Materials Transactions, Vol. 51, No. 2 (2010) pp.

371 to 376
#2010 Japan Foundary Engineering Society

Cold Crack Criterion for ADC12 Aluminum Alloy Die Casting*

Shuxin Dong, Yasushi Iwata, Yoshio Sugiyama and Hiroaki Iwahori
Materials Fundamental Research Division, Toyota Central R&D Labs., Inc., Aichi 480-1192, Japan
A cold crack criterion for JIS ADC12 aluminum alloy die casting is proposed. Through investigating the temperature dependence of the
fracture strain of JIS ADC12 aluminum alloy die casting, it was found that the fracture strain features a turning point at a temperature, Tc (we
called critical temperature to the ductility, about 573 K for the present composition), i.e. stays low while below Tc , rises rapidly to a high level
beyond Tc . Focusing on this character of the fracture strain, we analyzed the equivalent plastic strain ("c ) of the castings introduced below Tc in
casting processes by thermal stress simulations and compared with the occurrence of cold cracks in the die casting experiments. It was found that
the "c of the cracking positions in the castings exceeded, while the "c of the castings without crack were much lower than the fracture strain of JIS
ADC12 aluminum alloy die casting below Tc . That is to say, the occurrence of the cold crack in a die casting can be judged by comparing the "c
with the fracture strain below Tc . Based on this proposed criterion, it is possible to predict the appearance of the cold cracks in ADC12 die
castings by thermal stress simulations. [doi:10.2320/matertrans.F-M2009832]
(Received August 6, 2009; Accepted November 6, 2009; Published December 23, 2009)
Keywords: aluminum alloy, die casting, defect, cold crack, strain, simulation

1.

Introduction

The production of aluminum alloy die castings has been

seeing a continuous increase to answer the strong demand by
the automotive industry for higher strength and lighter
weight. The die castings for automotives are required of not
only high quality as of course but also thin walls and a high
dimensional precision, thus it is essential to provide advanced
casting technologies for the manufacturing of these castings.
The cold crack of castings, which was not considered as a
serious problem before, has become one of the most
important subjects together with the security of the dimensional precision in the die casting production.
It is crucial to predict the occurrence of cracks by
numerical simulations based on their formation mechanisms
in order to prevent such defects. Although many excellent
thermal and stress simulation techniques have been developed for die casting processes, we still have diculties in
predicting the occurrence of such cracks accurately.1) One
reason for such a situation can be attributed to the lack of the
knowledge of the cold crack formation conditions.
The cracks in die castings, according to their occurrence
time in the casting process, can be classied into 2 types, i.e.,
one forms during the solidication and another forms during
the cooling after the solidication.24) The former, called hot
crack or hot tear, happens at the low ductility temperature
range around the solidus temperature of cast alloys, while the
latter, named cold crack, appears at lower temperature in
the cooling process. The hot cracks show fractographies of
dendrites or solidied liquid metal lms, while the cold
cracks have fractographies like that of general mechanically
fractured metals. A large number of researches have been
made on the hot cracks encountered in continuous casting and
shape casting processes and several formation mechanisms
have also been proposed,59) while few researches can be
found on the cold crack in the literature. With the progress of
thinner walls and a higher dimension precision of the die
castings, the problem of the cold cracks resulting from the
*This Paper was Originally Published in Japanese in J. JFS

231.

81 (2009) 226

restraints of molds, inserts or casting itself is becoming more

and more obvious and the prediction technology for this
defect is much desired. Although the cold cracks occur in
solid metals of which the fracture theory has been fairly well
established at some denite temperature, the formation
condition is still left to be revealed. That is because these
cracks originate in the continuous cooling process of solid
metals in a temperature range of several hundred degrees.
Such a large range of temperature change makes it dicult
to understand the occurrence condition of the cold cracks
because the mechanical properties of metals, such as ultimate
strength, fracture strain, etc., may change radically.
In order to clarify the formation condition of the cold
crack occurring in JIS ADC12 aluminum alloy (hereinafter,
called ADC12 alloy) die castings, we produced the defects
by die casting experiments and investigated the temperature
dependence of the fracture strain of ADC12 alloy by tensile
tests, then traced the strain in the die casting processes by
numerical simulations of thermal stress.
2.

Experimental and Simulation Methods

2.1 Castings and casting method

The die casting shown in Fig. 1 was used for the
experiments producing the cold cracks. Two rings made of
SUS304 stainless steel (hereinafter, called SUS304) were set
in the inner side of the cylinder-shaped cavities of the mold as
inserts before casting. The die casting conditions are given in
Table 1. The mold was kept at room temperature (298 K) and
JIS AD12.1 aluminum alloy melt (hereinafter, called AD12.1
alloy melt) of the chemical composition shown in Table 2
was shot into the mold vertically from the lower side at
913 K. The distance between the SUS304 rings was changed
from 1 to 5 mm intentionally attempting to produce the crack
in the casting between the two rings. The outer surfaces of
rings which contact the melt in the casting were nished to
a surface roughness of 50 mm with lathe. To detect the
occurrence time of the cracks in the casting, a high
temperature strain gage was installed to the inner surface
(the surface contacts the mold) of one SUS304 ring in the
circumferential direction.

372

S. Dong, Y. Iwata, Y. Sugiyama and H. Iwahori

60

A
20
10

70

110

2- 34
2- 40
2- 52

16

Fig. 2 Die casting for tensile test specimens.

Distance
between
the rings

Unit: mm

A-A

24

Unit: mm

Casting

ADC12 Al alloy

10

Biscuit

Strain gage

Biscuit

20

Overflow
For tensile test

SUS304
insert rings

48
35

44
30

Fig. 1 Die casting for crack experiment.

Conditions for die casting experiment.

Shot velocity
(m/s)

Casting pressure
(MPa)

Melt temp.
(K)

Die temp.
(K)

0.4

25

913

298

(a) For below 473K

5

Table 1

35
60

20
60

Table 2 Chemical composition of AD12.1 aluminum alloy.

(mass%)
Si

Fe

Cu

Mn

Mg

Zn

Ni

Sn

Al

11.62

0.88

2.89

0.34

0.21

0.93

0.05

0.02

Bal.

2.2

Measurement of the mechanical properties of

ADC12 alloy die casting
AD12.1 alloy melt of a chemical composition identical to
that used for the die casting experiments was die cast to plateshaped castings for machining tensile test specimens. The
dimensions of the die casting for the tensile test specimens
are shown in Fig. 2. The casting conditions were the same
as that in section 2.1. Conrmed by transmission X-ray, the
castings without defects such as shrinkages, entrapped gases,
etc., were selected for machining the tensile test specimens.
The specimens were cut out from the positions shown in
Fig. 2 and machined by lathe to the nal shapes for the tensile
tests below 473 K and above 523 K respectively as illustrated
in Fig. 3. In the machining, specimens observed any
shrinkages or inclusions by visual inspection were excluded.
The tensile tests were carried out at seven temperatures
298, 373, 473, 523, 573, 673, and 773 K chosen from the
range of room temperature to the solidus temperature of
AD12.1 alloy. The strain rates during the cooling of the die
casting for the experiments producing the cracks in the
present study were estimated by simulation to be from 0 to
0.1 s 1 and the average strain rate, 0.05 s 1 was used as the
strain rate for the tensile tests. The strain in the tensile tests
was measured by a video camera and a dierential transformer strain gages for the test below 473 K and over 523 K
respectively. Each of the tensile test was started after the

(b) For above 523K

Fig. 3

Unit: mm

Tensile test specimens.

specimen was kept for 5 minutes at a predetermined

temperature. The temperature of the specimens was measured by a 0.1 mm diameter K-type thermocouple which was
welded to the center of the specimen by resistance-welding.
The mechanical properties of the SUS304 ring were also
obtained by tensile tests with the JIS G0567 II-10 specimens
machined from a 20 mm diameter bar under the same
temperatures and strain rate as that for the ADC12 alloy die
casting.
2.3 Thermal stress simulation of the die casting
To nd out the relation between the occurrence of the
cracks and the strain of the die castings, solidication and
thermal stress simulations were conducted for the casting
shown in Fig. 1. The materials of the die casting and the rings
are AD12.1 alloy and SUS304 respectively. The material
properties used for the simulation are given in Table 3 and
Table 4. The FEM model was taken as half of the die casting
because of its symmetric shape as shown in Fig. 4. The
coupled thermal mechanical analysis was performed with the
commercial structural analysis software MSC. MARC.
In the thermal analysis, heat transfer boundary condition
was applied to the casting/mold, casting/ring, and ring/mold
interfaces. Table 5 gives the heat transfer coecients for
each of the interfaces. The cavity of the mold at 298 K full of
AD12.1 melt at 913 K was assumed as the initial condition of

Cold Crack Criterion for ADC12 Aluminum Alloy Die Casting

373

Table 3 Properties of ADC12 aluminum alloy used for thermal stress simulation.
Modulus of
elasticity
(GPa)

Poissons
ratio
(-)

Density
(kg
m 3 )

76

0.3

2.67

Table 4

Coecient of linear
thermal expansion
(K 1 )
2:06  10

Specic
heat
(J
kg
K 1 )

Thermal
conductivity
(W
m 1
K 1 )

Latent heat
(kJ
kg 1 )

Liquidus
temp.
(K)

Solidus
temp.
(K)

962

121

390

853

803

Properties of SUS304 stainless steel used for thermal stress simulation.

Modulus of elasticity
(GPa)

Poissons ratio
(-)

Density
(kg
m 3 )

192

0.28

8.03

Coecient of linear
thermal expansion
(K 1 )
1:71  10

Specic heat
(J
kg
K 1 )

Thermal conductivity
(W
m 1
K 1 )

502

15.1

crack

10mm
Appearance
Fig. 5

Fig. 4
Table 5

Crack of ADC12 die casting.

FEM model for thermal stress simulation.

Crack

Interface heat conductance used for thermal stress simulation.

2

(W
m
K )
Castings (liquid)/
Dies

Castings (solid)/
Dies

Inserts/
Dies

Castings/
Atmosphere

41840

8368

8368

837

the thermal and stress analysis. The latent heat release of

AD12.1 melt was treated by the equivalent specic heat
method.
Both of the stress-strain behaviors of the ADC12 alloy die
casting and the SUS304 rings were treated as elastoplastic as
observed in the tensile tests and the stress-strain relations
obtained at various temperatures in the tensile tests of
section 2.2 were utilized in the stress analyses. Therefore, the
temperature dependence of the elasticity and the yield stress
gradients were taken into account for the ADC12 alloy die
castings and the SUS304 rings in the stress simulation. The
completely free sliding mechanical boundary condition was
applied to the interface between the die casting and the rings.
3.

50 m
Fractography

Results and Discussion

3.1 Occurrence of crack in the die casting

Cracks were observed as the die casting was made with
dierent distances between the two rings. One example of the
appearance and the fractography of the crack is shown in
Fig. 5. The crack occurred in the narrowest part of the casting
between the two rings and showed a dimple-covered fractography. Thus, the crack can be considered as a cold crack but

Crack

No crack
Critical distance
0.8

Fig. 6

2.4
2.8
1.2
1.6
2.0
Distance of insert rings, mm

Relation between crack and distance of insert rings.

not a hot crack according to the characters mentioned in

section 1. The relation between the crack occurrence and the
distance of rings is illustrated in Fig. 6. It can be seen that no
crack was observed when the distance of the rings exceeded
1.6 mm, while the cracks were well reproducible when the
distance between the rings was below 1.6 mm.
To know the occurring time of the crack, the output of the
strain gages installed on the inner surfaces of the rings was
drawn in Fig. 7 together with the pressure change of the
plunger in the die casting process. Compared with the strain
curve of the gage installed on the ring of the die casting
without a crack (Fig. 7(a), the distance between rings equals
to 4 mm), the strain curve of the cracked die casting showed
an abrupt jump towards the tensile strain side in 3 second
after melt injection. What this sudden change of the strain

S. Dong, Y. Iwata, Y. Sugiyama and H. Iwahori

20

-5

-10
-15
0

pressure

10

0
1 2 3 4 5
Time from shot, s

(a) Ring distance: 4mm

(No crack)
Fig. 7

Crack Strain 30

0
20

-5
-10

pressure

10

0
-15
0 1 2 3 4 5
Time from injection, s

Critical temp.
to ductility, (Tc)

20

10
0
200

(b) Ring distance: 1.4mm

(Crack)

400
600
Temp., K

800

Change of strains of the inner surfaces of rings.

means is considered as follows. The insert rings were

compressed by the surrounding ADC12 alloy casting due to
the thermal contraction in the solidication and cooling
process. Here, if the casting between the rings fractures, the
compressive force acting on the rings in the symmetric axial
direction of the casting will become weaker thus the rings
will deform elastically from the compressed state to give a
little elongation in the symmetric axial direction. Because the
gage was installed on the inner surface of the left side of the
right side ring as shown in Fig. 1, it will be pulled a little bit
when the casting between rings fractures. From the above
consideration, the crack of casting can be considered as
occurred in 3 second after melt injection.
3.2

The temperature dependence of the fracture strain

of ADC12 alloy die casting
Through the above experiments, the critical distance
between the rings to and the occurring time of the crack
were determined. To examine the occurrence condition of the
crack further, the fracture strains of ADC12 alloy die casting
at dierent temperatures were investigated by tensile tests.
The relation between the fracture strain and the test temperature was illustrated in Fig. 8. It was discovered from Fig. 8
that the fracture strain of ADC12 alloy die casting showed a
relatively low value and a minor variation in the temperature
range from room temperature to the vicinity of 600 K, while
it arises rapidly beyond 600 K with the increase of temperature. Focusing on this feature of the fracture strain, two
tangent lines were drawn along the fracture strain curve
and the temperature (573 K for the present composition of
ADC12 alloy) corresponding to the intersection point of the
two tangent lines was dened as the critical temperature to
the ductility (hereinafter, called Tc temperature).

3.3

Fracture strain, %

30

30
Plunger pressure, MPa

Strain

End of filling
End of presurization
40
10
Strain, X100

Strain, X100

End of filling
End of presurization
40
10

Plunger pressure, MPa

374

Thermal stress analysis and crack criterion for

ADC12 alloy die casting
The equivalent strains (hereinafter, called strain) of
castings both with and without the cracks were calculated
in the process of solidication and cooling and compared
with the fracture strain of ADC12 alloy die casting. The
thermal stress simulations were conducted for the casting
having a distance of 1.4 mm between the rings as a sample
casting with the crack and the castings having distances of
2 mm and 4 mm between the rings as sample castings without
the crack.

Fig. 8 Relation between fracture strain and temperature of ADC12 alloy

die casting.

(Crack)

(No crack)

(No crack) Equivalent

plastic
strain [%]
8
4

1.4mm

2mm

4mm

Distance of insert rings

Fig. 9

Strain distributions of ADC12 alloy die castings at 3 s after lling.

From the tensile test result in section 3.2, ACD12 alloy die
casting showed an elastoplastic behavior even at room
temperature and is not completely brittle. Therefore, it is
appropriate to consider the occurrence condition of the crack
for ACD12 alloy die casting in terms of the strain but not the
stress. To investigate the relation between the crack occurrence and the strain that generated in the die castings during
casting, the strain distributions in the die castings in 3 second
(the time at which the crack occurred in the die casting
having a distance of 1.4 mm between the rings) after melt
injection are illustrated in Fig. 9. The maximum strain for
each of the die casting is located at the same position, i.e., the
narrowest part between the rings at which the crack occurred
for the die casting of a distance of 1.4 mm between the rings.
Nevertheless, the maximum value of the strain for each die
casting is almost the same, about 78%, no matter the crack
occurred (The die casting with a distance of 1.4 mm between
the rings) or not (The die castings with distances of 2 mm
and 4 mm between the rings). At the same time, the strain
distributions also do not show much dierence.
To nd the exact cause for the occurrence of the crack,
the temperature dependence of the fracture strain of ADC12
alloy die casting was examined as related to the occurrence
of the crack in the die castings. The fracture phenomena of
the elastoplastic materials dealt with in fracture mechanics
are generally limited to a denite temperature or a narrow
range of temperature at which materials do not show large
changes in the mechanical properties. In such a case, it can

Cold Crack Criterion for ADC12 Aluminum Alloy Die Casting

(No crack) Equivalent

plastic
Strain [%]
3

2
1
1.4mm

2mm

4mm

Fracture strain of
ADC12
2
1

Cumulative
strain below Tc
2mm(No crack)

4mm(No crack)
0
300
400
500
600
Temp., K

Distance of insert rings

Fig. 10 Strain distributions of ADC12 alloy die castings. (cumulative
strain below Tc .)

be considered that a material will crack when the strain

generated in the material exceeds its fracture strain. However, the crack of die castings occurred in the cooling process
with a temperature variation as large as several hundred
degrees from the solidus temperature of ADC12 alloy to
room temperature. That is to say, the total strain of die
castings generated in the cooling process includes strains
taking place at dierent temperatures. At the same time, the
fracture strain of the die casting changes with temperature.
Therefore, it is not clear which strain during the cooling of a
casting is critical to the crack occurrence. At the same time,
it is not clear which fracture strain measured of the cast
material at dierent temperature should be used as the
criterion of the crack occurrence.
The relation between the fracture strain of ADC12 alloy
die casting and temperature, as shown in Fig. 8, is characterized of a turning point at Tc temperature. The fracture
strain staying at a comparatively low value below temperature Tc increases rapidly with the rise of temperature beyond
Tc . Focusing on this feature of the fracture strain of ADC12
alloy die casting, the cumulative equivalent plastic strains
generated in the die castings with a distance between the
rings of 1.4 mm, 2 mm and 4 mm when the temperature
changing from Tc to room temperature (hereinafter, called
"c strain) were calculated by thermal stress simulations and
the distributions of the "c strain were illustrated in Fig. 10.
The "c strain of the die casting with the occurrence of the
crack (The distance between the rings was 1.4 mm) was over
3%, while those of the die castings without the occurrence of
the crack (The distances between the rings were 2 mm and
4 mm) all were under 0.5%. It can be said that the occurrence
of the crack corresponds well with the relation between the "c
strain and the fracture strain of ADC12 alloy die casting. In
addition, the highest strain appeared at the midpoint of the
casting along the axial length of the ring in Fig. 9, while the
highest strain appeared at the areas of the casting near the
two ends of the axial length of the ring in Fig. 10.
Note that the strain in Fig. 9 includes the strain occurring
above Tc temperature which may not contribute to the crack
occurrence, while the strain in Fig. 10 considers only that
below Tc . The generating process of the cumulative strain "c
was further discussed in some detail in the following.
To conrm the strain variation quantitatively further, the "c
strains of the surface nodes at the center between the rings,

Distance between the

rings 1.4mm(Crack)

700

Fig. 11 Comparison of cumulative strain below Tc of die castings and

fracture strain of ADC12 aluminum alloy.

Fracture strain

(No crack)

Strain, %

(Crack)

375

Casting

RT

523K

Tc

Tensile test Temp.

673K

Tensile test specimens

50 m

Fig. 12 Fractographies of die castings and tensile test specimen.

i.e., the "c strains of the nodes showed the highest strain in
Fig. 10, were extracted and compared with the fracture strain
along the decrease of temperature from 573 K in the cooling
processes. As illustrated in Fig. 11, the "c strain of the die
casting with a distance of 1.4 mm between the rings was
much higher than the fracture strain of ADC12 alloy die
casting, while the "c strains of the die castings with distances
of 2 mm and 4 mm between the rings were much lower than
the fracture strain of ADC12 alloy die casting.
According to the result shown in Fig. 11, the "c strain of
the die casting with a distance of 1.4 mm between the rings
intersected the curve of the fracture strain of ADC12 alloy die
casting in the temperature range between 540 K and 500 K
in the cooling process. That is to say, the crack can be
considered to have occurred in the temperature range
between 540 K and 500 K. From the thermal simulation of
the die casting, the temperature of the node where the cold
crack occurred reached between 540 K to 500 K in about 3
seconds after melt lling, which is in good agreement with
the occurrence time of the crack as observed in the die
casting experiment. In Fig. 12, the fractography of the crack
occurred in the die casting with a distance of 1.4 mm between
the rings was compared with the fractographies of the

376

S. Dong, Y. Iwata, Y. Sugiyama and H. Iwahori

specimens tensile tested at various temperatures. It was

observed that the crack fractograph of the die casting was
most similar to that of the specimen tested at 523 K which is
just between 540 K and 500 K.
Through the above examinations of the die casting
experiments and the thermal stress simulations, it can be
said that the occurrence of the cold crack in ADC12 alloy die
castings can be judged by comparing the "c strain with the
fracture strain of ACD12 alloy die casting, i.e., the "c strain
exceeding the fracture strain of ADC12 alloy die casting can
be adopted as a criterion for the occurrence of the cold crack.
Therefore, it is possible to predict the occurrence of the
cold crack and its position in a ADC12 alloy die casting
by calculating the distribution of the "c strain, i.e., the
cumulative strain from Tc temperature to room temperature
in the casting process.
4.

Conclusions

The occurrence phenomenon of the cold crack in ADC12

alloy die castings and its occurrence criterion were investigated by die casting experiments, tensile tests and thermal
stress simulations. The following conclusions were obtained.
(1) The fracture strain of ACD12 alloy die casting
increases rapidly with the rise of temperature beyond
573 K, this temperature corresponding to the turning point
of the fracture strain was dened as the critical temperature to
the ductility of ADC12 alloy die casting (Tc temperature).
(2) The cold crack observed in ADC12 alloy die castings
can be explained by the following cold crack occurrence

criterion. The cold crack will occur when the cumulative

equivalent plastic strain generated below Tc (the "c strain)
exceeds the fracture strain of ADC12 alloy die castings.
(3) It is possible to predict the occurring positions of cold
cracks in ADC12 alloy die castings through calculating the
distributions of the equivalent plastic strain generated below
Tc (the "c strain) by thermal mechanical simulation.
Acknowledgment
We are deeply grateful to Dr. Eng. Eisuke Niyama for his
helpful comments and advice.
REFERENCES
1) D. G. Eskin and L. Katgerman: Metall. Mater. Trans. 38A (2007) 1511
1519.
2) K. Agatsuma: Kinzoku Kougaku Kouza 5, Kakou Hen I, Youkai
Chuzo
Imono, (Asakura Shoten, 1969) p. 285.
3) B. G. Thomas: ISIJ Int. 35 (1995) 737743.
4) Nihon Chuzo kougakai Daikasuto Kenkyubukai: Daikasuto no Chuzokekkan
Furyo oyobi Taisaku Jirei Shu, (Japan Foundry Engineering
Society, 2000) p. 45.
5) D. G. Eskin, Suyitno and L. Katgerman: Progress Mater. Sci. 49 (2004)
629711.
6) H. F. Bishop, C. G. Ackerlind and W. S. Pellini: AFS Trans. 60 (1952)
818833.
7) J. Vero: Met. Industry 48 (1936) 431455.
8) E. Niyama: Japan-US Joint Seminar on Solidication of Metals and
Alloys, Tokyo (Japan Society for Promotion of Science) (1977) pp. 271
282.
9) H. Fredriksson, M. Haddad-Sabzevar, K. Hansson and J. Kron: Mater.
Sci. Technol. 21 (2005) 521529.