Crack Formation in the Continuous

Casting of Steel
J. K. BRIMACOMBE AND K. SORIMACHI

This review examines the different types of internal and surface c r a c k s that can form
during the continuous casting of steel. For each c r a c k type, the operating and metallur-
g i c a l factors that are known t o influence c r a c k formation are assessed in the light of the
high temperature mechanical properties of s t e e l and a knowledge of the stresses g e n e r -
ated in the solidifying shell. The importance of two zones of low ductility in s t e e l is
demonstrated by this approach. One zone exists above 1340°C and probably accounts for
the formation of all internal c r a c k s and surface longitudinal cracks. The other zone
lies between 700 and 900°C and is related to the presence of soluble aluminum, niobium
and vanadium. Transverse surface c r a c k s in s l a b s can be related t o the latter zone.

K.,RACK formation has long been recognized as a sis of the possible strain-inducing events, leads to a
problem in the continuous casting of steel. Cracks c l e a r e r understanding of c r a c k mechanisms; and helps
have been observed at almost every conceivable l o c a - t o explain the known operating causes of, and solutions
tion in cast s t e e l sections as shown schematically in to, c r a c k formation.
Fig. 1. In the interior, c r a c k s may be seen near the
corners, at the centerline or diagonally between op- STRESSES AND STRAINS IN
p o s i t e corners. On the surface, transverse and longi- CONTINUOUS CASTING
tudinal c r a c k s may appear in both the midface and
c o r n e r regions. The nature of the stresses and strains that can give
The r e a s o n for the profusion of c r a c k types lies in rise t o c r a c k s in the solidifying s h e l l has been dis-
the nature of the continuous casting process itself. cussed by Lankford.3 In the mold, a x i a l stresses are
Continuous casting has achieved widespread popularity set up in the shell owing t o friction between the oscil-
lating mold surface and the s h e l l surface. These
because it is capable of extracting heat at a r e m a r k -
s t r e s s e s which are a combination of a x i a l and bend-
able rate with the combination of mold, s p r a y s and
radiant cooling. The rapid cooling, however, results ing stresses, are tensile when the mold moves upward
relative to the s h e l l and compressive when the r e l a -
in steep temperature gradients in the solid s h e l l that
tive motion of the mold is downward (negative stirip),
can change rapidly and generate t h e r m a l strains as
the s h e l l expands or contracts. In addition, because Fig. 2. The ferrostatic pressure provides the n o r m a l
f o r c e for friction which also generates stresses in the
the semisolid section is required t o move through the
machine, it is subjected t o a variety of mechanically transverse plane as the shell cools and shrinks. The
induced stresses caused by friction in the mold, roll transverse stresses are easiest t o visualize in the
pressure, ferrostatic pressure, machine misalignment, midface r e g i o n w h e r e they are tensile with the l a r g e s t
bending and straightening operations. Depending on stresses appearing near the surface. A s i m i l a r
their magnitude, any of these stresses and strains s t r e s s pattern is possible near the c o r n e r although
may result in c r a c k formation. the presence of a well defined air gap makes the na-
Much has been written about the causes of, and solu- ture of the stresses less certain. T h e r m a l stresses
tions to, these different types of cracks. However, in in and below the mold are generated by changing
this body of literature, which has included a number temperature gradients in the shell. They are normally
of reviews, 1'2 t h e r e has been a tendency t o treat each tensile at the surface which is coldest and compres-
sive near the solidification front. This profile can be
c r a c k type as a separate problem, largely from an
operating standpoint. Few attempts have been made reversed, however, should the surface temperature
to comprehensively review c r a c k formation in more rebound owing to a sudden d e c r e a s e in the rate of
fundamental t e r m s by relating the presence of cracks cooling.
not only to operating stresses and strains but also to Below the mold, ferrostatic pressure can give rise
the mechanical properties of steel at continuous cast- to bulging, particularly of the wide face of slabs if
ing temperatures. The present review has been under- support rolls are spaced too widely apart. As de-
taken to examine the cracking problem from this point picted in Fig. 3, the resultant stresses are tensile at
of view. What e m e r g e s is a unified concept of c r a c k the surface and compressive near the solidification
formation in which c r a c k s can be seen to form in two front. However this situation can change quickly if
distinct temperature ranges which depend on the duc- r o l l s lower in the machine a r e properly spaced s i n c e
tility of steel. This knowledge, combined with an analy- they will d e f o r m the bulged s h e l l and set up tensile
stresses at the solidification front. S i m i l a r stresses
can a r i s e if a set of r o l l s is too narrowly gapped.
J. K. BRIMACOMBE is Associate Professor, Department of Metal- Bending a n d / o r straightening of the strand introduces
lurgy, University of British Columbia, Vancouver, B.C., Canada,and l a r g e a x i a l strains in the solid shell. During straight-
K. SORIMACHIis Engineer, Kawasaki Steel Corporation, Chiba Works, ening these strains are tensile on the upper surface of
Chiba 280, Japan. the strand and compressive on the lower surface. The
Manuscript submitted January 10, 1977. strains are reversed, of course, for bending.

METALLURGICAL TRANSACTIONS B VOLUME 8B, SEPTEMBER 1977-489

 sprays. The latter will be discussed more fully l a t e r
in the paper in relation to different types of cracks.

HIGH TEMPERATURE MECHANICAL

PROPERTIES OF S T E E L
The mechanical properties of steel at elevated tem-
peratures are affected by s e v e r a l variables: tempera-
t u r e , s t e e l chemistry, structure, strain rate and t h e r -
mal history. With the number of variables involved it
is not surprising that the strength and ductility of
s t e e l are imperfectly known u n d e r continuous casting
conditions. Nevertheless the studies that have been
performed r e v e a l a picture of mechanical behavior
that is of considerable value in understanding c r a c k
Cracks in continuously cast steel formation. These investigations have employed a
Internal cracks variety of apparata: Gleeble and Instron type tension
I Midway machines, torsional machines and bending and impact
devices. Taken together they show that t h e r e are
2 Triple-point
three distinct temperature r a n g e s in which s t e e l has
3 Centreline low strength a n d / o r ductility and is therefore sus-
4 Diagonal ceptible t o cracking.
5 Straightening / bending
6 Pinch roll
High Temperature Z one: ~ 1340°C to Solidus
Surface cracks
7 Longitudinal, mid-face T h e r e is a m p l e evidence t o indicate that at tem-
8 Longitudinal, corner peratures above about 1340°C both the strength and
ductility of steel decline markedly. The r e g i o n of low
9 Transverse, mid-face
strength can be seen in Fig. 4 in which the ultimate
I0 Transverse, corner tensile strength is plotted against temperature .4 This
II Star graph has been constructed by extrapolating strength
F i g . 1--Schematic drawing of s t r a n d cast section s h o w i n g dif-
data measured at different strain r a t e s in four
f e r e n t types of cracks.
studies 5"~ t o a common strain rate of 10-3 s-1 which is
thought t o obtain in continuous casting. The depend-
ence of UTS on the logarithm of the s t r a i n rate has
been employed in the extrapolation. The ductility-to-
fracture of steel in this r a n g e of temperature appears
to be of the o r d e r of 0.2 to 0.3 pct according to Vom
Ende and Vogt.9
The low strength and ductility seem to ]~e due to the
presence of liquid f i l m s in the interdendritic regions
which do not f r e e z e until temperatures well below the
solidus are reached. 6'7 The liquid f i l m s apparently

F i g . 2--Schematic drawing of axial and bending~ s t r e s s e s i n

s o l i d s h e l l r e s u l t i n g from friction i n the m o l d . °

The other l a r g e a x i a l f o r c e that acts on the solid

s h e l l a r i s e s from the weight of the strand, the gravity
f o r c e . In the absence of the tractive f o r c e e x e r t e d by
the d r i v e r o l l s this f o r c e sets up an a x i a l tensile s t r e s s
in the shell. The d r i v e r o l l s can either contribute to
or r e d u c e this stress depending on whether they pull
on or act against the strand.
Stresses can also be generated from a variety of
other sources such as misalignment of the casting F i g . 3--Schematic drawing of axial s t r e s s e s i n s o l i d s h e l l r e -
machine and nonuniform cooling in the mold and s u l t i n g from bulging or narrowly gapped rolls.3

490-VOLUME 8B, SEPTEMBER 1977 METALLURGICAL TRANSACTIONS B

contain high levels of sulfur, phosphorus and other
elements which have a segregation coefficient less
than unity and which concentrate between the growing
dendrites. Adams,6 for example, has measured phos-
phorus concentrations of 0.2 to 0.5 pct at the g r a i n
boundaries whereas the m a t r i x l e v e l was only 0.02
pct. Fujii e t a l~° also have found strong segregation
of phosphorus associated with internal cracks, 0.18
pct, relative to a m a t r i x concentration of 0.02 pct.
Evidence of liquid f i l m s can be seen in Fig. 5 which
is a scanning electron micrograph of the interior
surface of a c r a c k that f o r m e d near the solidus tem-
perature. The smoothness of the surface, with no
s i g n s of fracture, is indicative of the presence of a
liquid film at the time of c r a c k formation.
In a c c o r d with this picture of mechanical behavior,
it has been found that both sulfur and phosphorus have
a harmful influence on the strength, ductility and
solidification cracking sensitivity of s t e e l at high

Fig. 6--Effect of phosphorus o n f r a c t u r e strength of SAE 4340

s t e e l .12

Fig. 4 - - E f f e c t of temperature o n u l t i m a t e t e n s i l e strength of

steel 4 temperatures, u'12 Of the two, Guglin e t a113 c l a i m
that phosphorus has the more deleterious effect. The
influence of phosphorus on fracture strength and duc-
tility of steel can be seen in Figs. 6 and 7, reproduced
from the paper by Sopher.12 S t e e l with a phosphorus
concentration of 0.039 pct has drastically reduced
strength and ductility compared to steel containing
0.017 pct phosphorus in the high temperature zone and
at lower temperatures as well. The effect of sulfur
on the strain-to-fracture and ultimate tensile strength
of s t e e l near the solidus temperature is shown in
Fig. 8.14 An i n c r e a s e in the sulfur concentration from
0.025 to 0.035 pct results in a reduction by half of the
mechanical properties. F i b r e 8 further indicates
that an increase in the Mn/S ratio has a beneficial
effect on the strain-to-fracture but apparently has lit-
tle influence on the ultimate tensile strength. A s i m i -
lar conclusion was reached by Kinoshita and KurokiIs
for s t e e l at 1350°C. They quantitatively investigated
the change of FeS and MnS in sulfide inclusions and
found that the FeS content decreased with increasing
manganese content, as seen in Fig. 9. The solidifica-
tion temperature of the sulfide inclusions was then
related to the manganese concentration using the FeS-
MnS d i a g r a m (Fig. 10); and it was shown that for man-
F i g . 5--Scanning electron micrograph of s u r f a c e of a c r a c k ganese levels less than 0.6 t o 0.7 pct, for example, the
f o r m e d near the s o l i d u s temperature. T h e i n c l u s i o n s are sulfides could e x i s t in the liquid form below the soli-
sulfides .27 dification temperature of 0.2 pct carbon steel (1480°C).

METALLURGICAL TRANSACTIONS B VOLUME 8B, SEPTEMBER 1977 491

 *C ratio and thermal history of the steel. Increasing
1370 1260 1150 1040
Mn/S increases the ductility of the s t e e l as shown in
Fig. 13. The influence of thermal history is con-
siderably more complicated however because s e v e r a l
- independent variables are inherently involved. T h e s e
75 .
•
=
-- ~ |
• • , 11" .L
include the question of melting/freezing in situ or
6 /i /Heat B
0.004%P
Heat F
0.017%P
I Heat G
0.039 % P
T ,," heating samples from room temperature, cooling
r a t e , isothermal holding temperature and isothermal
holding t i m e . An example of two t h e r m a l pretreat-
0.004 % S 0.006 % S O 007 % S
.C 50 0.40 % C 0.43 % C 0.38 % C ments employed by Wilber et a118 is shown in Fig. 14.
Nevertheless on the b a s i s of the studies by Wilber et
aI and Lankford, it can be said that the loss of duc-
tility is decreased with decreased cooling r a t e , de-

I I

S=0"
I I0
2700 2500 2300 2100 1900 1700
Test Temperature, F
F i g . 7 - - E f f e c t of phosphorus o n h o t d u c t i l i t y of SAE 4340
steel.12

I I I I I I

0"20 20

30~35 15~20 (D(D

0.15 ~5 5

O(3
0.10
E
"~ 8=i
0.05
8z~
0 I I I I I I 0
0,020 0"030 0.040 0 I I
Sulphur Content (%) 0 0-5 I-0
F i g . 8--Effect of s u l f u r o n strain-to-fracture and u l t i m a t e
t e n s i l e strength of s t e e l containing 0.19 t o 0.21 pct c a r b o n Mn ( % )
near the s o l i d u s . 14 F i g . 9 - - E f f e c t of manganese content i n s t e e l o n M n S / F e S i n
contained sulfide i n c l u s i o n s . 15
The carbon content also affects the mechanical
properties of s t e e l just below the solidus tempera-
ture. As shown in Fig. 11, Morozenskii et a114 have u 1600 I
found that the strain-to-fracture, 5~, and its plastic o
c o m p o n e n t , 5plastic, are a minimum for s t e e l contain- S = 0-01,',, 0"04 %
ing 0.17 to 0.20 pct carbon. S i m i l a r behavior as re-
g a r d s cracking sensitivity has been reported by
Guessier and Castro,~1 while Hall~6 recognized that
s t e e l containing 0.26 pct carbon is less ductile than
® 1500

/
E 1480*C -'-~-
0.1 pct carbon s t e e l above 1260°C. Morozenskii et al h-
c l a i m that this effect can be attributed t o t h e r m a l
e
shrinkage. .9
As will be seen in a following section, the high tem- 1400
perature zone of low strength and ductility is at the . I
tli
.~
root of most of the c r a c k s found in continuous cast-
ing. o
(/)
0"7 %
Intermediate Temperature Zone: 800 t o 1200°C 1300 I I
0 0.5 I-0 I-5
The second zone of low ductility in steel appears in
the temperature r a n g e of 800 t o 1200°C 3'7'17'~8 and can Mn ( % )
be seen in Fig. 12. The loss of ductility during cool- F i g . 10--Effect of manganese content i n s t e e l o n the s o l i d i f i -
ing below 1200°C is strongly dependent on the Mn//S cation temperature of sulfide i n c l u s i o n s , is

492-VOLUME 8B, SEPTEMBER 1977 METALLURGICAL TRANSACTIONS B

c r e a s e d holding temperature and increased isother- REDUCTION OF AREA, %
mal holding t i m e . Strain rate does not appear t o have
much influence on the ductility. Interestingly, the
strength of steel in this zone is not affected in the I00 -

same manner as the ductility (see Fig. 12).3'7 80 - 2000°~

Ill:I'~*~'~'~l}l
Matsubara17 has attributed the low ductility to ma-
trix hardening of austenite by the precipitation of
60
plate-like sulfides, which also lowers g r a i n boundary -

D •
cohesion. Lankford3 has proposed a s i m i l a r mecha-
nism by suggesting that the low ductility results from
the precipitation of liquid droplets of FeS in planar
a r r a y s at austenite g r a i n boundaries which are then
paths of easy c r a c k propagation. Steels with Mn/S
- \ ,ooo
ratios above 60 are not embrittled because the sulfur 0 , I I 1 I I I J
is tied t o the stable phase, MnS, which precipitates lO 20 30 40 50 60 70 80
in the m a t r i x , and not predominantly at g r a i n bound-
a r i e s . Lankford has further shown that slower cool- Mn:S RATIO
Fig. 13--Effect of Mn/S on ductility of steel after melting,
ing r a t e s improve the ductility even with low Mn/S casting, cooling at 14°C/s to 1425°C, then cooling at 5°C/s to
because manganese then has time to diffuse t o the test temperature of 2000°F (1095°C) or 1700°F (925°C).3
g r a i n boundaries.
The importance of this zone of low ductility from
the standpoint of c r a c k formation is difficult to a s s e s s . TEST T E M P E R A T U R E , °C
lOOO 1100 1200
0.5 ' I ' I ' I '
| I | I I -1.2

e-_ CYCLE C - I E~
§ 0.4
O.15
bJ
- cYCLe c-5 _ , =a
.u p- (I minute) ~ ~, -0.8
"' 0.3
a, ,C
O-lO ', - 0 . 6 ~'

6O ,.,0.2 _ CYCLE C - 5
(10 minutes)
" - , - . , ~ xo
~e
~
"--e
m
a¢
s r ~ plash¢
.
/" - 0.4 ~ -
F-
/ ¢J
t / HEAT TT-893 et
l / ~ o.I
l z LIL -0.2 "-
l /
l J'
0 i ~, me m i I
o O-I 0.2 0"3 0"4 05 O = I , I , I 0
1800 2000 2200
Carbon Content (%) I S O T H E R M A L HOLD a
Fig. ll--Effect of carbon content in steel on strain-to-frac- T E S T T E M P E R A T U R E , "F
ture (Sz) and its plastic component.~4 Fig. 14--Comparison of the ductility of steel after thermal
cycles C-1 and C-5. Cycle C-l: cast and free-cooled to test
temperature. Cycle C-5: cast, free-cooled to indicated tem-
=C peratures, held 1 or 10 rain and tested.18
FRACTURE ULTIMATE
DIAMETER, in. I000 1200 1400 LOAD, Ib Up t o the present t i m e , the ductility of s t e e l has not
~_-
I ! i
been m e a s u r e d under t h e r m a l history conditions that
r e l a t e directly to a given section of a continuously
0.4 4000 cast s h e l l in which c r a c k s f o r m . It is probably safe to
s a y , however, that most c r a c k s generated in the mold
are not a consequence of this low ductility zone be-

0.2 2"") •'%•"~0 2000

cause temperatures are normally above 1200°C. The
exception of course is the c o r n e r r e g i o n in the lower
part of the mold.

l
Low Temperature Zone: 700 t o 900°C
t
L Less is known about the t h i r d zone of low ductility,
O i I I I w'ee~'l " 0 which has been found in the r a n g e of 700 to 900°C; but
1200 1600 2000 2400 2800 there is g e n e r a l agreement that it is usually a s s o c i -
TEST TEMPERATURE, °F ated with soluble aluminum in the steel and the pre-
Fig. 1 2 - - F r a c t u r e d i a m e t e r and u l t i m a t e load for s t e e l a f t e r cipitation of A1N at g r a i n boundaries. 19-22 Iida e t a119
m e l t i n g , casting and f r e e c o o l i n g t o indicated t e s t t e m p e r a - have reported that AIN precipitation does not proceed
ture .3 appreciably d u r i n g cooling t o as low as 800°C but can

METALLURGICAL TRANSACTIONS B VOLUME 8B, SEPTEMBER 1977-493

 Fig. 15-Sulfur p r i n t of t r a n s -
v e r s e section of a 10.2 x 15.2
cm billet s h o w i n g midway
cracks.

take place rapidly during heating over the tempera- enough to pinpoint the probable location in the machine
ture r a n g e 700 to 1000°C. On the b a s i s of these find- where the c r a c k s form; and gives the direction of the
ings they suggest that repeated cooling and reheating tensile s t r e s s (normal to the c r a c k orientation). Pos-
cycles which obtain in the spray chamber should en- s i b l e sources of the tensile stresses then can be ex-
hance A1N precipitation and concomitant low ductility. amined and checked to determine if the stresses
In another experimental study, Mayrhofer2° has shown act on regions of the solid s h e l l that a r e weak and non-
that the deposition of A1N reaches a m a x i m u m rate at ductile, and would give rise to c r a c k s in the location
the A3 transformation temperature requiring only a observed. With respect t o mechanical properties it is
few minutes in aluminum killed, deep drawing steel. probably reasonable t o consider only the high and low
Steel chemistry has an influence on ductility in this temperature zones of low ductility s i n c e the intermedi-
temperature r a n g e . Obviously, increasing soluble ate temperature zone has not been adequately related
aluminum and nitrogen levels have a deleterious t o continuous casting conditions.
effect on ductility; niobium is another element that This g e n e r a l approach to c r a c k analysis has been
reduces ductility.21'22 The addition of titanium and applied in the following sections dealing with internal
vanadium to s t e e l has been reported t o be beneficial and surface c r a c k s .
in preventing the r e g i o n of low ductility although
zirconium proved t o be ineffective .23 Vanadium may
INTERNAL CRACKS
not be as beneficial as originally supposed, however,
since other workers c l a i m that it reduces ductility.22,24 The following internal c r a c k s are discussed in this
Embrittlement of steel, probably by a different review: midway cracks, triple point c r a c k s , center-
mechanism, over the temperature r a n g e 750 to 850°C line cracks, diagonal cracks, bending/straightening
has been observed by Robbins e t a125 in the two phase c r a c k s and pinch roll cracks. As will be s e e n , all six
ot + ~ r e g i o n of a 0.1 pct carbon steel. S i m i l a r find- types of c r a c k s r e s u l t from high tensile strains and
ings based on hot tension tests have been reported in s t r e s s e s acting on regions of the solid s h e l l that are
a separate s t u d y .26 in the high temperature zone of low strength and duc-
Owing to the low temperatures involved, the t h i r d tility.
zone of low ductility is seen t o be a factor only in the
formation of surface or subsurface cracks.
Midway C r a c k s
Midway cracks, or as they are sometimes called,
ANALYSIS OF CRACK FORMATION
halfway cracks, r a d i a l streaks or ghost lines, can be
The cause of a particular type of crack can be de- detected in sulfur prints and macroetches of t r a n s -
termined if the information in the two preceding sec- v e r s e sections, s i m i l a r t o that shown in Fig. 15. They
tions concerning stresses and mechanical properties appear as dark lines running n o r m a l t o a given face
is combined with practical knowledge obtainedfrom in a r e g i o n roughly midway between the surface and
a continuous casting operation. From plant observa- centerline. When viewed on a magnified s c a l e , the
tions the location and orientation of the cracks can be c r a c k s may be seen to be filled with sulfur-rich
ascertained, together with the operational factors m a t e r i a l (Fig. 16) although they may also be open. In
thought to cause and cure the cracks. This is often Fig. 17 an open midway c r a c k is shown, its surface

494-VOLUME 8B, SEPTEMBER 1977 METALLURGICAL TRANSACTIONS B

 b a r s . The numbers are temperatures in d e g r e e s
Centigrade. High tensile stresses can be seen next
to the solidification front in the nonductile zone. By
comparison to Fig. 15 it is evident that this r e g i o n is
the site of midway c r a c k s .
Strong reheating of the surface o c c u r s whenever the
rate of cooling decreases abruptly such as below the
spray chamber, between successive spray nozzles or
below the bottom of the mold. Obviously the magnitude
of the reheating is very dependent on spray cooling
conditions.
High casting temperature has a l a r g e effect on the
tendency to form midway c r a c k s because it influences
the cast structure. As shown in Fig. 20 a high casting
temperature favors an enlarged columnar zone extend-
ing inward from the surface at the expense of the cen-
tral equiaxed zone. Midway c r a c k s are able t o form
much more easily between the dendrites in the colum-
nar zone which run perpendicular to the tensile stress,
as compared t o the equiaxed zone.
The influence of sulfur and phosphorus levels on
midway c r a c k formation can be seen in t e r m s of their
deleterious effect on the strength and ductility of s t e e l

Fig. 16--Photomicrograph of a filled crack: numerous man-

ganese sulfide inclusions are visible; etched in nital.27

exhibiting the smooth topography characteristic of

c r a c k s or "hot t e a r s " formed in regions of liquid
films. MnS inclusions are evident on the surface as
well.
Many publications exist in the literature which dis-
cuss the factors that contribute t o midway c r a c k
formation. 9,27-31 In these studies it is generally a g r e e d
that excessive secondary cooling and a high casting
temperature are operating factors responsible for
midway cracks. In addition, the chemistry of the
s t e e l e x e r t s an important influence; sulfur and phos-
phorus in particular affect the c r a c k susceptibility
of steel as Vom Ende and Vogt9 have shown in Fig. 18.
Excessive spray cooling is a m a j o r factor in the
formation of midway c r a c k s because it leads t o subse-
quent reheating of the strand surface which provides
the driving force for cracking. 2v Reheating causes the Fig. 17--Scanning electron micrograph looking into an open
surface to expand and this imposes a tensile s t r a i n in midway crack in a billet section: note smooth surface and pro-
the interior r e g i o n of the solid s h e l l which, as seen truding dendrite branch.27
e a r l i e r , is weak and nonductile above about 1340°C
(the high temperature zone). Viewed in a transverse
plane, the tensile s t r a i n and stress run parallel to the Ni > 1°/o P > 0'030°/o _ l'<0.16°/o
surface, and thus c r a c k s will form perpendicular t o Cr >3°/o Mn :-1.0°/o (';10"25-0'60°/° Mo~.0.25o/o
the surface, depending on the magnitude of the strain.
The s t r e s s distribution that results from reheating S >0025°/o C [0'17-0"24°/o
[ >0.60o/o S;"=O' 50°/o
can be calculated roughly using finite element analy- S i~ 0.50°/o ]
sis. 4,32 F i g u r e 19 shows the stress distribution4 result- ! ! I !
ing from reheating below the spray chamber of the deteriorating neutral - - improving
10.2 by 15.2 cm billet shown in Fig. 15. The distri- I
bution corresponds to one quarter of a transverse internal crack susceptibility
section with tensile stresses represented by ordinary Fig. 18--Effect of alloying elements on internal crack sus-
lines and compressive stresses by lines ending in ceptibility of steels.9

METALLURGICAL TRANSACTIONS B VOLUME 8B, SEPTEMBER 1977-495

 I00
k g / c mz
Tension
I Compression

5.0 11o9÷1153+11,9+1247 ~- 1447 Liquid Core

l lO38 lO7O

-[ 1296 "4- 1345 ~ 1395

tlo34 1067 ~ 1106 /1149 ÷1195 +1242 1289 "P- 1336 "q~ 1382 ~ 1428 -~- 1470 1488
-- 1492

4-0 1431 1435-4rI

1026 i059 ~ i097 /1138 /3182 1227 1271 "~-- 1313 ~ 1353 --+--1388 --e-- 1416 __

1375 1379
1204 X 1244 ~ 1281 -- 1314 -- 1342 - - 1363 --
E0
3.0
1319 1323
I 999 ~ 1029 x1063 /1100 /1137 1175 ~ 1210 ~ 1242 ~ 1270 ~ 1293 -- 1310
x

1265
2'0
I 1142 / 1173 / 1201 1225 ~ 1244 -+ 1257
1268 1

1216
I
J
1214
I 961 ~ 987 / I016 /I047 /i078 ii08 ~ 1136 ...--4 1160 .----4 i180 ----4 1196 ~ 1207

1.0
1167 1169
[ ~ 940 ~ 964 X 991 "-A4019 : ,~48 1074 __-..---4 1099 ~ 1120 -- ~ 1138 t 1151 - 1 1161 , m 4

1126 1127
918 -.~94~ ~ 6 : 9~ 1y "-- :I0 2 :" ~" 1065 q 1084 .-4 ii00 11112 I 1120 ~ 4
I 41 I I I I I I I
O0 , i i , ,
I.O 2-0 3.0 4.0 5.0 6.0 7.0
Y-Axis (cm)
Fig. 19--Stress distribution resulting from reheating of the surface of a 10.2 × 15.2 cm billet below the spray zoneJ

in the high temperature zone. This was illustrated ures can markedly reduce the frequency of midway
e a r l i e r in F i g s . 6 t o 8. cracks.
From this analysis, it is c l e a r that midway c r a c k
formation can be prevented either by minimizing the Triple-Point C r a c k s
tensile strain, i . e . the surface reheating, or by in-
creasing the strength and ductility of the steel in the Triple-point cracks, such as shown in Fig. 21, a r i s e
high temperature zone. Minimization of the tensile d u r i n g the casting of slabs. Viewed in a transverse
strain can be achieved by paying close attention to section, they are oriented n o r m a l to the n a r r o w face
the design of the spray system to ensure that the rate within the V-shaped r e g i o n of the slab where the t h r e e
of cooling does not d e c r e a s e abruptly between mold solidification fronts m e e t . The zone of cracking,
and sprays, s p r a y s and radiation cooling or between according to Mori,4 lies between 3 and I0 cm from the
successive spray nozzles. The method by which a surface. Ozeki and Duke34 report that the c r a c k s may
spray system can be designed to attain minimum re- vary in length from I/4 (0.6 cm) t o s e v e r a l inches and
heating has been described by Brimacombe. 33 Main- in width from a thin dark line to an open c r a c k . They
tenance of the spray system is also an important con- have also found the susceptibility of s t e e l t o triple-
sideration because clogged or poorly positioned n o z - point cracking to be dependent on chemistry increas-
zles can c a u s e l o c a l reductions in cooling. Improve- ing with decreasing manganese levels below 0.9 pct
ment in the high temperature mechanical properties manganese. Steels with low manganese contents are
of steel can be attained by minimizing the pouring particularly prone t o c r a c k formation when Mn/S is
temperature and thereby decreasing the width of the less than 30.
columnar zone or by lowering the concentrations of T h e r e is g e n e r a l agreement that the c r a c k s are
phosphorus and sulfur t o below 0.02 t o 0.025 pct. In caused by bulging of the wide face of the slab owing
practice, it is well known that any one of these m e a s - t o insufficient containment of the solid shell. While

496-VOLUME 8B, SEPTEMBER 1977 METALLURGICAL TRANSACTIONS B

 to determine the position below the meniscus where
the c r a c k s formed. A s i m i l a r procedure can be a p -
plied to locate the position where any of the internal
cracks form.

Centerline Cracks
Centerline or core c r a c k s appear in the central
r e g i o n of a cast section and form toward the end
of solidification. Looking first at slabs, the c r a c k s
(see Fig. 22) can be found in all g r a d e s of s t e e l
independent of composition and superheat.34'35 The for-
mation of centerline c r a c k s is strongly influenced by
machine and operating variables such as spray water
intensity, roll alignment low in the strand and casting
speed.34'~5 Cracking can be reduced by regapping
r o l l s , reducing speed a n d / o r increasing spray cool-
ing. From this it can be deduced that the s t r a i n caus-
ing the c r a c k s is generated by bulging of the wide
face owing t o inadequate containment. The strain is
n o r m a l t~ the bulging face and acts on regions of low
F i g . 20--Variation of l e n g t h of columnar zone w i t h casting ductility near the solidus at the centerline. Increasing
t e m p e r a t u r e for s e v e r a l g r a d e s of s t e e l . 27 spray cooling has a beneficial effect because it results
in a cooler s h e l l that can more effectively withstand
the ferrostatic pressure. Reduction of the casting
speed has a s i m i l a r influence; and because the metal-
l u r g i c a l length is reduced proportionately, the center-
line temperature is lowered below the r a n g e of high
temperature low ductility before bulging occurs. That
s t e e l composition has little effect on c r a c k formation
is an indication that the strains are l a r g e r than the
ductility of most s t e e l g r a d e s above 1340°C.
Since bulging is the driving force for centerline
c r a c k s in slabs, maintenance of the correct roll gaps
is vital for c r a c k minimization. Ozeki and Duke34
have described two systems-load c e l l s mounted in
each roll station in the machine and a "gapping s l e d "
which can be attached to the s t a r t e r b a r - t o facilitate
the checking of roll gaps.
In the casting of billets, bulging p l a y s a l e s s e r role
in the formation of centerline cracks; and Van Drunen
et a127 have suggested that the sudden drop in center-
line temperature at the completion of solidification
Fig. 21--Triple-point crack in a continuously cast slab.34 generates the cracking strains. The abrupt d e c r e a s e
in centerline temperature is a natural phenomenon,
this situation can a r i s e if a set of r o l l s is too widely
gapped, it may also r e s u l t from r o l l s that are too
narrowly gapped.34 In the latter c a s e , bulging may oc-
cur at the next roll station.
Bulging of the wide face provides the tensile strain
for triple-point cracking because it leads to concavity
of the narrow face owing t o the relative high strength
of the colder corners. The tensile strain is then
greatest close t o the solidification front where the
s t e e l is the l e a s t ductile. Prevention of this type of
c r a c k clearly can be achieved by maintenance of the
proper gap between r o l l s throughout the metallurgical
length of the slab.
The set of r o l l s that is causing the c r a c k problem
can be located if the s h e l l profile adjacent to the nar-
row face is known either from break-out shells or
from computer calculations. The distance between the
inside tip of the crack, which is the position of the
solidification front at the time of c r a c k formation, and
the narrow face can be compared to the s h e l l profile Fig. 22--Centerline crack in a continuously cast slab.34

METALLURGICAL TRANSACTIONS B VOLUME 8B, SEPTEMBER 1977-497

 tension on the central r e g i o n by reheating will magnify the
compression
tensile stresses developing from r a p i d centerline
cooling. In the extreme, this situation could lead t o
I~1~329~1060 ]E1092 -1125 ~-1158 +1192 --1225--12~o~-12t~--1-~27---.l~342
] . / 271088
]'/ [1089
/ z1122
/ 1156 1189 1221 1252 1280 1303 1322 1333 1337 s e v e r e centerline c r a c k s .
• *" ." ." " / / _,Z 4 - -+4 Brovman e t a137 have proposed that the formation of
.1021 1051 1082 1114 1147 1179 1211 1241 1267 1290 130713181322t centerline c r a c k s could be suppressed by applying
I I I • - + . . . . ~- ÷ ÷, water s p r a y s at the point of complete solidification.
1012 1041 1071 1102 1133 1165 1194 1222 124,8 1269 1285129512ClC~ T h e i r suggestion would seem reasonable s i n c e then
l l I = . . . . . . . + :F~.
the surface temperature could be made t o drop at an
~gg ~027 1056 1085 1115 1144 1173 1199 1222 1242 125612c,~6, 128.54 equivalent rate to the centerline, thereby lowering
the central tensile stresses. The design of the s p r a y s
1o37 1%8, 17, 1,2, 1147 11;1 11,3 1,,1o 122,, 12321,1
N I i l to minimize these s t r e s s e s below the liquid pool has
I

been examined by G r i l l and Brimacombe.38

1 ~45 969 993 1018 1042 1066 1088 1109 1126 1141 1151 1158116J
Diagonal C r a c k s
924 9461 9 992 1 1038 1057 1075 1091 1104 1114
- ~i I'-' ~I ~ ~ ~ I~'-~ I This type of crack, normally seen in billets, is as-
0 1 2 3 4 5 6 7 sociated with rhomboidity. Diagonal c r a c k s usually
crn run between obtuse c o r n e r s of the rhomboid section
F i g . 23--Stress d i s t r i b u t i o n r e s u l t i n g from sudden d e c r e a s e as can be seen in Fig. 24. According to MoriI r h o m -
in centerline t e m p e r a t u r e below the pool b o t t o m of a 1 0 . 2
x 15.2 cm b i l l e t . 32
boidity is a g r e a t e r problem with s m a l l e r sizes of
billets and high pouring temperatures. It can be cor-
r e c t e d by installing c o r n e r r o l l s at the top end of the
roll apron as well as by maintaining symmetrical cool-
ing in the sprays.
Clearly, diagonal c r a c k s r e s u l t from distortion of
the billet which can a r i s e if two adjacent f a c e s a r e
cooled more rapidly than the other f a c e s1,39 in the
mold or secondary cooling zone. The contraction of
the s t e e l in the vicinity of the colder f a c e s generates a
tensile strain, oriented diagonally between these faces.
If sufficiently l a r g e , the s t r a i n causes distortion and
c r a c k s to form at r i g h t a n g l e s t o the s t r a i n axis; that
is between the obtuse corners. The c r a c k s form initi-
ally in the high temperature zone of low ductility but
may grow outward toward the corners depending on
the magnitude of the strain.
T o minimize diagonal c r a c k formation, care must
be taken to achieve equal cooling on each of the four
faces. This r e q u i r e s good alignment between the mold
and r o l l e r c a g e s and avoidance of plugged or bent
spray nozzles in the secondary cooling zone.

F i g . 24--Diagonal c r a c k i n a b i l l e t , s¢ I ,imi atio due'to e cee ing an elongation

3501/ of 03°/o on the solidification f r o n t

occurring in any casting, and a r i s e s when the final 300 / ~ a=26.Smm mi6tl2
E
t r a c e of latent heat has been extracted from the cen-
ter region. The drop in centerline temperature is con-
siderably more r a p i d than the d e c r e a s e in s u r f a c e
temperature with the result that the center r e g i o n
contracts. The center is constrained from contraction,
however, by the surrounding colder steel and thus is
put into tension. For illustrative purposes, the s t r e s s
distribution that results from these t h e r m a l events is
shown in Fig. 23 for the billet displayed in Fig. 15.
High tensile s t r e s s e s are evident in the central region. rr limitation due to slab surface temperature
It has been suggested that s e v e r e secondary cooling 5 0 being too low at straightening
may contribute t o the formation of t h e s e cracks.36 This (950°C a f t e r 15rain)
I I I I I I I 1
may be possible if, due to the arrangement of the 0 0-4 0.8 12 1,6 2.0 2-4 2.8 32
sprays, considerable reheating of the surface below CASTING SPEED~ m/rain
the secondary cooling zone coincides with the bottom
of the liquid pool. Then the tensile stresses imposed F i g . 25--Safe operating r a n g e s of c i r c u l a r - a r c machines. 9

498-VOLUME 8B, SEPTEMBER 1977 METALLURGICAL TRANSACTIONS B

F i g . 26--Sulfur p r i n t of longitudinal section from a 17.1 cm 2 billet s h o w i n g pinch r o l l c r a c k s , indicated b y s m a l l a r r o w s .32

T a b l e I. Internal C r a c k s *

Type Cause Influencing Factors Corrective Action

Midway cracks Surface reheatingin High casting temperature, S Adjust spray system to minimize re-
or below spray and P > 0.02 pct increase crack heating. Lower pouring tempera-
chamber formation ture, minimize P and S levels
Triple-point cracks Bulging ofwide face Crackingincreases with de- Regaprolls
of slabs creasing Mn below 0.9 pct Mn
and with Mn/S < 30
Centerline cracks In slabs, bulging of Spray water intensity, casting Regaprolls, reduce castingspeed or
wide face speed,roll alignment low in the increase spray cooling
strand.
In billets, rapid cooling Severe secondary cooling and Adjust secondary cooling near
of center regionbelow high pouring temperature may bottom ofpool
pool enhance crack formation
Diagonal cracks Asymmetrical cooling High pouring temperature and Install corner rolls at bottom ofmold;
in mold or sprays smallerbillet sizes increase check alignment between mold and
cracking roller apron. Look for plugged spray
nozzles
Stralgtening/ Excessive deformation Bending on liquid center. Reduce tensilestrain at solidification
bending cracks near solidification front front to less than 0.3 pct. Lower
due to straightening or castingspeed
bending
Pinch roll cracks Excessive pinchroll Squeezingon a liquid center Reduce pinch roll pressure
pressure
*All internal cracks form in high temperature zone oflow ductility, between 1340oC and the solidus temperature. Elementssuch as P, S and Cu worsen the crack
problem.

METALLURGICAL TRANSACTIONS B VOLUME 8B, SEPTEMBER 1977-499

 Straightening/Bending C r a c k s
Straightening or bending cracks can form if straight-
ening or bending operations are c a r r i e d out on a sec-
tion with a liquid center, or indeed when the center
is solid but above 1340°C. Straightening c r a c k s appear
above and n o r m a l to the withdrawal axis when viewed
in a longitudinal section, while bending c r a c k s appear
below the withdrawal a x i s . Because these c r a c k s
a g a i n form in the high temperature zone of low duc-
tility, casting conditions must be regulated to keep
the tensile strain near the solidification front below
roughly 0.3 pct if the cracks are to be eliminated.
Vom Ende and Vogt9 have applied this criterion t o
determine the m a x i m u m casting speed that can be
achieved using c i r c u l a r - a r c machines for different
strand thicknesses and r a d i i of curvature. T h e i r re-
sults can be seen as the sloping curves in Fig. 25.

Pinch Roll C r a c k s
Like straightening/bending cracks, pinch roll c r a c k s
(Fig. 26) can form if excessive roll pressures are
e x e r t e d on a strand at a point where solidification is
incomplete or the centerline is above 1340°C. Dahl
and Hengstenberg4° have proposed a mechanism of
c r a c k formation b a s e d on dendrite separation in the
columnar zone which is compatible with the view that
cracking occurs in the high temperature zone of low
ductility. Obviously, these c r a c k s can be prevented
by reduction of the pinch roll pressure.
Fig. 28--Effect of steel composition on longitudinal, mid-
For convenience, the information given here on in- face cracking.35
t e r n a l c r a c k s is summarized in Table I.

SURFACE CRACKS
From the standpoint of quality, surface c r a c k s nor-
mally p o s e more of a problem than internal c r a c k s
because, b e i n g exposed to air, the c r a c k surfaces
oxidize and do not reweld d u r i n g rolling. The follow-
ing surface c r a c k types are discussed h e r e : longitudi-
nal midface and c o r n e r cracks, transverse midface
and c o r n e r cracks, and star c r a c k s .

Longitudinal, Midface Cracks

This type of surface c r a c k , an example of which is
shown in Fig. 27, is a particularly serious problem in
the casting of slabs. They can vary in size from short,
shallow c r a c k s which can be removed by hand s c a r f -
ing, t o long, deep c r a c k s which may necessitate scrap-
ping of the slab. A very thorough investigation of
longitudinal, midface c r a c k s has been reported re-
cently by Irving and Perkins .35 They have shown that
steel composition has a strong influence on the fre-
quency and severity of cracking. As can be seen in
Fig. 28(a) carbon levels in the vicinity of 0.12 pct
are particularly susceptible t o c r a c k formation while
decreasing Mn/S and i n c r e a s i n g sulfur contents also
contribute t o the cracking problem (Fig. 28(b) and
(c)). Aluminum was found t o e x e r t little influence on
c r a c k formation except when present in low concen-
trations (0.004 pct) and in thin slabs. Niobium (colum-
bium) appears to have a deleterious effect s i n c e nio-
bium steels (0.02 to 0.04 pct) had poorer surface
Fig. 27--Longitudinal, midface crack in a slab.41 quality than niobium-free steels.

500-VOLUME 8B, SEPTEMBER 1977 METALLURGICAL TRANSACTIONS B

 nonuniform cooling: uneven distribution of mold pow-
1830M180mm Section
der, l o c a l air gaps, or uneven shell thickness owing t o
10350t 113/.0.t 8100t 29820t /.7700.t 13260t Tonnes erosion by the input m e t a l stream. Whatever t h e i r
origin, the hot s p o t s are regions of low ductility above
~ 90.
about 1340°C and are susceptible to cracking.
The cracking strains that develop in the s h e l l in the
mold have been analyzed by Sorimachi et a143 u s i n g
finite element analysis. They have shown that tensile
$
o strains sufficient t o c a u s e cracking in the surface hot
NTC s p o t s can a r i s e from a combination of:
[i] cooling of the wide face;
ZlO.,O,
Z 2 0.166
lOlO,
0.150
1o.o,61 o.o,61o.o9,1o.o,11
0-150 0-1/.1 0.1;1 0.133
Specific
coolina
rate{Ifkg) [ii] ferrostatic pressure acting on the narrow face;
i ~ = ~ ~ ~i Trial
number
[iii] peritectic phase transformation;
[iv] frictional f o r c e s acting in the transverse plane.
Fig. 29--Effect of top zone spray cooling on longitudinal, mid-
face cracking.35 This analysis indicates that the influence of carbon on
cracking, as observed by Irving and Perkins, is due t o
the 5 to 7 phase transformation which is accompanied
S e v e r a l operational factors have been reported to by a substantial shrinkage. E a r l i e r G r i l l and B r i m a -
i n c r e a s e the severity of longitudinal, mid/ace cracks: combe 44 had proposed that the same transformation
was responsible for a reduction in mold heat flux
[i]varying or increasing casting speed;35
[ii]casting with a high pouring temperature; 41 when casting 0.1 pct carbon steels. The effects of
mold powder, casting speed, slab width and s t e e l com-
[iii]casting wide slabs;41
position were also r e l a t e d to c r a c k formation with the
[iv] mold conditions-improper water cooling, 41 loss
finite element calculations.
of taper, i r r e g u l a r mold oscillation, improper
mold powder, worn molds;42 If longitudinal, mid/ace c r a c k s form predominantly
Iv] overcooling in the upper spray zones;35,41 in the mold, it would appear that excessive top zone
cooling or inadequate support below the mold have
[vii insufficient support below the mold; 35,4~
the effect of worsening the cracks. Both factors cause
[vii] poor alignment between mold and submold sup-
port system. l a r g e tensile strains at the surface that could deepen
and lengthen the existing cracks.
It has also been suggested that thin s l a b s are more
susceptible to c r a c k formation than thick slabs. This
Longitudinal Corner Cracks
behavior can be seen in Fig. 28. However, Irving and
Perkins have cautioned against drawing such a con- In blooms and billets, longitudinal c o r n e r cracks,
clusion because thinner slabs are normally cast at a as distinguished from off-corner c r a c k s to be dis-
h i g h e r speed.
The influence of top zone spray cooling on longi-
tudinal, mid/ace cracking can be seen in Fig. 29. In
this particular case a reduction in the top zone spray
w a t e r from 0.267 to 0.224 1//kg had a beneficial influ-
ence on surface quality. In another study4~ a reduction
of the spray cooling in the upper zones by as much as
35 pct also led t o an improvement in surface quality.
It has further been emphasized in this r e g a r d that
c l o s e attention must be paid to nozzle plugging as un-
even cooling gives rise t o more cracks.
The problem of inadequate support below the mold
in relation t o longitudinal, mid/ace cracks, has been
examined in s e v e r a l investigations. Irving and Per-
kins35 found that cooling g r i d s were preferable to cool-
ing plates for surface quality. L a r s e n and Moss 4~ re-
ported that solid, externally cooled r o l l s were less
susceptible t o breakage than internally cooled r o l l s ,
and therefore provided g r e a t e r support and better sur-
face quality on a continuing basis.
From an examination of these factors it is c l e a r
that mid/ace, longitudinal cracks form either in the
mold or upper spray zones. The mold, however, is the
most likely cracking site if the cracks form when the
surface is in the high temperature zone of low
strength and ductility. Cracking will then o c c u r within
about 6 s of the meniscus but may also take place
lower in the mold if localized hot s p o t s persist below
that level. The hot s p o t s can r e s u l t from nonuniform
cooling in the mold. A variety of phenomena can cause F i g . 3 0 - - L o n g i t u d i n a l , c o r n e r c r a c k i n a i 0 c m2 b i l l e t .

METALLURGICAL TRANSACTIONS B VOLUME 8B, SEPTEMBER 1 9 7 7 - 5 0 1

cussed later, tend t o appear directly on the c o r n e r s of mold and r o l l e r apron to reduce wear at the mold
as can be seen in Fig. 30. The c r a c k s have been re- bottom. Care should also be taken t o e n s u r e that
lated to a number of operational and metallurgical lubrication is adequate and uniform; for rapeseed oil
factors:l'45'4B lubrication this m e a n s checking that lubricant holes
a r e not clogged.
[i] a r e v e r s a l of mold taper owing to distortion
Regarding mechanism, longitudinal c o r n e r c r a c k s
a n d / o r wear of the mold face and corners;
clearly form in the mold. The fact that mold taper
[ii] with plate molds, c o r n e r gaps in e x c e s s of 0.8
p l a y s a critical role may further indicate that the
mm;
c r a c k s form a f t e r a c o r n e r of the billet has reheated
[iii] high tundish temperature;
with the surface temperature approaching the high
[iv] high casting speed;
temperature zone of low ductility. The combination
[v] incorrect foot r o l l e r settings;
of r e v e r s e taper and t h e r m a l shrinkage could c a u s e
[vii s t e e l containing 0.17 to 0.25 pct C, S > 0.035
the air gap near the c o r n e r to i n c r e a s e rapidly thereby
pct, P > 0.035 pct.
lowering the l o c a l rate of heat extraction and forcing
Aketa and Ushijima46 have suggested in an e a r l y work the surface temperature to rebound. The tensile
that the size of the c o r n e r radius in the mold is also strains at the surface in the c o r n e r would be g e n e r -
another factor that affects the frequency of longi- ated by a combination of t h e r m a l shrinkage and f e r r o -
tudinal, c o r n e r cracks. They found that a c o r n e r static pressure, and could e x c e e d the ductility of the
radius that is 1/10 of a side minimizes c r a c k f o r m a - s t e e l above 1340°C.
tion; but according t o Mori1 t h e r e are differences of Longitudinal, c o r n e r c r a c k s have also been re-
opinion r e g a r d i n g this conclusion. ported for the casting of slabs, but these tend t o be
The tendency toward r e v e r s e taper of the mold is situated just off the corners, as shown in Fig. 31.
possibly the most important of the factors listed Ozeki and Duke,34 for example, have described longi-
above. The c r a c k s form when the mold bulges toward tudinal c r a c k s located about 6 mm off the edge of slabs
the centerline at the meniscus l e v e l while at the same east with a 122 cm long mold. They found that the
time becoming worn near the bottom. In the case of c r a c k s were most prevalent when casting 0.15 t o 0.23
20 cm2 billets, for example, the frequency of c o r n e r pet carbon steels, but they could be eliminated by us-
c r a c k s apparently increases when the difference be- ing a shorter 90 cm mold. Interestingly, these authors
tween mold dimensions at the meniscus l e v e l and were also able to show, from an examination of b r e a k -
bottom exceeds 1.2 mm.1 Thus as the length of ser- out shells, that the c r a c k s formed high up in the mold,
vice of a mold increases in a campaign the frequency as close as 36 cm t o the meniscus. In addition, their
of longitudinal c o r n e r cracks is bound to increase. The study revealed that the c r a c k s were always a c c o m -
c r a c k s also appear increasingly at acute c o r n e r s if the panied by bulging of the slab edge, owing to inadequate
cast section is rhomboid. support of the shell in the e a r l y stages of solidifica-
Mori~ suggests the following measures to minimize tion. The strain generated by bulging probably acts
the formation of this type of crack: chromium plating on off-corner hot s p o t s that are above 1340°C. The
t o reduce mold w e a r , f a c i n g of the mold t o c o r r e c t generation of these hot s p o t s has been investigated by
the mold geometry, mold replacement, and adjustment G r i l l e t a147 in a theoretical examination of breakouts.

Fig. 31-Longitudinal,
corner crack in a slab.

502-VOLUME 8B, SEPTEMBER 1977 METALLURGICAL TRANSACTIONS B

 Transverse, Midface and Corner Cracks
Transverse cracks can be a s e v e r e problem in the
casting of slabs. They are often difficult to see on the
original slab surface, and to be revealed, require a
control-scarf pass. As can be seen in F i g s . 32 and 33,
the cracks appear at the base of the oscillation marks,
both in the midface r e g i o n and on or near the corners.
The d e p t h of transverse c r a c k s v a r i e s from 1 mm up
to a maximum of 10 to 15 mm although at the corners
it may be as much as 50 mm} 1,a%48 Thus in s e v e r e
c a s e s a scarfing d e p t h in e x c e s s of 10 mm may be
necessary to completely eliminate this defect.
One of the most important factors that affects the
severity of transverse cracks is steel composition.
S t e e l containing aluminum, niobium, vanadium and
over 1 pct manganese is particularly susceptible to
cracking. Hater e t a121 have reported that the cracks
o c c u r between austenite g r a i n boundaries which are
enriched in aluminum, probably in the form of A1N. Fig. 33--Transverse, corner crack in a slab.41
They found that the addition of 0.02 to 0.04 pct titan-
ium, to lower the soluble nitrogen level, had the ef-
fect of reducing the number of transverse cracks, al- sumably this helps to lower the soluble nitrogen con-
though over 0.15 pct was needed to completely e l i m i - tent.
nate them. At the same time, Japanese workersx have From these plant observations it is certain that
determined that the number of transverse c r a c k s is transverse cracking is related t o the low temperature
proportional to the aluminum and carbon contents in zone of low ductility. Because A1N precipitation is at
solution, and have observed A1N precipitates at open l e a s t partially responsible for this behavior, the
c r a c k surfaces. solution to transverse cracking lies both in the pre-
The longitudinal tensile strain that causes t r a n s - vention of the precipitation event and the minimization
v e r s e cracking is generated in the straightener or in of longitudinal tensile strain at the surface. This
the spray chamber. Straightening appears to r e s u l t m e a n s maintaining a surface temperature outside the
in c r a c k formation if the surface temperature is too 700 to 900°C r a n g e and reducing as much as possible
low, particularly in the r a n g e of 700 to 900°C. Spray repeated heating and cooling cycles in the spray c h a m -
cooling can lead t o cracking if l a r g e temperature ber. Reduction of spray cooling, as observed by H a t e r
gradients are set up in the longitudinal direction. et al, improved surface quality because it had the
Hater et a121 have demonstrated the latter effect by three-fold effect of increasing surface temperature,
stopping a cast, and after complete solidification driv- reducing the severity of heating/cooling cycles and
ing the strand upward. They observed transverse lowering the tensile strains at the surface. In c a s e s
c r a c k s at a l e v e l 130 to 190 cm below the m e n i s c u s - where the cracking strains result from the straighten-
well above the straightening section. By lowering the ing operation, a reduction in water s p r a y s will also
specific cooling water rate to 0.5 L/kg, and thereby re- have a beneficial effect provided that the surface tem-
ducing the temperature gradients, they obtained an im- perature at the straightener is g r e a t e r than 900°C
provement in surface quality. to minimize precipitation of A1N.
Ozeki and Duke34 found the solution t o their trans- In the casting of billets, transverse c r a c k s (Fig.
v e r s e cracking problem was to d e g a s c r a c k sensitive 34) can also form d u r i n g straightening if certain re-
g r a d e s of steel and add alloys under vacuum. Pre- sidual elements, particularly phosphorus and sulfur
are present in high concentrations. In addition, t r a n s -
v e r s e c r a c k s may result from excessive friction in
the mold a r i s i n g from inadequate lubrication or mis-
alignment.

Star Cracks
Star c r a c k s can be seen in Fig. 35. As the name
suggests, these c r a c k s appear in clusters, each
roughly in the pattern of a s t a r . The c r a c k s are usu-
ally less than 1.5 mm deep but may extend to 3 mm
in s e v e r e c a s e s . They can be linked t o the scraping of
copper from the mold walls .49 The copper penetrates
the surface and causes hot shortness 5° in localized re-
gions. Because, under n o r m a l cooling conditions the
surface is under tension the hot short a r e a s are
highly susceptible to cracking. This problem usually
can be solved by plating the inside mold walls with
Fig. 32--Transverse, midface cracks in a slab.34 chromium and realigning the mold and r o l l e r apron

METALLURGICAL TRANSACTIONS B VOLUME 8B, SEPTEMBER 1977-503

 Table II. Surface Cracks*

Type Cause Influencing Factors Corrective Action

Longitudinal, Tensile strain Crack frequency increased by: Adjust mold conditions to ensure
midface cracks generated in themold -carbon levelsof0.12 pct uniform cooling. Reduce cooling
and upperspray zones -increasing S and decreasing Mn/S in upper spray zones and check
-varying orincreasingcasting speed submold support system
-casting with high pouring temperatures
-casting wide slabs
-mold conditions-improper water cooling,
loss oftaper, irregular mold oscillation,
improper mold powder, worn molds
-overcooling in upper spray zones
-insufficient support belowmold
-poor alignment between mold and sub-
mold support system
Longitudinal, Nonuniform cooling in Crackingassociatedwith: Hate mold wallswith chromium.
corner cracks corner region -reversal ofmold taper owing to distortion Reface the mold. Check align-
o r wear ment and lubrication for uni-
-large corner gaps in plate molds formity
-high tundish temperature
-highcasting speed
-incorrect foot rollersettings
-steel containing 0.17-0.25 pct C, S >
0.035 pct, P > 0.035 pct
In slabs off-corner Long moldsand steel with 0.15 to Check mold support conditions.
cracks caused by bulging 0.23 pct C give rise to worse cracking
of narrow face in mold
Transverse, Large surface temperature Strongly influenced by steel composition, Reduce spray cooling and make as
midface and gradients in the spray zone AI, V, Nb, Mn ;> 1 pct being the most uniform as possible to minimize
corner cracks and straighteningwithin an important elements cooling]reheatingcycles and
unfavorable range oftemp- maintain surface temperature
erature, between 700 and above 900°C through to
900°C straightener
Star cracks Scraping of copper from Secondary cooling Plate mold wallswith chromium.
mold Adjust machine alignment

*With the exception oftransversecracks, surface cracks probably form in the high temperature zone oflow ductility.

Fig. 34--Transverse, m i d f a c e c r a c k i n a b i l l e t .

F i g . 35--Star c r a c k s o n the
s u r f a c e of a s l a b . 4a

504-VOLUME 8B, SEPTEMBER 1977 METALLURGICAL TRANSACTIONS B

to minimize scraping. Nickel mold linings have a l s o 9. H. Vom Ende and G. Vogt: J. Iron Steellnst., 1972, vol. 210, pp. 889-94.
been used successfully for the c a s t i n g of stainless 10. H. Fujii T. Ohashi,K. Hiromoto, and M. Oda: Tetsu ToHagane, 1975, vol.
s t e e l slabs .49 61, p. SS56.
11. A. Gueussierand R. Castro: RevueMet., 1960, vol. 57, pp. 117-34.
A s u m m a r y of the information on surface cracks is 12. R. P. Sopher: WeldingJ., 1958, vol. 37, pp. 481S-92S.
given in T a b l e II. 13. N. N. Guglin and B. B. Gulyaev:StalEng., 1961, no. 9, pp. 679-84.
14. L. I. Morozenskii, O. A. Mitenev, and V. K. Krutikov: Stal, 1965, no. 4,
pp. 272-76.
SUMMARY 15. T. Kinoshita and S. Kuroki: Imono, 1972, vol. 44, pp. 1080-89.
16. H. F. Hall: Second Report o f Steel Castings Research Committee, pp. 65-93,
This review has attempted to b r i n g together plant Iron Steel Inst., London, 1936.
experience on crack formation in continuously cast 17. K. Matsubara: Trans. Iron Steellnst. Jap., 1966, vol. 6, pp. 139-48.
s t e e l with fundamental knowledge of s t r e s s generation 18. G. A. Wilber, R. Batra, W. F. Savage, and W.J. Childs: Met. Trani A, 1975,
and the mechanical properties of s t e e l at e l e v a t e d tem- vol. 6A, pp. 1727-35.
19. Y. Iida, H. Moriwaki, T. Ueda, and Y. Habu: Tetsu To Hagane, 1973, vol.
peratures. It has shown that the cracks observed can
59, p. S89.
probably be linked to two zones of low ductility in 20. V. M. Mayrhofer: BergHuettenmaennMonatsh, 1975, vol. 120, pp. 312-21.
steel: a high temperature zone, above about 1340°C, 21. M. Hater, R. Klages, B. Redenz, andK. T/iffner: OpenHearthProc., 1973,
which appears to be responsible for most of the cracks, vol. 56, pp. 202-217.
including all internal cracks; and, a low temperature 22. S. Hasabe, T. Koga, T. Yamura, and Y. Sugikawa: Tetsu ToHagane, 1972,
zone, ranging between 700 and 900°C, which g i v e s r i s e vol. 58, p. 522.
23. S. C. Desai:J. Iron Steellnst., 1959, vol. 191, pp. 250-56.
to transverse cracking problems at the s u r f a c e . This 24. Y. Sugitani: Unpublished research, Sumitomo Metal Industries.
knowledge, combined with an understanding of the 25. J. L Robbins, O. C. Shepard, and O. D. Sherby: Trans. ASM, 1967, vol. 60,
s t r e s s e s generated in continuous casting, m a k e s it pp. 205-16.
p o s s i b l e to propose mechanisms of crack formation 26. A. Ito, Y. Setuji, and T. Sakumoto: Tetsu To Hagane, 1975, vol. 61, p. S132.
and relate them to operating and metallurgical vari- 27. G. Van Drunen, J. K. Brimacombe, and E. Weinberg: lronmakingSteelmaking,
1975, vol. 2, pp. 125-33.
ables. 28. L. Backer and P. Gosselin: Open Hearth Proc., 1970, vol. 53, pp. 145-56.
One of the most important, and perhaps obvious, 29. C. I. Miller, Jr.: Open Hearth Proc., 1971, vol. 54, pp. 316-21.
conclusions that e m e r g e s from the r e v i e w i s that to 30. J. W.Donaldson:J. Metals, 1965, vol. 17, no. 12, pp. 1338-43.
minimize crack formation, machine maintenance and 31. K. P. Korotkov: The Continuous Casting o f Steel in Commercial Use,
control of s t e e l c h e m i s t r y are absolutely essential. Pergamon, Oxford, 1960.
32. A. Grill, J. K. Brimacombe, and E. Weinberg: lronmaking Steelmaking, 1976,
Control of the s t e e l composition, v i e w e d in funda- vol. 3, pp. 38-47.
mental t e r m s , means control of high temperature 33. J. K. Brimacombe:Can. Met. Quart., 1976, vol. 15, pp. 163-75.
strength and ductility. Machine maintenance ensures 34. R. K. Ozeki and J. D. Duke: Proceedingsto be published, MetalsSociety/
that tensile s t r e s s e s and s t r a i n s , generated both t h e r - IRSID Conference on Continuous Casting, Biarritz, 1976.
35. W. R. Irvingand A. Perkins: Proceedingsto be published, MetalsSociety/
m a l l y and mechanically, are minimized.
IRSID Conference on Continuous Casting, Biarritz, 1976.
36. ISI Publ. 106, 1967.
37. M. Ya. Brovman: StalEng., 1967, vol. 1, p. 26.
ACKNOWLEDGMENTS
38. A. Grill and J. K. Brimacombe: Upublishedwork.
The authors are grateful t o s e v e r a l s t e e l companies 39. L. Schmidt and H. Eredriksson:IronmakingSteelmaking, 1975,vol. 2,
and to CONCAST for the provision of m a t e r i a l used in pp. 61-67.
40. W. Dahl and H. Hengstenberg: Z. Metallk., 1969, vol. 60, pp. 340-50.
the review. One of the authors (K. S.) wishes to thank 41. G. L. Larsen and T. E. Moss: Open Hearth Proc., 1974, vol. 57, pp. 211-25.
Kawasaki S t e e l Corporation for the granting of a two 42. E. L. Fogleman and R. T. Orie: Open Hearth Proc., 1974, vol. 57,
year study l e a v e at U.B.C. pp. 228-44.
43. A. S. BaUantyne, K. Sorimachi, and J. K. Brimacombe: Unpublishedwork.
44. A. Grill and J. K. Brimacombe: IronmakingSteelmaking, 1976, vol. 3,
REFERENCES pp. 76-79.
45. J. S. Webster and C. Graham: Proceedings to be published, MetalsSociety/
1. H. Mori: Tetsu To Hagane, 1972, vol. 58, pp. 1511-34. IRSID Conference on Continuous Casting, Biarritz, 1976.
2. H. Mori: Tetsu ToHagane, 1974, vol. 60, pp. 784-93. 46. Y. Aketa and K. Ushijima: Tetsu ToHagane, 1962, vol. 2, pp. 22-31.
3. W. T. Lankford: Met. Tran~, 1972, vol. 3, pp. 1331-57. 47. A. Grill, K. Sorimachiand J. K. Brimacombe: Met. Tran~ B, 1976, vol. 7B,
4. K. Sorimachiand J. K. Brimacombe: Unpublished research, University of pp. 177-89.
British Columbia, Vancouver, Canada, 1976. 48. S. Farkas, J. W. Shaw, and J. D. Gricol: Open Hearth Proc., 1971, vol. 54,
5. P. J. Wrayand M. F. Holmes: Met. Tran~ A, 1975,vol. 6A, pp. 1189-96. pp. 68-80.
6. C. J. Adams: OpenHearth Proc., 1971, vol. 54, pp. 290-302. 49. Y. Nakano, Y. Noguchi,F. Hoshi, and Y. Muranaka: Proceedingsto be
7. A. Fuchs: ESTEL.Berichteaus Forschungund Entwicklung unserer Werke, published, MetalsSociety/IRSID Conference on Continuous Casting, Biarritz,
1975, no. 3, pp. 127-35. 1976.
8. A. Nadai and M. J. Manjoine: J. Appt Meclz, 1941, vol. 63, pp. A77-A91. 50. E. T. Stephenson: J. Metals, 1974, vol. 20, no. 3, pp. 48-51.

METALLURGICAL TRANSACTIONS B VOLUME 8B, SEPTEMBER 1977-505