IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII l M 0 79> ”I 5
fly LIBRARY“-
Michigan State
University
This is to certify that the
thesis entitled
FRACTURE TOUGHNESS AS A FUNCTION OF PRE—EXISTING
MICROCRACKS IN YTTRIUM CHROMITE
presented by
TAE—GYO SUH
has been accepted towards fulfillment
of the requirements for
M.S. . Materials Science
degree 111
Major professor
07639 MSUis an Affirmative Action/Equal Opportunity Institution
IV1ESI_J RETURNING MATERIALS:
Place in book drop to
LJBRARJES remove this checkout from
—:——- your record. FINES will
be charged if book is
returned after the date
stamped below.
"’"VH'UHW DIAII- IlMlUEDoa-ru . n»- . _
FRACTURE TOUGHNESS AS A FUNCTION OF PRE—EXISTING
MICROCRACKS IN YTTRIUH CHROMITE
By
TAE—GYO SUH
A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
MASTER OF SCIENCE
Department of Metallurgy, Mechanics and Materials Science
1988
ABSTRACT
FRACTURE TOUCHNESS AS A.FUNCTION OF ERR-EXISTING
MICROCRACKS IN YTTRIUH CHROHITE
BY
TAB-GYO SUH
The fracture toughness and hardness of polycrystalline yttrium
chromite was experimentally measured as a function of microcrack density
parameter of pre-existing microcracks.
Elasticity measurements were performed by the sonic resonance
technique. The microcrack density parameters were determined by the
microcrack-elastic modulus theories from elastic moduli of
nonmicrocracked and microcracked polycrystalline yttrium chromite.
Fracture toughness and hardness were determined from the Vickers
indentation technique.
Fracture toughness of polycrystalline yttrium chromite decreased
linearly with microcrack density parameter. Fracture toughness changed
as a function of microcrack density parameter in a fashion that seems
consistent with microcrack link-up. Hardness decreased linearly with
increasing the microcrack density parameter.
ACKNOWLEDGEMENTS
I wish to express my deep gratitude to Dr. Eldon D. Case for his
advise and guidance in this study. I would like to thank the Division
of Engineering Research for providing the funds for this research.
I should express my sincere gratitude to my parents for their
encouragement and support. Finally I would like to thank my wife for
her understanding and encourgement, my brother and sisters for their
encouragement.
iii
TABLE OF CONTENTS
LIST OF TABLES
LIST OF FIGURES
1. INTRODUCTION
2. EXPERIMENTAL PROCEDURE
2.1 Specimens
2.2 Elasticity Measurements
2.3 Fracture Toughness Measurements
2.4 Annealing
3. THEORETICAL REVIEW
3.1 Theories Relating Dcrements in Young's Modulus to the
Microcrack Density Parameter
3.2 The Formation of Microcracks
3.3 Studies of Fracture Toughness of Microcracked Materials
4. RESULTS AND DISCUSSION
4.1 Experimental Results
4.2 Comparison of Fracture Toughness between Experimental Data
and Theoretical Equation from 3.3
5. CONCLUSIONS
6. REFERENCES
APPENDIX 1
APPENDIX 2
iv
Page
vi
14
20
23
23
27
32
37
37
58
64
66
7O
71
.— —.-.d-'u'
Table
LIST OF TABLES
The chemical compositions and dimensions of initial
yttrium chromite specimens
Mechanical properties of initial yttrium chromite
specimens
Polishing procedure used for each yttrium chromite
specimens
Annealing temperatures and times used in order to heal
microcracks in yttrium chromite specimens
Crack density parameter and mechanical properties for the
yttrium chromite specimens
The comparison of elastic moduli of unpolished specimens
with those of polished specimens which were adhered to
the mounting aluminium plate and heated at 3500 C for
1 hour to remove the plate after polishing
The crack density parameter without porosity correction
and the crack density parameter after porosity correction
Hardness, fracture toughness and Young' s modulus/hardness
ratio for each yttrium chromite specimen at each value of
indentation load
Page
16
21
38
39
41
44
Figure
10 (a).
10 (b).
11 (a).
11 (b).
12 (a).
12 (b).
LIST OF FIGURES
Page
Schematic of the experimental appatatus of sonic resonance
technique 10
Specimen suspension method for the sonic resonance
technique 11
Optical micrograph of surface of polished specimen l7
Indentation impression and radial crack geometry for
Vickers indentation 19
Fracture toughness versus crack density parameter according
to microcrack link-up and branching models 29
Schematic of the possible macrocrack-microcrack interaction
that could lead to microcrack link-up 30
Schematic of the possible macrocrack-microcrack interaction
that could lead to microcrack branching 31
Semi-infinite main crack collinear with two dimensional
microcrack 33
Young's modulus versus crack density parameter with porosity
correction 43
Hardness versus crack density parameter without porosity
correction for yttrium chromite. Error bars indicate
i 8 range where S - standard deviation. 47
Hardness versus crack density parameter with porosity
correction for yttrium chromite. Error bars indicate
i S range where S - standard deviation. . 48
Fracture toughness versus crack density parameter without
porosity correction for yttrium chromite. Error bars
indicate 1 S range where S - standard deviation. 49
Fracture toughness versus crack density parameter with
porosity correction fot yttrium chromite. Error bars
indicate i 8 range where S - standard deviation. 50
Young’s modulus/hardness ratio versus crack density parameter
without porosity correction for yttrium chromite. Error
bars indicate i S range where S - standard deviation. 53
Young's modulus/hardness ratio versus crack density parameter
with porosity correction for yttrium chromite. Error bars
indicate i S range where S - standard deviation. 54
vi
Figure
l3.
14.
15.
16 (a).
16 (b).
17 (a).
17 (b).
Page
Optical micrograph of yttrium chromite specimen Y0 indented
at 5.88N 55
Optical micrograph of yttrium chromite specimen Y4 indented
at 9.8N 56
Optical micrograph of yttrium chromite specimen YP21
indented at 4.9N 57
Fracture toughness as a function of microcrack density
parameter without porosity correction, e, for the
range 0 s e S 1 for three theoretical expressions
adapted from Rose (equations (40-42)). Exprermental
data on yttrium chromite is also included. Error bars
indicate i S where S - standard deviation. 60
Fracture toughness as a function of microcrack density
parameter with porosity correction, e, for the range
0 s e S 1 for three theoretical expressions adapted from
Rose (equations (40—42)). Experimental data on yttrium
chromite is also included. Error bars indicate i S where
S - standard deviation. 61
Fracture toughness as a function of microcrack density
parameter without porosity correction, e, for the
range 0 S e s 0.5 for three theoretical expressions
adapted from Rose (equations (40-42)). Experimental
data on yttrium chromite is also included. Error bars
indicate i S where S - standard deviation. 62
Fracture toughness as a function of microcrack density
parameter with porosity correction, e, for the range
0 s e s 0.5 for three theoretical expressions adapted
from Rose (equation (40-42)). Experimental data on yttrium
chromite is also included. Error bars indicate i S where
S - standard deviation. 63
vii
1. INTRODUCTION
Polycrystalline ceramic materials often exhibit excellent
refractory properties and resistance to chemical corrosion, but the
brittle nature of ceramics can lead to catastrophic failure under
thermal or mechanical loading. For many ceramics, the increase in
fracture toughness means an improved chance of producing viable
components with the materials. Hence considerable research effort has
been devoted to identifying and understanding physical mechanisms that
increase Kc’ the fracture toughness (also called the critical stress
intensity factor), which is a measure of the resistance to crack growth.
One mechanism for increasing fracture toughness is crack shielding.
In the crack shielding process, the microstructure of the specimen is
locally altered by the crack tip stress, which in turn changes the crack
tip stress field, effectively shielding the crack for the applied
loading [1]. Two examples of crack shielding mechanisms are martensitic
transformation and stress induced microcrack zone toughening [2,3].
Grains in a polycrystalline ceramic undergo a martensitic transformation
under the influence of localized crack tip stress fields [4-10].
However, toughening of ceramics by martensitic transformation is
limited, since the requisite martensitic transformation has been
documented only for ZrO2 [2-7] and HfO2 [9,10] with some limited
evidence that cordierite also may transform martensitically [10]. For
the development of a stress induced microcrack zone, localized stresses
(on the scale of the grain size) must be present, such as those induced
by thermal expansion mismatch in anisotropic ceramics [ll-l3].
In addition to crack shielding mechanisms, ceramics may be
toughened by a variety of crack interaction mechanisms, in which a
moving crack may interact directly with preexisting microstructural
features such as second phases or voids, where the increase in toughness
results from crack pinning or deflection by the second phase particles
[14-17].
A moving crack also may interact with pre-existing microcracks.
The interaction of a moving crack and a population of pre-existing
microcracks may be viewed in terms of the following two limiting cases.
First, microcracks could "link-up" ahead of the propagating crack,
dropping the fracture toughness by decreasing the fracture surface area
formed by the moving crack. This might be case if 1) the microcracks
were highly oriented, so that vectors drawn normal to the microcrack
planes point in a single direction, and 2) the plane of the moving crack
was coincident with the microcracks. Second, microcracks ahead of the
moving crack might produce multiple crack branching. In this case, the
fracture toughness could increase due to the contribution of the small,
branching cracks splitting off the moving macrocrack. This might occur
if each of the preexisting microcracks was oriented favorably for crack
branching (perhaps, for example, with the microcracks aligned
approXimately normal to the plane of the moving crack). In practice,
one would expect a polycrystalline ceramic to exhibt both microcrack
link-up and microcrack induced branching.
Residual stress fields existing near the tips of stationary
macrocracks tend to deflect the path of a moving crack away from the tip
of a nearly coplanar stationary crack [18]. A similar crack deflection
mechanism presumably would be operative for stationary microcracks, but
due to the size of the microcrack in ceramics (often microcrack radii
are of the order of the grain size [19]) direct observation of this
deflection mechanism would be experimentally difficult. Nevertheless,
the existence of a residual stress field near microcracks [6] and the
existence of a crack deflection mechanism for stationary moving
macrocrack pairs would indicate that such a mechanism ought to be
considered.
Both the crack deflection mechanism (that results from residual
stress fields) and the random orientation of microcracks would act to
suppress microcrack link-up. Thus a population of pre-existing
microcracks may increase the fracture toughness of a polycrystalline
ceramic by the enhancement of crack branching.
In order to study the effect of the pre-existing microcracks upon
fracture toughness, one needs a ceramic system in which the number of
preexisting microcracks can be varied in a systematic manner while other
microstructural features such as grain size and void size are unchanged.
For ceramics that microcrack due to thermal expansion anisotropy (TEA)
[11-13], the microcrack number density is a function of grain size. In
order to increase the microcrack number density in a ceramic that cracks
due to TEA, one can increase the grain size by annealing. However, a
grain growth anneal not only changes the grain size and morphology, but
also changes the void size, shape, and void fraction, along with factors
that affect grain boundary strength, such as solute segregation [20].
YCrO3 is, apparently, one ceramic material that can serve as an
appropriate model material in this study. Yttrium chromite is an
orthorhombic perovskite at room temperature. Yttrium chromite can be
sintered from powders at a sufficiently high temperature in a
environment with a low oxygen partial pressure. The low oxygen partial
pressure is required to prevent the volatilization of chromia during
sintering (For example, specimens prepared for this study were sintered
in flowing forming gas at 17500C.). Upon cooling, yttrium chromite
undergoes a phase transition at llOOOC which induces microcracking. By
annealing at a temperature lower than 110000, the microcracks can be
healed, as confirmed by elasticity and small angle neutron scattering
studies [19,21]. However, a subsequent thermal anneal above 11000C can
induce microcracking again. Also, it has been shown recently that the
microcrack number density in polycrystalline yttrium chromite can be
varied without altering other microstructural features such as the
porosity and the grain size [19,21,22]. Therefore, these properties in
yttrium chromite make it an excellent model material for studying the
change in a material parameter, such as fracture toughness, as a
function of the microcrack damage level.
The goal of this study is to determine the effect, if any, of pre-
existing microcracks on the fracture toughness and hardness of
polycrystalline yttrium chromite. In order to determine the level of
microcrack damage, measurements of elastic moduli were related to
microcrack parameters via known microcracking-elastic modulus theories.
The fracture toughness of the polycrystalline yttrium chromite specimens
also was measured as a function of the microcrack damage level.
Elasticity measurements were performed by the sonic resonance
technique. The microcrack density parameter of the specimens was
determined using equations (14) and (16) of microcrack-elastic modulus
theories in Theoretical Review from Young's modulus for microcracked
material, Young's modulus and Poisson's ratio for nonmicrocracked
material.
Fracture toughness was determined from the Vickers indentation
technique. A pyramidal shaped indentation impression was formed in the
specimen by a pyramid shaped diamond. Impression size and crack size
relate to hardness and fracture toughness, respectively. Hardness was
computed using equation (6) in Experimental Procedure from peak
indentation load and impression size. Fracture toughness was computed
by using equation (7) from Young’s modulus and hardness of the
specimens, the peak indentation load, and the post-indentation crack
length.
Briefly, this study found that the fracture toughness of
polycrystalline yttrium chromite decreased linearly with e, the crack
density parameter. Fracture toughness changed as a function of crack
density parameter changed in a fashion that seems consistent with
microcrack link-up ahead of a main crack. Hardness and Young's
modulus/hardness ratio were also studied as a function of crack density
parameter.
2. INEERIMENTAL PROCEDURE
2.1 Specimens
The chemical compositions and dimensions of initial rectangular
yttrium chromite bar specimens are listed in Table 1 and the mechanical
properties are listed in Table 2. The dopants were added to change
yttrium chromite from an insulator to a semiconductor as part of a
magnetohydrodynamic research project. Samples were fabricated by Trans-
Tek Inc., Adamston, MD, from intimately mixed powders of chromium and
yttrium [23]. The specimens were isostatically pressed at 207 MPa
(30,000 psi) [24] and sintered at 17500C in flowing forming gas (a 95%
nitrogen and 5% hydrogen mixture) in an electric furnace [19]. The
average particle diameter for the starting powder was 1.3 microns as
determined by sedimentation methods [23]. Details of powder mixing and
calcining processes are given elsewhere [24]. The average grain size of
sintered materials was approximately 6.0 microns, as determined from the
linear intercept technique on scanning electron micrographs of fracture
surfaces [25,26]. The density of green compacts was approximately 57%
of the theoretical density. X-ray diffraction analysis confirmed that
the specimens were orthorhombic (distorted perovskite structure) at room
temperature.
2.2 Elasticity Measurements
Elasticity measurements were performed by the sonic resonance
technique at room temperature in air. This technique was originated by
Table
l. The Chemical compositions and dimensions of initial
yttrium chromite specimens.
Specimens Component Mass Length Width Thickness Mass Density
(gm) (Cm) (Cm) (Cm) (cm/ems)
YPlO YCrOa 3.303 5.849 0.830 0.131 5.19
YPll YCrO3 3.447 5 849 0.830 0.135 5.16
Y0 YCao-OSCrO3 3.987 4.477 1.298 0.123 5.58
Y4 2.505 5.122 0.672 0.128 5.69
YCao o2oro3
Table 2. Medhanical properties of initial
yttrium Chromite specimens.
Specimen Young's Shear Poisson's Bulk
modulus modulus ratio modulus
(GPa) (GPa) (GPa)
YP10 66.9 29.2 0.146 31.5
YPll 61.2 27.3 0.122 26.9
Y0 245.5 98.5 0.246 161.4
Y4 260.5 105.9 0.236 160.8
Forster [27]. The experimental technique has been discussed in detail
by Spinner and Tefft [28]. A schematic of the experimental apparatus is
shown in Figure 1. A 2325A Synthesizer/Function Generator* was used to
generate a known sinusoidal electric signal which was converted into a
mechanical vibration of the same frequency by a high power model 62-1
piezoelectric driver transducer**. This mechanical vibration was
transmitted to a suspended specimen through cotton thread and conveyed
to the pick-up transducer through another suspension cotton thread at
the other end of the specimen. The mechanical vibration was reconverted
to an electrical signal which was amplified, filtered by 4302 Dual
240B/Octave Filter Amplifier***, and passed into an 8050A Digital
Multimeter#. The digital voltmeter aided in detecting the resonant
condition by giving a value of the amplitude at the pick up. A V-1100A
Oscilloscope## gave a visual indication of the amplitude on the screen,
so that the resonant frequency (maximum amplitude) could be estimated.
Precise determination of the resonant condition was made using the
digital voltmeter. The method of suspending specimens is shown in
Figure 2 [28].
Nodal points are the positions of zero displacement in the
vibration direction for each vibration mode. Nodal points were
determined to obtain the fundamental flexural and torsional resonant
frequencies. The fundamental flexural vibration has two nodal points,
*Hewlett Packard, Palo Alto, CA.
**Astatic Corporation, Conneaut, Ohio
***Ithaco, Ithca, NY.
#Fluke, Everett, WA.
##Hitachi, Tarrytown, NY
10
Oscilloscope
Frequency
Synthesizer
Driver
Voltmeter
Filter
Amplif
ier
Pickup
Specimen‘
Figure 1. Schematic of the experimental appatatus of sonic resonance
technique
11
Driver Pickup
Cotton
Support
Thread
./ W
Specimen
Figure 2. Specimen suspension method for the sonic resonance
technique
12
one located at 0.224 L and the other node located at 0.776 L where L is
the specimen length. The fundamental torsional vibration has one node
point at the center. When a high signal amplitude was obtained at a
frequency, a needle was put perpendicular to the length of the bar
specimen keeping the frequency. Then, the amplitude was recorded as the
position of the needle. When the amplitude at a frequency without a
needle was the same to the amplitude at the frequency with a needle at a
position, the position is the node point. The node points were
determined at the frequencies of each high amplitude in this way. If
the node points at a frequency with high amplitude are the same as the
node points of the fundamental flexural vibration, the frequency is the
fundamental flexural resonant frequency. Similarly, the fundamental
torsional resonant frequency was obtained. There may be a slight shift
of frequency when the thread is far-from the position of node. So, the
thread was positioned close to the nodes to obtain accurate results.
Elasticity measurements were made by varying the oscillator frequency
until the suspended specimen vibrated in a mechanical resonance
vibration. At the resonance condition, the amplitude of vibration
reached a maximum which was measured by the voltmeter and observed by
the osilloscope [29]. The flexural and torsional resonance frequencies
of each specimen were determined via the following equations.
Young's modulus, shear modulus and Poisson’s ratio were caculated
from the flexural and torsional frequency using the mass and dimension
of specimen [28,30]. Equation (1) relates Young's modulus and flexural
frequency for prisms of retangular cross section [30].
4 2 2
E = 0.94642d1 f T/t (1)
where
The correction factor
T = 1 + 6.585
2 4
8.340 ( 1 + 0.2023v + 2.173V .) (t/l)
' 2 2
1 + 6.338 ( 1 + 0.1408u + 1.53u ) (t/l)
where
13
= Young's modulus
- the mass density of specimen
= the length of specimen
- flexural resonant frequency
- cross-sectional dimension in the direction of
plane of vibration
= correction factor for the prisms of rectangular
cross section
T in equation (1) is in turn given by
2 2 4
(1 + 0.0752V + 0.819u ) (t/l) - 0.868 (t/l)
(2)
V = Poisson's ratio.
Equation (3) relates shear modulus to the torsional resonant
frequency of the specimen.
where
a ( 21f/n)2R (3)
shear modulus
the length of specimen
torsional resonance frequency
the order of the vibrational mode, which is unity
14
for the fundamental mode, two for the first
overtone, etc.
R = a shape factor which is a function of the shape of
the cross section of the specimen.
For the prisms of rectangular cross-section, the shape factor R in
equation (3) is given [28]
2 2 2
R = 1 + ( b/a1) ( 1+ 0.00851n b )
- 2
4 - 2.521(b/a) (1 - eV) 1
3/2 2
- 0.060 («b/l) ( b/a - 1 ) (4)
where v = 1.991
Hb/a
a and b the cross-sectional dimensions of prismatic specimens,
with restriction b g a.
Poisson's ratio for a homogeneous isotropic body is caculated by
V = E - l (5)
Using equation (5), one can calculate Poisson's ratio from
experimentally determined values of shear modulus and Young's modulus.
2.3 Fracture Toughness Measurements
Fracture toughness was determined from Vickers indentation
technique. Specimens were adhered to a 6.2 x 3.1 x 1.1 cm aluminium
15
plate with Super Glue (Super Glue Corp., Holn, NY) in order to make it
easier to hold the small specimens during the polishing process.
Specimens were polished using 240, 320, 400 to 600 grit silicon carbide
polishing papers (Mager Scientific Inc., Dexter, MI). Polishing was
continued using rotating polishing machines with 5, 0.3 and 0.05 micron
aluminium oxide powders (Mager Scientific Inc., Dexter, MI). Table 3
shows the typical time for each grit employed during thepolishing
procedure for each specimen. Total polishing time for an individual
specimen was about about 5 hours. Polished specimens had very small
flaws or pores as observed at 800 magnification by the Neophot 21
Optical Microscope (Leco, Warreendale, PA). An optical micrograph of
the surface of a polished specimen is shown in Figure 3.
After polishing, specimens were removed from the aluminium plates
by heating the specimens and aluminium plate in the wire-wound
resistance furnace (type 59344 Lindberg) at 350°C in air for 1 hour.
This heat treatment caused the glue to decompose to the point where the
specimens could be easily removed.
Elastic moduli were obtained for the polished specimens using the
sonic resonance technique. Elastic moduli were compared with those of
the polished specimens which were heated to remove the aluminum plate at
3500C for 1 hour (see Result and Discussion).
The specimens were again glued to an aluminium plate with the super
glue in order to perform the Vickers indentation test. The Vickers
indentation test employs a pyramid-shaped diamond, which in turn
produces a pyramidal-shaped indentation impression in the specimen.
Radial cracks typically extend from the corners of the intent
impression and propagate across the surface of the specimen (see Figure
4). The impression size and crack size relate to the hardness and the
Table 3. Polishing procedure used for each yttrium chromite specimen
Silicon carbide polishing Alumina polishing
paper (grit) powder (microns)
Specimen
240 320 400 600 600 5 .3 0.05
grit
(minutes) (minutes)
YP10 x x 10 10 20 30 80 100
YPll x x 10 10 20 30 80 100
Y0 10 20 20 20 30 50 60 80
Y4 10 20 20 20 30 50 60 80
x means that polishing was not done at this condition
17
100 microns
Figure 3. Optical micrograph of surface of polished specimen
18
fracture toughness of the indented specimen and to the applied load.
Figure 4 gives a schematic of the indentation impression and the
associated radial cracks. The hardness is computed from [31,32]
2
H = 0.47P / a (5)
where H = hardness
P = peak indentation load
2a = the length of the diagonal of the indent
impression
The fracture toughness is given by [33,34]
1/2 3/2 '
K = A ( E / H ) ( P/ C ) (7)
where A = materials independent calibration constant and
0.016 i 0.0004 for Vickers indentation [32]
E = the Young's modulus of the material
H a the hardness of the material
P = the peak indentation load
2c - the post-indentation crack length
The Vickers indentation test was done for yttrium chromite
specimens using the Ser. No. DV-5987 semi-macro indentor of Buehler LTD,
Lake Bluff, IL. Crack length and impression size were measured from the
same apparatus. The crack length after Vickers indentation test should
be immediately measured for the valid value of fracture toughness [32]
because the crack length can change as a function of time after the test
19
I
I---
Figure 4. Indentation impression and radial crack geometry for
Vickers indentation
20
owing to "the slow crack growth and crack-microstructure interaction."
The crack length and impression size after the indentation test were
measured within 20 and 40 seconds, respectively. The change of the
crack length after the indentation test was so small that it was
difficult to measure using the Ser. No. DV-5987 semi-macro indentor of
Buehler LTD, Lake Bluff, IL.
2.4 Thermal Annealing
Thermal annealing changed the microcrack damage state of the
specimens. The annealing times and temperatures are listed in Table 4.
The mirocracked yttrium chromite specimens were healed by thermal
anneals below the temperature of yttrium chromite phase transition
temperature (11000C). The thermal anneals were performed in a wire
wound resistance furnace of Lindberg type 59344 which had a maximum
temperature capability of 12000C. The inside dimension of furnace was
40 x 20 x 15 cm.
The furnace used for heat treatment was lined with refractory
bricks, so that precautions must be taken to assure that the specimen is
not contaminated during the heating process by either the floor of the
furnace or by "dusting" of the refractory brick onto the specimen. The
surface of refractory bricks can degrade during heating, by mechanical
and/or chemical processes. If the particles of "dust" freed by the
degradation falls onto the specimen, the specimen can be contaminated
during the annealing process. To minimize such contamination, the
following protection was provided for the specimen. The specimens were
annealed on a 5.2 x 1.0 x 0.1 cm alumina setter. The setter was placed
on a refratory brick of 22.5 x 11.3 x 6.2 cm dimension. The specimens
21
Table 4. Annealing temperatures and times used in order to heal
microcracks in yttrium chromite specimens
Specimen Specimen before annealing Temperature Time
0
( C) (hours)
YP10 YP10 x x
YP21 YP10 1000 12
1000 18
YP22 YPll 1000 18
1020 18
YP23 YP21 1035 50
1070 20
YP24 YP22 1035 20
1095 20
Y0 Y0 x x
Y4 Y4 x x
x means that thermal annealing was not done at this condition.
22
were covered with an alumina boat of 4.8 x 1.1 x 0.9 cm dimension. The
specimens, inside the alumina boats, were positioned 5 cm away from the
thermocouple.
The furnace temperature was slowly increased and decreased, step by
step, to prevent thermal shock damage. Heat treatment was done
according to the following schedule: 30 minutes at 300°C, 30 minutes at
500°C, 30 minutes at 700°C, 30 minutes at 9000C, and at the assigned
temperature and for the assigned time in Table 4. Cooling was done
according to the following schedule: 40 minutes at 900°C, 1 hour at
700°C, 2 hours at 500°C. The furnace power was then turned off, and the
furnace was allowed to cool for at least 3 hours before the door of the
furnace was opened. The temperature of the furnace was about 170°C when
the door was opened.
3. THEORETICAL REVIEW
3.1 Theories Relating Decrements in Young's Modulus to the Microcrack
Density Parameter
The microcrack damage state of a polycrystalline is related to
changes in the elastic moduli that accompany microcracking [21,35-38].
The studies of Walsh [35], Salganik [36], Budiansky and O'Connel [37],
Hasselman and Singh [38], and Kemeny and Cook [39] proposed that the
elastic moduli of a microcrocracked body is a function of e, the crack
density parameter.
3 .
e = N (8)
where e = crack number density parameter
N = microcrack number
= mean microcrack radius
These theories each predict a similar microcrack density-modulus
decrement behavior [21]. These theories [35-39] assumed a homogenous
isotropic body with a number density N of randomly oriented cracks of
mean radius . Walsh treated the case of elliptical cracks, Salganik
treated the disk shaped cracks, Budiansky and O'Connel treated
rectangular and circular cracks.
Kemeny and Cook assumed that a linear elastic, isotropic,
homogeneous material contains a random distribution of flat, open cracks
or external cracks where "external cracks" are identical to surface
23
24
breaking cracks. Kemeny and Cook treated the flat, external crack in
two dimensions, and penny-shaped crack and external cracks in three
dimensions. The relation between the intrinsic Young's modulus and the
effective Young's modulus was given for a solid containing microcracks,
loaded under a uniaxial stress a [40] (in an expression similar to that
of Budiansky and O'Connel [37]).
- 02V 02V + Ad (9)
21'0 2y
where Ad
the increase in strain energy due to the presence
of the voids (cracks or external cracks in this case)
V = the volume of the body containing the cracks
Y0 = intrinsic Young's modulus
Y = effective Young's modulus
Kemeny and Cook utilized an axisymetric extention of Irwin's
relation [41] between the energy release rate to extend a crack and the
crack tip stress intensity factors K1, K2 and K3. For a penny-shaped
crack in three dimensions, the strain energy and the stress intensity
factors are related by
2
2 c
- 2 2
U = (1 V0) |O[K1 + K2 + K3 ]2flCdC (10)
6 Y (1—u0)
where Ue = the additional strain energy due to a single crack
of length 2c in a three dimensional elastic body
V0 = the intrinsic Poission's ratio
25
K1, K2 and K3 = the crack tip stress intensity for
opening, shearing, and tearing modes of deformation,
respectively
2c a the length of penny-shaped crack
The stress intensity factors taken from Rice [42] are given by
2 1/2
K1 = 2 a sin 7(flC)
n
1/2
K2 - 4 a sinycosycosw(nc) c< a = 1/2,
2 2
7cos 7> = 2/15. Assuming that there are N penny
4
- 1/5 and , Kemeny
and cook gave
2 3 2 -
Ue a 8 Na ( 1-00) 10 3”° (12)
45Y 2 - V0
where Ue = the total strain energy due to a random
distribution of penny-shaped cracks
N = the number of penny-shaped cracks
a mean crack radius cubed
26
In an appendix 1, equation (12) is derived using equation (10) and K1,
K2 and K3 values from equation (11).
Substituting equation (12) in equation (9), Kemeny and Cook gave
the following relation
Y = Y, [ l-f(vo)e 1
where Y
the Young's modulus for a microcracked material
(13)
Y0 = the Young's modulus for a nonmicrocracked material
v0 = the Poisson ratio for a nonmicrocracked material
6 - crack density parameter
The function f(vo) is given by [36]
16(10-3vo)(l-V0)
f(vo) =
45(2-uo)
From equations(8) and (13)
3
AY = Y0 Y = f(uo)N
Yo
That is
Yo'Y
3
e = N = Y0 = AY
f(Vo) f(Vo)
From equation (16), the microcrack damage state of a specimen can be
(14)
(15)
(16)
determined from the change of Young's modulus that is induced by thermal
27
annealing. Solving for 6 uses the experimentally determined values of
Y, Yoand V0-
The experimental determination of the quantities N and 3 is
difficult. For example, when a specimen is observed in a microscope,
the microcrack state of the specimen surface can be different from that
of specimen bulk. The differences in the microcrack state can result
from surface damage by the manufacturing process. Also, the stress
state of a specimen surface is different from that of bulk. It is very F
difficult to experimentally determine both the inherent microcrack
number N and inherent microcrack size . However, the microcrack
density parameter can be determined directly from the decrement in
modulus, as shown in equation (16).
3.2 The Formation of Microcracks
Microcracks can be formed in a number of ways, including thermal
expansion mismatch and phase transformation. One example of thermal
expansion mismatch is thermal expansion anisotropy of a noncubic
material. Noncubic materials have different thermal expansion
coefficients along the crystalline axes. As a polycrystalline body
cools from high temperature, thermal expansion anisotropy can lead to
localized stresses that in turn induce stored elastic strain energy.
This stress can result in internal microcracking in brittle materials
[11,13,43]. Microcracking due to thermal expansion anisotropy can be
very pronounced for specimens with grain sizes larger than a critical
grain size, while specimens with a grain size smaller than the critical
grain size do not microcrack [ll-13, 43].
28
Thermal expansion mismatch also can occur in multi-phase materials
[12]. The stress owing to the difference of thermal expansion
coefficients between phases causes microcracks. Microcracking in multi-
phase materials is a function of the thermal expansion coefficient,
volume fraction, particle size, and elastic properties of each phase
[12]. Microcracks from the thermal expantion mismatch typically form at
the grain boundaries [44].
Microcracks also can result from the localized internal stresses
due to rapid volume changes of grains from phase transformation. Twins
are frequently formed by phase transformation in ceramics. Such twins
can reduce the macroscopic shape change of the transforming particle or
grain. Shear strains form at the twinned regions [45], so the large
stress concentrations can occur at twin interfaces and stress results in
microcrack nucleation.
The interaction of a moving crack with pre-existing microcracks can
be considered in terms of the following two cases.
(1). Microcracks may link-up ahead of a moving crack. Then, fracture
toughness may decrease as a function of increasing the microcrack number
density (Figures 5 and 6) by decreasing the fracture surface area formed
by the moving crack. Microcrack link-up may occur if a) the microcracks
are highly oriented, so that the vectors drawn normal to the microcrack
planes point in a single direction, or b) the plane of the moving crack
is coincident with microcracks.
(2). Microcracks ahead of the moving crack may produce multiple crack
branching. In this case, fracture toughness may increase with
increasing microcrack number density (see Figure 5 and 7) due to the
contribution of small, branching cracks splitting off the moving
29
mHopoE wcHsocmun can as-xcHH xomuoouows ou
maficuooom Homosmumm huHmSop xomuo msmuo> mmocnwsou ousuomum .m unawam
teem-2.1.5.101». mmufififi 1030
I 1] I: t. I |.q' Iivlvi (oil: .111. \l. I.“ I} Oiiii I‘ll-‘10.. .Irill‘. [IIIIx‘liJ’iui 111‘). lili‘uli’i‘llu .
. /
as... c.
b\ x w.
-l‘l‘liliiil
+
Hootm oz
\/
@Ezocctm
sseuufino] emu/002;]
"aim—.4
Q5-xswa xomuoouowa on wood padoo umfiu
cowuomuoucw xomuoouowe-xomuoouomfi oafiwmmom ofiu mo owumfiofiom .m oudwwm
mMUfiOOfiO—E
\.-..... \ /
30
It
\ moéoomofia
_
E = [ K“/<2«r>1/21 fa <9) (17)
”a2 a
where Kn = the nominal (mode 1) stress intensity factor
Rose considered that, "the problem is to determine the resulting
stress field", for the configuration in Figure 8. Rose obtained the
boundary conditions for the model in Figure 8 as follows.
0 = 0 = 0 (18)
for y 4 i 0 along -w < x < 0, and a < x < b;
33
O a 1)
>1“
MAIN CRACK MICROCRACK
Figure 8. Semi-infinite main crack collinear with two dimensional
microcrack
34
(r,6) » 00 n(r,o) for r>>b (19)
00,8
fl
Rose solved the problem using a conformal trasformation (z e g = J2) and
obtained, "the transformed boundary conditions in the f-plane
corresponding to those for two collinear cracks of equal length
perturbing a uniform stress field." Reverting the known solution to the
z-plane, Rose obtained
n 1/2
2(2) 8 K (2 - bC)/[2HZ(z-a)(Z-b)] (20)
C = E(k')/K(k') (21)
2 1/2 2 1/2
k = a/b k' = (l - a/b) = (l - k ) (22)
where K and E - the complete elliptic integrals of the first
and second kind, respectively
k and k' = the modulus and complementary modulus,
respectively, for the elliptic integrals
Rose obtained the stress intensity factor at the three crack tips from
equation (18) as follows.
K(x=0)/Kn = C/k (23)
K(x—a)/1 (27)
a/b - b - (b - a) = 1 - A (28)
b .
where A - the area fraction of the fracture plane corresponding
to pre-existing microcracks
2
= the average of the projected area of the microcracks
onto the fracture plane.
Regarding, "the linking up of the main.crack with this collinear
microcrack as an idealized repeat unit in the crack growth process and
the effect of all the other microcracks being accounted for by a uniform
reduction in the effective modulus" [46], Rose derived an estimate for
K0, the effective fracture toughness of a microcracked body, from
equation (23)-(25). Then, if one compares the magnitude of the stress
intensity factor, K, at the three crack tip positions (x=0, x=a, x=b) in
Rose's model, the relative magnitudes of the K's are: K(x=0) > K(x=a) >
36
K(x=b) of three crack tips. If it is assumed that, "crack extention
occur at either of these crack tips when the relevent K reaches a
definite critical value Kc’ then, under increasing Kn, crack extention
would begin from the main crack tip, and if Kn is kept constant, the
link-up of the main crack with the microcrack would proceed unstably,
rather than quasi-statically." If K0 is the Kn at which link-up
occurs, Rose derived.
Ko/KC - k/C (29)
where Ko - an estimate of the effective fracture toughness
of a microcracked body
KC - the intrinsic fracture toughness of the material
without microcracks.
Rose compared equation (29) with the estimate based on a rule of
mixtures for the work of fracture [2,47], which gave
1/2 1/2
Ko/KC - <1 - A) = unantoa m.w::o% .m ouswwm
coaoEcccd xtmcoo xocto
mic Nd To 0.0
_ L
. L .t mu
7
_r
o
co
/. 4.0m.
/ . Tom:
.5— 5.. nus—3-55 Ill:
/. roem
(DdO) smnpow s,6unoi
44
Table 8. Hardness, fracture toughness, and Ybung's modulus/hardness
ratio for yttrium chromite specimens at each value of
indentation load
Specimen Y Load Number of
1 2
(GPA) (N) (GPa) (MPam / ) indentations
YP10 66.9 2.94 5.815 1.367 11.50 6
YP21 106.6 2.94 6.662 1.374 16.00 6
4.9 6.950 1.562 15.34 6
YP22 133.8 2.94 8.173 1.277 16.37 6
4.9 8.107 1.517 16.51 6
YP23 147.0 2.94 8.327 1.541 17.65 8
4.9 8.584 1.449 17.12 8
YP24 178.0 2.94 9.260 1.601 19.22 6
4.9 9.451 1.771 18.83 6
Table 8 (continued)
45
Y0 245.5 4.9 12.31 1.717 19.94
9.8 13.04 1.828 18 83
Y4 260.5 4.9 13.04 1.757 19.98
9.8 13.80 1.888 18.88
46
crack sizes were measured at 2.94 N for YP10 specimen, 2.94 N and 4.9 N
for YP21, YP22, YP23 and YP24, and 4.9 N and 9.8 N for Y0 and Y4.
A plot of the hardness versus crack density parameter shows
hardness decreasing in an approximately linear fashion with increasing
crack density parameter (Figure 10 (a) and (b)). Thus, as the crack
density parameter increases, microcracks formed upon cooling the
specimen by the phase transformation in yttrium chromite at 11000C
apparently reduce the measured hardness. The hardness versus crack
density parameter data was fit to a linear equation of the form of
equation (35), using the least-square best fit method.
H = A - Be (35)
where H = measured hardness (GPa)
e = crack density parameter
A = 13.92 and 13.23 for the hardness versus crack
density parameter without porosity correction and
with porosity correction, respectively
B = 19.52 and 24.45 for the hardness versus crack
density parameter without porosity correction and
with porosity correction, respectively
Figure 11 (a) and (b) show the fracture toughness versus crack
density parameter for microcracked yttrium chromite. Fracture toughness
decreased with increasing crack density parameter. The fracture
toughness versus crack density parameter data was fit to a linear
equation using the least-square best fit
47
.cowumw>op cumpcmum u m muons owcmu m H
oumowpcw when Houum .ouwaouno Edwnuuh How aofiuoouuoo
Suwmouoe ufiofiuwz HouoEmumm huHmCop xomuo msmuo> newsman: .Amv 0H ousmwm
8008801 36:00 x080
N0 _._.0 0.0
L
0.0 v.0 0.0
p b p — L . _ IF *V
T
(1)
T0.
(Dds) sseupJoH
FN—
Em .53 3.5 l
.:
48
.cowumw>oc pumtamum I m ouo£3 owcmu m H
oncogene when uouum .oufisonso azauuux you coauoouuoo
haemouom £uw3 Houoemume mufimsop Momma mzmuo> mmocpumm .Anv 0H madman .
toeoEctcn. 33:00 x080
To 0.0 «.0 _ P0 0.0
h b. P L, b .*V
.. H
D
r W
m U
9
v. S
///. n0
/ )
/ TOP 0
/ .d
D
a /.\
//
a: .5... 533.? l
49
0.0
.COHUNfova
pudendum I m ouo£3 oowu m H concern“ when uouum
.ouwaouno Edwuuum How cowuoouuoo zuwmouoe unosufia
uouosoumm huamcop zoouo msmuo> mmocnmzou ouauomum .Amv Ha ouswam
poaoacncm 33:00 #920
#0
L
0.0
_
b
N0
IF
r
P0
0
.IOHHdEGfl m.flmcl
HE Hume gems:
(muxecm) sseuqfinol eanqoezd
50
.coaumw>op cumpcmum I m mamas owcmu m H oumowpce
when Houum .ouHEOHSQ Edwuuum com coauoouuoo haemouoe
Suwz wouoEmHmm Swanson xomuo mamuo> mmocswsou ouauomum .Anv Ha ouswwm
nouoEonom “0:980 #095
to no as to ad
_ b . L L 04—.
b
E..— amun 5493-935 l
23.53: «has. ll. IN. —
(z/medN) sseuqfinol amnioeig
51
KC- C - D6 (36)
1/2)
where KC Fracture toughness (MPam
m
I
crack density parameter
C - 1.853 and 1.815 for 6 without porosity correction
and with porosity correction, respectively
0
I
1.177 and 1.498 for 6 without porosity correction
and with porosity correction, respectively
Equation (36) implies that pre-existing microcracks formed by high
temperature phase trasformation of yttrium chromite decrease the
fracture toughness and the microcrack link-up ahead of a moving
macrocrack.
Young's modulus of the specimen is greater than 0 because Y s 0
means that specimen is falling apart. If we consider Young's modulus
from equation (13),
Y = Yo (I - f(vo)e) > 0
e < 1 (37)
£010)
For this study, Y - O at e - 0.573, so physically meaningful values of 6
must be less than 0.573. The expression for f(uo) is obtained from
equations (14) and (16), where V0 = 0.28. Similarly, if we consider
hardness from equation (35), the hardness of specimen must be greater
than 0. Thus, the crack density parameter must be less than 0.713 and
0.541 for 6 without the porosity correction and with porosity
correction. When we consider the fracture toughness from equation (36),
52
fracture toughness must be greater than 1.574 and 1.212 for 6 without
porosity correction and with porosity correction, respectively. If we
compare to e for Y - 0, H - 0 and KC = 0, the crack density parameters
for Y = 0 and H - 0 are similar. However, the crack density parameter
for KC= 0 is greater than those for Y = 0 and H = 0.
Figure 12 (a) and (b) show the ratio of Young’s modulus/hardness
versus crack density parameter for microcracked yttrium chromite. It is
difficult to determine a mathmatical expression for this relation. The
trend is that Young's modulus/hardness ratio is approximately constant
as a function of the crack density parameter for the range of low crack
density parameter. But, Young's modulus/hardness ratio decreases with
increasing the crack density parameter for the range of high 6.
Figures 13 and 14 are optical micrographs of the indent impression
and radial crack system for yttrium chromite specimens indented at 5.88
N and 9.8 N. For specimens having a low crack density parameter (that
is, high Young's modulus) such as Y0 and Y4 specimens, a clear indent
impression and crack form from the Vickers indentation test. This
phenomenon is shown in Figures 13 and 14 for specimens Y0 and Y4 which
have a low crack density parameter. However, for specimens having a
high crack density parameter (that is, low Young's modulus) such as YP10
and YP21, considerable chipping and many small cracks formed around the
indent impression at indentation loads higher than 4.94 N (see Figure
15). Cracks did not form at the surface of specimens for indentation
loads less than 1.96 N. Thus, it was difficult to obtain the only crack
emanated from the 4 corners of impression for specimen having a crack
density parameter higher than 0.27.
53
.coHu0H>0p pudendum I m
0u0£3 0wamu m H 0u00HUCH when uouum .0uuaouao
spauuuh mom coHuoouuoo haemouoa u3050H3 Houoasuma
Swanson zomuo mnmwo> canon unoccum£\m:H:toa 0.0::0» .Amv NH ouswwm
c0e0Eccod 36:00 x080
0.0 *0 0.0 «.0.
e _
hI
To
.b
Z
TNN
0N
sseupJoH/snlnpow s,bunoi
54
0.0 0.0
:I
.cowumfi>0p
pudendum u m 0H0£3 0wcmu m H 0U00Hpca mama uouum
.0uwsoufio Edwuuu% pom cowuoouuoo haemouoe £uw3 H0u0amumm
huwmcop xompo mamH0> owumu mm0flpu0£\m:H50oa m.wcao> .Aav NH 0uswwh
:000Eccod 36:00 x080
N0 P0 . 0.0
r
a .ll-
.b — . mw
.0-
,rF'I'II ‘1!.i fiu.lt-li.lllll.llllll\l I‘III‘II lilllln .ul'l I'll I ti-ll"7.ll'!.‘ll o’lgllliit 'Lr m m
sseupJoH/smnpow s,buno,(
55
100 microns
Vickers indentation
Yttrium Chromite
Load = 5.88 N
Figure 13. Optical micrograph of yttrium chromite specimen Y0 indented
at 5.88N '
56
100 microns
Vickers indentation
Yttrium Chromite
Load = 9.8 N
Figure 14. Optical micrograph of yttrium chromite specimen Y4 indented
at 9 8N
57
100 microns
Figure 15. Optical micrograph of yttrium chromite specimen YP21
indented at 4.9N
fl"
58
4.2 Comparison of Fracture Toughness between Experimental Data and
Theoretical Equation from 3.3
If we estimate that a is equal to c and that a is the same for all
the cracks, we can derive the following equation from equations (8) and
(27).
3
e - N (8)
2/3 2
A - N (27)
If all cracks have radius a, then = a and
3 2 6 2
A = N a = e
A a 62/3 - ' (38)
From equations (22), (28) and (38),
k = (1 - 52/3)1/2 (39)
From equation (39), equations (29), (30) and (31) can be written as
follows, respectively,
KO/KC = (1 ~ 62/3)1/2 (40)
C
2312
Ko/KC = (1 - e / ) / (41)
Ko/Kc = 1-52/3 (42)
59
Equations (40), (41), (42) and the experimental data are plotted in
Figures 16 (a) and (b), and 17 (a) and (b). The fracture toughness of
microcracked polycrystalline yttrium chromite specimens generally
corresponds to Rose’s equation which represents the fracture toughness
decrease due to microcrack linkvup. The fracture toughness corresponds
to Rose's theory of the following equation (see Figure 11 (a) and (b).
KC = s (1 - e 2/3)1/2 + F (43)
where K - fracture toughness
e = crack density parameter
E - 1.622 and 1.980 for the crack density
parameter without porosity correction and with
porosity correction, respectively
F = 0.291 and -0.099 for the crack density
parameter without porosity correction and with
porosity correction, respectively
.cowumw>0p cumccmum I m 0H0£3 m H 0umowccw
mama uouum .c0paaocfi Oman 0H 0uHaow£o Edwuuuh :o 0000
Hmuc0EHH00xm .AANq-0¢V mfioaumsv0v 000% Scam poummpm
mcoammouexo H00H00H00£u 00u£u How H W 0 W 0 0wC0u
0:» wow .0 .coauo0uuoo huwmouon uaonuw3 Hou0amu09
huwmcop Xumuoouowe mo aoHuossm 0 mm 000C£w:ou 0H3uomum .Amv 0H 0uswwh
68330.30 38:00 300000.825 0
0. F 0“0 0.0 ”.0 who 0.0
P . h l I o
. . . ma. "330500 - .. 0 0
x . . . . 3. dosage .. ..
I o I nil!
I . . . ca. 95250 .
z I I 1N C
II o I
x/ . . . 02:00.»
I; . . 3.305.898
O I I I
6 I I
I
I I
I I
I, I I
I, I I
I I
I I I
III
.coau0fi>0v cu0000um n m 0u0£3
m H 0u0owvsw mu0n uouum .00vaaocw omH0 ma 0quou£o
Edwuuuh so 000v H0ucoawu0mxm .AAN¢-ocv mcoau0sv0v 000m
scum 00u0000 mcowmm0unx0 H0owu0uo0£u 00050 How H W 0 W o
0m¢0u 0:» new .0 .aowuoouuoo huwmouoa spas u0u0a0u0m
huwmc0v Momuoouowa mo cowuocam 0 m0 mm0¢£msou 0uauo0um .Anv 0H 0usmfim
300030000 3:300 300.8835» 0
o; 00 who who «no od
L .. . . L «0 9.3330 .. - 0.0
x , . . . 3. 0030300 .. ..
z . . , 3. #330500 I...
I” o : o , IN.O
../ . . . , 33000
3.0250098
61
62
.0owu0w>00 00000000 I m 000:3 m H 00000000
0000 uouum .0000Ho0w omH0 00 00000050 E0wuuuh 0c 0000
H0u00aau00xm .AAN¢-O¢V mcowu0000v 000M 000m 0000000
m0owmm0uax0 H0Owu0uo0£u 000:» How m.o w 0 W o 0w00u
050 you .0 .0000000000 huwmouoa 0000003 000000000
huwm000 xo0uoouowa mo 0oHuoc0m 0 m0 mm00£w00u 00000000
$000033.“ 33000 #0030305 0
00 1o nno . «no rd
A3 5
ouswwm
0.0
m0 00300m0
3.. 0030000
3. 0030000
02000.0
- 3.3000393
H
0.0
1N6
51/51
63
0.0
.000000>00 00000000 I 0 00003 0 H
00000000 0000 00000 .00000000 0000 00 00000000 5000000
00 0000 0000000000xm .AAme-o0v 00000000v 0000 0000
0000000 00000000000 00000000000 00000 000 m.o W 0 W o
0m000 000 000 .0 .0000000000 00000000 0003 000000000
0000000 3000000000 00 00000000 0 00 0000£m000 00000000 .Anv m0 000w00
03000330 0000000 000000000005 0
who . No 0.0 . 0.0
NW 00000mW0 .. -
00. 00000000 .. ..
3. 00000000 I...
0000000
00000000000000
5. CONCLUSIONS
Young's modulus, shear modulus and Poisson’s ratio were determined
by the sonic resonance method for microcracked and nonmicrocracked
polycrystalline yttium chromite specimens.
The microcrack density parameter of each specimen was determined
using equation (14) and (16) from Poission's ratio and Young's modulus
in nonmicrocracked specimens, and the Young's modulus in microcracked
specimens. The crack density parameters after porosity correction of
the specimens were 0.029, 0.037, 0.113, 0.176 0.199, 0.255 and 0.335.
The hardness was determined from equation (6) using the indentation
impression size. The observed decrease in hardness of microcracked
polycrystalline yttrium chromite specimen was approximately linear with
increasing the crack density parameter over the entire 6 range. A
least-squares fit of the data yielded the relation.
13.92 and 13.23 for the hardness versus the crack
where A
density parameter without porosity correction and
with porosity correction, respectively
B = 19.52 and 24.45 for the hardness versus the crack
density parameter without porosity correction and
with porosity correction, respectively
64
65
Fracture toughness was computed from equation (7) using indentation
crack size. Fracture toughness decreased linearly with e, the crack
density parameter. The relating equation was by a least-square fit
where C = 1.853 and 1.815 for the fracture toughness
versus 6 without porosity correction and
with porosity correction, respectively
D = 1.177 and 1.498 for the fracture toughness
versus 6 without porosity correction and
with porosity correction, respectively
These results imply that pre-existing microcracks in yttrium chromite
link-up ahead of a moving macrocrack. Also, the observed trends in
fracture toughness is consistant with Rose's theory for microcrack link-
up with a moving macrocrack.
The relation between Young's modulus/hardness ratio and crack
density parameter was difficult to summarize mathematically. However,
Young's modulus/hardness ratio was approximately constant for low values
of crack density parameter. In contrast, the Young’s modulus/hardness
ratio decreased with e for crack density parameter values higher than
about 0.2.
10.
11.
12.
l3.
l4.
6. REFERENCES
K. T. Faber, "Toughening of Ceramic Materials", Ph D. Thesis,
University of California, Berkeley, CA (1982).
A. G. Evans and K. T. Faber, J. Amer. Ceram. Soc., 64 [7]: 394-398
(1981). 7
Y. Fu and A. G. Evans, Acta Metall., 39: 1619-1625 (1982).
F. F. Lange, J. Mater. Science, 11: 235-241 (1982).
D. L. Porter, A. G. Evans, and A. H. Heuer, Acta Metall., 21:
1649-1654 (1979).
A. G. Evans and A. H. Heur, J. Amer. Ceram. Soc., 6; [5-6]: 241-
248 (1980).
R. C. Garvie, R. R. Hughan and R. T. Pascoe, pp. 263 in "Processing
of Crystalline Ceramics," edited by H. Palmour, R. F. Davis and T.
M. Hare, Plenum Press, New York (1977).
R. C. Garvie, R. H. Hannick and R. T. Pascoe, Nature, gggz 703-706
(1975).
Y. Ikuma and A. Virkar, J. Mater. Science, 12: 2233-2238 (1984).
A. G. Evans and R. M. Cannon, Acta Meta11., 34 [5]: 761-800
(1986).
J. A. Kuszyk and R. C. Bradt, J. Amer. Ceram. Soc , 6 [8]: 420-
423 (1973).
R. W. Davidge and T. J. Green, J. Mater. Science, 3: 629-634
(1968).
E. D. Case, J. R. Smyth and 0. Hunter, Materials Science and
Engineering, i1: 175-179 (1981).
E. Horbogen and K. Friedrich, J. Mater. Science, 1;: 2175-2182
(1980).
66
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
67
R. P. Waki and B. Iluter, J. Mater. Science, 22 [4]: 875-885
(1980).
F. F. Lange, Philo. Mag., 22: 983—992 (1970).
D. J. Green, P. S. Nicholson and J. D. Embury, J. Mater. Science,
24: 1657-1661 (1979).
S. Melin, Int. J. Fracture, 22: 37-45 (1983).
E. D. Case and C. J. Glinka, J. Mater. Science, 22: 2962-2968
(1984).
W. D. Kingery, H. K. Bowen and D. R. Uhlmann, Chapter 5 in
Introduction to Ceramics, Second Edition, John Wiley and Sons, New
York (1976).
E. D. Case, J. Mater. Science, 22: 3702-3712 (1984).
E. D. Case, T. Negas and L. P. Domingues, unpublished data.
K. Hardman-Rhyne, N. F. Berk, and E. D. Case, "Porosity Study of
Sintered and Green Compact YCr03 Using Small Angle Neutron
Scattering Techniques," pp. 103-108, in Nondestructive Evaluation:
Application to Materials Processing, edited by 0. Buck and S. M.
Wolf, Amer. Soc. for Metals, Meta1s Park, Ohio (1984).
T. Negas and L. P. Dominges, in "Fourth International Meeting on
Modern Ceramics", edited by P. Vincenzini (Elsevier Interscience,
New York, pp. 993 (1979).
R. L. Fullman, AIME Trans., 2212: 447-452 (1953).
E. D. Case, J. R. Smyth and V. Monthei, Commun. J. Amer. Ceram.
Soc., 64: c24-c25 (1981).
F. Forster, Z. Metall, 22: 109-115 (1937).
S. Spinner and W. E. Tefft, ASTM Proc., 61: 1221-1238 (1961).
H. R. Kase, J. A. Tesk and E. D. Case, J. Mater. Science, 22: 524-
531 (1985).
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
68
G. Picket, ASTM Proc., 46: 846-865 (1945).
B. R. Lawn and D. B. Marshall, J. Amer. Ceram. Soc., _2 [7-8]:
347-350 (1979).
G. R. Antis, P. Chantikul, B. R. Lawn and D. B. Marshall, J. Amer.
a [9]: 533-538 (1981).
Ceram. Soc.,
B. R. Lawn and M. V. Swain, J. Mater. Science. 22 [1]: 113-122
(1975).
B. R. Lawn, A. G. Evans and D. B. Marshall, J. Amer. Ceram. Soc.,
62 [9-10]: 574-581 (1980).
J. B. Walsh, J. Geophysical Research, 24 [20]: 5249-5257 (1965).
R. L. Salganik, Mechanics of Solids, 2 [4]: 135-143 (1973);
English Translation.
B. Budiansky and R. L. O’Connell, Int. J. Solids Structures, 22:
81-92 (1976).
D. P. H. Hasselman and J. P. Singh, Ceram. Bull., 66: 856-860
(1979).
J. Kemeny and N. G. W. Cook, Int. J. Rock Mech. Min. Sci., _2 [2]:
107-118 (1986).
J. B. Walsh, J. Geophys. Res., 12: 399-411 (1965).
G. R. Irwin, J. Appl. Mech., 24: 361-364 (1957).
J. R. Rice, Mathematical Analysis in the Mechanics of Fracture, pp.
191- 311, Fracture, Vol 2 (edited by H. Libowitz) Academic Press,
New York (1968).
E. D. Case, J. R. Smyth and 0. Hunter, J. Mat. Sci., 26: 149-153
(1980).
A. G. Evans, Acta Meta11., 26: 1845-1853 (1978).
Y. Fu , A. G. Evans and W. M. Kriven, J. Amer. Ceram. Soc., 7 [9]:
626-630 (1984).
69
46. L. R. F. Rose, J. Amer. Ceram. Soc., 6_ [3]: 212-214 (1986).
47. W. Kreher and W. Pompe, J. Mater. Science, 26: 694-706 (1981).
48. A. G. Evans and K. T. Faber, J. Amer. Ceram. Soc., 61 [4]: 255-260
(1984).
49. J. K. Mackenzie, Proc. Phys. Soc. (Lond.), 63B: 2-11 (1950).
APPENDIX 1. Derivation of Equation (12) in Section 3.1
From equation (10) and (11) in Section 3.1 of this thesis,
(1 - u 2) 2 , 4 2 . 2 2 2
U - 0 IO [ 4 a Sln 7(wc) + 16 a Sin 7cos 7cos w
e Y o 2 2 2
2
_ 2 2 2 2
(WC) + 16 (1 V0) 0 sin 7cos 73in w (nc)] 2xcdc (44)
2 2
7r (2 " V0) (1 ' V0)
2 2 4
If we use the relations of - = 1/2, - 1/5 and
2 2
- 2/15, then equation (44) can be rewritten as
2 2 3 4 2 2 2
U = 8 (l - V0 )0 c [ sin 7 + 4 sin 7cos 7cos w
e
3Y 2
(2 - V0)
2 2 2
+ 4 (1 - vo)sin7 cosy sin w]
2
(2 ' V0)
2 2 3 2 2
= __§__ (1 - V0 )0 C [3 + 4/(2 - Vo) + 4 (1 - Vo)/(2 - V0) 1
45Y
2 3 2 -
-__8_ac(l-Vo) 1° 3"° (45)
ASY 2 " V0
If it is assumed that there are N penny-shaped cracks with mean crack
3
radius cubed in the body, equation (45) becomes
2 3 2 _
U - 8 Na (1 - V0 ) 10 3V0
45Y 2 ' V0
which is identical to equation (12) in Section 3.1 of this thesis.
70
APPENDIX 2. Rule of Mixtures Relations
Based on the assumption that "the reduction in load-bearing area"
will reduce fracture toughness via microcracking, Evans and Faber gave
[2]
Po - PC (1 - vm) (46a)
where F0 - the intrinsic resistance to crack of microcracked
body
PC - the matrix toughness
Vm - volume fraction of microcracks
Evans and Faber [2] also give the following relations between crack
toughness and elastic modulus
2
Kc — ECFC (47)
2
where Kc the intrinsic fracture toughness of the
nonmicrocracked body
Ko - the effective fracture toughness of the
microcracked body
E - Young's modulus of the nonmicrocracked body
Eo - Young's modulus of the microcracked body
Substituting equations (47) and (48) into equation (46a) gives
71
If“
72
2 2
K0 - KC (EC/EC) (1 - vm) (46b)
Taking the square root of equation (46b) and assuming E0 = EC, then for
dilute crack systems,
1/2
Ko - KC (1 - Vm) (49)
If we assume the volume fraction of microcracks, Vm = A, the area
fraction of the fracture plane corresponding to pre-existing
microcracks, and use equation (38), then equation (49) gives
2 a 1 2
K0 /Kc - (1 - e / ) / (41)
Rose refers to this equation as being based on a "rule of mixture for
the work of fracture."
If the reduction of fracture toughness in the presence of
microcracks is based on a rule of mixtures for fracture toughness Ko and
KC (assuming Vm{