 Colloidal processing approaches.

58

 

3.3.2 Processing of Sij/A1203 Composites in Aqueous Systems

Figure 26 shows SEM photomicrographs of green microstructures of 10 vol% Sij/
A1203 composite processed from suspensions of difi‘erent pH. Homogeneous Sij/A1203
green bodies with densities of 2.55 g/cm3 (~65% TD) were obtained. A comparison of
the SEM photomicrographs (Figure 26) shows that tmiform green microstructures were
obtained at pH 3, 4 and 11; agglomerated microstructm'es were obtained at pH 2, 8 and
10. These green microstructure results are in good agreement with the stability
predictions for suspensions at difi‘erent pHs. Figure 27 shows SEM photomicrographs of
sintered microstructures of 5 and 10 vol% SiWM203 composites sintered at 1650°C for ’

one hour. Densification was inhibited by the SiC whiskers. The sintered densities of 5

 

and 10 vol% Sij/Al203 were only 3.15 g/ cm3 (~80% TD) and 2.84 g/cm3 (~73% TD),
respectively, slightly better than the density values obtained by researchers at MIT for the
same composite (60-70% TD) [54]. From stability predictions and slip casting
experiments, pH 4 was the optimum processing condition. Figure 28 shows green
density changes at difl‘erent whisker loadings. The green density decreases as the
whisker loading increases. This can be explained by the whisker network efl‘ect. The
higher the whisker loading, the more the whisker network forms. The loose packed
whisker networks cause a decrease in the green density.

The stress induced by the presence of the whisker networks stopped further sintering
of the composites. Improvements in the sintering of the composites will only then occur
if the whisker network can be removed or reduced. Additions of certain alcohols to the
whiskers were formd to prevent the formation of large flocs [5 5]. Stability predictions

indicate that SiC whiskers in a mixture of ethanol and electrolyte should be stable at

59

   

Figure 26 SEM photomicrographs of green microstructures of 10 vol% Sij/A1203
composite processed from different pHs: a) pH 2, b) pH 3, c) pH 4, d) pH 8,
e) pH 10, and t) pH 11 (0.001N KN03, non-ball-milled, no polyelectrolyte
and sintering aids).

   

Figure 26 (cont’d).

61

 

Figure 26 (cont’d).

62

 

Figure 27 SEM photomicrographs of sintered microstructures: a) 5 vol% Sij/A1203
and b) 10 vol% Sij/A1203. Sintered at 1650°C for 1h, nitrogen atmosphere
(0.001N KN 03, non-ball-milled, pH 4, no polyelectrolyte and sintering aids).

63

Fights

70-

 

 

A 55 -
2‘1
5
E
a
'0

so -
55 l U *1
o 10 20 30

vol% Sij

Figure 28 Green density versus whisker loading (0.001N KNOg, non-ball-milled, pH 4,
no polyelectrolyte and sintering aids).

ill-light t
5....m

:lecrmlyi

Messed

pH<6. The results of efforts to weaken the Sij network, through effective processing of
the composite in a mixture of ethanol and electrolyte, are discussed in the following

section.

3.3.3 Processing of Sig/A1203 Composites in Nonaqueous System

To prevent the formation of large flocs, ethanol was added to the whiskers. It was
thought that this might alleviate the stress efiect of the whisker network on densification.
Stability calculations predicted that SiC whiskers should be stable in 99% ethanol and 1%
electrolyte at pH<6. Therefore, a mixture of ethanol and electrolyte was used during
processing. Figure 29 shows SEM photomicrographs of green microstructures of 10
vol% Sij/A1203 composites cast from a mixture of ethanol and electrolyte with different
ethanol/electrolyte ratios. Homogeneous SiWM203 green bodies with densities of 2.55
g/cm3 (~65% TD) were obtained. Figure 30 shows SEM photomicrographs of sintered
microstructures of 10 vol% Sij/A1203 composites cast from a) electrolyte and b) a
mixture of ethanol and electrolyte. From Figure 30 it was seen that the composite
processed from a mixture ofethanol and electrolyte (1:1) is better densified than the one
from only electrolyte. Figm'e 31 shows that the green density of 10 vol% SiWA1203
composites varies at difierent ethanol/electrolyte ratios.

Sintered densities of 10 vol% SijlA1203 composites slip cast from a mixture of
ethanol and electrolyte were only 2.92 g/cm3 (~75% T'D) due to the whisker network
efi‘ect. From Figure 30 it is seen that whisker networks inhibited the sintering in this
composite. Comparing with the results from the aqueous system, the sintered density
increased slightly 2.84:0.08 g/cm3 (~73% T'D) to 29210.08 g/cm3 (45% TD), which

indicated that ethanol was not able to sufficiently alleviate the whisker network.

65

 

Figure 29 SEM photomicrographs showing the green microstructures of
10 vol% Sij/A1203 composites: a) 100% electrolyte,
b) ethanol:electrolyte (1:4), c) ethanol:electrolyte (1:1),
d) ethanol:electrolyte (4:1) and e) ethanol:electrolyte (99:1)
(0.001N KN03, non-ball-milled, pH 4, no polyelectrolyte and sintering aids).

66

 

Figure 29 (cont’d).

67

 

6)
Figure 29 (cont’d).

68

a '1

.
'5
‘V

.;Y
\

‘1’
s u,
‘d

 

Figure 30 SEM photomicrographs of sintered microstructures of 10 vol% Sij/A1203.
a) electrolyte and b) ethanol:electrolyte (1:1) (0.001N KNOg, non-ball-milled,
pH 4, no polyelectrolyte and sintering aids, sintered at 1650°C for 1h, N; .

69

HEW

dgldu, (%)

 

 

50 I T l l l
0 20 40 60 80 1 00

ethanol (vol%)

 

Figure 31 Green density variation of 10 vol% Sij/A1203 at difl‘erent ratios of ethanol to
electrolyte (0.001N KNO3, non-ball-milled, pH 4, no polyelectrolyte and
sintering aids).

70

L11
i)!

pressure
317355 on
gm 3 111
lets [it
~l- Ev
@3315 ca
Alte
F351. “hi
33%ka
[fated 11)
Were adde
hr becon

aid-MEI

3.3 .4 Discussion and New Approaches

Although the composite suspension is stable and codispersion of Sij and A1203 in
electrolyte and a mixture of ethanol and electrolyte is good, densification was
inhibited by the SiC whisker network efiect. Monolithic alumina can be sintered at
1400°C to near full densification. However, the sintered density of 10 vol% SiWM203
composites in electrolyte or a mixture of ethanol and electrolyte were only 73-75%.
Although ethanol may prevent formation of large whisker flocs, the problem related to
whisker networks stops further sintering of the composites. Complete densification by
pressureless sintering is often impossible to achieve [5 5]. The whiskers exert a back
stress on the sintering matrix, which acts to impede densification [53,56]. The whiskers
form a network with tmequal distances between pairs of neighboring whiskers. This
causes nontmiform local shrinkage that produces stresses, which impede densification
[57]. Eventually, a continuous network of contacting whiskers can form, stopping further
densification [57,58].

Alternative approaches that were explored are discussed in the following chapter.
First, whiskers were ball milled to reduce the aspect ratio and to physically reduce the
network. Second, sintering aids were added to create a grain boundary liquid phase that
relaxed the back stresses that resulted from the network efi‘ect. Third, polyelectrolytes
' were added. It was thought that polyelectrolytes might assist in forming a surface coating

that becomes an interphase region between whiskers and particles to stabilize the mixed

suspensions.

71

Sign t
W I1:
1211:0115 54
tracks am

“Mimi

3.4 Densification of Sij/A1203 Composites through Improved Processing
Approaches

3.4.1 Colloidal Processing Combined with Mechanical Methods

The aspect ratio of the whiskers has a strong influence on the sinterability of the
composites because the problems associated with heterogeneous packing of the matrix
phase and interactions between the whiskers increase as the aspect ratio increases. Faber
and Evans [59] studied the toughening effect by using toughening models. Their results
suggest that little additional toughening be obtained from whiskers with aspect ratios
greater than 10-15. They calculated the toughness increment of composites containing
various second phases and considered the existence of crack deflection reactions between
cracks and second phases. The contribution of second phases to the toughening is almost
the same in the region after the aspect ratio becomes greater than 12. The whisker’s
initial aspect ratio in this research is from 10 to 40, which is provided by the
manufacturer. Therefore, the suspensions were ball milled for 25-50 hours before slip
casting. The microstructure of the composites showed a range of aspect ratios from 6 to
20 after ball milling.

Figures 32 and 33 show that green density and sintered density increase as the ball-
milling time increases. The physical reduction of the whisker networks enhanced the
densification of the composites by allowing better packing of shorter whiskers and
reducing back stress.

While the aspect ratio was reduced to prevent whisker networks, higher solids
loadings were used to avoid segregation of powders and whiskers by difi'erential settling

of the slip during slip casting. The risk of differential sedimentation of the components in

72

70 -
. —l:l— 10vol%
. + 5vol%

 

 

 

 

dg’dm (%)

 

 

 

 

ball milling (hour)

Figure 32 Green density of 5 and 10 vol% Sij/A1203 as a function of ball milling time
(0.001N KNO3, ball-milled, pH 4, no polyelectrolyte and sintering aids).

73

 

 

 

 

 

 

 

 

1°° ‘ —e—10vol%
—)K— 5vol%
5

1: .
o\° ,'

.. h- ...

<

70 Y I T l I I T I ' l
o 10 20 30 4o 50

ball milling (hour)

Figure 33 Sintered density of 5 and 10 vol% Sij/A1203 as a function of ball milling
hour (0.001N KNOg, ball-milled, pH 4, no polyelectrolyte and sintering aids,
sintered at 1600°C for 1h, nitrogen atmosphere).

74

the liquid processing of ceramic composites is reduced if the suspensions are dense [60].
Processing of ceramic composites from dense liquid suspensions, containing up to 50
vol% solids, can reduce the settling rates of the components and avoid the risk of
difimfial sedimentation of whiskers and matrix powders during processing. Figures 34
and 35 show that the green density and sintered density of 5-10 vol% Sij/A1203 at pH 4
changes at different solids loadings. Increasing the solids loading from 34 vol% to 50
vol% increased the green density due to less segregation and better packing. However,
the sintered density’did not show a significant difi‘erence as the solids loading changed.
The whisker aspect ratio reduction had a more dramatic efl‘ect in achieving dense

composites in 20 vol% SiC whisker composites in comparison to 5 or 10 vol% SiC

 

whisker composites. High densities with 10 vol% SiC whisker and sintering aids were 1...-..
attainable without an aspect ratio reduction (see Figure 43 in Chapter 3.4.3).
Ball milling the aqueous suspension to control the whisker aspect ratio and break
agglomerates resulted in better densification. However, densification was still inhibited
by the presence of the whisker networks. The presence of a bormdary liquid phase was
predicted to reduce the whisker’s network efi‘ect during sintering [53]. Therefore, the

next step was to promote liquid phase sintering by using sintering aids.

75

70-

 

, —<>— 1 0vol°lo
. + 5vol%

 

 

 

«I . ’lr‘h‘fi‘d “I
\

03
0|

 

dgldu. (%)

 

 

 

SOIITITITI‘I
“3638404244464850

solids loading (vol%)

Figure 34 Green density of 5 and 10 vol% SiCa/A1203 as a filnction of solids loading
(0.001N KNO3, non-ball-milled, pH 4, no polyelectrolyte and sintering aids).

76

 

100 -

 

 

 

 

 

 

-6— 1 Dvol%
—)IE- 5vol°lo . -
a
3
5 E
3 90 . r1
9 --.
i
L___
80 r r T . 4a
30 35 4o 45 50 55

solids loading (vol%)

Figure 35 Sintered density of 5 and 10 vol% SijlA1203 as a fimction of solids loading
(0.001N 10103, non-ball-milled, pH 4, no polyelectrolyte and sintering aids,
sintered at 1600°C for 1h, nitrogen atmosphere).

77

3 .4.2 Sintering Aids

Theoretical calculations indicate that the presence of a liquid boundary phase will
increase the stress relaxation rate and increase densification [5 3]. Studies by O. Sudre
and coworkers [61,62] have demonstrated that the factors controlling the sintering of
ceramic particulate composites involve the packing of the matrix phase and interactions
between the inclusions, which lead to a constraining network. Sintering aids may
therefore reduce these factors and increase sinterability. I r .-

Figure 36 shows an A1203- Y203-SiC; phase diagram [63]. At above 1600°C, liquid
forms in the Si-rich area such as the SiC whisker surface with A1203 and Y203 particles

[63]. Tokai whiskers have a high surface oxygen content where the oxide resembled

 

SiOz [64]. The bulk oxygen is 0.23 wt% from fusion analysis whereas the surface
oxygen is 12.1 at% as measured by x-ray photoelectron spectroscopy [64]. Therefore,
additions of Y203 will react with the SiOz and A1203 powders to form a glass grain
bomdary phase that improves sinterability.

Figure 37 shows plots of green density of 5 and 10 vol% SijlA1203 versus Y203
content. Green density stays relatively constant after the addition of 2-4 wt% Y203
although the green density decreased at the addition of 0.1-0.5 wt% Y203. Figure 38
shows plots of sintering density of 5 and 10 vol% Sij/A1203 versus Y203 content.
Sintered densities increased after introducing Y203 content. For 5 and 10 vol%
Sij/A1203, respectively, the sintered densities remained almost constant at Y203
additions greater than 0.5 and 2 wt%, respectively. Sintering aids improved the sintering
of the composites, however, the suspensions were slightly agglomerated and experienced

some degree of dimculty in slip casting. The following section describes attempts to

78

  
  
     

nos-25.02

2Y203-35i02 swam-25.6z

-Isoo‘x‘ I I

v o I 20 40‘ so, so I Al 0
2 3 - \ imam-lingo3 ' 3

Figure 36 The A1203-Y203-Si02 phase diagram [63].

79

 

 

 

 

—A— 10vol%
+ 5vol%

 

 

 

 

 

 

 

58 I I I r I I I fl
0 l 2 3 4
Y203 (Mo/0)

Figure 37 Green density of 5 and 10 vol% Sij/A1203 as a function of Y203 content
(0.001N KNO3, ball-milled, pH 11, 2.5 vol% polyelectrolyte).

80

 

’1'“ -v. u
5

 

100 —A- 10vol‘lo

5%

 

 

 

 

 

Y203 (wt%)

Figure 38 Sintered density of 5 and 10 vol% Sij/Alzog as a function of Y203 content
(0.001N KNO3, ball-milled, pH 1 1, 2.5 vol% polyelectrolyte, sintered at
1600°C for 1h, nitrogen atmosphere).

81

 

further stabilize the composite suspensions in the presence of sintering aids by adding a
polyelectrolyte to make stable homogeneous microstructures to finally achieve good

densification and mechanical properties.

3.4.3 Polyelectrolyte

A polyelectrolyte stabilizes suspensions to promote a tmiform spatial distribution of
the multiphases. A phase between the matrix and inclusion phases may be required to
control the strength of the bond between the two phases [55]. A variety of coating and
inclusion combinations have been demonstrated using dispersants to control the surface
potential of the materials [55]. Polyelectrolytes are powerful additives that are able to
reverse the zeta potential or to shift the iso-electric point of ceramic powders. Charge
reversal mechanisms include ion change, surface complexing, ion-solvent interaction,
adsorption of difi‘erent species and steric stability of dispersant using polymeric
carboxylic salts. It is important to select dispersants that are compatible in mixed
systems [54]. The selected polyelectrolyte for this research is an ammonium salt of a
polymeric carboxylic acid. It was reported that CA, tricarboxylic acid, might be the
interface coating to stabilize the suspensions and also interact with particles
electrostatically [26]. An ammonium salt of a polymeric carboxylic acid was also used
successfully in stabilizing the suspensions [65].

Tables 4 and 5 list stable pH ranges for different components in aqueous media and
in ethanol and allow a compatibility range to be identified. In aqueous media, the

compatible processing range is pH 10-1 1.

82

 

Table 4. Compatibility of stable pH ranges for difi‘erent
components in aqueous media

Stable
in 7 10-11
in electrol 5 6-11
between Sij 9-11

in
0-11
6 12
1

 

Table 5 Compatibility of stable pH ranges for different
components in 99% ethanol and 1% electrolyte

 

 

 

 

 

 

 

 

 

Components Stable pH range ‘

A1203 in 99% ethanol and 3-5
1% electrolyte
Sij in 99% ethanol and 1% 3-8
electrolyte

teraction between Sij and 3-5
A1203 in 99% ethanol and
1% electrolyte
lPolyelectrolyte[66] 10-1 1
Y203[67] 8-12
Compatible processing range none

 

 

 

83

 

Polyelectrolyte dispersants can alter the A1203 surface charge. Figure 39 shows the
zeta potential of A1203 versus pH for varying concentrations of polyelectrolyte. Figure
39 shows that the zeta potential at pH 4 changes from positive to negative after adding
the polyelectrolyte, consistent with Cesarano H1 and Aksay’s results [30]. In Figure 39,
the increase in polyelectrolyte concentration caused the iso-electric point shifting fiom
pH 8.8 to 3.2. The iso—electric point at pH 3.2 did not decrease further with additional
increases in polyelectrolyte above 0.23 vol%. The absolute value of the zeta potential
decreased at polyelectrolyte concentrations greater than 0.23 vol%. This is consistent
with depletion flocculation at excess polyelectrolyte [30]. Table 6 lists the iso-electric

point for A1203 at difi'erent polyelectrolyte concentrations. Table 7 shows the

 

comparison of the iso-electric point for A1203 determined in this research and that - ~- ....
published. The iso-electric point values of A1203 for suspensions with addition of
difi‘erent polyelectrolytes: citric acid, PMMA and ammonia salt of polymeric carboxylic
acid, respectively, are within 0.5 pH. It is believed that the addition of these three
difl'erent polyelectrolytes results in the formation of a negative charge on the alumina
particles. As a result, the iso-electric point of an alumina suspension shifts towards pH 3.
Additions of excess polyelectrolyte impact hardly any further change on the
electrokinetic properties of the particles [71]. The saturation of the surface that is evident
from the ESA measurements (Figure 39) is in very good agreement with the results of
other researchers (Table 7).

Polyelectrolyte adsorbs onto the surface of A1203, leading to a change in the surface
charge as well as the magnitude of the zeta potential of the particles. It is possible to con-

trol the interaction potential between the powder particles by varying the polyelectrolyte

84

 

 

 

 

 

 

 

60 + 0 vol%
50 —Cl-—0.03vol%
' +0.23vol%
; 40 —)<—0.45vol%
g 30
g 20 I
g. 0 ' - I -
8 .10 I
S I; --
.20 - ‘“!! “‘56:?
40 ~
40 . v
3 5 7 9 11

.pH

Figure 39 Zeta potential for A1203 versus pH at different polyelectrolyte concentrations

(0.001N 105103).

85

 

 

Table 6 Comparison of polyelectrolyte addition to iep and

 

 

 

 

 

C—potential for A1203
Polyelectrolyte iso-electric point of Zeta potential versus
A1203 JDH plots
0 8.8 Original
0.03 vol% 4.2 Shifted
0.23 vol% 3.2 Saturated
0.45 vol% 8.2 Saturated

 

 

 

 

Table 7 Comparison of iso-electric point of A1203 between
this research and published data [26,30]

 

 

 

 

 

 

 

Authors iso-electric iso-electric point in
point in electrolyte with
electrolyte polyelectrolyte addition
Hidber, Graule and 9.2 04 wt% CA: 3.2
Gauckler [26]
Cesarano II] and 8.7 enough PMAA: 3.4
Aksay [30]
This research 8.8 0.23 vol% NH4PCA: 3.2

 

Note: CA= tricarboxylic acid PMAA= polymethacrylic acid
NHaPCA= ammonia salt of polymeric carboxylic acid

process [65,68].

86

concentration, thus enabling the isoelectric point (iso-electric point) of the A1203 as well

as the viscosity of the suspension to be adjusted to the specific requirements of a given

Addition of 2.5 vol% polyelectrolyte stabilized the 10 and 20 vol% SijlA1203 sus-
pension in increasing the‘composite densities. However, further additions of polyelectro-
lyte past the adsorption saturation limit left excess polyelectrolyte in suspension. This

excess polyelectrolyte lowered the densities of the composites due to depletion

 

 

flocculation in which the excess polyelectrolyte remaining in the solution leads to an
increase of the ionic strength [71]. The presence of a large amount of rmadsorbed
polyelectrolyte in the solution results in a decrease of Debye length and, subsequently, an
increase of the flocculation. Figures 40 and 41 show plots of green and sintered densities
of the composites versus polyelectrolyte concentration. The sintered density increased as
the polyelectrolyte was introduced. Without polyelectrolyte addition the suspension
contained hard agglomerates and experienced some degree of dificulty in slip casting.
Polyelectrolytes had a powerful efl‘ect in stabilizing the mixed suspensions. The optimum
addition was formd at a polyelectrolyte concentration of 2.5 vol%. Increasing the
concentration beyond 2.5 vol% caused a slight decrease in the sintered density.

The addition of a polyelectrolyte (polymeric carboxylic salt as a dispersant-
ammoniurn salt of a polymeric carboxylic acid) in controlling the dispersion and stability
of Sij/Alzog in the presence of sintering aids in aqueous suspension proved successful.
The carboxylic salt molecules absorbed on the surface not only influenced the surface
charge of the alumina particles but also created a steric barrier that acted to inhibit the
mutual approach of the individual particles [10].

Figures 42 and 43 are plots of the green and sintered densities of 10 and 20 vol%
Sij/A1203 with sintering aids and polyelectrolyte at pH 11 as a function of changes in
ball milling time. Because sintering aids relax the back stress of whisker networks and
polyelectrolytes stabilize the suspensions, the whisker network effect due to a high aspect
ratio is reduced. Ball milling had no apparent effect on green density (Figure 42).
Sintered density improved slightly only at high whisker loadings (Figure 43). By

controlling the surface chemistry to eliminate inhomogeneities and choosing suitable

87

 

 

 

 

 

 

‘5 ‘ +10vol%
—Cl—20vol%
.. so .
i5
5
a
\
a
1:
55 -
so I r I T r I
o 1 2 3 4 s s

polyelectrolyte (vol%)

Figure 40 Green density of 10 and 20 vol% Sij/A1203 as a function of polyelectrolyte

concentration (0.001N 10103, ball-milled, pH 11, 0.5 wt% MgO and
2 Wto/o Y203).

88

 

—o— 20vol%
100 - -a—10voI°/o

 

 

 

%dth

 

 

80urrr
0 2 4 6

polyelectrolyte(vol%)

Figure 41 Sintered density of 10 and 20 vol% Sij/A1203 as a function of polyelectrolyte
concentration (0.001N KNO3, ball-milled, pH 11, 0.5 wt% MgO and
2 wt% Y203, sintered at 1600°C for 1h, nitrogen atmosphere).

89

 

65 -
I 10vol%
CI 20vol%

 

 

 

d,/d,.,(%)
+

 

60 -
0 48

ball milling (hour)

Figure 42 Green density of 10 and 20 vol% Sij/A1203 as a flmction of ball milling hour
(0.001N KNO3, ball-milled, pH 11, 2.5 vol% polyelectrolyte, 0.5 wt% MgO

and 2 wt% Y203).

90

 

 

 

 

 

 

 

10° ‘ I 10vol%
3 20vol%
so -
5
'U
o\°
so -
ml

 

0 48
ball milling (hour)

Figure 43 Sintered density of 10 and 20 vol% Sij/A1203 as a function of ball milling
hour (0.001N KN03, ball-milled, pH 11, 2.5 vol% polyelectrolyte,
0.5 wt% MgO and 2 wt% Y203, sintered at 1600°C for 1h, nitrogen
atmosphere).

91

sintering aids to alleviate the back stress of whisker networks, densification of 10 vol%

Sij/A1203 with high aspect ratio whiskers can be achieved.

3.4.4 Discussion

The combined use of difi‘erent approaches has been successful in producing dense
Sij/Aleg composites in aqueous suspensions. pH 11 was determined to be the
optimum processing pH value for SijlA1203 composites with 2 wt% Y203, 0.5 wt%
MgO and 2.5 vol% polyelectrolyte (ammonia salt of polymeric carboxylic acid).

The stable pH range for composites with sintering aids and polyelectrolyte and the
stable pH range for the composite suspension in a mixture of ethanol and electrolyte are
not compatible (see Table 5). From stability predictions the stable pH ranges for A1203,
Sijand the interaction between A1203 and Sij in 99% ethanol and 1% electrolyte are
pH 3-5, pH 3-8 and pH 3-5, respectively. However, the stable pH range for the
polyelectrolyte as reported by the manufacturer is pH 10—1 1, and since Y203 dissolves
only in acid, it can only be used at pH>7. Thus, there is no compatible processing pH
range for A1203 particles, Sij whiskers Y203 and polyelectrolyte in 99% ethanol and
1% electrolyte. Therefore, the composites which were slip cast from a mixture of ethanol
and electrolyte were devoid of Y203 as a sintering aid and the ammonium salt of a
polymeric carboxylic acid as a stabilizing polyelectrolyte to improve densification.

The efl‘ect of non-aqueous media in ceramic colloidal suspensions has not been
clearly identified. For nonaqueous suspensions, it is a challenge to find suitable sintering
aids and appropriate polyelectrolytes to match the Sij/A1203 suspension systems. The

current processing effort focused on aqueous suspensions. For nonaqueous suspensions,

92

further research should be directed to creating compatible processing ranges by choosing

suitable sintering aids and an appropriate polyelectrolyte to match the stable pH range of

A1203 and Sij in the non-aqueous suspension.

3.5 Improved properties of SiWM203 Composites
3.5.1 Sintered density
A bulk density of 3.79g/cm3 (97% TD) for 10 vol% Sij/A1203 was achieved at a

sintering temperature of 1600°C under nitrogen flow for one hour. Densification was
inhibited by the presence of the SiC whiskers. Bulk densities of3.80 g/crn3 (96% TD),
3.79 g/cm3 (97% TD), and 3.40 g/cm3 (89% TD) were obtained at 1600°C for composite
samples containing 5, 10 and 20 vol% SiC whiskers, respectively. Figure 44 depicts the
green and sintered densities for different pH values. Bulk densities for the 10 vol% Sij/
A1203 composites were 3.79 g/cm3 at pH 11 and 3.66 g/cm3 at pH 4, respectively. At pH
7, the “creamy” slip failed to cast. Figure 45 is a comparison of the sintered density for
SiWMzog composites for this study and other published results fi'om Tiegs and Becher
[3] at Oak Ridge National Laboratory, Sacks, Lee and Rojas [73] at the University of
Florida and Barclay, Fox and Bowen [54] at Massachusetts Institute of Technology. The
bulk sintered densities achieved in this study on average were higher at lower sintering
temperatures and sintering time shorter by an average of 60 min. Tiegs and Becher [3]
ball milled the whiskers, used sintering aids and sintered at 1700-1800°C. HIP at 1600°C
and l70MPa in an argon atmosphere were also used along with the pressureless sintering.
Sacks, Lee and Rojas [73] used suspension processing, polyelectrolytes, slip casting, and

sintering at 1600°C for 4 hours in a flowing nitrogen atmosphere. Barclay, Fox and

93

 

El green density
I sintered density

 

 

 

  
     
 

  

 

100 -
95 -
90 -
85 ‘ creamy slip
80 _ fail to cast
5
13 75 - l
.\
70 -
65 -
60 -
55 -
50 d
4 7 ' 11
pH

Figure 44 Green and sintered density of 10 vol% Sij/A1203 as a function of pH
(0.001N KNO3, ball-milled, 2.5 vol% polyelectrolyte, 0.5 wt% MgO and
2 wt% Y203, sintered at 1600°C for 1h, nitrogen atmosphere).

94

 

+1111: research (16000.11!)

-0— Oak Ridge National Lab (1700C,1h)
4-1 ‘ + University of Florida (1 BOOCAh)
-)(- MIT (1 800C,1 5h)

 

 

 

d, (glcm’)

 

2.7 -'

 

 

2.5 r r A r a
0 5 1 0 1 5 20

vol% of SIC“,

Figure 45 Comparison of sintered density of Sij/A1203 composites between
this research and other published results [3,54,73].

95

Bowen [54] used yttrium isopropoxide as a sintering aid, polymeric dispersants and
pressureless sintering at 1800°C for one and a halfhours in a flowing argon atmosphere.
The composites of this research used polyelectrolyte, sintering aids, ball milling and

pressm'eless sintering at 1600°C for one hour under nitrogen flow. By combining
difiemnt methods together this research achieved high densification. Compared with hot
pressing, the pressureless sintering is less expensive, is able to produce complicated
shapes and is isotropic in mechanical properties. However, the mechanical properties are
not as good as those for hot pressing specimens because the densification of A1203-SiC
Whisker composites by pressureless sintering is inhibited as a result of whisker
interference with particle rearrangement and composite shrinkage.

Figure 46 shows the relationship between the sintered density and the green density.

The higher green density tends to have the higher sintered density. Better packed green
body has less inhibition in densification.

96

100 l
95 a
90 r
85 r
80 4
75 r
70 i

deth (‘70)

65-:
60‘

55 l j f l
61 62 63 64 65

 

 

dgldu. (%)

Figure 46 The sinterd density versus the green density (5-20 vol% Sij/A1203,
0.001N KNO3, ball-milled, le l, 2.5 vol% polyelectrolyte, 0.5 wt% MgO
and 2 wt% Y203, sintered at 1600°C for 1h, nitrogen atmosphere).

97

3.5.2 Microstructure and EDX analysis
The final microstructure of 10 vol% Sij/A1203 composites revealed homogeneous
near fully densified composites as evidenced by the whiskers being completely sur-
rounded by the alumina matrix without space between them. Compositions of the
composites were characterized by EDX. Figure 47 shows the Energy Dispersive X-ray
Spectra of 10 and 20 vol% Sij/A1203 composites, respectively. Table 8 lists the EDX

composition of 10 and 20 vol% Sij/A1203 composites (at%).

Table 8 EDX composition of 10 and 20 vol%

 

 

 

SiC“/A1203 composites
Element 10vol% 20vol%
A1 (at%) 89.8 72.0
Si Qt%) 10.2 28.0

 

 

 

 

 

The Si at% may be in error because of some nearby peaks of yttrium, aluminum and
some impurities. The energy of the silicon peak is at 1.8keV, that of yttrium is at
1.95keV. The energy of the aluminum peak is at 1.5~1.6keV. The peaks of yttrium and
aluminum are close to that of silicon. The yttrium peak is too small to be identified.
Fm'thermone, since all the peaks have a width as big as 0.3keV, some overlaps between
those of aluminum, yttrium and silicon are expected, and probably cause the measured
Silicon composition to be somewhat higher than the actual composition.

Figures 48 to 50 are SEM photomicrographs of the fracture surfaces of 5, 10, 20
vol% SijlAle; ball-milled composites. The fracture surfaces were taken in the Hitachi

S‘2500C SEM at an accelerating voltage of lSkV fi'om broken chips made by a hammer.

98

 

 

 

 

 

2oo— A“
4. :2
1 .al .
a A, ;
'3 lg :

m“. ’i z
i L- I
J i ;
I 1! .
.' -: |
i ’1 i
1. I! i

1 ' 3| ;

00—} H g

I: g
4 I!

E’ i
4 .1 1
.' g...
i O ' . I
o; 3 "SI ;
u . S- .
lefllu‘ J ‘lf’L :

e .wwt 3...” .rflTf Ti ,

1 2 3 0
WNW
8)

cps
1. A.

““2 i
1 ”r
l a

1'1 ;

”1 1i .
‘t 2'; 2
7'1 l

50—-; .. I :
4 :‘zéi !
. ,o _£ '2”. g
1;”: f “.... l
°#'ZW/ K - my . I i

1 2 3 4
EnemOzeV)
b)

Figure 47 EDX composition spectra: a) 10 vol% Sij/A1203, b) 20 vol% Sij/A1203
(0.001N KNO3, ball-milled, pH 11, 2.5 vol% polyelectrolyte, 0.5 wt% MgO
and 2 wt% Y203, sintered at 1600°C for 1h, nitrogen atmosphere).

99

Figure 48 shows cleavage fracture and transganular fracture in the composite containing
5 vol% Sij whiskers. The photomicrograph of Figure 49 shows only transganular
flame and whisker pullout in the composite containing 10 vol% Sij whiskers. The
composite is very densified around SiC whiskers. Figure 50 shows transgranular fracture
and pores. Figure 51 is a SEM photomicrogaph of the fractme surface of 20 vol% '
Sij/A1203 composites that were not ball-milled. Figure 51 shows the mixed
transganular, interganular, cleavage fixtures for composite that were not ball-milled.

The thermal expansion coeficient of the SiC whisker is only one half that of alumina.

 

This thermal mismatch causes a tensile stress in the alumina matrix when the composites

are cooled after sintering is complete. Back and Kim [69] believe that this stress weakens

 

the alumina matrix. They reported that dislocations were observed in the alumina matrix
around the SiC whiskers in composites. The dislocations weakened the alumina matrix.
Therefore, the appearance of transganular fracture is an indication that dislocations
aromd the SiC whiskers have weakened the alumina matrix.

Table 9 lists the difl‘erent gain sizes for the 5 and 20 vol% Sij loadings. The gain
sizes were measured on the SEM photomicrogaphs by the intercept method [72]. A
Straight line of a known length L was drawn on a photomicrogaph of magnification M.
The number of gains, g, which were intersected, by the straight line, was counted. The
following equation was then used to compute the grain size:

(3.8 = L/(gM) (21)
This method was repeated three times and an average value for the gain size was then
coUntamed.

100

 

 

Figure 48 SEM photomicrogaph of the fracture surface of 5 vol% Sij/A1203
(0.001N 10103, non-ball-milled, pH 4, no polyelectrolyte and sintering aids,
sintered at 1600°C for 1h, nitrogen atmosphere).

101

 

Figure 49 SEM photomicrogaph of fracture surface of 10 vol% Sij/A1203
(0.001N KNO3, ball-milled, pH 11, 2.5 vol% polyelectrolyte, 0.5 wt% MgO
and 2 wt% Y203, sintered at 1600°C for 2h, nitrogen atmosphere).

102

 

Figure 50 SEM photomicrogaph of fracture surface of 20 vol% Sic.../A1203
(0.001N KN03, ball-milled, pH 11, 2.5 vol% polyelectrolyte,
0.5 wt% MgO and 2 wt% Y203, sintered at 1600°C for 4h, nitrogen
flow).

103

 

Figure 51 SEM photomicrogaph of fracture surface of 20 vol% Sij/A1203
(0.001N 10103, non-ball-milled, pH 1 1, 2.5 vol% polyelectrolyte,
0.5 wt% MgO and 2 wt% Y203, sintered at 1600°C for 4h, nitrogen
atmosphere).

104

Table 9 Difi'erent gain sizes for the 5 and 20 vol% Sij loading.

 

 

 

 

 

 

 

Sij loading (vol%) 5 20 20
(ball milled) (ball milledL (without ball millingL1
Grain size(pm) 2.7 0.8 0.6

 

Figures 52 and 53 show SEM photomicrogaphs of fracture surfaces of 5 and 20
vol% SiCdAlzog, respectively, sintered at 1600°C for 4 hours 1mder nitrogen flow. A
comparison of Figures 52 and 53 shows that the gain size has decreased as the whisker
loading increased. The gain size decreases from 2.7m to 0.8m as the whisker loading
increases from 5 to 20 vol%. Smith, Singh and Scattergood [70] believe that gain
boundary pinning is responsible for the decrease in gain size at higher whisker loadings.

Figures 54 and 55 show SEM photomicrogaphs of the fracture surfaces of 20 vol%
Sij/A1203 with and without ball milling, respectively. 0.8 pm and 0.6 um gains were
measured using the intercept method. A comparison of Figures 54 and 55 shows that ball
milling the SiC whiskers serves to increase the size of grains slightly.

Another feature that is shown in the photomicrogaphs of the composites is the
tagged fracture appearance. Back and Kim [69] believe that the ragged fracture surfaces

reSult from a crack deflection mechanism.

105

 

Figure 52 SEM photomicrogaph of the fracture surface of 5 vol% Sij/A1203
(0.001N IQVO3, non-ball-milled, pH 4, no polyelectrolyte and sintering aids,
sintered at 1600°C for 1h, nitrogen atmosphere).

106

 

Figure 53 SEM photomicrogaph of the fracture surface of 20 vol% Sij/A1203
(0.001N KNO3, ball-milled, pH 11, 2.5 vol% polyelectrolyte, 0.5 wt% MgO
and 2 wt% Y203, sintered at 1600°C for 4h, nitrogen atmosphere).

107

 

Figure 54 SEM photomicrograph of the fracture surface of 20 vol% Sij/A1203
(0.001N KNOg, ball-milled, pH 11, 2.5 vol% polyelectrolyte,
0.5 wt% MgO and 2 wt% Y203, sintered at 1600°C for 4h, nitrogen flow).

108

 

Figure 55 SEM photomicrogaph of the fracture surface of 20 vol% Sij/A1203
(0.001N KNO3, non-ball-milled, pH 11, 2.5 vol% polyelectrolyte,
0.5 wt% MgO and 2 wt% Y203, sintered at 1600°C for 4h, nitrogen
atmosphere).

109

3.5.3 Hardness and Fracture Toughness

The critical value of the stress intensity factor required to initiate fi'acture, Kc is a
measure of the resistance of a material to fracture, i.e., of its toughness. Figures 56 and
57 show the fracture toughness and hardness for the composites as a fimction of ball
milled whisker loadings with 2.5 vol% polyelectrolyte at pH 11 and sintered at 1600°C,
one hour under nitrogen flow. Figure 56 shows a comparison of fracture toughness
between this research and other published results [3, 74]. This research achieved good
We toughness at lower temperature without pressure assisted consolidation. The
high fracture toughness results from Oak ridge National Lab [3] were obtained by high
pressure consolidation. .
The fracture toughness and hardness were measured by the indentation technique
[49] . The equations for calculating both the hardness and fracture toughness are
discussed in the Experimental Procedure section. Although the hardness decreased as the
Whisker loading increased due to the whisker’s ability to inhibit densification, the fracture

tOl-Ighness increased as whisker loading increased fi'om 3.8102 MPa-m”2 to

6.3 :02 MPa-nn"2 user adding 20 vol% ball milled SiC whiskers.
Figures 58 and 59 are comparisons of fracture toughness and hardness of

SiCe/A1203 composites with and without ball milling, respectively. Figure 58 shows that

the fiacture toughness decreased from 7.302 MPa-m"2 to 63:02 MPa-m"2 afier ball

milling 48 hours. This is due to the whisker pullout toughening mechanism in which the

10tiger whisker length contributes to higher fracture toughness.

The fracture toughness of A1203 improved with increasing whisker content. The

fracture toughness of the 20 vol% Sij/A1203 composite was twice that of the

110

 

- + This research
1 (1600C,1h)

g —O—Oak Ridge National Lab
1 (1600-1700C. HIP)

' —-o—- University of Florida
(1700.1800C)

 

 

K,C (MPa.m“2)

 

 

Sij (vol%)

Figure 56 KC of composites versus whisker loading showing the comparison
between this research and other published results [3, 74].

111

19-
18¢
17~

16‘

14 . ..
/\
13 1 ‘- _

12 r ..

Hardness (GPa)

 

 

 

11‘

 

 

10 r 1
0 10 20

810., (vol%)

Figure 57 Hardness of composites versus whisker loading (0.001N KN03, ball-milled,
pH 11, 2.5 vol% polyelectrolyte, 0.5 wt% MgO and 2 wt% Y203, sintered at

1600°C for 1h, nitrogen atmosphere).

112

 

8a
7..

'E.

a 6-

m

.5.

0

s2 5'
4-
3

   

 

0 48
ball milling (hour)

Figure 58 Comparison of ch of 20 vol% Sij composites with ball milling and without
ball milling (0.001N 10103, pH 11, 2.5 vol% polyelectrolyte, 0.5 wt% MgO
and 2 wt% Y203, sintered at 1600°C for 1h, nitrogen atmosphere).

113

 

I 10vol%

El 20vol%

 

 

 

 

Hardness (GPa)

 

 

0 48
ball milling (hour)

Figure 59 Comparison of Hardness of 10 and 20 vol% Sij composites with ball milling
and without ball milling (0.001N KNO3, pH 11, 2.5 vol% polyelectrolyte,
0.5 wt% MgO and 2 wt% Y203, sintered at 1600°C for 1h, nitrogen
atmosphere).

114

A1203 matrix. Back and Kim [69] reported that the whisker length influenced the

whisker protruding length of the fracture surface. Figure 58 shows that the composite
without ball milling where the whisker lengths (5-1 Sum) are longer has a higher fracture
toughness in comparison to the composite having 20 vol% SiC whiskers (5-10um) which
were ball milled. Figure 60 shows SEM photomicrogaphs of the fracture surface for the
10 vol% SiWM203 showing long whisker protruding length (a) and deep whisker pull-
out holes (b).

The experimental data with theoretical prediction by Smith, Singh and Scattergood
[70] showed thm whisker pullout, whisker bridging, and crack deflection each 1
contributed to the toughening increase. The present results agee with the conclusion of
Smith, Singh and Scattergood as evidenced in the SEM photomicrogaphs of the fracture
Surfaces that show evidence of whisker pullout and crack deflection.

Figure 61 is a gaph of hardness of the 10 vol% Sij/Al203 as a flmction of pH. At
pH 11, the better densified composite has a higher hardness when compared to the com-
posite processed at pH 4. Figure 62 is a plot of the hardness of 10 vol% Sij/A1203 at
pH 11 as a function of Y203 content (sintered at 1600°C for one hour) which shows that
the hardness increases as Y203 content increases. Hardness increases linearly at Y203
cOlltent less than 2 wt%, probably because the density of composites increases as Y203
cOntent increases. However, when the Y203 content is increased over 2 wt%, the hard-
ness value decreases slightly. Figure 63 is a plot of hardness of 10 vol% SiCo/A1203 as a
function of density (sintered at 1600°C for one hour). Figure 64 is a plot of hardness of
1 0 vol% SiC,JA1203 at pH 4 as a function of ball milling time (sintered at 1600°C for one

how). Hardness also increases as the ball milling time increases. Figure 65 is a plot of

115

 

b)
Figure 60 SEM photomicrogaphs of the fracture surface of 10 vol% Sij/A1203
showing (a) long whisker protruding length and (b) deep whisker pullout holes
(0.001N KNOg, ball-milled, pH 11, 2.5 vol% polyelectrolyte, 0.5 wt% MgO
and 2 wt% Y203, sintered at 1600°C for 1h, nitrogen atmosphere).

116

 

19 -
18 -
17 -
16.-
15 -
14 -

Hardness (GPa)

 

 

pH

Figure 61 Hardness of 10 vol% Sij/Aleg as a function of pH (0.001N KNO3,
ball-milled, no polyelectrolyte and sintering aids at pH 4 and
2.5 vol% polyelectrolyte and 2 wt% Y203 at pH 1 l, sintered at 1600°C for 1h,
nitrogen atmosphere).

117

18-

 

 

 

16‘
E 1
Q, 4
O
0
2
u 12“
g t-
10‘“
8- . a . fl
0 1 2 3 4 5

Y203 (W35)

Figure 62 Hardness of 10 vol% SiWM203 as a fimction of Y203 content
(0.001N 10103, ball-milled, pH 11, 2.5 vol% polyelectrolyte, sintered at
1600°C for 1h, nitrogen atmosphere).

118

18 a

 

 

 

16 ~
7:?
o. 14 -
9.
m
in
0
-§ 12-
l-
N
1:
10 -
8 - r r
85 90 95 100

ds/dth(%)

Figure 63 Hardness of 10 vol% Sij/A1303 as a function of density (0.001N KNO3,

ball-milled, pH 11, 2.5 vol% polyelectrolyte, 0.5 wt% MgO and 2 wt% Y303.
sintered at 1600°C for 1h, nitrogen atmosphere).

119

 

E1 5vol°/o
I 10vol%

 

 

 

20 -
19 -
18 -
17 -
16 -
15 -
14 -
13 -
12 -
11
10

 

Hardness (GPa)

 

 

 

   

 

36 50
ball milling (hour)

Figure 64 Hardness of 10 vol% SiWN203 as a function of ball milling time
(0.001N 10103, ball-milled, pH 4, no polyelectrolyte and sintering aids,
sintered at 1600°C for 1h, nitrogen atmosphere).

120

hardness of 10 vol% Sij/Alzog at pH 11 as a function of polyelectrolyte concentration
(sintered at 1600°C for one hour). Hardness increases from 14.9il .0 GPa to 16.3:t1.0
GPa as the polyelectrolyte concentration increases from 1.2 to 2.5 vol%. Further
increasing the polyelectrolyte concentration to 5 vol% causes a decrease in hardness from
16.3:t1.0 GPa to 14521:] .0 GPa. Figure 66 is a plot of hardness of 10 vol% Sij/A1203 at
pH 4 as a function of solids loading (sintered at 1600°C for one hour). Hardness
increases as the solids loading increases. Higher solids loading prevents the alumina and
SiC whiskers from difl‘erential settling during slip casting and the resulting homogeneous
distribution of powders and whiskers results in a higher hardness value.

In general, the hardness of composites changes closely as density changes. The
higher density results in a higher hardness value. The processing which produces the best
densification of composites also produces the highest hardness. However, there is no
apparent relation between the fracture toughness and the density of the composites. The
fi‘acture toughness only depends on the volume percent and the length of the SiC
whiskers and their interaction with the alumina matrix. The fracture toughness values are
higher when the composites have higher non-ball-milled whisker loadings.

Figure 67 shows plots of Km and hardness of sintered composites versus the geen
dfifthsity of their geen bodies, respectively. The higher geen density results in better
densification that contributes to the higher hardness value. However, the higher geen

density is not a factor to achieve a higher fracture toughness value.

121

 

1 8 - + 10vol%
" -l— 20vol%

 

 

17~

 

16*

 

15-

 

 

 

14 ‘ «L

 

13‘

Hardness (GPa)

12-

11-

 

10 r r r
O 2 4 6

polyelectrolyte (vol%)

 

Figure 65 Hardness of 10 vol% Sij/A1203 as a function of polyelectrolyte
concentration (0.001N KN03, ball-milled, pH 11, 0.5 wt% MgO
and 2 wt% Y203, sintered at 1600°C for 1h, nitrogen atmosphere).

 

+ 5vol°/o
—D— 10vol%

 

 

 

20-
19“
18-
17‘
16“
15‘
14-
13-
12-
11-

10 —"l" T ‘1
30 40 50 60

Hardness (GPa)

 

 

solids loading (%)

Figure 66 Hardness of 10 vol% Sij/A1203 as a function of solids loading
(0.001N KNO3, non-ball-milled, pH 4, no polyelectrolyte and sintering aids,
sintered at 1600°C for 1h, nitrogen atmosphere).

 

+ Fracture toughness

 

 

 

 

 

 

-A-Hardness
1° ‘ __ 18
—> ~- 16
s" 8‘ - 14 E
E _. 12 g-
m 6 4 re
0- <— ‘- 10 a
é .. 2'3
2 4 - 8 ~e
x -- o e
2 . -~ 4
-- 2
o r r . 0
61 62 63 64 65
dg/du1 (%)

Figure 67 ch and hardness of sintered composites versus the green density of their green
bodies, respectively (5-20 vol% Sij/A1203, 0.001N KNO3, ball-milled,
le l, 2.5 vol% polyelectrolyte, 0.5 wt% MgO and 2 wt% Y203, sintered at
1600°C for 1h, nitrogen atmosphere).

124

CHAPTER 4 CONCLUSIONS

As the ethanol content increases, the absolute values of the zeta potential for both
' SiCw and A1203 decrease. The iso—electric point likewise shifts for Sij to more basic
pHs and to more acidic pHs for A1203 because ethanol does not ionize. This leads to a
decrease in the predicted stability for both materials, as verified by sedimentation
experiments. Ethanol and a mixture of ethanol and electrolyte reduce the dielectric
constant of both suspensions. The zeta potential decreases due to the compression of the
double layer. The stability ratio shows that the stability generally decreases as the
ethanol content increases. But at pH<7, stability predictions indicate the interactions
between Sij/Sij are more stable. This can be explained by assuming that the van der
Waal’s attractive energy, VA, is higher in ethanol than in water [20]. Green body
microstructure shows the SiCw and A1203 to be well dispersed in the electrolyte and in
the ethanol:electrolyte (1 :1). The partially sintered microstructure again shows good
dispersion of SiC...

Although the composite suspension is stable and codispersion of Sij and A1203 in
the electrolyte and in a mixture of ethanol and electrolyte is good, densification was
inhibited by the SiC whisker network effect. Comparing these observations with the
resmlts from the aqueous system, the sintered density of 10 vol% SiCaJAlzog composites
which were slip cast from a mixture of ethanol and electrolyte increased slightly fi-om
2-341008 g/cm3 (~73% ID) to 2.92:0.08 g/cm3 (~75% TD). This small change in

density after adding ethanol indicated that ethanol was not sufficiently able to eliminate

the formation of the whisker networks.

125

The combined use of different approaches has been proved successful in producing
dense SiCa/Ale3 composites from aqueous suspensions. First, whiskers were ball
milled to reduce the aspect ratio and to reduce the occurrence of whisker networks.
Second, sintering aids were added to create a gain bormdary liquid phase that relaxed the
back stresses resulting fi'om the network efiect. Third, a polyelectrolyte was used to
control the surface chemistry and to stabilize the mixed Sij/Ale3 suspension and
sintering aids, Y203 and MgO. Polyelectrolyte is able to reverse the zeta potential or
Shift the iso-electric point of alumina powders as shown by zeta potential results. For
nonaqueous suspensions, further research should be directed to creating compatible
processing ranges by choosing suitable sintering aids and an appropriate polyelectrolyte
to match the stable pH ranges of A1203 and Sij.
Homogeneous SiWAleg green bodies with densities of 2.55ro.07 g/cm3 (~65%
TD) were obtained. Bulk densities of 3.80aro.06 g/cm3 (96% TD), 3.79:0.06 g/cm3 (97%
TD), and 3.40:0.07 g/cm3 (89% TD) were obtained at 1600°C under nitrogen flow for
composite samples containing 5, 10 and 20 vol% SiC whiskers, respectively. Bulk
densities of the 10 vol% Sij/A1203 composites were 3.79i0.06 g/cm3 (97% TD) at pH
1 1 and 3.66i0.07 g/cm3 (94% TD) at pH 4, respectively. pH 11 was determined to be the
optimmn processing pH value for Sij/Al203 composites with sintering aids (2 wt%
Y203 and 0.5 wt% MgO) and 2.5 vol% polyelectrolyte (ammonia salt of polymeric
carboxylic acid). pH 11 was selected according to the compatibility for all the
Components and the stability for the suspension system. The final microstructures
l‘e‘w'ealed homogeneous and near fully densified composites containing 10 vol% SiC

whiskers (97% TD).

126

The efl‘ects of the SiC whisker aspect ratio, the content of Y203 and polyelectrolyte
were examined. The whisker aspect ratio reduction had a more dramatic effect on
achieving dense composites in 20 vol% SiC whisker in comparison to 5 or 10 vol% SiC
whisker composites. High densities with 10 vol% SiC whisker were attainable without
aspect ratio reduction. Sintered densities increased as Y203 content increased. For 5 and
10 vol% Sij/A1203, respectively, the densities remained almost constant at Y203
addition geater than 0.5 wt% and 2 wt%, respectively. 2.5 vol% polyelectrolyte
stabilized the suspensions and resulted in high densities of the composites. However,
further addition of polyelectrolyte past the adsorption saturation limit served to leave
excess polyelectrolyte in suspension and this excess polyelectrolyte lowered the densities
of the composites due to depletion flocculation.

The hardness of composites changes closely as density changes. The higher density
results in a higher hardness value. The processing which produces the best densification
of composites also produces the highest hardness. The fi'acture toughness only depends
on the vol% and the length of the SiC whiskers and their interaction with the alumina
matrix. The fracture toughness values are higher when the composites have higher non-
ball-milled whisker loadings. Although the hardness decreased as whisker loading
increased due to the whisker’s ability to inhibit densification, the fracture toughness
increased as the whisker loading increased. The fracture toughness of A1203 was
remarkably improved with increasing whisker content. The fracture toughness of the 20
vol% Sij/Ale3 composite was twice the fiacture toughness of the A1203 matrix. The

fracture toughness of the 20 vol% Sij/A1203 composite decreased from 73:02

lvrPa-m“2 to 6.35:0.2 MPa-mm after ball milling SiC whiskers.

127

REFERENCES

128

REFERENCES

[1] P. F. Becher and G. C. Wei, Am. Ceram. Soc. Bull., 63, C-267-69 (1984).
[2] G. C. Wei and P. F. Becher, Am. Ceram. Soc. Bull., 64[2], 298-304 (1985).
[3] T. N. Tiegs and P. F. Becher, Am. Ceram. Soc. Bull., 66[2], 339-42 (1987).

[4] R. Rice, p126-135 in Advanced Ceramic Processing and Technology, Edited by
J. G. P. Binner, Vol.1, NOYES PUBLICATIONS, Park Ridge, New Jersey, 1990.

[5] F. F. Lange, B. 1. Davis and I. A. Aksay, J. Am. Ceram. Soc. 66, 407-08 (1983).

[6] S. Lio, M. Watanabe, M. Matsubara and Y. Matsuo, J. Am. Ceram. Soc. 72[10],
1880-84 (1989).

[7] B. V. Deryaguin and L. Landau, Acta Physicochim. URSS, 14, 633 (1941).

[8] E. J. W. Verwey and J. T. G. Overbeek, Theory of the Stability of Lyophobic
Colloids, Elsevier, Amsterdam, 1948.

[9] J. T. G. Overbeek, in H. R Kruyt (Ed.), Colloid Science, Vol. I, Elsevier, Amsterdam,
1952.

[10] J. Lyklema, Advan. Colloid Interface Sci., 2, 65-114 (1968).
[l l] P. Sarkar and P. S. Nicholson, J. Am. Ceram. Soc., 79[8], 1987-2002 (1996).

[12] S. N. Heavens, p262 in Advanced Ceramic Processing and Technology, Edited by
J. G. P. Binner, Vol.1, NOYES PUBLICATIONS, Park Ridge, New Jersey, 1990.

[13] G. Gouy, J. Phys., 9(4), 457, (1910); Ann. Phys., 7(9), 129 (1917).

[14] D. L. Chapman, Phil. Mag, 25(6), 475 (1913).

[15] R. Hogg, T. w. Healy and D. w. Fuerstenau, Trans. Faraday Soc., 62, 1638 (1966).
[16] B. A. Wilson and M. J. Crimp, Langnuir, 9, 2836-43 (1993).

[17] B. v. Derjaguin, Kolloid-z, 69, 155 (1934).

[18] B. v. Derjaguin, Acta Physicochim. URSS, 10, 33 (1939).

[19] H. Kallmann and M. Willstatter, Naturwiss., 20, 952 (1932).

129

[20] A. Bleier, J. Am. Ceram. Soc., 66, C79 (1983).

[21] J. T. G. Overbeek, In Colloid Science Volume 1: Irreversible Systems; Elsevier,
New York, 1952.

[22] R. Hogg, T. W. Healy and D. W. Fuerstenau, Trans. Faraday Soc. 62,1638 (1966).

[23] R. J. Hunter, Foundation of Colloid Science, Vol.1, p419, Clarendon Press,
Oxford, 1987.

[24] R. J. Htmter, Zeta Potential in Colloid Science, Academic Press, New York, 1981.
[25] D. E. Brooks, J. Colloid Interface Sci. 43, 670, 687, 700, 714 (1973).

[26] P. C. Hidber, T. J. Graule and L. J. Gauckler, J. Am. Ceram. Soc., 78[7], 1775-80
(1995).

[27] R. w. O’Brien, J. Fluid Mech. 71, 190 (1986).
[28] R w. O’Brien, J. Fluid Mech. 81, 212 (1990).

[29] M. J. Crimp, Colloidal Characterization of Silicon Carbide and Silicon Nitride.
M.S. thesis, Case Western Reserve University, 1985

[30] J. Cesarano HI and I. A. Aksay, J. Am. Ceram. Soc., 71[f12] 1062-67 (1988).

[31] J. Homeny, W. Vaughn and M. K. Ferber, Am. Ceram. Soc. Bull., 66[2], 333-38
(1987).

[32] J. R. Porter, F. F. Lange and A. H. Chokshi, Am. Ceram. Soc. Bull., 66[2], 343-47
(1987).

[33] S. M. Smith, J. P. Singh and R. O. Scattergood, J. Am. Ceram. Soc., 76[2], 497-502
(1993).

[34] H. Liu, M. E. Fine and H. S. Cheng, J. Am. Ceram. Soc., 77[1], 179-85 (1994).
[35] T. N. Tiegs and D. M. Dillard, J. Am. Ceram. Soc., 73[5], 1440-42 (1990).
[36] C. Hu and M. N. Rahaman, J. Am. Ceram. Soc., 76[10], 2549-54 (1993).

[37] M. D. Sacks, H. Lee and o. E. Rojas, J. Am. Ceram. Soc., 71(5), 370-79
(1988)

[38] A. Takashi, T. Yasuhiro, M. Yohtaro and Y. Eiichi, J. of Ceram. Soc. of Japan,
100[l 167], 1297-303 (1992).

130

[39] C. Cheng-Chih, Y. Fu-Su, Y. S. C. Max, Proc. Third 93 Int. Ofi‘shore Polar Eng.
Conf. 380-85, 1993.

[40] A. L. Stuijts, p1 in Ceramic Microstructure’76; Edited by R. M. Fulrath and
J. A. Pask, Wesview Press, Boulder, 1977.

[41] G. Y. Onoda and L. L. Hench, Ceramic Processing Before Firing, Wiley,
New York, 1978.

[42] A. Roosen and H. K. Bonen, J. Am. Ceram. Soc.,71, 250 (1988).

[43] A. O. Boschi and E. Gilbart, p75 in Advanced Ceramic Processing and
Technology, Edited by J. G. P. Binner, Vol.1, NOYES PUBLICATIONS, Park
Ridge, New Jersey, 1990.

[44] Data provided by the manufacturers (T okai Carbon Co. of Tokyo, Japan,
Sumitomo Industries, New York and Rhone-Poulenc Basic Chemicals Company)

[45] B. A. Wilson, M. S. Thesis, “Prediction of Colloidal Suspension Stability for
SiC/Si3N4 and F eAl/A1203 Fiber Systems Using Material and System Parameters”,
Michigan State University, 1992.

[46] E. D. Case, Private Commrmication.

[47] G. R. Anstis, P. Chantikul, B. R. Lawn, and D. B. Marshall, J. Am. Ceram. Soc.,
64 [9], 533-43 (1981).

[48] A. G. Evans and E. A. Charles, J. Am. Ceram. Soc., 59 [7-8], 371-72 (1976).

[49] A. G. Evans, pp. 112-35 in Fracture Mechanics Applied to Brittle Materials.
Edited by S. W. Freiman. Am. Soc. Test. Mater. Spec. Tech. Publ. 678, 1978.

[50] J. P. Singh, K. C. Goretta, D. S. Kupperman, J. L. Routbort and J. F. Rhodes,
Advanced Ceramic Materials, 3[4], 357-60 (1988).

[51] D. W. Richerson, Modem Ceramic Engineering: Properties, Processing and
Use in Design, M. Dekker, New York, 1992.

[52] B. 1. Lee and U. Paik, Ceramic International, 19, 214 (1993).
[53] W. H. Tuan, E. Gilbart and R. J. Brook, J. Mat. Sci. 24, 1062-68 (1989).
[54] S. J. Barclay, J. R. Fox and H. K. Bowen, J. Mat. Sci. 22, 4403-06 (1987).

[55] T. Garino, J. Am. Cer. Soc. 75[3], 514-18 (1992).

131

 

[56] R. K. Bordia and G. W. Scherer, pp872-86 in Ceramic Transactions, Vol. 1,
Ceramic Powder Science 11, B. Edited by G. L. Messing, E.R. Fuller, and
H. Hausner. American Ceramic Society, Westerville, OH, 1988.

[57] F. F. Lange, J. Mater. Res., 2[1], 59-65 (1987).

[58] J. V. Milewski, Adv. Ceram. Mater., l[1], 36-41 (1986).

[59] K. T. Faber and A. G. Evans, Acta. Metall., 31[4], 565-76 (1983).

[60] K. J. Konsztowicz, Ceram. Eng. Sci. Proc., 11[7-8], 695-708 (1990).

[61] O. Sudre and F. F. Lange, J. Am. Ceram. Soc., 75[3], 519-24 (1992).

[62] O. Sudre, G. Bao, B. Fan, F. F. Lange, and A. G. Evans, J. Am. Ceram. Soc.,
75[3], 525-31 (1992).

[63] E. M. Levin, C. R. Robbins and H. F. McMurdie, Phase Diagams for Ceramists,
The American Ceramic Society, Ohio, pp165, Figure 2586, 1969.

[64] K. n Karasek, s. A. Bradley, J. T. Donner, H. C. Yeh, J. L. Schienle and
H. T. Fang, J. Am. Ceram. Soc. 72(10], 1907-1309119).

[65] T. Graule, P. Hidber, and L. J. Gauckler, pp. 299-305 in Euro-Ceramics, Vol. 11,
Basic Science and Processing of Ceramics. Edited by G. Ziegler and H. Hausner.
Deutsche Keramische Gesellschafi e. V., Koeln, Germany, 1993.

[66] Information provided by the manufacturer (Rohm and Haas Company).

[67] CRC Handbook of Chemistry and Physics, ed. by D. R. Lide, CRC Press, 1990

[68] P. Hidber, T. Graule, and L. J. Gauckler, pp. 247-54 in Handbook on
Characterization Techniques for the Solid-Solution Interface. Edited by J. A. Adair,
J. A. Casey, and S. Venigalla. American Ceramic Society, Westerville, OH, 1993.

[69] Y. K. Back and C. H. Kim, J. Mat. Sci., 24, 1589-93 (1989).

[70] S. M. Smith, J. P. Singh and R. O. Scattergood, J. Am. Ceram. Soc., 76[2], 497-502
(1993).

[71] P. C. Hidber, T. J. Graule, and L. J. Gauckler, J. Am. Ceram. Soc.,
79[7], 1857-67 (1996).

[72] Dept. of Materials Science and Mechanics, Michigan State University,
MSM 250 Lab Manual, 1995.

132

[73] M. D. Sacks, H. W. Lee and O. E. Rojas, Ceram. Eng. Sci. Proc. 9 [7-8],
741-54 (1988).

[74] H. W. Lee and M. D. Sacks, J. Am. Ceram. Soc., 73[7], 1894-1900 (1990).

133

MICHIGAN $7an UNIV. LIBRQRIES
11111111111111111111111111111111111111111111111
31293018017313