Chilling Tendency and Chill of Cast Iron

TSINGHUA SCIENCE AND TECHNOLOGY
ISSN 1007-0214 11/20 pp177-183
Volume 13, Number 2, April 2008
Chilling Tendency and Chill of Cast Iron
E. Fraś ** , M. Górny, W. Kapturkiewicz, H. López †
AGH University of Science and Technology, Reymonta 23, 30-059 Cracow, Poland;
† University of Wisconsin-Milwaukee, P.O. Box 784, Milwaukee, WI 53201, USA
Abstract: An analytical expression is presented for the susceptibility of liquid cast iron to solidify according to
the Fe-C-X metastable system (also known as the chilling tendency of cast iron, CT). The analysis incorpo-
rates the nucleation and growth processes associated with the eutectic transformation. The CT is related to
the physicochemical state of the liquid, the eutectic cells in the flake graphite, and the number of nodules in
nodular cast iron. In particular, the CT can be related to the critical wall thickness, scr, or the chill width, Wcr,
in wedge shaped castings. Finally, this work serves as a guide for understanding the effect of technical fac-
tors such as the melt chemistry, the spheroidizing and inoculation practice, and the holding time and tem-
perature on the resultant CT and chill of the cast iron. Theoretical calculations of scr and Wcr compare well
with experimental data for flake graphite and nodular cast iron.
Key words: chill; chilling tendency; gray cast iron; nodular cast iron
Introduction
The susceptibility of liquid cast iron to solidify according
to the Fe-C-X metastable system (i.e. the chilling
tendency,CT) of cast iron dictates its subsequent performance
in many applications. In particular, cast irons
possessing a high chilling tendency are prone to develop
zones of white or mottled iron. Considering that
these regions can be extremely hard, their machinability
can be severely impaired. Alternatively, if white
iron is the desired structure, a relatively small chilling
tendency f1avours the formation of grey iron which in
turn leads to poor hardness and wear properties for the
cast components. Hence, considerable effort has been
made to correlate the various factors affecting the chill
of cast iron such as chemical composition [1-4] , pouring
temperature [2] , spheroidization and inoculation treatment
[2,3,5,6] , casting geometry [7] , plate thickness [2,3,7] ,
mold material [8] , and nodule count [1] .
**
Received: 2007-06-06
To whom correspondence should be addressed.
E-mail: edfras@agh.edu.pl
Furthermore, some works [4,5] have given qualitative
descriptions of the influence of the chemical composition
and spheroidization and inoculation practice on
the CT. These experimental relationships are very useful
but are limited in their physical meaning. Accordingly,
this work presents analytical expressions that
explain the chill formation mechanism. For the sake of
simplicity, this paper focuses on phenomena and results
which still hold when one ignores the segregation
effects of alloying elements, such as silicon, manganese,
including the so-called inverse chill.
1 Analysis
A theoretical analysis of the solidification of cast
iron [9,10] showed that the critical wall thickness, scr,
(below which, the chill is formed) and the width of the
total chill, W (ASTM A367-55T Standard) can be
given by
s = 2pCT (1)
cr
4np
W = CT
(2)
cos /2
( α )

178
where for flake graphite cast iron,
or
5 3
s
e
3
φ
2
ef
1/6
a ⎛ 2 T ⎞
p = ⎜ ⎟ (3)
π ⎝L c ⎠
⎡ 1
⎛ b ⎞⎤
CT = ⎢ exp
3 8 ⎜ ⎟⎥
⎢Ns( 1 fγ) µ T ∆T
⎣
− ∆ sc ⎝ sc ⎠⎦⎥
⎡ 1
CT = ⎢
⎢⎣
N ( 1 − f ) µ ∆T
and for ductile cast iron,
or
⎛
p = a ⎜
⎝4
π
3 8
cr γ sc
T
3
s
5 2 2 2 4
β B Lez c
⎤
⎥
⎥⎦
⎞
⎟
⎠
1/6
1/6
1 ⎡ 1 ⎛ b ⎞⎤
CT = exp
1/2 ⎢ 2 ⎜ ⎟⎥
D ⎢⎣Ns β ∆Tsc ⎝∆Tsc ⎠⎥⎦
1/3
1/3
1/6
(4)
(5)
(6)
(7)
1 ⎡ 1 ⎤
CT = 1/2 ⎢
2
D ⎣Ncr β ∆Tsc⎦
⎥ (8)
∆ Tsc = Ts− Tc
(9)
φ = cef B1+ cB2
(10)
Tsinghua Science and Technology, April 2008, 13(2): 177-183
T T T
B= ln , B = ln , B2=
ln
T T
i l
1
1 s
L
γ
cef = c+ Tlγ−Ts i
T 1
(11)
(12)
z=0.41+0.93B (13)
In these equations CT is the chilling tendency of the
cast iron; Ns and b are the nucleation coefficients for
flake graphite or nodular cast iron; fγ is the proeutectic
austenite volume fraction; a is the material mould abil-
ity to absorb heat; cef is the effective specific heat of
the proeutectic austenite; Ti is the initial metal tem-
perature just after filling the mould; φ is the heat transfer
coefficient; α is the wedge angle; ∆ T = T −T
and
sc s c
c, D, Le
, Lγ , Tl , Tlγ , Ts , Tc , µ , and β are defined in
Table 1; Ncr is the critical cells or nodule count at tem-
perature T≈Tc; and n is the wedge size coefficient
which expresses the influence of wedge size on the
chill width. The ProductLog [y]=x is the Lambert func-
tion, also known as the omega function which can be
easily calculated by means of the instruction Product-
Log [y] in Mathematica.
Table 1 Selected thermophysical data
Parameter Value
Latent heat of graphite eutectic L e=2028.8 J/cm 3
Latent heat of austenite L γ=1904.4 J/cm 3
Specific heat of cast iron c=5.95 J/(cm 3 · ℃ )
9.2 6.3Si 10 −
Growth coefficient of graphite eutectic µ= ( )
0.25 6
− × cm/(℃ 2 ·s)
Material mould ability to absorb heat a=0.10 J/(cm 2 ·s 1/2 · ℃ )
The diffusion coefficient of carbon in austenite D = 3.9 10 −6 cm 2 /s
Coefficient related with the slopes of the solubility lines JE’, E’S’, and BC’
in Fe-C system
β = 0.001 55℃ –1
Liquidus temperature for austenite T l=[1636 –113(C+0.25Si+0.5P)]℃
Formation temperature for cementite eutectic Tc=[1130.56+4.06(C–3.33Si–12.58P)]℃
Graphite eutectic equilibrium temperature T s=[1154.0+5.25Si–14.88P]℃
Carbon content in graphite eutectic C e=(4.26–0.30Si–0.36P)%
Maximum carbon content in austenite at Ts
Liquidus temperature of austenite for austenite composition C γ
C γ=(2.08–0.11Si–0.35P)%
Weight fraction of austenite g γ=(C e–C)/(C e–C γ)
Austenite density ρ γ=7.51 g/cm 3
Melt density ρ m=7.1 g/cm 3
T lγ=[1636–113(2.08+0.15Si+0.14P)]℃
Volume fraction of proeutectic austenite f γ=ρ m g γ/[ρ γ+g γ(ρ m–ρ γ)]
Note: C, Si, and P indicate content of carbon, silicon, and phosphorus in the cast iron, %.

E. Fraś et al:Chilling Tendency and Chill of Cast Iron 179
The theoretical predictions agree qualitatively with
the available data in the literature. In particular, it is
well known that the chill of cast iron decreases with
increasing eutectic cell or nodule count N (and in consequence
Ncr) as a result of inoculation, as well as decreased
heating times and decreased bath superheating
temperature [11] . It is also known that increasing C and
Si in the cast iron reduces the chill [11] . An increase in
the carbon (1) increases the eutectic cell count in
uninoculated cast iron or the number of nodules
(Ncr) [11] , (2) decreases the proeutectic austenite fraction
fγ (Table 1), (3) slightly narrows the ∆Tsc range (Eq. (9)
and Table 1) and (4) decreases the austenite liquidus
temperature Tl. Points 1 and 2 are the main reason for
the reduction in the chill by carbon (Eqs. (1) and (2)).
An increase in the silicon (1) reduces the growth coefficient
for graphite eutectic µ (Table 1), (2) increases
the cell or the number of nodules (Ncr) [11] , (3) decreases
the proeutectic austenite fraction fγ (Table 1), (4)
widens the ∆Tsc range (Eq. (9) and Table 1), and (5)
decreases the austenite liquidus temperature Tl. The effects
of increases in the number of cells or nodules and
of the proeutectic austenite fraction are the most important
and as a result silicon reduces the chill (Eqs. (1)
and (2)).
In addition, the critical plate wall thicknesses, scr,
and the wedge chill widths, W, according to Eqs. (1),
(2), (3), and (6) increase as the ability of the mold to
absorb heat, a, increases. scr and W also depend on B,
B1, and B2 (Eq. (11)) and, hence, on the initial temperature,
Ti, of the cast iron just after pouring into the
mould. Higher pouring temperatures, Tp, will result in
higher Ti. Therefore, decreasing Tp reduces φ parameter
(Eq. (10)), thus increasing scr and W (Eqs. (1) and (2)).
A schematic diagram showing the role of the various
factors on the chilling tendency and the chill of cast
iron is given in Fig. 1.
Fig. 1 Schematic representation of the effect of various factors on the chilling tendency

180
2 Experimental Procedure
2.1 Flake graphite cast iron
Experimental melts were made in a 15-kg capacity
crucible of an electric induction furnace. The raw materials
were pig iron, steel scrap, commercially pure
silicon, and Fe-P and Fe-S alloys. Melting was followed
by liquid iron superheating up to 1420℃ and
inoculation using FOUNDRYSIL with a 0.2-0.5-cm
granulation and added as 0.5 % of the total charge
weight. After various time intervals (1.5, 5, 10, 15, 20,
and 25 min) from the instant of inoculation, the cast
iron was poured into plate shaped molds of s=0.6, 1.0,
1.6, 2.2, and 3.0 cm in thickness. The average chemical
composition of cast iron was 3.18% C, 1.91% Si,
0.13% Mn, 0.092% P, and 0.064% S. In all cases, the
plates had a common gating system. The foundry
molds were prepared using conventional molding sand.
In addition, they were instrumented with Pt/PtRh10
thermocouples enclosed in quartz sleeves. The thermocouple
tips were located in the geometric center of
each mold cavity. An Agilent 34970A electronic module
was used to record the cooling curves which were
used to determine the initial metal temperature Ti just
after the mold filling and the maximum undercooling,
2.2 Ductile cast iron
Tsinghua Science and Technology, April 2008, 13(2): 177-183
∆Tm, at the onset of eutectic solidification. After cooling,
specimens were taken for metallographic examination
from the geometric centers of the plates. Metallographic
examinations were made on polished and
etched (stead reagent) specimens to show the graphite
eutectic cell boundaries. The planar microstructure was
characterized by the average number of eutectic cells
NF per unit area (cell count) [12]
N
N + 0.5 N
F
+ 1
i w
F = (14)
where Ni is the number of eutectic cells inside a rectangle
S, Nw is the number of eutectic cells that intersect
the sides of S but not their corners and F is the surface
area of S. The average number of eutectic cells N
per unit volume (volumetric cell count) was given
by [13]
N ≅ 0.568 N
(15)
3/2
F
Wedges with Bw=1.25 cm and α=28.5 o and samples
for chemical composition measurement were also cast
simultaneously with the plates. Figure 2 shows typical
planar microstructures of the wedges. The width, W, of
the total chill and the critical cell count, Ncr, were
measured at the junction of the gray structure with the
first spot of cementite as shown in Fig. 2.
Fig. 2 Exhibited microstructure in wedge-shaped of flake graphite castings
The test melts were made in an electric induction furnace
having 8000 kg capacity. The raw materials were
iron scrap, steel scrap, and commercially pure silicon.
After melting and preheating at 1485 ℃ , the cast iron
was poured into a casting ladle where it was spheroidized
using the cored wired injection method. Different
inoculants in various amounts were used. The aim
of using different inoculants and inoculation processes
was to induce different maximum undercoolings, ∆Tm,
and various nodule counts N. The average chemical
composition of the nodular iron was 3.69% C, 2.63%
Si, 0.42% Mn, 0.02% P, 0.02% S, and 0.04% Mg. The
nodular cast iron was poured into the same molds as
for the gray cast iron. The cooling curves and metallographic
structure were examined in the same way as
for the gray cast iron. In the nodular cast iron the

E. Fraś et al:Chilling Tendency and Chill of Cast Iron 181
graphite nodules were characterized by Raleigh distributions
[14] so the volumetric nodule count, N, could
be related to the planar nodule count, NnF, using the
Wiencek equation [15]
N
3
NF
= (16)
f
where fgr is the volume fraction of graphite at room
temperature, fgr≈0.11-0.14.
3 Results and Discussion
3.1 Flake graphite cast iron
The experimentally determined ∆Tm and N were used
with [16,17]
Case
No.
td
s
gr
⎛ b ⎞
N = N exp −
⎜
⎝ ∆Tm
⎠ ⎟
∆Tsc/℃
(17)
Table 2 Experimental and calculated results
Nucleation coefficients
Ns/cm −3
I/1 — 35.2 1.6×10 5
I/2 0.06 51.1 6.1×10 6
I/3 0.20 52.3 6.8×10 6
I/4 0.40 52.3 4.4×10 6
I/5 0.60 51.5 5.4×10 6
I/6 0.80 53.0 1.3×10 6
I/7 1.00 52.9 8.4×10 5
to determine the nucleation coefficients Ns and b by
statistical methods (shown in Table 2). In all cases, the
correlation coefficients were high (0.86-0.99) so the re-
lationship between N and ∆Tm described by Eq. (17) is
in good agreement with the experimental evidence.
The nucleation coefficients Ns and b can be related to a
dimensionless time t = t/ t , where t is the time calcu-
d
r
lated from the instant in which the inoculant was intro-
duced into the melt and tr
is the time when the observed
changes in the cell count were negligible (tr=25 min).
b/℃
Note: Mean temperature (just after pouring) of wedge Ti ≈ 1270℃, n=0.877
The chilling tendency, CT, and the total width of the
chill, W, can be estimated from Eqs. (2) and (4) as
functions of the chemical composition of the cast iron,
the nucleation coefficients, the mean initial temperature
of wedge, Ti, the wedge size coefficients n in the
note of Table 2, and thermophysical data in Table 1.
The results in Table 2 show that the theoretical predictions
agree well with the experimental results.
3.2 Ductile cast iron
The experimental data for ductile iron ∆Tm and N can
be used with Eq. (17) to calcutate the melt nucleation
coefficients, Ns and b, listed in Table 3. The correlation
coefficients for all the melts are quite high (0.84-0.98).
It is not surprising that the nucleation coefficients Ns
and b for each melt differ. However, in each case, N
( )
−3
N = 6.5 −0.8 t − 5.3 t × 10 cm ,
2 6
s d d
b= (96.9 + 122.6 t − 59.2 t ) ℃ (18)
Measured
total width of
chill (mm)
d
Calculated
total width of
chill (mm)
2
d
CT/( s 1/2 ·℃ –1/3 )
76.8 7.9 8.8 1.10
104 3.3 3.6 0.43
119 3.6 3.6 0.44
135 4.0 4.1 0.50
154 4.5 5.4 0.65
152 4.8 5.2 0.63
162 5.5 5.8 0.71
and ∆Tm are related as in Eq. (17), thus confirming that
the theory is in good agreement with the experimental
results.
Table 3 Nucleation coefficients, temperature ranges, ∆T sc,
and chilling tendencies, CT, for graphite
Melt
Nucleation coefficients
No. Ns/cm −3
1 4.13×10 7
2 2.95×10 8
3 5.69×10 7
4 4.01×10 7
5 4.24×10 7
6 5.16×10 7
7 4.85×10 7
b/℃
Note: CT is calculated by Eq. (7).
∆Tsc/ ℃ CT/(s 1/2 ·℃ –1/3 )
58 59.7 1.14
100 51.9 0.90
60 56.9 1.09
42 58.9 1.07
20 61.2 0.90
20 60.9 0.84
43 59.2 1.00

182
The chilling tendency CT can be calculated from Eq.
(7) as a function of the chemical composition of the
cast iron, Ns and b, listed in Table 3, the mean temperature
of the metal just after pouring, Ti≈1260℃, and any
relevant thermophysical data (shown in Table 1).
Table 4 shows results for plate-shape castings reported
in the literature [4] with various chemical compositions
and wall thicknesses, as well as the number of
nodules and the cementite fraction. In these cases, the
nucleation coefficients b and Ns are not known, so CT
is calulated using Eq. (8). The results in Table 4 show
that in melts I and II, chill starts to occur in walls with
thicknesses between 3 and 6 mm, while in melt III chill
starts to occur for wall thicknesses between 1.5 and 2
mm. scr must be calculated using Eq. (1) to compare
these results with the theoretical predictions. These
calculations assumed that a=0.11 J/(cm 2 ·s 1/2 · ℃ )
and Ti=1250℃ with other information taken from Table
1. The results in Table 4 show that in melt I the
Tsinghua Science and Technology, April 2008, 13(2): 177-183
change of the wall thickness, s, from 6 mm to 3 mm is
closely linked to the change in the number of nodules
from 588 to 1039 mm −2 . As a result, an average nodule
count of NF,cr=813 mm −2 was used in this work. Similar
determinations were made for melts II and III with NF,cr
values of 945 and 1959 mm −2 , respectively. scr for
melts I, II, and III were then 3.4 mm, 3.1 mm, and
1.8 mm as shown in Table 4. In addition, comparison
of s and scr in Table 4 indicates that the predictions
from the theoretical analysis are in good agreement
with the experimental determinations of the wall thickness
at which chill occurs. Table 4 shows the CT for
the three melts calculated using Eq. (8) and data in Table
1. The results in Table 4 show that as the chilling
tendency increases from 0.34 to 0.68 s 1/2 ·℃ –1/3 , the
critical wall thickness, scr, increases from 1.8 to
3.4 mm.
Table 4 Chemical composition, wall thicknesses, nodule count, cementite fraction, and chilling tendency for three
ductile cast iron melts
Melt
No.
C/% Si/% P/%
I 3.40 2.70 0.046
II 3.45 2.91 0.044
III 3.31 4.42 0.051
4 Conclusions
Wall thickness (mm)
Experimental s Calculated scr
(1) A simple theoretical analysis is presented which
predicts the chilling tendency and chill in flake graphite
and ductile cast iron based on
The chemical composition of the casting which
for flake graphite cast iron is characterized by ∆Tsc,
fγ, T, µ, Ncr or Ns, and b, and for ductile cast iron is
characterized by ∆Tsc, D, Tl, Ncr or Ns and b;
The inoculation and spheroidization method, the
superheating temperature, and bath holding
Nodule count
(NF/mm −2 )
Fraction of
cementite (%)
CT/( s 1/2 ·℃ –1/3 )
6.0 588 0.0 0.68
3.4
3.0
1039 9.0 —
2.0 1380 24.0 —
1.5
1311 34.0 —
6.0 854 0.0 0.60
3.1
3.0
1037 7.3 —
2.0 1100 24.0 —
6.0 1127 0.0 —
3.0
1726 0.0 —
2.0 1890 0.0 0.34
1.8
1.5
2027 9.3 —
times as characterized by Ncr or Ns and b;
Selected thermophysical data of the metal and
mold material as well as the metal temperature
just after pouring into the mold.
(2) Theoretical calculations of the width of the total
chill in wedges for flake graphite and of the critical
wall thickness for nodular cast iron compare well with
experimental data.
(3) The absolute values of the chilling tendency of
cast iron are calculated to range from 0.3 to
1.1 s 1/2 ·℃ –1/3 .

E. Fraś et al:Chilling Tendency and Chill of Cast Iron 183
References
[1] Javaid A, Thompson J, Davis K G. Critical conditions for
obtaining carbide-free microstructures in thin-wall ductile
irons. In: Transactions of the American Foundry Society
and the One Hundred Sixth Annual Casting Congress.
Kansas City, USA, 2002: 889-898.
[2] Javaid A, Thompson J, Sahoo M, Davis K G. Factors af-
fecting the formation of carbides in thin-wall DI castings.
In: One Hundred Third Annual Meeting of the American
Foundrymen’s Society. Rosemont, USA, 1998: 441-456.
[3] Ruxanda R E, Stefanescu D M, Piwonka T S. Microstruc-
ture characterization of ductile thin wall iron castings. In:
Transactions of the American Foundry Society and the
One Hundred Sixth Annual Casting Congress. Kansas City,
USA, 2002: 1131-1147.
[4] Giacopini A, Boeri R E, Sikora J A. Carbide dissolution in
thin wall ductile iron. Mats. Sci. Tech., 2003, 19(12):
1755-1760.
[5] Labrecque C, Gagne M, Javaid A, et al. Production and
properties of thin-wall ductile iron castings. Int. J. Cast
Metals Research, 2003, 16(1-3): 313-317.
[6] Choi J H, Oh J K, Choi C O, et al. Effect of Bi on forma-
tion of microstructure and mechanical properties of ductile
iron castings with thin-wall section. AFS Transactions,
2004, 112: 831-840.
[7] Mampaey F, Xu Z A. Mold filling and solidification of
thin-wall ductile iron casting. AFS Transactions, 1997,
105: 95-103.
[8] Showman R E, Aufderheide R C. Process for thin-wall
sand castings. AFS Transactions, 2003, 111: 567-578.
[9] Fraś E, Górny M, López H F. The transition from gray to
white cast iron during solidification: Part I. Theoretical
background. Metallurgical and Materilas Transactions A,
2005, 36: 3075-3082.
[10] Fraś E, Górny M, López H F. Eutectic transformation in
ductile cast iron. Part I. Theoretical background. Metal-
lurgy and Foundry Engineering, 2005, 31: 113-136.
[11] Merchant H D. Recent Research on Cast Iron, Gordon and
Breach. New York: Science Publishers, 1968.
[12] Rys J. Stereology of Materials. Cracov: Fotobit, 1995.
[13] Osher J, Lorz U. Quantitative Gefuengenanalysie. DVG:
Leipzig-Stuttgard, 1994.
[14] Wojnar L. Effect of graphite size and distribution on
fracture and fractography of ferritic nodular cast iron. Acta
Stereologica, 1986, 5(2): 319-324.
[15] Wiencek K, Rys J. The estimation of Fe3C particle density
in steel by simple counting measurements made in plane
sections. Materialas Engineering, 1998, (3): 396-399.
[16] Fraś E, Górny M, López H F. Nodule count in ductile iron:
Theoretical model based on Weibull statistics. Interna-
tional Journal of Cast Metals Researches, 2005, 18(3):
156-162.
[17] Fraś E, Górny M, Tartera J. Nucleation and grains density
in grey cast iron—A theoretical model and experimental
verification. International Journal of Cast Research, 2003,
16: 99-104.