ISIJ International, Vol. 53 (2013), No. 2, pp.

382–384

bution and plastic strain distribution directly ahead the crack

Note
Physical Interpretation of Grain Refinement-induced
Variation in Fracture Mode in Ferritic Steel
Hai QIU,1)* Linning WANG,2) Toshihiro HANAMURA1)
and Shiro TORIZUKA1)
1) Research Center for Strategic Materials, National Institute for
Materials Science, 1-2-1 Sengen, Tsukuba, Ibaraki, 305-0047 Japan.
2) Tianjin Pipe Group Corporation, Tianjin, China.
(Received on June 7, 2012; accepted on October 3, 2012)
KEY WORDS: work-hardening; ductile fracture; brittle
fracture; ferrite; grain size.
1. Introduction
To utilize the effect of grain refinement in enhancing
strength and toughness, people have been making effort to
decrease ferrite grain size, and recently succeeded in devel-
oping Si–Mn ultrafine-grained steels with ferrite grain size
even less than 1 μ m. The toughness of ferritic steel has been
confirmed to increase with a decrease in ferrite grain size.1–5)
In all the parameters applied to evaluate toughness, such as
fracture toughness (stress intensity factor, crack tip opening
displacement, J integral, etc.) and impact toughness
(absorbed energy, ductile-brittle transition temperature
(DBTT), and so on), only DBTT directly reflects the varia-
tion in fracture mode. Morrison6,7) used “Yoffee diagram” to
interpret DBTT, in which the fracture mode transition from
ductile to brittle is recognized to occur at the condition
where yield stress (or local stress) rises above the brittle
fracture stress. It has been found from Charpy impact tests
that DBTT is related to the ferrite grain size D by4,5)
DBTT = T0 − kD −1/ 2 ......................... (1)
where T0 and k are constants (k > 0).
There are two basic brittle fracture modes: intergranular
separation and transgranular cleavage. For a well-controlled
alloy, transgranular cleavage is dominant.6) It should be not-
ed that “brittle fracture mode” in the present study means
the transgranular cleavage. Equation (1) indicates that
decreasing ferrite grain size lowers the DBTT. Although the
decreased DBTT indirectly reflects the effect of grain refine-
ment on fracture mode, Eq. (1) doesn’t answer why grain
refinement can affect fracture mode. The correlation between
grain refinement and fracture mode is not yet physically and
systematically discussed, and thus we try to interpret it from
the viewpoint of fracture mechanics in the present study.
tip along crack direction are schematically shown in Fig. 1.
Under a certain loading level, the crack begins to propagate.
We reveal the effect of grain refinement on fracture mode
through analyzing the effect of grain size on the conditions
required for the initiation of cracking.
Because of the presence of crack, in addition to stress
concentration and strain concentration, stress singularity and
strain singularity exist near the crack tip. For a given crack,
stress or strain concentration is related to the material’s
work-hardening ability. The effect of work-hardening is sys-
tematically discussed in Ref. 8). For plane strain state, alter-
ing work-hardening exponent (n) will change the local
stress field ahead of a crack tip. If we take remote flow
stress σ 0 as σ 0/E=0.0025 (where E is Young’s modulus), the
local peak stress (as shown in Fig. 1) can reach ~3σ 0, ~3.6σ 0
and ~5σ 0 for n=0, 0.1 and 0.2, respectively.8) This result
shows that decreasing work-hardening exponent (i.e., work-
hardening ability) reduces the level of local stress ahead of
a crack. As for the corresponding near-tip equivalent plastic
strain distribution, it is essentially independent of work-
hardening ability.9)
It has been known that ductile fracture and brittle fracture
are strain-controlled and stress-controlled, respectively.8)
The critical condition for the initiation of brittle fracture (or
ductile fracture) is that over a characteristic distance l0, as
shown in Fig. 1, the local tensile stress perpendicular to the
crack plane σ yy (or local equivalent plastic strain ε p) exceeds
the critical fracture stress σ cr (or the critical fracture strain
ε cr). Which mode a material fractures in is depended on
which critical condition is first attained: if the condition for
ductile fracture is first satisfied, ductile fracture occurs, oth-
erwise, brittle fracture takes place.
Otsuka et al.10) used a schematic diagram to interpret the
correlation of stress-strain curve with the fracture conditions

2. Physical Models for Ductile Fracture and Brittle

Fracture
Suppose a stationary crack subjected to a load of Mode I
(i.e., opening mode). The opening tensile stress (σ yy) distri-
* Corresponding author: E-mail: QIU.Hai@nims.go.jp Fig. 1. Schematic diagrams of microscopic fracture criteria for
DOI: http://dx.doi.org/10.2355/isijinternational.53.382 brittle fracture and ductile fracture.

© 2013 ISIJ 382

 ISIJ International, Vol. 53 (2013), No. 2
(σ cr, ε cr) for all fracture modes in tensile tests, as shown in
Fig. 2. Three typical cases are given in Fig. 2. In Case A and
Case C, as stress and strain increase, only one critical value
(σ cr in Case A and ε cr in Case C, respectively) is exceeded,
accordingly the fracture mode for Case A and Case C is brit-
tle and ductile modes, respectively. In case B, the ductile
fracture condition is first satisfied, and followed by the crit-
ical fracture stress being exceeded, and thus the material
first fractures in ductile and then in brittle mode.
We select the minimum stress or strain point ahead of the
crack tip along the crack direction within the characteristic
distance l0 in Fig. 1, and depict the relationship between σ yy
and ε p at this point in Fig. 3 with the same way used in Fig.
2. Three typical cases (solid lines) are also given in Fig. 3
as those in Fig. 2. We assume the solid lines correspond to
the ferrite grain size Di. When ferrite grains are refined from
Di to Di+1 (Di > Di+1), the critical values (σ cr and ε cr) and
the equation relating stress and strain will change.
Qiu et al.11) investigated the work-hardening of ferritic
steels with carbon content blow 0.75%, and found that
decreasing grain size (D) lowers the work-hardening expo-
nent (n). When carbon content is below 0.22%, n and D are
Fig. 2. Schematic description of the fracture initiation for ductile,
ductile-brittle, and brittle fracture.
Fig. 3. Schematic interpretation of the effect of grain refinement
on fracture mode. Solid and dashed lines corresponding to
the grain size of Di and Di+1, respectively (Di > Di+1).
383
related by
n = 0.3301 − 0.2401D −1/ 2 (D in μ m) ........... (2)
The valid range of D in Eq. (2) is 1.30 to 37 μ m. Within this
range, n linearly decreases with D–1/2, and when D is less
than 1 μ m n reduces to zero.11) As aforementioned, decreas-
ing n reduces the local stress concentration ahead of a crack,
but has little effect on the local strain distribution. There-
fore, the decrement in n induced by grain refining from Di
to Di+1 merely lowers local stress, and as a result the three
solid lines will drop down as shown in Fig. 3.
Cottrell analyzed the critical fracture stress for transgran-
ular cleavage with dislocation theory,12) and found that the
critical fracture stress σ cr is in proportion to D–1/2. For steels
and irons, the relationship between σcr and D in the range
of 10 to 10 000 μ m is given in the form13)
σ cr = 351.5 + 3.30 × 103 D −1/ 2 (D in μ m; σ cr in MPa)
.......................................... (3)
Therefore, grain refinement elevates the σ cr, as shown in
Fig. 3.
To investigate the dependence of critical fracture strain on
ferrite grain size, tensile tests were performed at room tem-
perature and at a crosshead rate of 0.4 mm/min with speci-
mens of 4 mm in diameter and 10 mm in gage length. The
used steels and samples are as follows: (1) 0.10C–0.3Si–
1.45Mn (in wt%) ferritic steels, grain size: 0.9, 4.6, 6.2 and
9.6 μ m; (2) 0.16C–0.3Si–1.45Mn ferritic steels, grain size:
0.9, 2.9, 4.5 and 7.1 μ m; (3) 0.45C–0.3Si–1.45Mn ferritic
steels, grain size: 0.46, 0.60, 4.83 μ m; (4) 0.75C–0.3Si–
1.45Mn ferritic steels, grain size: 0.58, 0.80 and 2.01 μ m.
All the samples are only composed of ferrite as well as sec-
ond phase particles (cementite particles). Figure 4 shows
the microstructure of the samples of 0.10C–0.3Si–1.45Mn
steel. Ferrite grains are equiaxial. The fracture strain of the
tensile specimens, ε f, was given by the equation
ε f = ln( A0 / A f ) ............................ (4)
where A0 and Af are the area of the initial and fractured
cross-section, respectively. The results are given in Fig. 5.
The results show that decreasing ferrite grain size reduces
fracture strain. Carbon content, i.e., cementite volume frac-
Fig. 4. Microstructures of the samples of 0.10C–0.3Si–1.45Mn
steel. (a) 0.9 μ m, (b) 4.6 μ m, (c) 6.2 μ m, and (d) 9.6 μ m.4)
© 2013 ISIJ
 ISIJ International, Vol. 53 (2013), No. 2
Fig. 5. The dependence of fracture strain on ferrite grain size in
tension tests.
tion, also affects the fracture strain; higher carbon content is
harmful to fracture strain, which accords with Torizuka’s
result.14)
Round smooth tensile specimens were applied in the pres-
ent study, and thus before the onset of necking the speci-
mens are in uniaxial stress state, but after necking stress
state turns into triaxial state. For a sample with a crack,
because the crack tip is sharp, the stress triaxiality in the
region ahead of the crack tip is higher than that in the round
smooth tensile specimen. Therefore the fracture strain
obtained from tensile tests is usually larger than the critical
fracture strain in the local region ahead of the crack.
Although the exact values of fracture strain in the two cases
are different, if fracture mode remains unchanged, the ten-
dency of fracture strain associated with grain size in both
cases is identical. Therefore, the critical fracture strain in
Fig. 3 also declines when ferrite grain size decreases from
Di to Di+1. It should be noted that Fig. 3 merely shows the
trend of the grain refinement-induced variations in local
stress, strain, σ cr, and ε cr, instead of exact values.
Figure 3 schematically interprets the effect of grain
refinement on fracture mode. In the brittle fracture region
(for example, Case A), the multiplier effects of elevated σ cr
and decreased local stress increase the difficulty of attaining
the critical condition of σ yy > σ cr. This means that in the brittle
fracture region, grain refinement increases the fracture tough-
ness of ferritic steels, which has been verified by Armstrong’s
work15) in which fracture toughness KIC has been shown
to be proportional to D–1/2 ( K IC = 12.8 + 3.04 D −1/ 2 ; KIC in
© 2013 ISIJ
MPam–1/2, D in mm). If the multiplier effects are significant
enough, fracture mode can be changed from brittle to duc-
tile. This type of transformation has been confirmed in the
ferritic steels with the composition of (0.10C or 0.16C)–
0.3Si–1.45Mn (in wt%).4)
In the ductile fracture region (e.g. Case C), for a given
stress, relatively large strain is produced for fine ferrite grain
rather than for coarse ferrite grain. This increased strain as
well as the decreased critical fracture strain induced by grain
refinement makes ductile fracture easily occur at a relatively
low load, which indicates that grain refinement has a negative
impact on toughness. The experimental results that ferritic
steels with fine grains have lower values of absorbed energy
than those with coarse grains agree with this conclusion.4)
For Case B, grain refinement changes the fracture mode
from ductile/brittle to ductile fracture.
3. Summary
In summary, grain refinement in ferritic steels lowers the
work-hardening ability, and accordingly decreases the local
stress ahead of a crack. The critical fracture stress for brittle
fracture and the critical fracture strain for ductile fracture
are correspondingly increased and decreased, respectively.
These variations are responsible for the effect of grain
refinement on fracture mode. Grain refinement positively
impacts on toughness in brittle fracture region while nega-
tively in ductile fracture region.
REFERENCES
1) T. Hanamura, M. C. Zhao, H. Qiu, F. X. Yin and K. Nagai: Tetsu-
to-Hagané, 95 (2009), 71.
2) H. Qiu, K. Enami, Y. Kawaguchi, K. Hiraoka, K. Nagai and Y.
Hagihara: J. Mater. Sci., 43 (2008), 1910.
3) H. Qiu, Y. Kawaguchi and C. Shiga: J. Mater. Sci., 39 (2004), 3733.
4) H. Qiu, R. Ito and K. Hiraoka: Mater. Sci. Eng. A, 435–436 (2006),
648.
5) T. Hanamura, F. X. Yin and K. Nagai: ISIJ Int., 44 (2004), 610.
6) J. W. Morris, Jr.: Mater. Res. Soc. Symp. Proc., 539 (1999), 23.
7) J. W. Morris, Jr.: 1st Int. Conf. on Super-High Strength Steels,
Associazione Italiana di Metallurgia and CSM - Centro Sviluppo
Materiali, Italy, (2005).
8) R. O. Ritchie and A. W. Thompson: Metall. Trans. A, 16 (1985), 233.
9) R. M. McMeeking: J. Mech. Phys. Solids, 25 (1977), 357.
10) A. Otsuka, T. Miyata, S. Nishimura and M. Ohashi: Trans. Jpn. Soc.
Mech. Eng. A, 76 (1981), 286.
11) H. Qiu, L. N. Wang, T. Hanamura and S. Torizuka: Mater. Sci. Eng.
A, 536 (2012), 269.
12) A. H. Cottrell: Trans. Met. Soc. AIME, 212 (1958), 192.
13) A. W. Armstrong: Fracture 1977, Vol. 4, ed. by D. M. R. Taplin,
Pergamon Press, Oxford, (1977), 83.
14) S. Torizuka, E. Muramatsu, S. V. S. N. Murty and K. Nagai: Scr.
Mater., 55 (2006), 751.
15) R. W. Armstrong: Eng. Fract. Mech., 28 (1987), 529.
384