Third International Conference on Sustainable Construction Materials and

Technologies http://www.claisse.info/Proceedings.htm

Investigation of the Cause of Reinforcing Steel Fracture

Induced by Alkali-Silica Reaction and Study on
Maintenance
Kazunori Sasaki1, Misa Fujibayashi2, Masakazu Tona3, Akinori Sato4, Yoshio Hisari5 and
Toyoaki Miyagawa6
1
Osaka Construction Dept., Construction Management Headquarters, Hanshin Expressway
Company Limited, 1-2-1-1900 Benten, Minato-ku, Osaka 552-0007, Japan. kazunori-
sasaki@hanshin-exp.co.jp
2
Osaka Business & Maintenance Dept., Hanshin Expressway Company Limited, 3-1-25
Ishida, Minato-ku, Osaka 552-0006, Japan. misa-fujibayashi@hanshin-exp.co.jp
3
Osaka Business & Maintenance Dept., Hanshin Expressway Company Limited, 2-11-12
Kosei, Minato-ku, Osaka 552-0023, Japan. masakazu-tona@hanshin-exp.co.jp
4
Sakai Construction Dept., Construction Management Headquarters, Hanshin Expressway
Company Limited, 2-3-20 Minami-Hanadaguchicho, Sakai-ku, Sakai 590-0075, Japan.
akinori-sato@hanshin-exp.co.jp
5
Planning and Investigation Division, Hanshin Expressway Management Technology
Center, 4-5-7, Minami Honmachi, Chuo-ku, Osaka 541-0054, Japan. hisari@tech-
center.or.jp
6
Dept. of Civil & Earth Resources Eng., Kyoto University, Nishikyo, Kyoto 615-8540,
Japan. miyagawa@sme.kuciv.kyoto-u.ac.jp

ABSTRACT
There are an increasing number of reports of reinforcing steels fractured or cracked at the
bend with the progress of alkali-silica reaction-induced deterioration in old structures. This
study investigated the cause of reinforcing steel fracture, with the focus placed on material
properties of reinforcing steels, stress as external force and hydrogen embrittlement. The
reinforcing steel fracture was likely to occur by the following mechanism: (1) tensile residual
stress occurs in the inside of the bend of reinforcing steels due to bending, initiating cracks
depending on the rib shape or bending radius; (2) fracture toughness decreases with work
hardening or strain aging; (3) corrosion of reinforcing steels causes occlusion of diffusible
hydrogen, increasing the risk of hydrogen embrittlement; and (4) cracks propagate due to
alkali-silica reaction-induced expansion and tensile residual stress, ultimately causing
fracture of reinforcing steels. This paper also includes proposals on maintenance of
structures with a risk of reinforcing steel fracture.

Keywords. Alkali-silica reaction, Fracture of reinforcing steel, Residual stress, Fracture

toughness value, Hydrogen embrittlement cracking
1. INTRODUCTION
Cracks in concrete attributable to alkali-silica reaction (hereinafter referred to as “ASR”)
were found in T-shaped reinforced concrete piers on the Hanshin Expressway in Japan in
1982. A study committee was formed immediately to investigate the cause and
countermeasures of ASR-induced cracks which were then almost unknown. They established
maintenance guidelines and focused their efforts on controlling the progress of deterioration
and understanding the behavior of affected structures through follow-up inspections. Since
fracture of reinforcing steels was found in severely deteriorated bridge piers in 2000, more
efforts have been made in light of the urgency of the issue to determine the cause of the
problem and carry out repair and strengthening work. However, there have been similar
reports of reinforcing steel fracture to date. It is highly likely that fractured reinforcing steels
remain undetected in structures which were constructed at similar time to or in the same
project with those already found to have the problem. This possibility should be taken into
account in a new maintenance concept.

The purpose of this study is to determine the cause of ASR-induced fracture of reinforcing
steels and present some proposals on maintenance. With the focus placed on materials,
external force (stress) and environment, the authors investigated the following subjects: (1)
fracture of reinforcing steels; (2) effects of bending on the material properties of reinforcing
steels; (3) effects of residual stress in reinforcing steels from bending, as well as effects of
ASR-induced expansion force; and (4) possibility of hydrogen embrittlement cracking in
reinforcing steels due to deterioration of concrete. The mechanism of reinforcing steel
fracture was estimated based on the investigation results. Moreover, proposals were made on
maintenance of the structures with a risk of reinforcing steel fracture, including
recommendations on selection of structures to be monitored and inspection and examination
on them.

2. REINFORCING STEELS FRACTURED DUE TO ASR

Characteristics of reinforcing steels fractured due to ASR are summarized below.

2.1 Appearance. Figure 1 shows a typical example of fractured stirrup. This sample had
corrosion in the bent section, being corroded more severely in the inside than in the outside of
the bend. Three ribs were found crushed on the inside of the bend, and corrosion was in the
vicinity of the crushed ribs. It was likely that these ribs had been crushed during the bending
process. Cracks initiated at the bases of the ribs on the inside of the bend.

2.2 Chemical Composition. Table 1 shows the chemical composition analysis results.
The values shown in the table are the mean values of fractured reinforcing steel samples
taken from an existing pier on the Hanshin Expressway, example values from an existing
structure on the Noto Toll Road (Tarui and Torii, 2010), and measurements of two
reinforcing steels manufactured in 1980s (Toyohuku, et al., 1988). All of these reinforcing
steels were assumed to be SD295A or SD295B of JIS G 3112 steel bars for concrete
reinforcement and were found satisfying the specification values. Contents of Cu, Ni, Cr and
N which were not specified in the specifications were higher in these samples as compared to
blast furnace steel bars. From the similarity in the chemical composition with the electric
furnace steel bars of 1980s, the fractured steel samples from the existing road structures were
considered to be electric furnace steel bars.

2
 節 の ribs
Crushed つ

Crack

Figure 1. An Example of Appearance of Fractured Reinforcing Steel

Table 1. Chemical Composition Analysis Results (Mass %)
C Si Mn P S Cu Ni Cr N
Hanshin Expressway,
0.27 0.25 0.77 0.029 0.029 0.31 0.11 0.14 0.0109
average
Noto Toll Road, example
0.28 0.18 0.73 0.018 0.026 0.45 0.10 0.18 0.0103
(D32)
Electric furnace steel bar
0.25 0.18 0.90 0.028 0.032 0.31 0.08 0.14 0.0109
of 1980s
Blast furnace steel bar of
0.25 0.29 1.39 0.029 0.025 0.01 0.02 0.03 0.0042
1980s
JIS G 3112 SD295A – – – £0.050 £0.050 – – – –
JIS G 3112 SD295B £0.27 £0.55 £1.50 £0.040 £0.040 – – – –

2.3 Shape of Ribs of Reinforcing Steels. Specifications of the inner radius of the bend
(hereinafter referred to as “bending radius”) in the JIS G 3112 vary depending on the
diameter of reinforcing steel: for instance, at least 1.5 times the nominal diameter for D16 or
at least two times the nominal diameter for D22. However, some samples were found to have
as small bending radii as 1.2d (d = nominal diameter), and all of them had cracks in the inside
of the bend.

Figure 2 shows examples from the rib shape measurement. Although the rib spacing and
height were satisfying the JIS G 3112 requirements, there were significant variations in the
radius of base curvature of the ribs. The values tended to be smaller on the fractured
reinforcing steels as compared to the currently marketed products. In the shown examples, it
was 2.85 mm (5.7 mm in diameter) and 3.3 mm (6.6 mm in diameter) on a fractured
reinforcing steel (D22), while being 7.85 mm (15.7 mm in diameter) on a product currently on
the market (D22). No exact numbers are specified for the radius of base curvature of the ribs
in the JIS G 3112 which only requires the rib base to have a shape which allows less
concentration of stress.

2.4 Work Hardening (Vickers Hardness). Figure 3 shows Vickers hardness

measurement results at the bend and the straight part of reinforcing steels. The inside and
outside hardness values were about 40 to 60 (HV10) higher than those at the center of the
bend or about 90 (HV10) higher than those in the straight part, showing increases due to the
work hardening. According to the approximate conversions between Vickers hardness and
tensile strength (Japanese Standards Association, 2009), Vickers hardness values of 240 to

3
260 (HV10) on the inside and outside of the bend were assumed to be equivalent to tensile
strengths of 765 to 825 N/mm2.

(a) Example of fractured reinforcing steel (D22) (b) Example of currently marketed product (D22)

Figure 2. Rib Shape Measurement Results

Measured Vickers hardness (HV10)

d: nominal diameter of
the reinforcing steel

Straight part
inside d/8 × Bend, Pier A
× inside d/4 × inside d/8 Bend, Pier B
× ceter d/2 × inside d/4 Straight part, Pier A
× outside d/4 × center d/2 Straight part, Pier B
× outside d/8 × outside d/4
Bend × outside d/8
d/8 d/4 d/2 d/4 d/8
Inside Center Outside
Measurement positions

Figure 3. Vickers Hardness Measurement Results

2.5 Properties of Fracture Surface. Figure 4 shows a macro view of the fracture
surface of a fractured reinforcing steel. The flat and granular fracture surface with very little
plastic deformation suggests a brittle fracture. Cracks initiated at the base of rib on the inside
of the bend and, as shown by the radial patterns on the surface, propagated toward the outside
of the bend. Minor shear lip was observed at the end point of fracture on the outer periphery.
Figure 5 shows the fracture surface observation results by a scanning electron microscope.
Cleavage or quasi-cleavage cracks were predominant in the crack propagation region starting
from the crack initiating point as shown in Figure 5-(a). Cleavage cracks were predominant in
a region about 12 mm inside the crack initiating point as shown in Figure 5-(b). Figure 5-(c)
shows the region of shear lip at the crack end point. The elongated dimples in the surface are
characteristic to ductile fracture, suggesting a failure by shear.

3. FRACTURE TOUGHNESS VALUE OF REINFORCING STEEL

Work hardening or strain aging in the bent section may lead to toughness reduction in the
reinforcing steel. If cracks are present, it is highly likely that resistance against fracture is
much affected. The authors investigated the fracture toughness value of fractured reinforcing
steels. Samples for the determination of fracture toughness in the bend must be taken from
the bent section of the reinforcing steels. However, required samples were not available from

4
appropriate portions. Instead, samples of the bend were prepared by using samples from the
straight part and introducing compressive strain in them to reproduce the strain occurring in
the inside of the bend, thereby simulating the effect of the bending process. Fracture
toughness test was carried using the specimens taken from these samples.

Inside of the bend Crack initiating point

Shear lip

Figure 4. Macro View of Fracture Surface of Fractured Reinforcing Steel (D16)

(a) Fracture surface near the (b) Fracture surface about 12 (c) Fracture surface at the shear
crack initiating point mm inside lip region

Figure 5. Fracture Surfaces of Fractured Reinforcing Steel

3.1 Samples. The test samples were fractured primary reinforcing steels of D32 taken from
a beam of an existing reinforced concrete bridge pier (hereinafter referred to as “fractured
reinforcing steels”). Control samples for comparison were standard reinforcing steels
complying with the current JIS requirements (hereinafter referred to as “JIS-complying
reinforcing steels”). The samples were confirmed to have required chemical composition of
JIS G 3112 steel bars for concrete reinforcement, satisfying all specification values. Table 2
shows the descriptions of the samples.

3.2 Test Method. Fracture toughness test was carried out at a room temperature (20°C) in
compliance with ASTM E1820 [Elastic-Plastic Fracture Toughness (JIC) Testing, Standard
Test Method for Measurement of Fracture Toughness]. The J value for stable ductile crack
growth was determined by using an electro-hydraulic servo fatigue tester with a maximum
capacity of 50 kN. The unloading compliance method was used to determine J values for
stable ductile crack growth (JQ). Details of the test method are not shown here due to the
page limit.

3.3 Test Results. Figure 6 shows the relationship between the fracture toughness value and
the amount of pre-strain in the fractured reinforcing steels and the JIS-complying reinforcing
steels. In the cases with no pre-strain (unprocessed) the JIS-complying reinforcing steel

5
exhibited a higher value than that of the fractured reinforcing steel. The fracture toughness
value decreased significantly in the both reinforcing steels with the introduction of
compressive strain simulating the strain in the bend. These results suggest that fracture
toughness value at the bend decreases due to work hardening, and that resistance against
fracture may be much affected if cracks are present.

Table 2. Test and Control Samples

Locations Reinforcing steel Specimen Hardness Amount of strain
represented sample types symbols (HV10) introduced
Straight part Fractured F-1 170 equivalent 0%
(unprocessed) JIS-complying S-1 170 equivalent 0%
Fractured F-2 250 equivalent 20.8% compression
Bend (processed)
JIS-complying S-2 250 equivalent 20% compression

■ JIS-complying reinforcing steel (unprocessed)

Fracture toughness value,

● Fractured reinforcing steel (unprocessed)

□ JIS-complying reinforcing steel
Kc(J) (MPa×Öm)

(20% compressive strain introduced)

○ Fractured reinforcing steel
(20.8% compressive strain introduced)

Pre-strain (%)

Figure 6. Fracture Toughness Values versus Pre-strain Amounts

4. STRESS IN REINFORCING STEEL
Residual stress from bending and stress from ASR-induced concrete expansion were
estimated to understand the change in stress in the reinforcing steels and investigate its effect
on the initiation and propagation of cracks.

4.1 Residual Stress in the Bend of Reinforcing Steel. Stress occurring in the
reinforcing steel was determined by FEM analysis simulating the bending process with a steel
bender. The change in stress during bending was understood as residual stress. The analysis
took into account the spring back phenomenon or elastic strain recovery after the removal of
confinement at the completion of bending on the reinforcing steel.

Elastic-plastic finite element analysis was adopted in this study, using the analysis
application of Abaqus/Standard version 6.5-3.

Reinforcing steel of D16 was modeled into a symmetrical model about the center of bending,
using about 23,000 trilinear hexahedral elements. Bending radius was 1d (d = nominal
diameter), and the properties of the reinforcing steels were as follows: Young's modulus =
210 GPa; Poisson's ratio = 0.3; and yield stress = 305 MPa.

Figure 7 shows an example of axial stress distribution in a reinforcing steel at the completion

6
of bending. Compressive stress occurred in the inside of the bend from the center of the cross
section, and tensile stress occurred in the outside of the bend. Figure 8 shows an example of
axial stress distribution after complete release of the load. The tensile stress regions in the
outside of the bend almost disappeared, and tensile stress occurred in the inside surface layer
of the bend.

Axial stress showed a general change from tensile stress to compressive stress in the outside
of the bend, and a change from compressive stress to tensile stress in the inside surface layer
of the bend during the period from completion of bending to complete release of the load.
Tensile residual stress of about 300 MPa remained at the rib base, which was considered to
be the cause of crack initiation and propagation.
10
Tension ① (2)(3)
(1) ②
③④(4)
8
Outside of the bend

Reinforcing steel diameter

6

鉄筋径（D16鉄筋中心＝0）
sec. (1)

(D16, center = 0)
4 断面①
sec.
断面②(2)
2 断面③(3)
sec.
断面④
0 sec. (4)
Compression -2
Stress (MPa) -4

-6

-8
Inside of the bend
-10
-500 -400 -300 -200 -100 0 100 200 300 400 500
鉄筋軸方向の応力（MPa）
Axial stress in reinforcing steel (MPa)
(a) Stress contour diagram (b) Stress distribution

Figure 7. Axial Stress in Reinforcing Steel at the Completion of Bending Process

10
Tension ① (2)(3)
(1) ②
(4) 曲げ外側of the bend
③④ Outside 8
Reinforcing steel diameter

sec.
断面①
(1)
6
sec.
断面② (2)
鉄筋径（D16鉄筋中心＝0）
(D16, center = 0)

断面③
4 sec. (3)
断面④
2
sec. (4)

Compression -2

Stress (MPa) -4

-6

-8
曲げ内側
Inside of the bend
-10
-500 -400 -300 -200 -100 0 100 200 300 400 500

Axial stress鉄筋軸方向の応力（MPa）
in reinforcing steel (MPa)
(a) Stress contour diagram (b) Stress distribution

Figure 8. Axial Stress in Reinforcing Steel after Complete Release of the Load
4.2 Stress in Reinforcing Steel Induced by Concrete Expansion. Stress in
reinforcing steels induced by concrete expansion was estimated using strain values in
reinforcing steels in concrete specimens prepared. Deterioration in an actual bridge pier was
successfully reproduced in these specimens through outdoor exposure.

The specimens were prestressed concrete beams of 750×750×5000 mm. Weldable strain
gauges were attached to shear reinforcing steels for measurement. Strain exceeded 15,000μ in

7
the middle part on 900 days (Sasaki et al., 2009). Stress corresponding to the strain of
15,000μ was estimated from the stress-strain curves of the shear reinforcing steels used. The
estimation revealed that stress in some reinforcing steels could have reached a yield stress
level of 350 MPa. These results suggest that actual stress including residual stress may be
extremely high in reinforcing steels.

5. HYDROGEN EMBRITTLEMENT OF REINFORCING STEEL

Hydrogen embrittlement cracking is likely to be one of the causes of reinforcing steel fracture
induced by ASR. Since the major causal factor of hydrogen embrittlement is diffusible
hydrogen, analysis of hydrogen absorbed in steel is extremely important to discuss the
possibility of hydrogen embrittlement. The authors investigated the amount of diffusible
hydrogen occluded in reinforcing steels in ASR-damaged concrete and determined the
possibility of hydrogen embrittlement cracking from the relationship between the tensile
strength of steel and the amount of diffusible hydrogen occluded in the steel.

5.1 Occluded Hydrogen. Samples of bent sections were taken from three stirrups in an
existing bridge pier constructed 35 years ago, and occluded hydrogen was measured using an
atmospheric pressure ionization-mass spectrometer (API-MS). Figure 9 shows a comparison
with separate measurement results on concrete specimens. It was revealed that more
diffusible hydrogen was occluded in the inside and outside of the bend than at the center of
the bend. About 0.4 ppm of occluded hydrogen was present in the specimen with heavier
corrosion. The dependence of the amount of occluded hydrogen on the degree of corrosion
was similar in pattern between the existing pier and the concrete specimens in the separate
measurement, with more occluded hydrogen present in reinforcing steels with more severe
corrosion.

Figure 9. Comparison of Diffusible Hydrogen Occluded in Reinforcing Steels

between the Test Specimens and Existing Bridge Pier
5.2 Possibility of Hydrogen Embrittlement. Risk of hydrogen embrittlement cracking
was evaluated in the relationship between the steel strength and environmental severity
(Figure 10), using the measured amounts of occluded hydrogen in the reinforcing steels and
the tensile strength values in cracking regions calculated from the hardness measurement
results. All reinforcing steel types in this study were found to fall on the border between the

8
safe and unacceptable zones. Kobayashi et al. (2010) investigated the possibility of hydrogen
embrittlement cracking in reinforcing steels by experiment. In their examination on the effects
of strain rate and hydrogen charging current on embrittlement cracking, embrittlement was
recognized even with a slight amount of hydrogen charging current when strain rate was
adequately slow in the simulated environment of ASR-damaged concrete during the slow
strain rate tensile test. This suggests the possibility that embrittlement may occur in the
presence of an extremely small amount of hydrogen generation which accompanies corrosion
reaction, when strain rate is as slow as that in ASR expansion. Consequently, it is highly
possible that hydrogen embrittlement was involved in the fracture of reinforcing steels.

100.00 Bar
Blister zone
H2S environment

No. 1

No. 2
Line
pipes Unacceptable
10.00 No. 3
Diffusible hydrogen amount (ppm)

zone
700 MPa

High-tension
steel, LPG
tanks, oil well Practical safety limit in
Under water,

pipes stress concentration

1.00
salt water
Natural atmospheric
environment

0.10 Safe High-tension

zone bolts
Prestressing
Concrete

steel

Safety limit with Aircrafts

preexisting cracks
0.01
0 500 1000 1500
Tensile strength (MPa)

Figure 10. Risk of Hydrogen Embrittlement Cracking in the Relationship

between the Steel Strength and Environmental Severity (Matsuyama, 1989)
6. MECHANISM OF ASR-INDUCED REINFORCING STEEL FRACTURE
Reinforcing steel fracture induced by ASR was considered to occur by the mechanism
described in Figure 11.

7. PROPOSALS ON MAINTENANCE
Fracture of reinforcing steels induced by ASR affects the load carrying performance of a
structure. Therefore, early detection of reinforcing steel fracture is essential to proper
maintenance of ASR-affected structures. Selection of structures with a risk of reinforcing
steel fracture basically requires understanding of the characteristics in appearance
degradation or other changes over time. It is also important to inspect other sections or
structures constructed at similar time to or in the same project with those already found
affected. Non-destructive testing is recommended for the diagnosis of reinforcing steel
fracture. Technological development is expected to improve the precision and accuracy of

9
non-destructive testing techniques. In consideration of the mechanism of reinforcing steel
fracture, the diagnosis may be basically focused on the bend of reinforcing steels.

(1) Bending (2) Strain aging

 Cracks may occur at the rib base.

 Fracture toughness decreases
 Strain aging promotes
with work hardening.
embrittlement.
 Tensile residual stress occurs.

(3) ASR-induced cracking (4) Spring back force due to ASR

 ASR-induced cracks occur.  Spring back force occurs in
 Reinforcing steel corrosion occurs due to the reinforcing steel due to
penetrating corrosive factors or chloride ions ASR expansion in addition to
contained in sea sand of fine aggregate. the tensile residual stress in
the inside of the bend.
 Amount of occluded
 Stress is concentrated in
hydrogen increases in the
micro parts with cracks.
reinforcing steel.
 Fracture occurs due to crack
(Hydrogen embrittlement ASR
propagation.
cracks may occur and expan-
cause fracture.) sion

Figure 11. Proposed Mechanism of Reinforcing Steel Fracture

REFERENCES
Japanese Standards Association. (2009). JIS Handbook: Ferrous Materials & Metallurgy I-
2009, Tokyo, 1951-1953.
Kobayashi, M., Nishikata, A. and Tsuru, T. (2010). “Effect of Stain Rate and Hydrogen
Charging Current on Hydrogen Embrittlement of Carbon Steel in High Alkaline
Chloride Environment.” zairyo-to-kankyo, Japan Society of Corrosion Engineering,
59(4), 129-135. (in Japanese)
Matsuyama, S. (1989). Delayed Fracture, Nikkan Kogyo Shimbun Ltd., Tokyo. (in
Japanese)
Sasaki, K., Hisari, Y., Igarashi, H. and Miyagawa, T. (2008). “Analysis of the stress and
strain on bent section of reinforcing bar by finite element method.” Proceedings of
Japan Concrete Institute, JCI, 30(1), 987-992. (in Japanese)
Sasaki, K., Matsumoto, S., Kuzume, K., Kanaumi, S. and Miyagawa, T. (2009). “Expansion
Behavior of Prestressed Concrete Beams Deteriorated by Alkali-Silica Reaction in
Long Term Exposure Test.” Proceedings of 4th International Conference on
Construction Materials: Performance, Innovations and Structural Implications, JCI,
739-746.
Tarui, T. and Torii, K. (2010). “Fracture Mechanism of Steel Bar by Alkali Silica Reaction.”
zairyo-to-kankyo, Japan Society of Corrosion Engineering, 59(4), 143-150. (in
Japanese)
Toyohuku, T., Yoshioka, H. and Yoshimura, Y. (1988). “Actual quality of electric furnace
steel bars.” Nihon Doro Kodan Laboratory report, Nihon Doro Kodan Laboratory,
25, 59-71. (in Japanese)

10