materials

Article
Evaluation of Fracture Strength of Ceramics
Containing Small Surface Defects Introduced by
Focused Ion Beam
Nanako Sato 1 and Koji Takahashi 2, * ID

1 Graduate School of Engineering, Yokohama National University, 79-5 Tokiwadai,

Hodogaya, Yokohama 240-8501, Japan; sato-nanako-nw@ynu.jp
2 Faculty of Engineering, Yokohama National University, 79-5 Tokiwadai,
Hodogaya, Yokohama 240-8501, Japan
* Correspondence: takahashi-koji-ph@ynu.ac.jp; Tel.: +81-45-339-4017




Received: 9 February 2018; Accepted: 19 March 2018; Published: 20 March 2018


Abstract: The aim of this study was to clarify the effects of micro surface defects introduced by the
focused ion beam (FIB) technique on the fracture strength of ceramics. Three-point bending tests
on alumina-silicon carbide (Al2 O3 /SiC) ceramic composites containing crack-like surface defects
√
introduced by FIB were carried out. A surface defect with area in the range 19 to 35 µm was
introduced at the center of each specimen. Test results showed that the fracture strengths of the
√
FIB-defect specimens depended on area. The test results were evaluated using the evaluation
√
equation of fracture strength based on the process zone size failure criterion and the area parameter
model. The experimental results indicate that FIB-induced defects can be used as small initial cracks
for the fracture strength evaluation of ceramics. Moreover, the proposed equation was useful for the
fracture strength evaluation of ceramics containing micro surface defects introduced by FIB.

Keywords: Al2 O3 /SiC; focused ion beam; surface defect; fracture strength; process zone size failure
√
criterion; area parameter model

1. Introduction
Ceramics have excellent mechanical, tribological, and thermal properties in comparison with
metallic materials. However, there is a risk that failure initiates from defects due to the low fracture
toughness of ceramics. Therefore, it is very important to quantitatively evaluate the influence of defects
on fracture strength for maintenance management of ceramic components.
It is reported that the fracture toughness for small flaws is lower than that for large cracks. Thus,
conventional linear elastic fracture mechanics (LEFM) is not applicable for predicting the fracture
strength of ceramics containing small flaws [1]. Many authors have proposed various models for
the relationship between defect size and fracture strength [2–6]. Ando et al. proposed the process
zone size failure criterion and reported that the fracture strength of ceramics containing small surface
cracks could be evaluated by this criterion [7]. This model can predict the fracture strength of a
material having a semicircular crack introduced perpendicular to the tensile stress. Murakami and
√
Endo proposed the area parameter model to predict the fatigue limit of metals having a surface crack
√
with arbitrary shape [8]. Here, the area is the square root of the area of a surface defect projected
√
onto the direction of the maximum tensile stress. The area parameter model is widely applied to
evaluate the fatigue limit of metals [9,10]. In a previous study, we proposed a new evaluation equation
√
for the fracture strength of ceramics combining the process zone size failure criterion and the area
parameter model [11]. This equation is useful for the evaluation of fracture strength of ceramics having
a surface crack with arbitrary shape and orientation introduced by the indentation method.

Materials 2018, 11, 457; doi:10.3390/ma11030457 www.mdpi.com/journal/materials

Materials 2018, 11, 457 2 of 9

However, it is known that tensile residual stress is generated around the indentation when a
pre-crack is introduced on the surface of a ceramic material by the indentation method [12]. Therefore,
elimination of the tensile residual stress by surface polishing or annealing in inert gas is necessary [13].
If the tensile residual stress is not eliminated, the influence of residual stress on the fracture strength
should be considered to predict the fracture strength with high accuracy [13]. The dimensions of a
defect at the origin of a fracture in an actual ceramic material are equivalent to the grain size, which is
less than 20 µm in a typical ceramic. However, the dimensions of defects introduced in most of the
previous studies were larger than 20 µm because it is difficult to introduce small surface cracks by the
indentation method [14–16].
In this study, we introduced crack-like surface defects on the surface of ceramics by the focused
ion beam (FIB) technique. The FIB technique is a high-precision working method that uses the effect of
sputtering by ion beam scanning on the surface of a material. Sakamoto et al. reported that the fatigue
limits of FIB-notched specimens were equivalent to the fatigue limits of annealed fatigue-cracked
specimens of carbon steel [17]. Rokio and Solin evaluated small crack initiation and propagation from
small FIB-defect in high strength steel and reported that a FIB notch to initiate small cracks is a useful
way to test small crack growth in high strength steels [18]. However, to our knowledge, the influence
of FIB-introduced surface defects on the fracture strength of ceramics has not yet been studied. In this
paper, the effects of microscopic surface defects introduced by FIB on the fracture strength of an
Al2 O3 /SiC composite are clarified. Regarding the relationship between fracture strength and defect
size, the consistencies of the test results and the values predicted by the proposed evaluation equation
were verified.

2. Evaluation Equation of Fracture Strength

In this section, an outline of the proposed equation for fracture strength evaluation is
√
described [11]. The equation is based on the process zone size failure criterion [7] and the area
parameter model [8,19].
Ando et al. proposed the following equation to predict the fracture strength (σC ) of ceramics with
a small surface crack focusing on the process zone size at the crack tip and [7]
 2    
π KIC πσC
= ae sec −1 , (1)
8 σ0 2σ0

where KIC (MPa·m1/2 ) is the fracture toughness, and σ0 (MPa) is the fracture strength of a smooth
specimen assuming a defect is sufficiently small, and σC (MPa) is the fracture strength of a specimen
containing a surface crack. The equivalent crack length, ae (µm) is defined as
 2
1 KIC
ae = , (2)
π σc

On the other hand, Murakami showed that the maximum stress intensity factor (Kmax ) at the edge
of a surface crack with arbitrary shape under tensile stress (σ∞ ) could be approximated within 10%
error by [19] q √
Kmax ∼
= 0.650σ ∞ π area, (3)
√
where area is the square root of the area of a surface defect projected onto the direction of the
maximum tensile stress. When σC is applied to a surface crack, the Kmax value reaches KIC . Equation (3)
leads to q √
KIC = 0.650σC π area . (4)
Materials 2018, 11, 457 3 of 9

√
The2018,
Materials relationship between
11, x FOR PEER fracture
REVIEW strength σC and the crack size area is expressed as follows
3 of 9
based on Equations (1), (2) and (4)
3.38 √ √  2
= cos 3.38σ0 2 area ∙2σ .

(5)
σC = cos−1 2 3.38 2
√
√ · 0 . (5)
π ·KIC + 3.38σ0 area π
This is the equation to evaluate the fracture strength of a ceramic material containing a surface
This is the equation to evaluate the fracture strength of a ceramic material containing a surface
defect with arbitrary shape. Figure 1 shows the relationship between σC and √ √ of alumina
defect with arbitrary shape. Figure 1 shows the relationship between σC and area of alumina
reinforced by silicon carbide (Al2O3/SiC) [11]. The dotted line indicates the σC predicted based on
reinforced by silicon carbide (Al2 O3 /SiC) [11]. The dotted line indicates the σC predicted based on
conventional LEFM. The predicted values of σC based on LEFM are overestimate for √ √ less than
conventional LEFM. The predicted values of σC based on LEFM are overestimate for area less than
20 µm. Thus, conventional LEFM is not applicable to ceramics with small cracks. On the other hand,
20 µm. Thus, conventional LEFM is not applicable to ceramics with small cracks. On the other hand,
the solid line indicates the σC predicted based on Equation (5). Although the σC predicted based on
the solid line indicates the σC predicted based on Equation (5). Although the σC predicted based on
Equation (5) were slightly smaller than those of the experimental results, they agreed relatively well
Equation (5) were slightly smaller than those of the experimental results, they agreed relatively well
with the experimental results in the range of 5–20 µm. In this study, the relationship between√√
with the experimental results in the range of 5–20 µm. In this study, the relationship between area of
of an FIB defect and fracture strength was evaluated using KIC and σ0 as the basic mechanical
an FIB defect and fracture strength was evaluated using KIC and σ0 as the basic mechanical properties
properties of the sample material.
of the sample material.

√
Figure 1. Relationship
Figure 1. Relationship between
between fracture
fracture strength and √area ofof
strength and AlAl
2 O2O
3 /SiC
3/SiC[11].
[11].

3. Experimental
3. Experimental Procedures
Procedures

3.1.
3.1. Test
TestSpecimen
Specimen
Alumina
Alumina reinforced
reinforced bybysilicon
siliconcarbide
carbide(Al
(Al22O
O33/SiC) wasselected
/SiC) was selectedas
as aa test
test material
material in
in this study.
The
TheAl Al22O
O33 powder (AKP-50, Sumitomo Chemical Industry Co. Ltd., Tokyo, Tokyo, Japan)
Japan) used
used in
in this
this study
study
had a mean particle size of 0.2 µm. The SiC powder (ultrafine grade, Ibiden Co. Ltd.,
had a mean particle size of 0.2 µm. The SiC powder (ultrafine grade, Ibiden Co. Ltd., Gifu, Japan) Gifu, Japan) used
had
useda mean
had a particle size of 0.27
mean particle size µm. Theµm.
of 0.27 samples were prepared
The samples using a mixture
were prepared using aofmixture
Al2 O3 with
of Al152Ovol %
3 with

SiC powder.
15 vol % SiC To this mixture,
powder. To isopropanol
this mixture, was added, and was
isopropanol the mixture
added, was
andblended thoroughly
the mixture for 24 h
was blended
by ball milling.
thoroughly for After drying
24 h by and comminution,
ball milling. theand
After drying mixture was hot-pressed
comminution, at 1973was
the mixture K and 35 MPa for
hot-pressed at
21973
h in K
anand
N2 atmosphere
35 MPa for 2[20].h inThe
an hot-pressed
N2 atmosphere plates were
[20]. Thecut into bending
hot-pressed test were
plates specimens measuring
cut into bending
× 4specimens
3test × 20 mm.measuring
The surface3 of× 4each
× 20 specimen
mm. The was polished
surface of eachtospecimen
a mirror-like
wasfinish.
polished to a mirror-like
finish.
3.2. Specimen Preparations
3.2. Specimen Preparations
To heal surface cracks induced during machining, all the bending test specimens were heat-treated
at 1573
ToKheal
for 1surface
h in air,cracks
which induced
are the optimum healing conditions
during machining, all the for surfacetest
bending cracks in Al2 Owere
specimens 3 /SiCheat-
[20].
treated at 1573 K for 1 h in air, which are the optimum healing conditions for surface cracks in
Al2O3/SiC [20]. These specimens are called “smooth” specimens. The σ0 value was evaluated by the
strength distribution of the smooth specimens.
An FIB defect was introduced at the center of each smooth specimen by using a multi beam SEM-
FIB system JIB-4501 (by JEOL Ltd., Tokyo, Japan). These specimens are called the “FIB-defect”
Materials 2018, 11, 457 4 of 9

These specimens are called “smooth” specimens. The σ0 value was evaluated by the strength
distribution of the smooth specimens.
An FIB defect was introduced at the center of each smooth specimen by using a multi beam
SEM-FIB system JIB-4501 (by JEOL Ltd., Tokyo, Japan). These specimens are called the “FIB-defect”
specimens. Figure 2 shows the detail of the FIB technique process. A test specimen was fixed on a jig
Materials 2018, 11, x FOR PEER REVIEW 4 of 9
with carbon tape. The FIB defect was processed while monitoring the processing state using
Materials 2018, 11, x FOR PEER REVIEW 4 of 9
the SEM.
Figure 3a schematically shows the shape of an FIB defect. The surface lengths
specimens. Figure 2 shows the detail of the FIB technique process. A test specimen was fixed on a jig (2c) ranged from 20 to
specimens.
60 µm andwith
thecarbon
depths Figure 2
(d)The
tape. shows
ranged the detail
from
FIB defect was of the
19processedFIB
to 24 µm. technique
whileThe process.
conditions
monitoring A test specimen
of the FIB
the processing was fixed
technique
state on a jig
are presented
using the SEM.
with carbon tape. The FIB defectthe
was processed while monitoring the processing state using the SEM.
in Table 1.Figure
The area3a schematically
of a defectshowswas shape
evaluated of an FIB
based defect.
on theThefracture
surface lengths
surface (2c) ranged from
observation. 20 to
Figure
60 µm and 3a schematically
the depths (d)showsrangedthe shape
from of 24
19 to anµm.
FIB The
defect. The surface
conditions lengths
of the (2c) ranged
FIB technique are from 20 to
presented
For comparison,
60 µm and the “pre-crack
depths (d) + polished”
ranged from 19 to specimens
24 µm. The were prepared
conditions of the FIB in the following
technique are presented procedure.
in Table 1. The area of a defect was evaluated based on the fracture surface observation.
First, a Knoop indentation
in Table
For 1. The area was
comparison, madewas
of“pre-crack
a defect at the center
evaluated
+ polished” of each
based
specimens on thesmooth
fracture
were specimen.
surface
prepared Figure 3b
observation.
in the following shows the shape
procedure.
of a Knoop First,For comparison,
pre-crack.
a Knoop “pre-crack
The indentation
indentation was made+ polished”
loads specimens
at the ranged
center were
of from
each prepared
9.8
smooth 49 inN.theNext,
tospecimen. following 3b procedure.
Figurethese specimens
shows the were
First, aofKnoop
shape a Knoop indentation was
Themade
pre-crack.slurry at the center
indentation loadstheof eachfrom
ranged smooth9.8 specimen.
to 49 N. Figure
Next, 3b specimens
these shows the
polished using a 1-µm diamond to eliminate tensile residual stress around the indentations.
shapepolished
were of a Knoop pre-crack.
using a 1-µmThe indentation
diamond slurryloads ranged from
to eliminate 9.8 to 49residual
the tensile N. Next, thesearound
stress specimensthe
The amount of polished
were materialusing removed
a 1-µm from
diamond the surface was
slurry tofrom 4 to 6the
eliminate times
tensilethe indentation depth.theThe ASTM
indentations. The amount of material removed the surface was 4 residual
to 6 timesstress around
the indentation
standard C1421
depth. suggests
indentations.
The ASTM removing
The standard
amount of 4.5suggests
material
C1421 to 5removed
times the
removing indentation
from thetosurface
4.5 depth
5 timeswas 4 to[21].
6 times
the indentation Thus, the
[21].removed
the indentation
depth Thus, depth
depth.
in this study The
the almost
removedmet ASTM standard
depththe C1421
standard.
in this suggests removing
study almost met the standard. 4.5 to 5 times the indentation depth [21]. Thus,
the removed depth in this study almost met the standard.

Figure 2. Photographs
Figure 2. Photographs of theofFIB
the FIB process:(a)
process: (a)fixing
fixing method
method of test specimen;
of test (b) inside
specimen; (b)ofinside
device.of device.
Figure 2. Photographs of the FIB process: (a) fixing method of test specimen; (b) inside of device.

Figure 3. Schematic illustrations of surface defects: (a) FIB-defect specimen; (b) pre-crack + polished
Figure 3. Schematic
Schematic
Figure 3. specimen. illustrations of
illustrations of surface
surfacedefects: (a) FIB-defect
defects: specimen; (b)specimen;
(a) FIB-defect pre-crack + polished
(b) pre-crack +
specimen.
polished specimen.
Table 1. FIB processing conditions.
Table 1. FIB processing conditions.
Ion Source
Table 1. FIB processingGa Liquid Metal Ions
conditions.
Ion Source
Accelerating voltage Ga Liquid Metal Ions
30 kV
Accelerating voltage
Prove current 30 kV
3000 pA
Ion Source
Prove current
Dose Ga Liquid pAMetal
3000nC/µm
190–280 2 Ions
Dose 190–280 nC/µm2
Accelerating voltage 30 kV
Prove current 3000 pA
Dose 190–280 nC/µm2
Materials 2018, 11, 457 5 of 9
Materials 2018, 11, x FOR PEER REVIEW 5 of 9

3.3. Measurement
3.3. Measurement of
of Fracture
Fracture Toughness
Toughness
The fracture
The fracture toughness
toughnessKKIC
IC value
value was
was evaluated
evaluated byby the
the indentation
indentation fracture
fracture (IF)
(IF) method.
method. Vickers
Vickers
indentations were made on the polished surface of a bending test specimen by the application
indentations were made on the polished surface of a bending test specimen by the application of of an
an
indentation load
indentation loadof
of49
49NNfor
for15
15s.s.Then,
Then,KKICIC(Pa
(Pa·m
·m1/2)) was
1/2
was calculated
calculated using
using the following equation
the following equation inin
accordance with the Japan Industrial Standards
accordance with the Japan Industrial Standards [22] [22]

1 1
= 0.026E 2 P 2 a , (6)
KIC = 0.026 3 , (6)
c 2

where E (Pa) is the Young’s modulus, P (N) is the indentation load, a (m) is half the diagonal length
where E (Pa) is the Young’s modulus, P (N) is the indentation load, a (m) is half the diagonal length
of the indentation, and c (m) is half the surface crack length. The units of calculated values of KIC were
of the indentation, and c1/2(m) is half the surface crack length. The units of calculated values of KIC
converted into “MPa·m ”. Crack length was measured using an optical microscope with total
were converted into “MPa·m1/2 ”. Crack length was measured using an optical microscope with total
magnifications of 580× and higher to avoid the measurement error of KIC values [23].
magnifications of 580× and higher to avoid the measurement error of KIC values [23].
3.4. Measurement
3.4. Measurement of of Fracture
Fracture Strength
Strength
TheσσC values
The values of
of each
each specimen
specimenwere
weremeasured
measuredbybyaathree-point
three-pointbending
bending test,
test, asas shown
shown inin Figure
Figure 4,
C
4, at room temperature in air. The span length was 16 mm, and the crosshead speed was
at room temperature in air. The span length was 16 mm, and the crosshead speed was 0.5 mm/min. 0.5 mm/min.
After the
After the bending
bending test,
test, the
the fracture
fracture surfaces
surfaces were
were observed
observed byby laser
laser microscopy,
microscopy, and
and the
the size
size of
of the
the
defects at the fracture origins was measured.
defects at the fracture origins was measured.

Figure
Figure 4. Schematicdiagram
4. Schematic diagramofofthethethree-point
three-point loading
loading system
system andand
sizesize
andand dimensions
dimensions of
of test
test specimen.
specimen.

4.
4. Results
Results and
and Evaluation
Evaluation of
of Fracture
Fracture Strength
Strength

4.1. Mechanical Properties of the Sample Material

4.1. Mechanical Properties of the Sample Material
In this section, the calculated results of KIC and σ0 are presented. An average and a standard
In this section, the calculated results of KIC and σ0 are presented. An average and a standard
deviation of the KIC value of four samples measured by the IF method were 3.50 ± 0.05 MPa1/2 ·m1/2 .
deviation of the KIC value of four samples measured by the IF method were 3.50 ± 0.05 MPa·m . The
The scatter of the KIC value is quite small.
scatter of the KIC value is quite small.
The σ0 value of the sample material was estimated using the strength distribution of the smooth
The σ0 value of the sample material was estimated using the strength distribution of the smooth
specimens. The two-parameter Weibull function is expressed by
specimens. The two-parameter Weibull function is expressed by
  α 
σC
1−
F (σC ) = = 1 − exp−− β
exp ,, (7)
(7)

where F(σ
where F(σCC)) is
is the
the cumulative
cumulative failure
failure probability
probability based
based onon the
the median
median rank
rank method,
method, αα isis the
the scale
scale
parameter, and β is the shape parameter. Figure 5 shows the two-parameter Weibull
parameter, and β is the shape parameter. Figure 5 shows the two-parameter Weibull distribution of distribution
of
thethe smooth
smooth specimens.
specimens. Theresults
The test test results were plotted
were plotted as the relationship
as the relationship betweenbetween ln (σ
ln (σC) and C ) and
ln{ln(1/(1
ln{ln(1/(1
− F)} led by−Equation
F)} led by(7).
Equation (7). The
The second secondand
horizontal horizontal
verticaland
axisvertical
show σCaxis
andshow σC and
F values, F values,
respectively.
respectively. Two parameters were determined using the maximum likelihood
Two parameters were determined using the maximum likelihood method based on JIS R1625 method based on[24].
JIS
R1625 [24]. The α value was 8.81, and the β value was 871 MPa. In Figure 5, a thick solid
The α value was 8.81, and the β value was 871 MPa. In Figure 5, a thick solid line and thin solid lines line and
thin
showsolid
the lines show the approximation
approximation line and the 90% line and the 90%interval,
confidence confidence interval, respectively.
respectively. The 90%
The 90% confidence
interval values of α and β are (αL, αU) = (6.72, 11.6), (βL, βU) = (838, 910). In the Weibull distribution,
Materials 2018, 11, 457 6 of 9

Materials 2018, 11, x FOR PEER REVIEW 6 of 9

Materials 2018, 11, x FOR PEER REVIEW 6 of 9
confidence interval values of α and β are (αL , αU ) = (6.72, 11.6), (βL , βU ) = (838, 910). In the Weibull
the F(σC) value
distribution, increased
F(σC ) value
theincreased as the defectasdimensions in the specimens decreased. Because theBecause
size of
the F(σC) value as increased the defect dimensions
the defect dimensions in the decreased.
in the specimens specimens Because
decreased. the size of
defects in a smooth
the sizeinofa defects the specimen with F(σ ) = 99% considered
F(σC ) = 99% to be sufficiently small, the σ value at
smoothin thea specimen
smooth the specimen withconsidered considered to small,
be sufficiently small,
C C
defects with F(σC) = 99% to be sufficiently the σC value at
F(σCσ)C= value
the 99% isatregarded as σ0. This
F(σC ) = 99% evaluation method of σ0 showed a good estimation results of σC as
F(σ C) = 99% is regarded as σ0is. This
regarded as σ0 . This
evaluation evaluation
method method
of σ0 showed of σ0 estimation
a good showed a good estimation
results of σC as
shown of
results in σFigure
as 1 [11].
shown Substituting the values of α and β of the smooth specimen and F(σ ) = 99% in
1 [11]. Substituting the values of α and β of the smooth specimen and F(σC) = 99%and
in Figure 1 [11]. Substituting the values of and of the smooth specimen
C α β C
shown in Figure in
Equation
F(σ C ) = 99%(7), σ0Equation
= 1036 MPa was calculated.
Equation (7),inσ0 = 1036 MPa (7),was
σ0 =calculated.
1036 MPa was calculated.

Figure 5. Two-parameter Weibull

Two-parameter Weibull distribution of
Weibull distribution
distribution of fracture strength
strength of the
the smooth specimens.
specimens.
Figure 5.
Figure 5. Two-parameter of fracture
fracture strength of
of the smooth
smooth specimens.

4.2. Fracture Strength of FIB-Defect Specimens

Fracture Strength
4.2. Fracture of FIB-Defect Specimens
All specimens
All specimens containing
containing FIB-induced
FIB-induced surface
surface defects
defects fractured
fractured at at the
the FIB-defect
FIB-defect area
area as
as aa result
result
of the three-point bending tests. Figure 6a,b show laser microscope images of fractured surfaces of a
of the three-point bending tests. Figure 6a,b show laser microscope images of fractured surfaces of a
FIB-defect specimen
FIB-defect specimen anda pre-crack
a pre-crack + polished specimen, respectively. The surface length 2c, the
√specimenand and a pre-crack + polished
+ polished specimen, respectively.
specimen, The surface
respectively. lengthlength
The surface 2c, the2c,
depth
the
depth d, and theof√the FIB-defects
of the FIB-defects and pre-cracks after polishing were measured by these
d, and the
depth the √
d, andarea and pre-cracks
of the FIB-defects after polishing
and pre-cracks afterwere measured
polishing werebymeasured
these observations.
by these
observations. Table
Table 2 shows Table 2 shows
the results the
of the results of the bending tests for the FIB-defect specimens and the pre-
observations. 2 shows thebending tests
results of thefor the FIB-defect
bending tests forspecimens and the
the FIB-defect
√
pre-crack
specimens + polished
and the pre-
crack + polished
specimens.
crack
specimens.
The sizes
+ polished
The
of the FIB
specimens.
sizes
Thedefects
of the
sizes ofwere
FIB defects
in the
the FIB
were
following
defects were
in the
range: following
in the following
range:
area = 19–35 √ The σ=C= 19–35
µm.√
range:
19–35
value
µm. The σ √
C value of the FIB-defect specimens decreased with increasing √ . Similar to the pre-
of the
µm. FIB-defect
The σC value specimens decreased
of the FIB-defect with increasing with area.increasing
Similar to√the pre-crack
. Similar +topolished
crack + polished specimens, σ √ specimens decreased
C showed strong √ -dependency in the FIB-defect specimens.
the pre-
specimens,
crack σC showed
+ polished strong
specimens, area-dependency
σC showed strong √ in the FIB-defect specimens.
-dependency in the FIB-defect specimens.

Figure 6. Laser microscope

Lasermicroscope images
microscopeimages ofoffracture
imagesof surfaces:
fracture (a) FIB-defect
surfaces: specimen, √ √ area
= 19.1 µm, σC
Figure 6.6. Laser fracture surfaces: (a)(a) FIB-defect
FIB-defect specimen,
specimen, √ = 19.1
= 19.1 µm,µm,
σC
= 687 MPa; (b) pre-crack + polished specimen, √ √ = 64.6 µm, σ C = 400 MPa.
= 687 MPa; (b) pre-crack + polished specimen,
= C687 MPa; (b) pre-crack + polished specimen, √
σ area = 64.6 µm, = 400
= 64.6 µm, σC = C400 MPa.
σ MPa.

Table 2. Bending test results.

Table 2. Bending test results.
Surface Fracture
Surface Depth √ Fracture
Type of Defect Length Depth √ Strength
Type of Defect Length d (µm) (µm) Strength
2c (µm) d (µm) (µm) σC (Mpa)
2c (µm) σC (Mpa)
FIB-defect test 20.8 19.4 19.1 687
FIB-defect test 20.8 19.4 19.1 687
specimen 20.0 23.2 20.4 715
specimen 20.0 23.2 20.4 715
Materials 2018, 11, 457 7 of 9

Table 2. Bending test results.

Surface Length 2c √ Fracture Strength

Type of Defect Depth d (µm) area (µm)
(µm) σ C (Mpa)
Materials 2018, 11, x FOR PEER REVIEW 7 of 9
20.8 19.4 19.1 687
20.0 23.2 20.4 715
33.4 21.1 25.3 597
33.4 21.1 25.3 597
FIB-defect test
32.332.3 22.5
22.5 26.2
26.2 557 557
specimen
32.832.8 23.8
23.8 27.4
27.4 592 592
46.246.2 24.1 32.6
32.6 530 530
60.160.1 21.3 35.2
35.2 508 508
Pre-crack
Pre-crack + + 121 121 44.0
44.0 64.6
64.6 400 400
polished
polished test 124 124 47.5
47.5 68.0
68.0 322 322
test specimen
specimen 228 228 106 137
137 253 253

4.3. Comparison
4.3. Comparison Between
Between Predicted
PredictedFracture
FractureStrength
Strengthand
andTest
TestResults
Results
TheσσC of
The of the
thesample
samplematerial
materialwas waspredicted
predicted substituting
substituting KIC= =3.50
KIC 3.50 MPa·m
MPa
1/2 and σ0 = 1036 MPa
·m1/2 and σ0 = 1036 MPa
C
into Equation
into Equation (5).(5).Figure 7 shows
Figure 7 showsthe predicted σC values
the predicted σC and the bending
values and thetest results. test
bending For the pre-
results.
crack + polished specimens, the predicted σ C values matched the experimental results. Thus, the
For the pre-crack + polished specimens, the predicted σC values matched the experimental results.
usefulness
Thus, of this equation
the usefulness of thiswas confirmed.
equation wasThe errors between
confirmed. the experimental
The errors between the and predicted results
experimental and
predicted results were generally within ±15%. It is thought that the prediction errors are caused
were generally within ±15%. It is thought that the prediction errors are mainly by the
mainly caused
approximation
by the approximationerrors errors
of Kmaxofby Equation
Kmax (3). For(3).
by Equation theFor
FIB-defect specimens,
the FIB-defect the experimental
specimens, σC are
the experimental
between the predicted σ C based on LEFM and Equation (5). The predicted values of σC based on LEFM
σC are between the predicted σC based on LEFM and Equation (5). The predicted values of σC based
and
on Equation
LEFM (5) give (5)
and Equation thegive
upper
the and
upperlower limit, limit,
and lower respectively. Similar
respectively. tendency
Similar tendencythatthat
Equation
Equation(5)
gives
(5) lower
gives lowerσCσwas observed in the results shown in Figure 1. The results indicate that a FIB-defect
C was observed in the results shown in Figure 1. The results indicate that a FIB-defect
can be used as a small initialcrack
can be used as a small initial crackfor forfracture
fracturestrength
strengthevaluation
evaluationof ofceramics.
ceramics.ItItwas
wasfound
foundthat
thatthe
the
evaluation equation was useful for evaluating the fracture strengths of ceramics
evaluation equation was useful for evaluating the fracture strengths of ceramics containing micro containing micro
surfacedefects
surface defectsintroduced
introducedby byFIB.
FIB.

Figure 7. Comparison
Figure 7. Comparison between
between the
the predicted
predicted fracture
fracturestrength
strengthand
andthe
thebending
bendingtest
testresults.
results.

5. Conclusions
5. Conclusions
This study
This study investigated
investigated the
thefracture
fracturestrengths
strengthsofofAlAl
2O2O /SiC samples in which
3 3/SiC which micro
micro surface
surface
defectshad
defects hadbeen
beenintroduced
introducedby bythe
theFIB
FIBtechnique.
technique. The
The test
test results
results were
were evaluated
evaluated using
using the
the proposed
proposed
evaluationequation
evaluation equationofoffracture
fracturestrength.
strength.The
Theresults
resultsare
aresummarized
summarizedas asfollows:
follows:
(1) The micro surface defects in the size range of √ = 19–35 µm were introduced on the surface
of Al2O3/SiC by the FIB technique.
(2) The fracture strengths of the FIB-defect specimens showed √ dependency. This result is
similar to that obtained for the pre-crack + polished specimens.
(3) FIB-defect can be used as a small initial crack for fracture strength evaluation of ceramics.
(4) The fracture strengths of the FIB-defect specimens predicted by the evaluation equation matched
Materials 2018, 11, 457 8 of 9

√
(1) The micro surface defects in the size range of area = 19–35 µm were introduced on the surface
of Al2 O3 /SiC by the FIB technique.
√
(2) The fracture strengths of the FIB-defect specimens showed area dependency. This result is
similar to that obtained for the pre-crack + polished specimens.
(3) FIB-defect can be used as a small initial crack for fracture strength evaluation of ceramics.
(4) The fracture strengths of the FIB-defect specimens predicted by the evaluation equation matched
the test results. It was confirmed that the proposed equation for fracture strength evaluation was
useful for ceramics containing micro surface defects introduced by FIB.

Acknowledgments: This study was supported in part by a grant for Scientific Research (B) (16H04231) from the
Ministry of Education, Culture, Sports, Science, and Technology of Japan.
Author Contributions: N. Sato and K. Takahashi conceived and designed the experiments; N. Sato performed the
experiments and evaluated the data; N. Sato and K. Takahashi wrote the paper.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Usami, S.; Kimoto, H.; Takahashi, I.; Shida, S. Strength of ceramic materials containing small flaws.
Eng. Fract. Mech. 1986, 23, 745–761. [CrossRef]
2. Hoshide, T. Grain fracture model and its application to strength evaluation in engineering ceramics.
Eng. Fract. Mech. 1993, 44, 403–408.
3. Li, C.W.; Lui, S.C.; Goldacker, J. Relation between strength, microstructure, and grain—Bridging
characteristics in in situ reinforced silicon nitride. J. Am. Ceram. Soc. 1995, 78, 449–459. [CrossRef]
4. Taylor, D. Predicting the fracture strength of ceramic materials using the theory of critical distances.
Eng. Fract. Mech. 2004, 71, 2407–2416. [CrossRef]
5. Munz, D. What can we learn from R-curve measurements? J. Am. Ceram. Soc. 2007, 90, 1–15. [CrossRef]
6. Askes, H.; Susmel, L. Understanding cracked materials: Is linear elastic fracture mechanics obsolete?
Fatigue Fract. Eng. Mater. Struct. 2015, 38, 154–160. [CrossRef]
7. Ando, K.; Kim, B.A.; Iwasa, M.; Ogura, N. Process zone size failure criterion and probabilistic fracture
assessment curves for ceramics. Fatigue Fract. Eng. Mater. Struct. 1992, 15, 139–149. [CrossRef]
8. Murakami, Y.; Endo, M. Quantitative evaluation of fatigue strength of metals containing various small
defects or cracks. Eng. Fract. Mech. 1983, 17, 1–15.
9. Saberifar, S.; Mashreghi, A.R.; Mosalaeepur, M.; Ghasemi, S.S. The interaction between non-metallic
inclusions and surface roughness in fatigue failure and their influence on fatigue strength. Mater. Des.
2012, 35, 720–724. [CrossRef]
10. Beretta, S.; Romano, S. A comparison of fatigue strength sensitivity to defects for materials manufactured by
AM or traditional processes. Int. J. Fatigue 2017, 94, 178–191. [CrossRef]
11. Sato, N.; Takahashi, K. Prediction of fracture strength for ceramics containing a surface defect with arbitrary
shape. Trans. Jpn. Soc. Spring Eng. 2018, 63. in press.
12. Marshall, D.B.; Lawn, B.R. Residual stress effects in sharp contact cracking: Indentation fracture mechanics.
J. Mater. Sci. 1979, 14, 2001–2012. [CrossRef]
13. Chou, I.A.; Chan, H.M.; Harmer, M.P. Effect of Annealing Environment on the Crack Healing and Mechanical
Behavior of Silicon Carbide-Reinforced Alumina Nanocomposites. J. Am. Ceram. Soc. 1998, 81, 1203–1208.
[CrossRef]
14. Li, Z.; Ghosh, A.; Kobayashi, A.; Bradt, R.C. Indentation fracture toughness of sintered silicon carbide in the
Palmquist crack regime. J. Am. Ceram. Soc. 1989, 72, 904–911. [CrossRef]
15. Takahashi, K.; Iwanaka, K.; Osada, T.; Koike, H. Increase in strength of partially stabilized zirconia after shot
peening. J. Mater. Eng. Perform. 2015, 24, 3573–3578. [CrossRef]
16. Ćorić, D.; Renjo, M.M.; Ćurković, L. Vickers indentation fracture toughness of Y-TZP dental ceramics. Int. J.
Refract. Met. Hard Mater. 2017, 64, 14–19. [CrossRef]
Materials 2018, 11, 457 9 of 9

17. Sakamoto, J.; Takahashi, Y.; Aono, Y.; Noguchi, H. Method for Assessing Applicability of an Artificial Flaw
as a Small initial crack for fatigue limit evaluation and its application to a drill Hole and an FIB processed
sharp notch in annealed 0.45% carbon steel. J. Test. Eval. 2013, 41, 194–199. [CrossRef]
18. Roiko, A.; Solin, J. Measurement of small cracks initiating from inclusions, Focused Ion Beam notches and
drilled holes. Int. J. Fatigue 2014, 62, 154–158. [CrossRef]
19. Murakami, Y. Analysis of stress intensity factors of modes I, II and III for inclined surface cracks of arbitrary
shape. Eng. Fract. Mech. 1985, 22, 101–114. [CrossRef]
20. Osada, T.; Nakao, W.; Takahashi, K.; Ando, K.; Saito, S. Kinetics of Self-crack-healing of alumina/Silicon
carbide composite including oxygen partial pressure effect. J. Am. Ceram. Soc. 2009, 92, 864–869. [CrossRef]
21. ASTM International. Standard Test Methods for Determination of Fracture Toughness of Advanced Ceramics at
Ambient Temperature; ASTM C1421-10; ASTM International: West Conshohocken, PA, USA, 2010.
22. Japanese Standards Association. Testing Methods for Fracture Toughness of Fine Ceramics at Room Temperature;
JIS R1607; Japanese Standards Association: Tokyo, Japan, 2015.
23. Miyazaki, H.; Yoshizawa, Y. Correlation of the indentation fracture resistance measured using high-resolution
optics and the fracture toughness obtained by the single edge-notched beam (SEPB) method for typical
structural ceramics with various microstructures. Ceram. Int. 2016, 42, 7873–7876. [CrossRef]
24. Japanese Standards Association. Weibull Statics of Strength Data for Fine Ceramics; JIS R1607; Japanese
Standards Association: Tokyo, Japan, 2010.

© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).