Materials Science and Engineering A 528 (2010) 583588

Contents lists available at ScienceDirect

Materials Science and Engineering A

journal homepage: www.elsevier.com/locate/msea

Effect of cooling rate on microstructure and mechanical properties

of gray cast iron
M.M. Jabbari Behnam a, , P. Davami a,b , N. Varahram a,b
a
b

Department of Materials Science and Engineering, Sharif University of Technology, P.O. Box 11365, Tehran, Iran
Razi Metallurgical Research Center, No. 8, Fernan St., Sorkhehesar Road, Km 21 Karaj Makhsous Road, Tehran, Iran

a r t i c l e

i n f o

Article history:
Received 4 July 2010
Received in revised form 29 August 2010
Accepted 10 September 2010

Keywords:
Gray cast iron
Cooling rate
DAS
SDAS
e
HB

a b s t r a c t
This paper presents the results obtained and the deductions made from a series of microstructural studies
and mechanical tests involving gray cast iron which was sand cast using a variety of modules. The effect
of cooling rate on the primary dendrite arm spacing (DAS), secondary dendrite arm spacing (SDAS),
thickness of ferritecementite layer (e ) and the hardness (HB) were evaluated. Results show that the
both DAS and SDAS and also e are highly dependent on the cooling rate, and they decreases as the cooling
rate increases. More attempts were also done to correlate the HB with DAS, SDAS and e . It was found
that HB increases as DAS, SDAS and e decreases.
2010 Elsevier B.V. All rights reserved.

1. Introduction
Gray cast iron (GCI) remains the most important casting material with over 70% of the total worlds production tonnage [1]. It is
traditionally chosen in many industrial applications because of its
exibility of use, good castability, low-cost (2040% less than steel)
and wide range of achievable mechanical properties. The structure
of GCI depends on chemical composition before the casting process, inoculants and cooling conditions [2]. The microstructure of
GCI is characterized by graphite lamellas dispersed into the ferrous
matrix. Foundry practice can inuence nucleation and growth of
graphite akes. So that size and type of graphite akes or (them),
enhance the desired properties. The amount of graphite and size,
morphology and distribution of graphite lamellas are critical in
determining the mechanical behavior of GCI [24].
The as-cast microstructure is governed by the solidication
process and solid state transformation (eutectoid reaction). The
Matrix microstructure depends on the conditions under which the
eutectoid reaction occurs. Among the variables that inuence the
mechanism of the eutectoid reaction are the chemical composition
and the cooling rate through the eutectoid temperature range. The
result of eutectoid transformation have key role in determining the
mechanical properties of cast iron. Thus, the effect of alloying ele-

Corresponding author. Tel.: +98 9143107797; fax: +98 21 66036012.

E-mail address: jabbaribehnam@gmail.com (M.M. Jabbari Behnam).
0921-5093/$ see front matter 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.msea.2010.09.087

ments on mechanical properties of iron might be related to their

inuence on eutectoid transformation [57].
Although there is sufcient information available on
microstructural characteristics of cast iron, however effect of
them on mechanical property (hardness) is not formulated. This
has led to a series of experiments to nd the effect of cooling
rate on DAS and SDAS, and also e . Besides the effect of these
parameters on mechanical property (HB) is evaluated.
The effect of high cooling rates in producing ne structures
results in development of high-strength cast alloys. The undercooling of a melt to a lower temperature increases the number
of effective nuclei relative to the growth rate, the latter being
restricted by the rate at which the latent heat of crystallization can
be dissipated [8]. The rening inuence of an enhanced cooling rate
applies both to primary grain size and to substructure. Thus, there
is a marked effect upon DAS, SDAS and e over a wide range of
cooling rates.
The eutectic solidication initiates at certain locations and
continues by radial growth with the simultaneous separation of
graphite and austenite from the melt. Furthermore, the lower the
temperature of formation, i.e. the greater the amount of undercooling the ner is the graphite formed. Sun and Loper [9] showed that
intensive supercooling both constitutional and thermal lead to
the development of cellular dendritic growth of the graphite into a
form recognized as exploded graphite. The microstructures of gray
iron and ductile irons are determined by cooling rate, composition,
nucleation and growth conditions existing during solidication and

584

M.M. Jabbari Behnam et al. / Materials Science and Engineering A 528 (2010) 583588
Table 1
Nominal composition of melt used in this work.
Element

Si

Mn

Cu

Amount (wt%)

3.43.6

1.21.3

0.130.15

0.101

0.64

0.48

equation:
HB = ecp(5.01 + %pearlite)
Simple equations were suggested by Svensson et al. [16] to correlate harness of iron to Si content (between 1.7 and 4.9). These
equations are written below:
Si
Si
HB = HB
f + HBpe
(1 f )
Si
= 54 + 37Si
HB
Si
HBpe
= 167 + 31Si

Fig. 1. Plot of the logarithm of the diffusion coefcient versus the reciprocal of
absolute temperature for several metals [12].

the transformation deeds of austenite. DAS in alloys that solidied

dendritically have been investigated by Zhang et al. [10]. The results
of these investigations are in general agreement. Pande et al. [11]
concluded from their investigation that SDAS can easily be related
to cooling rate and have the same behavior of DAS in relation with
cooling rate.
As the development of GCI microstructure is in the nucleation
and growth classication, so the temperature and its variation
has the key effect on it. Temperature has a most remarkable
inuence on the diffusion coefcients and rates. For example,
for the self-diffusion of Fe in -Fe, the diffusion coefcient
increases approximately six orders of magnitude (from 3.0 1021
to 1.8 1015 m2 /s) in rising temperature from 500 to 900 C. The
temperature dependence of diffusion coefcients is related to temperature according to [12]:

 Q 
d

D = D0 exp

RT

where D0 is temperature-independent preexponential (m2 /s) and

Qd is the activation energy for diffusion (J/mol). This phenomenon
is illustrated in Fig. 1.
So as the undercooling decreases (by decrease of cooling rate)
diffusion rate increases and the carbon atoms can diffuse more
extensively by the time. This results in coarse grains (akes).
Mechanical properties of cast in room temperature are more
dependent to solidication microstructure and related phases in
matrix, although different parameters such as chemical composition, size, shape and distribution of graphite, pearlite/ferrite ratio in
matrix, have signicant effects [13]. Venugopalan and Alagarsamy
[14] evaluated the effect of alloy elements and conducted an equation for calculating mechanical properties (TS1 and YS2 ) versus
phase fraction. Yu and Loper [15] investigated the effect of pearlite
amount on Hardness (Brinell) in ductile iron and come to the below

1
2

Tensile strength.
Yield stress.

Goodrich and Shaws [17] investigations results were an equation which calculate the ultimate tensile strength. Their formula
includes not only composition of elements, but also cooling condition in the form of cast bar radius. Yang et al. [18] were the rst
ones who included the cooling rate of alloy in the formulation of
mechanical property (Brinell hardness). This cooling rate (R) was
in the special temperature (900 C). Their formula was summarized
below:
HB = A + f (%alloy elements) + f (R900 )
2. Experimental procedure
Gray iron specimens with the composition shown in Table 1
were cast at 1420 C. To obtain different cooling rates, a step form
sample designed which is illustrated in Fig. 2a. To calculate real
cooling rates, ve thermocouples embedded in the middle of spatial geometry of each step. These K-type thermocouples connected
to data acquisition system (DAQ) in personal computer by A/D3
transformer (schematically showed in Fig. 2b). The body is molded
in CO2 mold box. After casting, the bulk cut into metallographic
samples. These samples are cut in places which thermocouples are
embedded. Fig. 2a and b show the location of samples in body.
All these samples then are analyzed by commercial image analyze
software (aquinto) to obtain microstructure parameters, and also
hardness Brinell test was done. Details of each step are summarized
in Table 2.
Cooling curves obtained from thermocouples are shown in Fig. 3.
Related cooling rates are summarized in Table 3
for each step. It is clear that by increase of module, the cooling rate decreases. Cooling rates are calculated just a little below
the eutectic transformation temperature and in the constant range
(11001000 C).
3. Results and discussion
3.1. Dendrite characteristics (DAS and SDAS)
Effect of cooling rate on DAS and SDAS are illustrated in Fig. 4a
and b respectively. As the cooling rate increases, both DAS and SDAS
decrease. Rapid cooling produces ne dendrites, while slow cooling results in large and coarse dendrites. Thus solidication over a
temperature range is the primary requirement for dendrite growth.

Analog to digital.

M.M. Jabbari Behnam et al. / Materials Science and Engineering A 528 (2010) 583588

585

Fig. 2. (a) Designed sample for getting different cooling rates, (b) schematic illustration of A/D conversion and data acquisition system (DAQ).

DAS = (78.48)R0.61

1300

1100

Temperature (.c)

Primary austenite dendrites readily grow from the liquidus below

the eutectic temperature. Growth of dendrites may also continue
alongside with the eutectic as the temperature decreases through
the eutectic range to the solidus. Therefore, undercooling in thick
section may lead to longer dendrites and higher interaction. DAS
and SDAS are shown in Fig. 5 for S4.
The correlation between DAS and SDAS whit cooling rate (R) is
written below:

900

700

SDAS = 2.47 Ln(R) + 93.30

3.2. Ferritecementite layers

500
0

The austenite-FG eutectic solidies with the formation of eutectic cells that are more or less spherical in shape. It is generally
thought that each eutectic cell is the product of a nucleation
event. The eutectic cell is made of interconnected graphite plates
surrounded by austenite. The degree of ramication of graphite
within the cell depends on undercooling, Higher undercooling
results in more graphite branching. The leading phase during the
eutectic growth is the graphite. Graphite spacing is determined
by the same parameters as for regular eutectics, with branching
occurring as a response to interface instability. In turn, interface
instability is determined by localized changes in composition, convection currents, crystallographic orientation different from the
heat extraction direction, and a change in temperature gradient.
Phase changes that occur upon passing from the region into the
+ Fe3 C phase eld are relatively complex and similar to those for
the eutectic systems. The microstructure for eutectoid that is cooled
through the eutectoid temperature consists of alternating layers or
lamellae of the two phases ( and Fe3 C) that form simultaneously
during the transformation. In this case, the relative layer thickness
is approximately 81. The thick light layers are the ferrite phase,
and the cementite phase appears as thin lamellae most of which
appear dark. Many cementite layers are so thin that adjacent phase
Table 2
Details of each step in designed model.
Step number
1
2
3
4
5

Label
S1
S2
S3
S4
S5

100

200

300

me (s)
Fig. 3. Cooling curves of cast body obtained from thermocouples.

Table 3
Calculated cooling rates for each step.
Step

S1

S2

S3

S4

S5

Cooling rate ( C/s)

1.90

3.14

8.87

13.53

17.67

boundaries are indistinguishable, which layers appear dark at high

magnication. Mechanically, pearlite has properties intermediate
between the soft, ductile ferrite and the hard, brittle cementite.
Effect of cooling rate on eutectoid layer thickness is shown in
Fig. 6. As it seen from the diagram, increase of cooling rate redounds
in decrease of thickness in eutectoid. Results are shown for four
points, since in the S5 section there was no good detectable metallographic picture for eutectoid. Behavior of eutectoid thickness on
the basis of cooling rate formulated as below:
e = 0.863R0.16
Fig. 7 shows the metallographic pictures of image analysis for
calculating the thickness of ferritecementite layers.
3.3. Hardness evaluations

Geometric
module (cm)

Dimensions
(L mm W mm H mm)

1.5
1
0.75
0.5
0.25

150 83.5 55
150 52.5 24.2
150 51.5 15.8
150 51.5 10
150 48.5 5

Hardness is a measure of a materials resistance to localized

plastic deformation (e.g., a small dent or a scratch). Hardness tests
are performed more frequently than any other mechanical test for
several reasons; they are simple and inexpensive (ordinarily no
special specimen need be prepared, and the testing apparatus is
relatively inexpensive), the test is nondestructive (the specimen is

586

M.M. Jabbari Behnam et al. / Materials Science and Engineering A 528 (2010) 583588

60

a
SDAS(m)

50

DAS (m)

40
30
20

6
5
4
3

10

2
0

10

15

20

R (.c/s)

10

15

20

R (.c/s)

Fig. 4. Effect of cooling rate on (a) dendrite arm spacing (DAS) and (b) secondary dendrite arm spacing (SDAS).

0.9

e (m)

0.8

0.7

0.6

0.5
0

12

16

R (.c/s)
Fig. 5. Metallographic picture of step 4 showing DAS and SDAS.

Fig. 6. Effect of cooling rate on thickness of ferritecementite layers.

neither fractured nor excessively deformed; a small indentation is

the only deformation) and other mechanical properties often may
be estimated from hardness data, such as tensile strength.
Fig. 8 shows the effect of cooling rate on hardness of the samples. The cooling rate is highest at the S5 step. This results in the
formation of white iron, a mixture of iron carbide and pearlite that
is considerably harder than gray iron. When the cooling rate has
decreased sufciently to allow the formation of some graphite, a
mottled zone appears. The mottled zone, which is a mixture of gray
and white iron, has a lower hardness than the white iron tip. As
the section size increases, the white iron gradually disappears, and
there is a corresponding drop in hardness. As the white iron disappears, the microstructure becomes a mixture of ferrite and type D
graphite, resulting in the low hardness. A further decrease in the
cooling rate (8.513.5 C/s) results in low increase in hardness
(constant to somehow) as the microstructure shifts from type D to

type A graphite and the matrix converts from ferrite to pearlite.

As the cooling rate decreases, the hardness is reduced because of
a gradual conversion of the pearlite to ferrite and the formation
of a coarser graphite structure. Correlation between hardness and
cooling rate is as below:
HB = 170.9 0.067R + 0.147R2
Effect of DAS, SDAS and thickness of ferritecementite layers on
hardness is laid in Figs. 9 and 10 respectively. Increase of DAS, SDAS
and e results in decrease of hardness. When all these parameters
decrease, in fact sample faces with ne structure.
Kumruoglu [5] evaluated the mechanical and microstructure
properties of chilled cast iron camshaft. To investigate the hardness
values obtained in this study to apply them to industrial elements
and limit the cooling range, cooling rates and related hardness values from the Kumruoglus work were obtained using point value

Fig. 7. Metallographic pictures for (a) S2 with R = 3.14 C/s, and (b) S5 with R = 1.90 C/s.

M.M. Jabbari Behnam et al. / Materials Science and Engineering A 528 (2010) 583588

186

220

182

200

Hardness (HB)

Hardness (HB)

587

180

160

178
174
170

10

15

20

R (.c/s)
166
Fig. 8. Effect of cooling rate on hardness.

DAS

Hardness (HB)

220

0.5

0.6

0.7

0.8

0.9

e (m)

SDAS

Fig. 10. Relation between hardness with thickness of eutectoid layers.

200

Cooling rate ranges

4.3

27.4

180
1.9

from literature

17.67

this paper

160
0

10

20

30

40

50

60

DAS, SDAS(m)
Fig. 9. Relation between hardness with DAS and SDAS.

measuring by GetData Graph Digitizer software. Fig. 11a and b illustrates the cooling curve and hardness value of lobe in camshaft.
Note that the hardness values are in Rockwell scales, which they
were converted to Brinell scale using ASTM E140-97.
To limit our proposed formulation in industrial use we compare our data with the Kumruoglus work. The cooling rate ranges
comparison is shown in Fig. 12.

10

15

20

25

30

R (.c/s)
Fig. 12. Comparison of cooling rate ranges between industrial results and proposed
model.

From Fig. 12 we come to this point that the industrial range for
our proposed formula, is between 4.3 and 17.67 C/s. In this range,
we compare the hardness value between those two data, which is
shown in Fig. 13.

Fig. 11. (a) Cooling curve and (b) hardness values of lobe in camshaft which was measured by Kumruoglu [5].

588

M.M. Jabbari Behnam et al. / Materials Science and Engineering A 528 (2010) 583588

proposed formula

3. Thickness of ferritecementite layers in the gray cast iron

increased as the cooling rate increased

from literature

220

Hardness (HB)

e = 0.863 R0.16
200

4. The hardness Brinell (HB) decreases as DAS, SDAS and e

increase. Also it shows the below behavior whit cooling rate:

180

HB = 170.9 0.067R + 0.147R2

160

Acknowledgment
0

10

15

20

R (.c/s)
Fig. 13. Comparison of hardness value between proposed formula and literature in
the subscription cooling rate range.

This work has been nancial supported by the Razi Metallurgical Research Center (RMRC) for advanced manufacturing and data
acquisition system.
References

4. Conclusion
Evaluation of gray cast iron with different cooling condition
(rate) shows that the cooling rate has remarkable effect on DAS,
SDAS, e and hardness Brinell (HB). It is found that HB decreases as
DAS, SDAS and e increase. The HB however, decreases as the cooling rate decreases, screening a direct relationship. But this behavior
becomes reverse in a special cooling rate range (8.513.5 C/s)
because of change in graphite type in the structure. All these results
are clear and a lot of piece of data are in different literature, but their
behavior was formulated. To limit cooling rate ranges for industrial
uses, proposed formula was compared with experimental data in
literature for camshaft. The valid range is 4.317.67 C/s and the
hardness value in this range for proposed formula is in good agreement with experimental data in literature. So these formulas can be
used in industrial scales. In addition, the presented experimental
method can be used for other commercial alloys. The formulations
were drawn as follow.
1. Dendrite arm spacing (DAS) in the gray cast iron decreases as the
cooling rate increased.
DAS = (78.48)R0.61
2. Secondary dendrite arm spacing (SDAS) in the gray cast iron
decreases as the cooling rate increased:
SDAS = 2.47 Ln(R) + 93.30

[1] D.M. Stefanescu, Mater. Sci. Eng. A A413414 (2005) 322333.

[2] L. Collini, G. Nicoletto a, R. Konecna, Mater. Sci. Eng. A 488 (2008)
529539.
[3] J.R. Davis, ASM Metals Hand Book, vol. 9, ASM International, 1985, pp.
127135.
[4] K. Mills, et al., ASM Metals Hand Book, vol. 9, ASM International, 1985, pp.
219228.
[5] L.C. Kumruoglu, Mater. Des. 30 (2009) 927938.
[6] M.M. Haque, J Mater. Process. Technol. 191 (2007) 360363.
[7] F.R. Salazar, M. Herrera-Trejo, M. Castro, N.J. Mndez, T.J. Torres, N.M. Mndez,
J Mater. Eng. Perform. 8 (1999) 325329.
[8] K.W.H. Seah, J. Hemanth, S.C. Sharma, Mater. Des. (1995) 175179.
[9] G.X. Sun, C.R. Loper, Trans. Am. Foundry Soc. (1983) 91841.
[10] L.Y. Zhang, Y.H. Jiang, Z. Ma, S.F. Shan, Y.Z. Jia, C.Z. Fan, W.K. Wang, J. Mater.
Process. Technol. 207 (2008) 107111.
[11] K.S. Pande, S.K. Pande, P. Premkumar, Indian Foundry J. (1979) 2540.
[12] W.D. Callister, Fundamentals of Materials Science and Engineering, Department of Metallurgical Engineering, the University of Utah, 5th edition,
2001, 126146.
[13] X. Guo, D.M. Stefanescu, L. Chuzhoy, M.A. Pershing, G.L. Biltgen, AFS Trans. 97
(1989) 214223.
[14] D. Alagarsamy, A. Venugopalan, AFS Trans. 98 (1990) 395400.
[15] C.R. Yu, S.K. Loper, AFS Trans. 96 (1988) 811822.
[16] I.L. Svensson, M. Wessen, A. Gonzalez, in: V. Voller, L. Katgerman,
T.S. Piowonka (Eds.), Modeling of Structure and Hardness in Nodular
Cast Iron Castings at Different Silicon Content, Proceeding of Modeling
and Casting; Welding and Advanced Solidication Process VI, 1993, pp.
2936.
[17] G.M. Shaw, W.F. Goodrich, AFS Gray Iron Research Committee 5-H, Cast Facts.
[18] Y. Yang, A. Louvo, T. Rantala, The effect of alloying and cooling rate on the
microstructure and mechanical properties of low-alloy gray iron, 57th World
Foundry Congress, September 2328, 1990.