UDC 669.141.241.2 : 539.52 : 669.112.228.1 : 669.112.227 .1 : 620.186.8 : 539.4.

015

Relation between Hot Ductility and Grain Boundary

Embrittlement of Low-carbon Killed Steels·

By Kazuo YAMANAKA,·· Fukunaga TERASAKI,.* Hiroo OHTANI,.*

Mitsuo ODA" and Masahiro YOSHIHARA**

Synopsis 2. Hot T ensile Test

The hot ductility of low-carbon killed steels has been Jtudied by the Hot tensile specimens with a 40 mm gauge length
hot tensile tests, microscopic observations, and the frac ture surface ob- and 5.0 mm dia. were taken from the cast ingots
servations .
transverse to the casting direction. The sh ape of the
The ductility during cooling after solution treatment at 1 350°C decreased
hot tensile test specimen is shown in Fig. 1. The
remarkably in the temperature range between 700° and 800°C and the
hot tensile tests were carried out using a n INSTRON
fracture surface showed the intergranular ductile fracture accompanied
with AlN or MnS precip itates. This loss of ductility is related to the
tensile machine; samples being heated by the in-
formation of the primary ferrite along the austenite grain boundaries and frared heating apparatus (SHINKU-RIKO RH-L-
the precipitation of AlN or MnS particles at the grain boundaries, and E2) in a n argon atmosphere. The test pieces were
becomes very large when it occurs concurrently. The intergranular ductile given a homogenizing treatmen t for 10 min at 1 350°C
fracture can be induced by the micro-void coalescence nucleated at the before cooling rapid ly d own to a test temperature in
grain boundary precipitates as the result of strain concentration at the .film- the range between 500° a nd 1 OOO°C a nd strained at
like primary f errite formed along the austenite grain boundaries . a rate of 2.1 X 1O-3jsec. The schematic diagram of
the a ppa rat us is shown in Fig. 2. Fracture diameter
I. Introduction was m easured after hot tensile test and fracture strain
A loss in the hot ductility and surface cracks of the cJ = ln Ao j AJ (where Ao is initia l diameter and AJ is
continuous casting sla b or the ingo t cracking are well- fractured dia meter) was ob tained .
known to occur in certain low-alloyed steels treated T he fracture surface was observed with the JSM-2
with a luminium.1 - 7 ) The cracks can occur a long type scanning electron microscope. The microstruc-
the a ustenite g rain boundaries. Thus the occurrence ture a nd the precipitation of the second phase were
of such intergra nular fracture seems to be closely
related with the transformation cha racteristics caused
by thermal history, together with the grain boundary
embrittlement accompanied by the precipitation of
second phase particl es at the a ustenite grain bound a-
ries when cooling from hig h temperatures. H ow-
ever, the mecha nism of the intergranula r hot-cracking
h as not been fully understood . Thus the aim of the
Fig. I . Shape of the hot tcnsile tes t specimen.
present investigation is to reveal the mechanism of
the loss in hot du ctility o f low-carbon kill ed steels.
load cell
II. Experimental Procedure
1. Materials +""=<-it-thermo-couple

The alloys were m elted to a llow study of the in-

r-----· -- ---,
fluence of Al a nd S in a base composition of 0.22 C- I I
I I
0.35Si- 1.45Mn- 0.020P. Three steels were examined , I
I
I
I
I
materials being from ingot production and laboratory I
I
I - - ; - i nfrared imoge furnoce
casts. The laboratory melts were cast as 50 kg hig h- I
I
frequency induction m elts. The ingot production ' - tensile speci men
Ir-t- - - i -
were cast into a 10 ton mould at a temperature of
r----.;... QUARTZ tube
abo ut I 550°C. The chemical compositions are shown
in Table 1.

Table 1. Chemical co mpositions of steels. (wt % )

woter-_~-=

Steel C Si Mn P S sol. Al N

H 0.23 0.31 1.48 0 . 026 0 .020 0 .084 0.0111

H-O 0.22 0.37 1.42 0 .016 0 .005 0 .0048
H-I 0 . 21 0.39 1.45 0 . 018 0.031 0.0059
Fig. 2. Schematic diagram of the a ppara tus.

* Originall y published in Tetsu-to-Hagane, 65 (1979), 1410, in J a pa nese. Engl ish version receivcd O ctober 5, 1979.
** Central R esearch Laboratories, S umi tomo M etal Industries, Ltd., ishinagas u-hondori , Amagasa ki 660.

( 810 ) Research Article

 Transactions ISIJ, Vol. 20, 1980 ( 811 )

examined by both optical and electron microscopies. 900°C. T h e measurement of the strain in both phases
The optical microscopic observation was carried out of ferrite a nd a ustenite was carried out by means of
after etching in 2% nita l. The specimens for electron optical a nd scanning electron microscopy.
microscopy were prepared as carbon replicas extracted
in 3% Br- alcohol after etching in 2% picral a nd were III. Results
examined in a Hitachi HU-IID electron microscope
operating at 100 KV. 1. Hot Tensile Test
Figure 4 shows the results of the hot tensile tests.
3. The Model Experiment The fracture strain increased with increasing of the
The model experiment was performed , using the test temperature up to 700°C, decreased in the tem -
ferrite-austenite two phase structures which were perature range between 700° a nd 800°C and increased
artificially made, to clarify the mechanism of the loss again above 800°C. The fracture strain of AI-killed
in hot ductility of low-carbon killed steels. This two steel (Steel H ) d ecreased remarkably in the tem -
phase material preparation is shown schematically in perature range between 700° and 800°C. On the
Fig . 3. other ha nd , the ductility loss of Si-killed steels (Steel
Ferrite phase is pure iron or 2vol % alumina doped H-I and H-O ) in the same temperature range was
iron and austenite phase is 18%Cr- 12 % Ni austenite small as compared with AI-killed steel. In this case
stainless steel. The reason why the artificial ferrite- the degree of ductility loss is much smaller in Steel
austenite two phase structure was composed of pure H-O than in Steel H-l.
iron or 2vol % alumina doped iron a nd 18 % Cr- Even at temperatures above 800°C the fracture
12 % N i a ustenite stainless steel is as follows: strain of AI-killed steel was lower than that of Si-
(1) Stable b.c.c. and f.c.c. crystallographic struc- killed steels.
ture can b e obtained in spite of the tempera ture a nd
holding time at temperatures below about 900°C. 2. Fracture Surface Observations
(2 ) Austenite stainless steel shows simila r duc- The fracture surface after the hot tensile test was
tility ch aracteristics with the austenite phase of carbon observed with the scanning electron microscope
steel at hig h temperatures. (SEM ). Figure 5 shows the results of the fracture-
The width of the ferrite phase is both 45 ~ 95 fl surface observation by SEM. The fracture surface
a nd 200~270 fl. Alumina doped iron was m ad e by showed the intergranular fracture in the temperature
the method that a lumina particles were artificially range between 700° a nd 800°C correspond ing to the
mixed with electrolytic iron powd er a nd hot-rolled d ecrease in du ctility a nd showed the ductile fracture
after sintering in a hydrogen a tmosphere. Alumina in the temperature below 700°C or above 800°C.
pa rticle size was revealed smaller than 5 fl. Tensile Steel H, showing the most remarka ble ductility loss,
test specimens were taken at 90° and 45° as a n a ngle showed almost all the intergranula r fracture in the
of ferrite phase to the tensile axis direction as shown temperature range between 750° and 775°C and the
in Fig. 3. Hot tensile tests were carried ou t on a n ductile fracture with partially intergranular fracture
I NSTRON tensile machine at a strain rate of 3.3x even at temperatures of 725 ° and 800°C. On the
1O- 3 jsec in the temperature ra nge between 500° and other hand , Steel H-l with hig h sulfur showed abo ut
60 % intergranular fracture, and Steel H-O with low
sulfur showed o nly a bout 20 % intergranula r fracture
in the temperat ure ra nge between 750° a nd 775 °C.
austenite stainless steel
( 18Cr-12Ni) Microscopic observation of sections of the fractured

pure iron ar 2 % alum ina

doped iron
heol pOllern : /'6.82

l
1250 CX2h

*
for~.d and hOI-roiled
~:it 2. rmT"~'I~:g
10m," 1,/I,
II
/ ,,
+ c
II
lenslle les

UJ as cos, kI.,,:':

"!'-I.oi ";',;1'\- y1//// 0

~ /? ,,~-~ ~
-/j
H -0

maC hin e4 \~:-' ~

O ~--~--~--~~--~--~~
:00 600 700 800 900 1000
Fig. 3.
The me thod of the ferrite-
test temperature (Oel
a usteni te two phase structure Fig. 4. The results o[ the measurement o[ fracture strai n
8=9 0° 8= 45° preparation . by hot tensile tes ts.

Research Article
( 812 ) Transactions ISIJ, Vol. 20, 1980

speCImens revealed the primary ferrite formed along and wide (200-27011 ) width in the temperature
the austenite grain boundaries in the temperature range between 600° and 900°C. In this case the
range between 750° and 775°C showing a severe strain concentration was larger in the narrow than
ductility loss as shown in Photo. 1. in the wide width. And also the strain concentration
Detailed observation by SEM indicated that the was larger at the 90° inclined ferrite phase to the
intergranular fracture surface was composed of in- tensile axis than at the 45°.
numerable dimples caused by micro-void formation In the case where the ferrite phase was ·alumina
as shown in Photo. 2. Such intergranular fracture doped iron, the temperature variation of the strain
accompanied with dimples is called the intergranular concentration was considerably less in comparison to
ductile fracture. The precipitates extracted from the the case where the ferrite phase was pure iron as shown
intergranular ductile surface of Steel H were mainly in Fig. 7.
id entified as Al as shown in Photo. 3. And Steel
H-O and H-l with no Al addition were identified as IV. Discussion
a- MnS. At temperatures above 800°C the austenite
phase only was formed, but the precipitation of AlN 1. M echanism of the Loss in Hot Ductility
still occurred at the austenite grain boundaries in According to the hot tensile tes ts, the fracture
Steel H, as shown in Photo. 4. strain d ecreased in the temperature range between
700° and 800°C, and extremely d ecreased in the
3. Hot T ensile T ests oj the Artificial T wo Phase Steels
Hot tensile tests of the artificial two phase steels
were carried out in the temperature range between
500° and 900°C and measured the strain in both
phases of ferrite and austenite, respectively. Figure
6 showed the temperature variation of the strain rate
of ferrite- a ustenite two phase calc" where the ferrite
phase is pure iron. The strain concentration in the
ferrite phase occurred in both narrow (45-9511 )

123] : ductile fracture (100%)

~ : intergranular fracture (100 %) "-

Photo . I . The mjcro-structure of the hot tensile test specimen
k ;'~i,. . duct i Ie + intergranular of Steel H showing the ferrite network along the
. fracture austenite grain boundaries . Etched by 2% nita!.
Fig. 5. The results of the fracture surface observation by (Test specimen was heated for 10 min at I 350°C,
scanning electron microscope. cooled rapidly to 750°C and then drawn at 750°C.)

Photo. 2.
The fracture su rface of the hot tensile
tes t specimen of Steel H showing the in-
tergranula r ducti le fracture. (Test speci-
men was heated for 10 min at I 350°C,
cooled rapid ly to 750°C and then drawn
at 750°C.)

Research Article
 Transactions ISIJ, Vol. 2 0 , 1980 ( 813 )

grain boundary
a I ~

)
,(
•

a: brigh t fi eld image

b: elec tron diffraction pa ttern
c: its schema ti c representation
Photo. 4. AIN precipit a tes ex tracted from Steel H treated
isothermally for 20 min at 900°C after soluti on-
trea tment at I 350°C.

o. 6 - 45-
0 . : 8 - 90-
50 phos e : narrow wid rh (45 ' 95 ~ )
• • • rj..
o o:c( phose : wid e width (200"2701-1)

40 .I

i:\ t
...
~
I \
30 / .........
c o
UJ

r~\
20
i\
.. . I \
\
10
--;;.~ -:i /·°"0 j---
O ____
!\:-:;i
o? S- o" ----6 0 _

o 500 600 700 800 900

test temperat(!re (oC)

Fig. 6. T emperature d ependence of the strain ratio of
ferrite- austenite two ph ase ca/cp where the ferrite
a: bright field image
phase is p ure iron. (0 is an angle of ferrite phase
b: electron diffraction pattern
to the tensile axis direction .)
c: its schematic representation
Photo. 3. AIN precip itates extracted from the intergranular
°o•. 1-45"
ducti le surface of Steel H. (Test specimen was
heated for 10 min at I 350°C, cooled rapidly to 750°C
and then drawn at 750°C.)
..
° 0
, 0-90"
((phose : narrow widfh (4S ..... 951J)
,,,phose , wide width 1200~270 ~ )

temperature range between 750° and 775°C. In this ...

W
20

ductility loss region, as shown in Photo. 2, the fracture

surface showed th e intergranular ductile fracture "o
W

caused by micro-void formation. A IN or MnS pre-

10 ~,
cipitates existed at the bottom of micro-voids , and the
in terval of each void was roughly consistent with the ~---.;:~
interval of precipitates. And also the fi lm -like pri-
mary ferrite formed a long the austenite grain bound-
~___

~o
,
~~1~' I
---8 a0
I

aries in the temperature range corresponding to the o 500 600 700 800 900
decrease in ductility. Therefore, the intergranular te st temperature (OC)
ductile fracture can occur by micro-void coalescence Fig. 7. Temperature d ependence of the strain ratio of
nucleated at the g rain boundary precipitates as the ferrite-austenite two phase ca/c r' where th e ferrite
result of strain concentration at the film-like primary phase is 2% a lumina doped iron. (0 is a n angle
ferrite formed along the austenite grain boundaries. of the ferrite phase to the tensile axis direction .)

Research Article
( 814 ) Transactions ISIJ, Vol. 20, 1980

The main factor of the loss in hot ductility can be induced by the micro-void coalescence nucleated at
considered as follows: the grain boundary precipitates, such as AlN or
I) The formation of the film-like primary ferrite MnS, as the result of strain concentration at the
along the austenite grain boundaries. film-like primary ferrite formed along the a ustenite
2) The precipitation of the second phase particles grain boundaries.
at the austenite grain boundaries.
Figure 8 shows a schematic diagram of the relation 2. Consideration by the Model Experiment
between the formation of the primary ferrite a long The characteristics of surface cracks of the con-
the austenite grain boundaries and the fracture strain. tinuous casting slab or ingot casting can be sum-
Hot ductility depends on the ferrite formation at the marized as:
austenite grain boundaries (I-V) as shown in Fig. 8. (1) Steels containing aluminum a re highly sus-
I t is considered that the allstenite single phase (I in ceptibl e to cracking. 7 )
Fig. 8) shows good ductility a nd the small amount of (2) The cracks occurred along the coarse austenite
ferrite formation at the austenite grain boundaries g rain boundaries near the surface which formed the
(II and III in Fig. 8) still shows comparatively good film -like primary ferrite. 4 ,7)
du ctility by plastic restraint of the austenite phase (3) The precipitates of the second phase particles
around the ferrite phase. And it can be considered such as carbides or nitrides,IO) and non-me tallic in-
that the formation of the film -like primary ferrite clusions 4 ) existed at the a ustenite grain boundaries.
a long the austenite grain boundaries (IV in Fig. 8) (4) From the d egree of decarbonization and inter-
shows the remarkable ductility loss as the result of nal oxid ation near the crack surface cracks may occur
strain concentration at the ferrite phase, whereas the at temperatures below a bout 1 OOO°C.1ll
growth of primary ferrite at the a ustenite grain bound- Thus, it can be considered that the cracks are
a ries (V in Fig. 8) recovers the ductility loss due to closely related to the grain boundary embrittlem ent,
the relaxation of strain concentration. due to the formation of the film-like primary ferrite
Thus, it can be concluded that the formation of and the precipitation of the second phase particles
the film-like primary ferrite a long the austenite grain at the austenite g rain bound a ries. The m odel ex-
boundaries leads to the loss in hot ductility and, in periment was performed to d emonstra te the loss in
addition, the precipitation of the second phase par- hot ductility due to the primary ferrite forma tion a t
ticles a t the austenite grain boundaries furthers the the austeni te grain boundaries, using the artificial
loss in ductility. In this case, the loss in ductility is ferrite- austenite two phase steels. H ere the situation,
domina ted by the contents of sol AI, Nand S in the second phase particles sllch as AIN or MnS being
the case where alloying elements of C, Si and Mn precipitated at the austenite grain boundaries, was
are the sam e contents. This can be explained by the modeled by using alumina d oped iron as the ferrite
fact that fracture strain d epends on the volume frac- phase.
tion of the second phase particles in the ductile The stress- strain relation of the ferrite- austenite
fracture .8 ) duplex structure varies with the distribution, volume
Figure 9 shows a schematic diagram of the inter- fraction, a nd strength of each phase of ferrite a nd
granula r ductile fracture m echanism. It can be con- austenite. Assuming that the total strain contributed
sidered that the interg ranular ductile fracture can be ind epend ently to each phase of ferrite and austenite:
in the case o f simple ferrite- austenite two phase struc-
ture as shown in Fig. 10, a nd the total strain is given
9 roin groin
bound ary boundar y as:
I I

']
"' 1
"

"

," '
where, strain of ferrite phase
, "
, I ", '
'

: " ' : ' ," , ','

Co :
cr : strain of austenite phase
. I .. . ... :.:,'.: : . . . . <. Va: volum e fraction of ferrite phase
v IV II I Vr: volume fraction of a ustenite phase

Iv m U
V ---1t+IIH'I- r -

temperature
befo re fracture atter fracture
Fig. 8. Schema tic di agram of the relation between the
formation of the primary ferrite along the austenite Fig. 9. Schema tic diagram of the intergranula r ductile
grain boundary and the fracture strain . fracture m echanism.

Research Article
 Transactions ISIJ, Vol. 20, 1980 ( 815 )

t
r

r
Fig. 10. Simple ferrite-austenite
two phase structure.

V. + V, = 1 .
Transforming Eq. (1) , Eq . (2) is obtained: t~st t~mperature ('c)

CT = C. {V. + R,v,) ........ ............ .(2) o. : a phase: narrow width (45-95/1)

where, R,= c,/c •.

o. : a phase: wide width (2 00-2701')
o 0: a phase : pure iron
R , is a parame ter which depends on the distribution •• : a phase: 2% alumina doped iron
and volume fraction of ferrite- austenite two phase d . d . : 0=45°
structure, temperature, deformation degree and so o• 6 e: 0=90°
on. Assuming that C7' and c. at fracture is d efined Fig. II . Variation of the calcu lated fracture strain with
as c~ and c~ respectively, within a range of Ca Va~CT' temperature.
}" .
cT IS glven as:

cf' = c; (V. + R,v,) ............... ...... (3) 1 . 5r-~---~--~--,

c~is corresponding to the fracture strain and governed

by the volume fraction and independent of species
or size of th e second phase particles. According to
Gurland and Plateau,9) the fracture strain is given as: •
•
1- ]
c; = k ] - ..... .... .... ... ........ (4)

where, f:volume fraction of the second phase par-

ticles.
k: a constant.
Using Eq. (4), Eq. (3) is transformed as follows:
O~70~O~--~80~O--~9~O~O~
test temperature ('c)
"' _ k l - ] Va+ RV
C1' - 7 -{ , ,) .... ... ........... (5)
• 0: measurement ( . : Stee l H, ®: austenite
steel)
Equation (5) represents the product of k{(l - ])/] ) 60: calc ulation (6: 0=90°, 0: 0 = '~5°)
which depends on the volume fraction of the second Fig. 12. Comparison of the calcu lated fract ure strain with
phase particles and V. + R, V, which depends on both the experimental value for Steel H.
the volume fraction of the ferrite phase and the strain
ratio of ferrite- austenite two phase. This just means alumina doped iron than in pure iron. This just
that the fracture strain is governed by both the means that the loss in hot ductility was larger in the
formation of primary ferrite and the precipitation of narrow than in the wide width of the primary ferrite
the second phase particles at the austenite grain formation along the austenite grain boundaries, and
boundaries as already mentioned in Section IV. I. so as in AI-killed than rimmed steels. Furthermore,
Figure 11 shows the tempera ture variation of the it should be noted that the fracture strain showed a
calculated fracture strain of the artificial ferrite- larger d ecrease at the 90° inclined ferrite phase to the
austenite two phase steels using Eq. (5). In Eq. (5), tensile axis direction than in the 45°. This means
Va is 0.047 for narrow width and 0.133 for wide width that the grain boundaries situated at the 90° are easier
of the ferri te phase, V, is 1- Va, ] is 0.001 for pure iron to fracture in consequence of the strain concentration
and 0.02 for alumina doped iron, k is 0.0023 for pure than at 45° to the tensile stress direction.
iron and 0.0214 for alumina doped iron obtained from Using Eq. (5), fracture strain was calculated for
Eq. (4), R, is c,/c. used for data of Figs. 6 and 7. It the Steel H. In Eq. (5), Va is determined by optical
can be seen that the fracture strain decreased in the microscopy, V, is 1- Va, ] is determined by electron
temperature range between 700° and 800°C in both microscopy, R, is used for d ata of Fig. 6 beca use of
the narrow and the wide width of the ferrite phase. the difficulty in measuring directly. Figure 12 shows
Here the fracture strain showed a larger decrease in the comparison of the results of the fracture strain
the narrow than in the wide width, and that in calculation with the experimental value for the Steel

Research Article
( 816 ) Transactions ISIJ, Vol. 20, 1980

H. It can be seen that th e calculations of the fracture (3) The fracture strain was d ominated by the con-
strain were in good agreement with the experimental tents of sol AI , Nand S in t he case where alloying
value in the temperature range between 700° and elements of C, Si and Mn were the same contents.
800°C corresponding to the ferrite- austenite two (4) The loss in hot ductility is related to the
phase region. However, at temperatures a bove 800°C formation of the primary ferrite along the a ustenite
the calculated value showed considerably lower than grain boundaries, a nd the precipitation of AIN or
the exp erim ental value. This discrepancy can pro- MnS particles at th e grain bounda ries, and the duc-
bably be explained in the m ain by considering the tility drops m arkably when they occur concurrently.
fact that Steel H exhibits a single austenite phase This was confirmed by the results of the model ex-
although the artificial two phase steel remains ferrite- perim ent using the artificia l ferrite- austenite two
austenite two phase structure even above 800°C. As phase steels.
mentioned above, it should be emphasized tha t the
model experiment can well d emonstrate the phenom- Ackn owledgem ents
enon of the hot ductility loss of low-carbon kill ed Sincere tha nks are du e to Dr. N. Oda, General
steels in the temperature range between 700° a nd M a nager, Dr. T. N ishioka, Assistant General Manager
800°C due to both the formation of the primary and Dr. T. Kunitake, Assistant to General M a nager
ferrite and the precipitation of the second phase par- of the Central R esearch Labora tories, Sumitomo
ticles at the austenite grain boundaries. M etal Industries, Ltd. , for permission to publish this
work.
v. Conclusions
The h ot ducti lity of low-carbon killed steels was
investigated by means of hot tensile tests, micro- REFERENCES
structural observations a nd fracture surface observa- I) S. H asebe : T etsll-to-Haganc, 48 (1 962), 76 l.
tions . The following results were obtained: 2) C. G. Carlsson: J emkont. Alln., 148 ( 1964), 152 .
( I ) The fracture strain d ecreased in the tempera- 3) L. E ricson: Scand. j. M eL, 6 (1977 ), 11 6.
ture range between 700° a nd 800°C. Particularly 4) L . Schmidt a ndA . ] osefsso n: Scand.j. Met ., 3 ( 1974), 193.
5) K. Irvine an d F . B. Pickering: Iron Steel, 30 ( 1957), 2 19.
the frac ture strain remarka bly decreased in the tem-
6) S. C. Desa i: JISI, 191 (1959), 250.
perature r a nge between 750° and 775 °C where the
7) M. Oda, H. Ohtani, M. Yosh ih ara, A. Nashiwa, Y. Murata
film -like primary ferrite was formed along the a us- and K. O zaki: T etsu-to-Haganc, 63 ( 1977), S 110.
tenite grain boundaries. 8) L. R oesch : Mem . Sci. Rev. M et., 66 ( 1969), 29.
(2) The fracture surface showed the intergranular 9) ]. Gurland a nd] . Plateau: Trans. ASM, 56 ( 1963), 442.
ductile fracture accompa nied with Al T or MnS 10) K. Ma tsumoto a nd C. Ouchi: T etsu-to-Hagallc, 64 ( 1978),
precipitates in the temperature range corresponding S257.
to the d ecrease in ductility. II ) K. Yaman aka a nd F. T erasaki: Private communica tion.

Research A rticle