Materials Transactions, Vol. 45, No. 3 (2004) pp.

936 to 941
#2004 The Japan Institute of Metals
EXPRESS REGULAR ARTICLE

Fracture Toughness of JLF-1 by Miniaturized 3-Point Bend Specimens

with 3.37.0 mm Thickness
Hiroaki Kurishita1 , Takuya Yamamoto1 * , Takuya Nagasaka2 , Arata Nishimura2 ,
Takeo Muroga2 and Shiro Jitsukawa3
1

The Oarai Branch, Institute for Materials Research, Tohoku University, Oarai, Ibaraki 311-1313, Japan
National Institute for Fusion Science, Toki, Gifu 509-5292, Japan
3
Radiation Eects and Analyses, Department of Materials Science, Japan Atomic Energy Research Institute,
Tokai-Mura, Ibaraki 319-1195, Japan
2

A small specimen test technique is required to evaluate the fracture toughness values of several millimeter thick plates of structural
materials and to maximize the use of very limited space for materials irradiation in intense neutron sources like IFMIF. In view of several
advantages of three-point bending (3PB) over compact tension (CT), miniaturized 3PB specimens with 7.0, 5.0 and 3.3 mm thickness were
prepared from a Japanese low activation ferritic steel, JLF-1, which is a candidate rst wall and fusion blanket material.Elastic-plastic fracture
toughness tests by the unloading compliance method at room temperature and plane-strain fracture toughness tests at 77 K were conducted in
general accordance with the ASTM standards. Emphasis was focused on the determination of the actual J-value for crack initiation, JIN , for
reliable fracture toughness evaluation with the 3PB specimens. The obtained values of JIN at room temperature and KIC at 77 K were 100
120 kJ/m2 and 2022 MPam1=2 , respectively, exhibiting little dependence on specimen size. By combining the experimentally obtained data
with the plane-strain FEM analysis, a method was proposed to estimate JIN from a load-displacement curve measured for a single specimen. The
method is applicable to heavily irradiated materials with little ductility.
(Received October 8, 2003; Accepted January 20, 2004)
Keywords: small specimen test technique, neutron source, irradiation, low activation, ferritic steel, crack initiation, elastic-plastic fracture
toughness, plane-strain fracture toughness, three-point bending, nite element analysis

1.

Introduction

Fracture toughness is a key engineering property of

structural materials. The dimensions of standard size specimens used for fracture toughness testing are as large as
62:5
60
25 mm for compact tension (CT) and 250

50
25 mm for 3-point bending (3PB). However, the rst
wall and blanket of fusion reactors is a thin wall structures,
only several millimeters thick. Furthermore, irradiation
volumes available in intense neutron sources like International Fusion Materials Irradiation Facility, IFMIF1,2) are
very limited to approximately 500 cc for displacement
damage above 20 displacements per atom (dpa)/year. This
volume of 500 cc corresponds to that occupied by only six
standard CT specimens or two standard 3PB specimens for
fracture toughness testing. In addition, there will be a
considerable gradient in irradiation temperature and neutron
ux over the volume. Overcoming these limitations requires
miniaturization of fracture toughness test specimens.
A reduced activation ferritic steel, JLF-1 (Japanese low
activation ferritic steel), is a primary candidate as a rst wall
and fusion blanket material.3) For fracture toughness evaluation of the steel, the elastic-plastic fracture toughness testing
should be mainly used. Many attempts of specimen miniaturization have mainly concentrated on the CT specimens,
although there have been several reports on miniaturized 3PB
specimens.4,5) This is probably because the load-line displacement can be correctly measured with a clip gage for the
CT specimen, but not for the 3PB specimen. However, as
shown in Table 1, 3PB specimens have many advantages
*Present

address: Department of Chemical Engineering, UCSB, Santa

Barbara, CA 93106-5080, USA

Table 1 Comparison of 3PB and CT specimens. The marks of } and

indicate very good and good, respectively.
CT specimen

3PB specimen

Data precision
Data availability

}
}

Specimen miniaturization

Specimen preparation

Specimen setting for testing

Control of testing condition

over CT specimens, and hence it is desirable to develop test

techniques for reliable evaluation of fracture toughness using
miniaturized 3PB specimens.
On the other hand, one should consider that the number of
fracture toughness test specimens subject to irradiation in
IFMIF may be limited and the irradiated specimens may
exhibit severe radiation embrittlement. This suggests that it is
not easy to carry out elastic-plastic fracture toughness testing
by the single-specimen (unloading compliance) or multiplespecimen method. Therefore, it is desirable to develop a
method that allows us to evaluate J-values from a loaddisplacement curve measured for a single specimen.
In this study 3PB miniaturized specimens of JLF-1 with
thickness 7.0, 5.0 and 3.3 mm were tested using parameters J
and K under the opening mode (Mode I) loading condition.
The tests were conducted at room temperature and 77 K.
FEM analysis was applied to develop a method to estimate
the actual J-value for crack initiation, JIN , from a loaddisplacement curve measured for a single specimen.

Fracture Toughness of JLF-1 by Miniaturized 3-Point Bend Specimens with 3.37.0 mm Thickness

2.

Experimental

2.1 Material
A 300 kg heat of JLF-1 was produced, as a 30 mm thick
plate with a composition, in weight percent, 0.098 C, 0.05 Si,
0.5 Mn, 0.002 P, 0.004 S, 8.92 Cr, 2.00 W, 0.20 V, 0.098 Ta,
0.0149 N, 0.0001 B, and balance iron. The plate was
normalized at 1323 K for 3.6 ks (air cooling) and tempered at
1053 K for 3.6 ks (air cooling).
2.2 Specimens and test methods
The plate of JLF-1 was machined into four types of 3PB
specimens and one type of tensile specimens. Figure 1 shows
the conguration of the 3PB specimens having the following
dimensions, where B is the thickness, W is the width and L is
the length: B7:0 mm
W10:0 mm
L45 mm, B5:0 mm

W5:0 mm
L25 mm, B3:3 mm
W5:0 mm
L25 mm, and
B3:3 mm
W3:3 mm
L18:3 mm. The magnitude of span,
S, for these specimens was approximately 4 W, i.e., 40.0,
20.3 and 13.2 mm. The original crack length, a0 , was
controlled between 0.50 W and 0.59 W. A straight-through
precrack starter notch terminating in a semi-circle with a
radius of 0.06 mm was introduced by spark discharge. The
depth of starter notch was between 0.6 and 3.0 mm depending
on specimen width and fatigue precrack length was not less
than 1.3 mm. For specimens tested at RT side grooves with
the depth of 0.4B were machined to increase the constraint
for high fracture toughness of JLF-1,6) whereas for specimens
tested at 77 K no side grooves were machined in most cases.
The dimensions of the tensile specimen were B0:5 mm

W4 mm
L16 mm with a gage section of B0:5 mm

W1:2 mm
L5:0 mm.
A fatigue testing machine (Shimadzu Servopulser of 50kN capacity equipped with a 5-kN shear-type load cell) was
used to introduce a well-dened crack and perform fracture
toughness tests. For pre-cracking, stress intensify factor,
K,
was controlled from the initial values of 1423 MPam1=2 to
the nal values of approximately 413 MPam1=2 depending

on specimen size, test temperature and notch preparation.

Because the fatigue pre-cracking should meet the requirements that for the nal 3% of fatigue precrack extension
Kfmax 3% < 0:6
YS1 =
YS2 KQ and for the nal 50% of
fatigue precrack extension the maximum fatigue load, Pmax ,
is no larger than Pf = 0.5Bb0 2
Y /S (b0 W  a0 ). Here,

YS1 and
YS2 are 0.2% oset static yield stresses at
temperatures where fatigue precracking and fracture toughness testing were conducted, respectively.
Y is the eective
yield strength, dened as
Y
YS
TS =2, where
TS is
the ultimate tensile strength. Fatigue cycling was conducted
at room temperature with a sinusoidal waveform of 20 Hz and
the ratio of minimum and maximum fatigue loads of 0.095.
Some typical data on pre-cracking for each type of 3PB
specimens are listed in Table 2.
A temperature control bath, 3PB test xtures with dierent
spans of 40.0, 20.3 and 13.2 mm and miniature clip gages
with working ranges of 2.0 and 1.5 mm were designed to
perform fracture toughness tests at room temperature and
below. A linear variable dierential transformer (LVDT) of
2:5 mm was installed on the actuator head for precise
measurements of load line displacement. Figure 2 shows the
test xture with a 3PB specimen and clip gage. In the gure
the clip gage is seated on an attachable knife-edge cemented
to the notched side (Fig. 1(b)). However, cementing of the
knife-edge would not be appropriate for both higher-temperature irradiation and low-temperature fracture toughness tests
because of loss of the adhesion. Therefore, for the 3PB
specimens with 10.0 and 5.0 mm widths, the knife-edge was
machined on the notched side of each specimen (Fig. 1(a)).
The depth of the knife-edge, F, was approximately 0.7 mm.
Elastic-plastic fracture toughness tests by the single
specimen method at room temperature and plane strain

WC/Co ball
Specimen

(b)

Guide

45

0.2B

0.2B

(a)

Clip gauge
W

a0

a0N

45

a0

937

5W

5W
Fig. 1 Conguration of 3PB specimens with (a) W 10 and 5 mm, and (b)
W 3:3 mm. F 0:7 mm, G 2:0 mm, Z 1:0 mm.

Fig. 2

Test xture for 3PB specimen.

Table 2 Typical values of specimen dimensions and fatigue precracking data for four types of 3PB specimens of JLF-1.
W

a0 N

a0

b0

Kf

Kfmax

Kfmax (3%)

(mm)

(mm)

(mm)

(mm)

(mm)

(MPam1=2 )

(MPam1=2 )

(MPam1=2 )

10.01

7.00

2.98

5.57

4.44

5.25

5.82

5.95

0.037/0.382

828768

5.03

4.99

1.50

2.92

2.11

5.03

5.58

5.58

0.017/0.172

555170

5.01

3.29

1.01

2.77

2.24

5.06

5.65

5.65

0.012/0.122

599998

3.33

3.32

0.65

1.96

1.37

3.95

4.39

4.80

0.007/0.071

794564

Pfmin =Pfmax

Number of
cycles

938

H. Kurishita et al.

fracture toughness tests at 77 K were conducted in general

accordance with the ASTM standards7,8) at a cross-head
speed of 0.013 mm/s and in some cases at 13 mm/s. A
program for the unloading compliance method including load
frame compliance corrections was used for testing the 3PB
specimens. The elastic-plastic fracture toughness tested
specimens were heat-tinted and then fractured at 77 K to
measure the initial and nal crack lengths.
In order to develop the method that allows us to evaluate
the J-value for crack initiation, JIN , from a load-displacement
curve measured for the single specimen, two dimensional
plane-strain FEM analyses with 1274 8-node elements were
used, assuming that the crack-tip radius is 1 mm and the precrack does not grow. Tensile tests were conducted at room
600

Vc= 0.013 mms-1

Load, P/N

295 K
400

77 K
200

0.5 mm
0

Clip Gage Displacement, D/mm

Fig. 3 Load-displacement records for pre-cracked 3PB specimens with
B3:3 mm
W3:3 mm of JLF-1 tested at 295 K (left) and 77 K (right) at
0.013 m/s.

temperature and 77 K at initial strain rates of 1
103 and

10 s1 by the above testing machine.
3.

Results and Discussion

Figure 3 shows an example of the load versus clip-gage

displacement records measured at 295 and 77 K at
0.013 mm/s for the smallest specimen, B3:3 mm

W3:3 mm of JLF-1. Figure 4 shows SEM micrographs of
their fracture surfaces. From the gures, even in this
miniaturized specimen, a slow, stable extension of the precrack at 295 K and unstable crack extension at 77 K can be
seen. The tests with dynamic loading at 13 mm/s gave
essentially the same records. At 77 K the miniaturized 3PB
specimen fractured in a completely brittle manner at a very
low load and the plane-strain fracture toughness, KIC , was
estimated. Table 3 lists the values of KIC measured for four
types of the 3PB specimens, which satisfy the criterion of B,
W-a0 > 2:5KIC =
YS 2 . The tensile test results are listed in
Table 4. The values of KIC lie in the range 2022 MPam1=2
regardless of specimen size.
Since slow, stable crack extension was observed at room
temperature, the elastic unloading compliance method was
applied to measure the J-integral for JLF-1. Figure 5 shows
the J-integral versus crack extension curves for three types
of specimens: B7:0 mm
W10:0 mm, B3:3 mm
W5:0 mm
and B3:3 mm
W3:3 mm. The number of the data points in
the gure indicates the sequence order of unloading. The
blunting line was obtained by calculating 2
Y
a. According
to the ASTM standards, the J value is dened as the
intersection of the J-R curve with the 0.2 mm oset line,

Machined
notch

Fatigue
pre-crack
Side
groove
Crack front

Fig. 4 SEM micrographs of fracture surfaces of pre-cracked 3PB specimens with B3:3 mm
W3:3 mm of JLF-1 tested at 77 K (left) and
295 K (right) at 0.013 mm/s.

Table 3

Plane strain fracture toughness, KIC , at 77 K for four types of 3PB specimens of JLF-1.

BN

PQ

KQ KIC

0:6
YS1 =
YS2 KQ

2:5KIC =
YS 2

B, b0 >

(mm)

(mm)

(mm)

(kN)

(MPam1=2 )

(MPam1=2 )

(mm)

2:5KIC =
YS 2

10.0

7.0

7.0

1.14

21.0

6.01

1.00

5.0

5.0

5.0

0.51

20.1

5.74

0.91

5.0

3.3

3.3

0.42

22.4

6.40

1.13

3.3

3.3

3.3

0.28

20.6

5.89

0.96

Fracture Toughness of JLF-1 by Miniaturized 3-Point Bend Specimens with 3.37.0 mm Thickness
Tensile properties of JLF-1 tested at room temperature and 77 K.

Test temperature Initial strain

(K)

rate (s1 )

286
286

3

77
77

YS

TS

(MPa) (MPa)
501
604

1
10
10

3

1
10
10

1052
1296

623
688

"u

(MPa)

(%)

BN

a0

JQ ASTM

JQ

JIN

7.97
7.50

(mm)

(mm)

(mm)

(mm)

(kJ/m2 )

(kJ/m2 )

(kJ/m2 )

10.0

7.0

4.2

5.18

431

251

117

4.70

5.0

5.0

3.0

2.89

467

273

123

0.67

5.0

3.3

2.0

2.82

443

261

116

3.3

3.3

2.0

1.84

336

222

103

562
646

1111

1082

1330

1313

(a)
J - Integral (kJ/m2)

Table 5 Elastic-plastic fracture toughness test results on dierently sized

3PB specimens of JLF-1 at room temperature.

Y
YS
TS =2

(b)
7.0 x 10.0 mm
2F a

a=0.2mm

J - Integral (kJ/m2)

Table 4

3.3 x 5.0 mm
2F a

a=0.2mm

Crack Extension, a/mm

Crack Extension, a/mm

(c)
J - Integral (kJ/m2)

939

(d)
3.3 x 3.3 mm
2F a

Crack
extension
a=0.2mm

Crack Extension, a/mm

Pre-crack
Notch

Fig. 5 Examples of J-integral versus crack extension curves for 3PB specimens with (a) B7:0 mm
W10:0 mm, (b)
B3:3 mm
W5:0 mm and (c) B3:3 mm
W3:3 mm and (d) fracture surface of (b) of JLF-1 tested at room temperature and
0.013 mm/s.

which is shown here as JQ ASTM . However, it should be noted

that the initial part of the J-R curve is much steeper than the
calculated blunting line. Therefore, the best t straight line
was drawn through the initial portion of the data and the
second line was drawn parallel to the rst but oset by an
amount corresponding to a crack extension of 0.2 mm.9) The
candidate toughness value JQ was determined by the
intersection of the data with this oset line. In addition, it
is obvious from the gure that clear jumps in the crack
extension occur about at data point number 79, which
should correspond to the onset of pre-crack extension.
Therefore, the value of JIN , dened here as the J-value at
the onset of pre-crack extension, was determined from the
intersection of the straight line and a linear approximation of

the data falling between approximately 0.1 and 0.3 mm of

crack extension.
The values of JQ and JIN are listed in Table 5, together with
those of JQ ASTM . The values of JQ ASTM , 431467 kJ/m2 , are
close to those reported for the 1CT and 1/2 CT specimens of
JLF-1, respectively (419 and 431 kJ/m2 ),6) and it is obvious
that JQ ASTM > JQ > JIN for each size of specimens. It should
be noted that the value of JIN lies between 100 and 120 kJ/m2
and is almost independent of specimen size.
The reliability of the fracture toughness values obtained in
this study will be discussed below.
In order to examine the possibility of local deformation of
specimen regions in contact with the loading and supporting
pins during fracture toughness testing, the data obtained by

940

H. Kurishita et al.

1000

Load (N)

800
600

JLF-1 (5 x 5 x span20)
286 K

400

Experiment
FEM

200

* t=2.48mm equivallent
0
-0.5

0.5

1.5

2.5

3
Loading pin

Displacement (mm)

Fig. 6 Comparison between the load-displacement curves obtained by the plane-strain FEM analysis and by the experiment for 3PB
specimen with B5:0 mm
W5:0 mm of JLF-1 tested at 286 K and 0.013 mm/s.

unloading compliance method were analyzed with and

without conventional load frame compliance corrections. It
was found that the dierence between the load frame
compliance corrected data and uncorrected data is negligibly
small, less than one percent. Careful observations of specimen surfaces after fatigue precracking showed that the
contact regions exhibited a slight indent due to plastic
deformation. The present result of no appreciable dierence
between the load frame compliance corrected data and
uncorrected data indicates that the indent did not grow by the
subsequent fracture toughness testing due to work hardening
caused during pre-cracking. This is because the magnitude of
the load applied during fatigue pre-cracking was larger than
that during fracture toughness testing. For example, for the
specimen size of B3:3 mm
W5:0 mm
L25 mm, the applied load during the initial 525% fatigue pre-cracking was
approximately 0:70:9 KN,10) whereas that during fracture
toughness testing was approximately 0.55 KN at room
temperature and approximately 0.42 kN at 77 K.
The determination of JIN by the compliance method as
described above seems dicult to apply to static tests as well
as to dynamic tests for highly irradiated specimens with less
ductility. Therefore, a method to estimate JIN is proposed in
this study and described below.
In this method, at rst, static or dynamic 3PB loading was
conducted for a pre-cracked specimen to measure the load
versus both load-line and clip gage displacements records.
Next, the tested specimen was heat-tinted and fractured at
77 K to determine the original and nal crack lengths, a0 and
a0
a, respectively. Then, the plane-strain FEM analysis
was applied to reproduce the measured load-displacement
record. Finally, the measured record and the FEM curve were
compared.
The FEM analysis assumes no extension of the pre-crack.
Thus, the curve obtained by FEM should reproduce the
measured record up to the point where the pre-crack starts to
extend. Beyond that point, the FEM curve should stay higher
than the measured record. In other words, the pre-crack may
start to extend at the point where appreciable deviation of the

FEM curve from the measured record starts to occur. This

means that the area, A, under the load-displacement record up
to the deviation point is related to JIN for crack initiation by
JIN 2A=b0 BN ;

where BN is the net specimen thickness.

Figure 6 shows the comparison of the load-displacement
curves calculated by the plane-strain FEM analysis with the
measured load versus displacement records for the
B5:0 mm
W5:0 mm specimen tested at 286 K and
0.013 mm/s. Here, the eective specimen thickness,  t, was
assumed to be 2.48 mm so that the FEM curve can well
reproduce the measured record. It should be noted that the
value of 2.48 mm is quite close to the thickness for the normal
fracture to occur, approximately 2.5 mm, determined by a
fracture surface examination. Comparison of the two curves
shows that there is a good agreement until the displacement
reaches about 0.7 mm, and above 0.7 mm appreciable deviation of the FEM-curve from the measured one becomes
noticeable. Substitution of the calculated area A (Fig. 6) in
eq. (1) gives a JIN value of approximately 130 kJ/m2 , which
is nearly equal to the value of JIN shown in Table 5. For the
B3:3 mm
W3:3 mm specimens essentially the same result
was obtained.
In this study, two-dimensional plane-strain FEM analysis
was used. Three-dimensional FEM analysis, which will be
needed to demonstrate the validity of this approach, is now in
progress.
4.

Conclusions

For the miniaturized 3PB specimens with thickness of 7.0,

5.0 and 3.3 mm of JLF-1 steel, elastic-plastic fracture
toughness tests by the elastic unloading compliance method
at room temperature and plane-strain fracture toughness tests
at 77 K were performed to develop a small specimen test
technique for reliable evaluation of fracture toughness. The
main results are as follows.
(1) The elastic-plastic fracture toughness tests gave JQ

Fracture Toughness of JLF-1 by Miniaturized 3-Point Bend Specimens with 3.37.0 mm Thickness

values ranging between 222 and 273 kJ/m2 and JIN

values of 103 to 123 kJ/m2 . JIN , which is dened as the
J-value at the onset of pre-crack extension, is almost
independent of specimen size. The plane-strain fracture
toughness tests at 77 K gave KIC of 20-22 MPam1=2
regardless of specimen size.
(2) A method to estimate JIN from a load-displacement
curve measured for a single specimen is proposed,
which is based on the nding that the two-dimensional
plane-strain FEM analysis can reproduce the measured
load-displacement record until the pre-crack starts to
extend. The method is applicable also to heavily
irradiated materials with less ductility.
Acknowledgements
The authors would like to express their gratitude to Drs. A.
Kimura and R. Kasada for their supply of a JLF-1 plate, to Dr.
H. Suwarno for his experimental assistance to a part of this
work and to Dr. S. Matsuo for his reviews. The present work
was performed as the activity of International Fusion
Materials Irradiation Facility Key Element Technology
Phase (KEP) supported by National Institute for Fusion
Science in Japan, which is greatly appreciated.

941

REFERENCES
1) T. Kondo: J. Nucl. Mater. 258263 (1998) 4755.
2) K. Noda, K. Ehrlich, S. Jitsukawa, A. Moslang and S. Zinkle: J. Nucl.
Mater. 258263 (1998) 97105.
3) A. Kohyama, Y. Kohno, K. Asakura and H. Kayano: J. Nucl. Mater.
212215 (1994) 684689.
4) S. Jitsukawa, A. Naito and J. Segawa: J. Nucl. Mater. 271272 (1999)
8791.
5) S. Tahtinen, P. Moilanen, B. N. Singh and D. J. Edwards: J. Nucl.
Mater. 307311 (2002) 416420.
6) A. Nishimura, N. Inoue and T. Muroga: J. Nucl. Mater. 258263 (1998)
12421247.
7) ASTM E1820-99a, Standard Test Method for Measurement of Fracture
Toughness, Annual Book of ASTM Standards, (2000).
8) ASTM E399-90, Standard Test Method for Plane-Strain Fracture
Toughness of Metallic Materials, Annual Book of ASTM Standards,
(1990).
9) D. J. Alexander, J. E. Pawel, M. L.Grossbeck, A. F. Rowclie and K.
Shiba: ASTM STP 1270, D. S. Gelles, R. K. Nanstad, A. S. Kumar and
E. A. Little, Eds., 1996, pp. 945967.
10) In this study the initial 525% of total fatigue precrack was introduced
under the load of 1:11:2Pf (Pf 0:75 kN for b0 4:0 mm, and
Pf 0:62 kN for b0 3:625 mm). Because no appreciable dierence
was found between the specimens precracked under 1:11:2Pf and
those under 0:951Pf for the initial 525% of total fatigue precrack.