INVESTIGATION OF THE DIFFERENCE BETWEEN CARBON EQUIVALENT

FROM CARBON SATURATION DEGREE AND THAT FROM LIQUIDUS

Shigeki Koriyama , Toshitake Kanno, Yuki Iwami, and Ilgoo Kang

Kimura Foundry Co., Ltd, 1157 Nagasawa, Sunto-gun, Shimizu-cho, Shizuoka-Ken 411-0905, Japan

Copyright Ó 2020 American Foundry Society

[Link]

Abstract
Carbon equivalent of cast iron is known as the following [%P] - 71 [%S] - 2 [%Ni] - 7 [%Cr]. Dividing this
equation: CE = [%C] ? (1/3) [%Si]. This equation is equation by carbon coefficient, the carbon equivalent
basically calculated from the carbon solubility in hyper- equation of liquidus (hereinafter referred to as CEL) from
eutectic composition. On the other hand, liquidus from hypoeutectic to eutectic composition is obtained as follows:
hypoeutectic to eutectic composition in Fe–C phase dia- CEL = [%C] ? 0.23 [%Si] - 0.03 [%Mn] ? 0.32
gram is well known. Liquidus temperature (hereinafter [%P] ? 0.64 [%S] ? 0.02 [%Ni] ? 0.06 [%Cr]. This
referred to as TL) can be measured easily with thermal CEL is different from the commonly used CE since it is
analysis by focusing primary crystallization temperature. determined from liquidus. We investigated how the differ-
In this study, the effects of elements on TL and carbon ence of carbon equivalent (CE/CEL) application in the
equivalent in cast iron were examined from the view point range of hypoeutectic to eutectic composition affected on
of cooling curve, carbon flotation, and internal shrinkage, the cast iron and revealed that CEL = [%C] ? 0.23 [%Si]
and a more accurate equation for calculating carbon should be applicable.
equivalent is suggested. It is revealed by the experiment
that the relationship between TL and elements from Keywords: primary crystallization temperature, CE,
hypoeutectic to eutectic composition is as follows; TL CEL, carbon solubility, liquidus, Fe–C phase diagram
(°C) = 1625 - 110 [%C] - 25 [%Si] ? 3 [%Mn] - 35

Introduction coefficient is the silicon activity coefficient which is con-

cerned with the carbon solubility in hypereutectic compo-
Primary crystallization temperature is measured by thermal sition. Generally, as the element’s solubility is related to its
analysis, and carbon equivalent is calculated from this activity coefficient, the carbon solubility shows a strong
temperature. There is a similar word to the carbon equiv- relationship with 1/3 of silicon activity coefficient in
alent, which is called carbon saturation degree (hereinafter hypereutectic composition. However, the liquidus from
referred to as SC). P.A. Heller and H. Jungbluth suggested hypoeutectic to eutectic shows the temperature at which
the equation for SC by considering the influence of Si and P solidification starts and that gradient is called solidification
on carbon equivalent in 1955.1 F. Neumann and W. Pat- temperature drop. Cast iron is practically made from
terson suggested the following equation for carbon equiv- hypoeutectic to eutectic chemical composition. Therefore,
alent: CE = [%C] ? 0.31 [%Si] ? 0.33 [%P] ? 0.4 it is a contradiction to use activity coefficient in the case of
[%S] - 0.028 [%Mn],2 when investigating the carbon solidification temperature drop.
solubility in hypereutectic composition. W. Patterson also
suggested similar equation for carbon equivalent in 1961.3 On the other hand, there were several studies which
From these references, it can be assumed that 1/3 of silicon investigated the liquidus within the range of hypoeutectic
composition (Figure 1b).4 For example, Dietert suggested
the relationship concerning CEL and TL when investigated
This paper is an invited submission to IJMC selected from presen- the influence of pouring temperature and chemical com-
tations at the 2nd Carl Loper 2019 Cast Iron Symposium held position on flowability.5 L. F. Porter investigated the
September 30 to October 1, 2019, in Bilbao, Spain.

774 International Journal of Metalcasting/Volume 14, Issue 3, 2020

Figure 1. Fe–C diagram which additionally indicates Figure 2. Principle of measuring primary crystallization
carbon solubility.4 temperature TL from cooling curve.

changed the additional amount of target elements with

Table 1. Range of Change in Each Element. (mass %). basic composition of molten metal (3.1%C, 1.7%Si,
0.75%Mn, 0.07%P, and 0.05%S). For adjusting the com-
C Si Mn P S Ni Cr
position, graphite electrode and the following Fe-alloys
1 2.4–4.0 1.7 0.75 0.07 0.05 0 0 were used: Fe-75%Si,Fe-73%Mn,Fe-26%P,Fe-35%S.
2 3.1 0–5.0 0.75 0.07 0.05 0 0 Holding the molten metal at 1430 °C ± 10 °C in the
induction furnace, CE cup samples were taken at
3 3.1 1.7 0–4.0 0.07 0.05 0 0
1400 ± 10 °C. Simultaneously, chill specimens for emis-
4 3.1 1.7 0.75 0.1–1.0 0.05 0 0 sion spectroscopic analysis (SPECTROLAB) were sam-
5 3.1 1.7 0.75 0.07 0–0.22 0 0 pled with 700 g of graphite cup. The CE cup consisted of a
6 3.1 1.7 0.75 0.07 0.05 0–40 0 shell mold of 30 mm diameter and 50 mm in height, and a
7 3.1 1.7 0.75 0.07 0.05 0 0–6 chromel–alumel thermocouple was protected by 0.6 mm
diameter of quarts tube. The weight of the melt poured into
the cup was 250 ± 10 g. By pouring it into the cup until
the melt flowed out, same amount of melt remained in the
cup.
liquidus temperature and derived the experimental equa-
tion.6 Daido steel reported the liquidus calculation equa-
We investigated the various elements content and TL, with
tion as follows (as cited in ISIJ Tekko Benran, 1981): TL
the following methods. Table 1 shows the range of change
(°C) = 1536 - 91 [%C] ? 21 [%Si].7 Dividing this by 91
in each element over the basic composition of 3.1%C,
(carbon coefficient) makes the following equation: CEL =
1.7%Si, 0.75%Mn, 0.07%P, and 0.05%S. With changing
[%C] ? 0.23 [%Si]. As the results of the papers which
only one element content among the basic composition, the
investigated the carbon equivalent from liquidus, the sili-
effect of each element on TL was investigated. As shown in
con coefficient in the carbon equivalent is around 0.22 to
Figure 1, there is almost a linear relationship between
0.25.
carbon content and solidification temperature drop within
hypoeutectic (2.4%C) to eutectic composition. Therefore,
In this study, we investigated the influence of various
we changed the carbon content in the molten metal from
elements on the primary crystallization temperature and the
2.4 to 4.0%, poured it into the CE cup and measured TL.
way of calculating the carbon equivalent. In addition, we
considered the carbon equivalent equation for calculating
Figure 2 shows the principle of measuring TL from the
the accurate eutectic composition from the following
cooling curve.7,8 As shown in Figure 2b, on the case of
viewpoints: (i) cooling curve of thermal analysis, (ii) car-
slow cooling (large cup), the peak of the exothermic curve
bon flotation result, and (iii) internal shrinkage test result.
moves to lower temperature. Therefore, both TL and the
finishing point of primary crystallization temperature TF
Methodology become lower as the cooling rate decreases. In addition,
Wang et al. suggested that more stable thermal analysis
Melting was performed in a 50 kg, 3000 Hz of silica-lined result was obtained by using ATAS cup (NovaCast) than
high-frequency induction furnace. As shown in Table 1, we using our type of thermal cup.9 Although the starting point

International Journal of Metalcasting/Volume 14, Issue 3, 2020 775

Table 2. Chemical Analysis Result of the Test Changing Carbon Content and Primary Crystallization Temperature TL

C% Si % Mn % P% S% Ni % Cr % TL Measured TL Modifieda

2.31 2.07 0.78 0.017 0.017 0.027 0.055 1321 1325.6

2.42 2.09 0.78 0.017 0.017 0.026 0.055 1311 1316.1
2.49 2.08 0.78 0.017 0.017 0.026 0.055 1299 1303.8
2.59 2.11 0.78 0.017 0.016 0.026 0.055 1288 1293.5
2.73 2.08 0.78 0.018 0.016 0.026 0.055 1276 1280.9
2.79 2.06 0.78 0.017 0.015 0.025 0.054 1267 1271.3
2.93 2.06 0.78 0.018 0.016 0.027 0.055 1257 1261.4
3.06 2.05 0.78 0.018 0.016 0.028 0.055 1242 1246.1
3.11 1.62 0.63 0.054 0.038 0.000 0.086 1240 1237.3
3.20 1.63 0.64 0.052 0.038 0.000 0.086 1222 1219.6
3.21 2.05 0.78 0.018 0.016 0.027 0.054 1226 1230.1
3.32 2.04 0.78 0.017 0.015 0.029 0.054 1210 1213.8
3.42 2.03 0.78 0.019 0.018 0.030 0.055 1197 1200.8
3.47 1.61 0.64 0.054 0.038 0.000 0.086 1203 1199.9
3.53 2.01 0.78 0.018 0.016 0.030 0.058 1188 1191.2
3.66 1.59 0.64 0.055 0.038 0.000 0.086 1184 1181.1
3.86 1.57 0.64 0.056 0.037 0.000 0.086 1163 1159.5
a
Effect of each element is modified to base composition (1.7% Si, 0.75% Mn, 0.07% P, and 0.05% S)

Table 3. Chemical Analysis Result of the Test Changing Silicon Content and Primary Crystallization Temperature TL

C% Si % Mn % P% S% Ni % Cr % TL Measured TL Modifieda

3.08 0.01 0.50 0.062 0.031 0.006 – 1285 1282.4

3.06 0.91 0.52 0.064 0.033 0.005 – 1265 1261.2
3.00 1.20 0.51 0.062 0.031 0.006 – 1265 1254.6
2.98 1.42 0.51 0.063 0.032 0.006 – 1257 1244.8
3.11 1.64 0.51 0.065 0.034 0.005 – 1251 1252.3
2.85 1.93 0.51 0.059 0.031 0.005 – 1251 1224.9
2.93 2.18 0.52 0.064 0.036 0.006 – 1245 1228.0
2.92 2.39 0.51 0.062 0.032 0.006 – 1242 1223.6
2.90 2.60 0.52 0.063 0.031 0.006 – 1235 1214.2
2.84 3.28 0.51 0.060 0.032 0.005 – 1226 1199.3
2.82 3.52 0.51 0.066 0.036 0.005 – 1219 1190.6
2.89 3.69 0.51 0.066 0.033 0.005 – 1214 1192.9
2.82 4.26 0.50 0.060 0.033 0.006 – 1208 1179.5
a
Effect of each element is modified to base composition (3.1% C, 0.75% Mn, 0.07% P, and 0.05% S)

of primary crystallization temperature TS is stable and The relationship between the various elements content and
rarely changes, TS was not used as primary crystallization the primary crystallization temperature was investigated by
temperature in this study. The reason is as follows: (a) in multiple regression analysis. Finally, divide the equation of
the foundry field, the inflexion point of cooling curve (TL) TL by carbon coefficient to obtain the accurate carbon
is used as primary crystallization temperature, (b) since it equivalent (CEL) equation within the range of hypoeutectic
becomes an extreme value in the exothermic curve, it is to eutectic composition.
easy to distinguish, and (c) our thermal analysis apparatus
applies TL as primary crystallization temperature.

776 International Journal of Metalcasting/Volume 14, Issue 3, 2020

Figure 3. Relationship between C content and TL. Figure 5. Relationship between Mn content and TL.

Figure 4. Relationship between Si content and TL. Figure 6. Relationship between P content and TL.

Experimental Results and Analysis Then, the raw data of every test result (e.g., the result of
changed C and Si content is shown in Tables 2, and 3),
First, we analyzed the relationship between TL measured other than changed element, were adjusted by using the
by thermal cup and chemical compositions measured by coefficients of Eqn. 4 to the base material composition
spectrometer with multi regression analysis as follows: shown in Table 1. The modified TL shown in Table 2 are
calculated as follows:
ðxi1 ; . . .; xi7 ; yi Þ; i ¼ 1; . . .; n:
y^ ¼ b0 þ b1 x1 þ . . . þ b7 x7 Eqn: 1 TL modified ð C Þ ¼ TL measured 25 ð1:7  ½%SiÞ
X n
þ 3 ð0:75  ½%MnÞ  35 ð0:07  ½%PÞ
RSS ¼ ðyi  y^Þ2 Eqn: 2
i¼1  71 ð0:05  ½%SÞ  2 ð0  ½%NiÞ
oRSS  7 ð0  ½%CrÞ
¼ 0; k ¼ 0; . . .; 7: Eqn: 3
obk Eqn: 5
As a result, the following equation for the relationship The relationship between modified TL and five main
between TL and various elements is obtained: alloying elements: carbon, silicon, manganese,
TL ð C Þ ¼ 1625  110 ½%C  25 ½%Si phosphorus, and sulfur contents, is shown in Figures 3, 4,
þ 3 ½%Mn  35 ½%P  71 ½%S  2 ½%Ni  7 ½%Cr 5, 6 and 7. The purpose of applying modified TL is to make
Eqn: 4 the effect of changing element on TL clear by fixing the

International Journal of Metalcasting/Volume 14, Issue 3, 2020 777

 the MnS formation is not shown. The reason why the
transition resulted from the MnS formation is not shown
and the reason of Mn increasing solidification temperature
are uncertain.

Figure 8 shows the influence of each element on TL. The

order of elements that lower liquidus temperature of iron is
C [ S [ P [ Si [ Cr [ Ni.

Equation 4 shows the influence of various elements on TL,

in the range of hypoeutectic to eutectic composition.
Therefore, CEL which is calculated from the liquidus of the
phase diagram can be shown with dividing the coefficient
of various elements in Eqn. 4 by carbon coefficient of
- 110.

CEL ¼ ½%C þ 0:23 ½%Si  0:03 ½%Mn þ 0:32 ½%P

þ 0:64 ½%S þ 0:02 ½%Ni þ 0:06 ½%Cr
Figure 7. Relationship between S content and TL.
Eqn: 6
Here, ‘‘L’’ means that the carbon equivalent is not
calculated from the carbon solubility in hypereutectic
composition, but it is calculated from liquidus in the range
of hypoeutectic to eutectic composition. When applying
this equation to common gray or ductile iron casting,
CEL = [%C] ? 0.23 [%Si] is enough for use, because the
influence of other terms is small with usual chemical
compositions.

Discussions

There is a problem that (1/3) of silicon coefficient in the

conventional equation of CE and 0.23 from CEL are dif-
ferent. In order to investigate which one is applicable, we
examine it based on our past experimental results. In this
discussion, the following equations are applied for CE and
CEL.
Figure 8. Influence of additional elements on TL to the
base material (3.1% C, 1.7% Si, 0.75% Mn, 0.07% P, and CE ¼ ½%C þ ð1=3Þ ½%Si Eqn: 7
0.05% S).
CEL ¼ ½%C þ 0:23 ½%Si: Eqn: 8
variation of other elements to the base composition shown
in Table 1. The carbon liquidus gradient is -117 K/ % in
the reference of M. Hansen 10 and that is - 104 K/ % in Cooling Curve of Thermal Analysis
the reference of Okamoto.4 In this study, the gradient is
almost middle of the both results (- 110 K/ %). The Figure 9 shows cooling curves from hypoeutectic com-
intercept of this study (1625 °C) is also intermediate position to hypereutectic composition, with changing CE or
between those of M. Hansen (1650 °C) and H. Okamoto CEL. In the case of (c) cooling curve, even though the
(1600 °C) in their phase diagrams. primary crystal point (a sign of hypoeutectic composition)
Figure 5 shows the manganese have an effect rising the appears, CE is 4.41 ([ 4.3), suggesting a hypereutectic
TL. Except for manganese, all the elements lower TL are composition. On the other hand, CEL is 4.2 (\ 4.3) and it
shown in Eqn. 4. According to the report of Kanno,11 in the suggests hypoeutectic composition. That is, CEL is more
molten metal of 0.05%S, Mn bonds S to form MnS until accurate. In the case of cooling curve of (d), it is considered
0.45%Mn. During Mn bonding S to form MnS, it is thought as eutectic composition because the super cooling is large
that the solidification drop does not occur, because both Mn and primary crystallization point is not shown. However,
content and S content reduce at the same time. However, as the calculated CE (= 4.5) suggests hypereutectic compo-
for TL shown in Figure 5, the inflexion point resulted from sition ([ 4.3). On the foundry site, CE of 4.5 is reported as

778 International Journal of Metalcasting/Volume 14, Issue 3, 2020

 Figure 9. Relationship between carbon equivalent (CE = C 1 (1/3) Si, CEL = C 1
0.23 Si) and cooling curve of thermal analysis (size: Dia. 30 3 H50 mm). 12

Figure 10. Relationship between carbon equivalent (CE = C 1 (1/3) Si, CEL = C 1
0.23 Si) and graphite structure.12 (CE cup: Dia. 30 3 H50 mm).

Figure 11. Relationship between carbon equivalent (CE = C 1 (1/3) Si, CEL = C 1
0.23 Si) and shrinkage volume.12

International Journal of Metalcasting/Volume 14, Issue 3, 2020 779

 the test piece was measured by filling water into the
shrinkage cavity in the test piece.

Metallurgically, shrinkage is minimized at eutectic com-

position, and it is minimum at a little less than 3.8 of C as
shown in Figure 11. Therefore, eutectic composition is
considered as 4.49 in the case of CE, and 4.28% in the case
of CEL. It can be said that the eutectic composition is not
4.3 of CE as generally said, but 4.28 of CEL also from the
viewpoint of shrinkage.

Change of Si Coefficient by the Amount of Si

Content

When we produced Hi-Si cast iron, we found that the TL

Figure 12. Relationship between Si amount and Si coef- from thermal analysis and calculation by Eqn. 8 and 9 were
ficient, calculated from this study TL. not matched. Therefore, we recalculated Si coefficient a by
using following equation transformed from Eqn. 8 and 9;
an ideal composition that minimizes shrinkage defects.
TL ð CÞ ¼ 1625  110 ½CEL  Eqn: 9
Here, CEL is 4.28, which is consistent with the eutectic
composition at graphite eutectic temperature (dotted line) 1625  TL  110C
shown in Figure 1. Primary graphite temperature appears a¼ Eqn: 10
110  Si
on the cooling curve of (e). Owing to the primary graphite
formation, super cooling temperature is high and the where TL is the value measured by thermal cup, C was
eutectic solidification temperature is high. Considering the analyzed by CS meter, and Si was analyzed by
result of cooling curve of thermal analysis, it is clear that spectrometer.
the carbon equivalent calculated from CEL is more accu- As shown in Figure 12, Si coefficient is constant as 0.23
rate than that calculated by CE. until 3.65% of Si content, but it increases at the range of
Hi-Si castings. The relationship equation of TL in this study
Carbon Flotation is linear approximation of liquidus curve at the carbon
content of base material composition; deviation may occur
Figure 10 shows the relationship between carbon equivalent if the element content exceeds the approximate range. And
and graphite structure of CE cup. These samples were used also the carbon coefficient is fixed as 110 in this calculation;
for the cooling curves in Figure 9. Even though no carbon therefore the deviation causes the change in Si coefficient.
flotation is observed in the upper side of sample (a), 4.41 of CE
([ 4.3) assigns that it is hypereutectic composition. In sample
(b), carbon flotation is also not observed, and 4.28 of CEL Conclusions
assigns it is around eutectic composition. In sample (c), carbon
flotation is observed in the upper side, and 4.49 of CEL also The influence of various elements on primary crystalliza-
assigns it is hypereutectic composition. Regarding carbon tion temperature and carbon equivalent is investigated, and
flotation, it can be said that carbon equivalent calculated with the following results are obtained:
CEL is more accurate than that calculated with CE. For your
reference, sample (c) in which carbon flotation occurs at the 1. The relationship between various elements content
upper surface, the carbon and silicon content at the bottom are and TL within the range of hypoeutectic to eutectic
3.77% and 2.19%, respectively. That means the structure of composition is as follows: TL (°C) = 1625 – 110
bottom is considered almost eutectic composition when cal- [%C] - 25 [%Si] ? 3 [%Mn] - 35 [%P] - 71
culated with CEL (= 4.27). [%S] - 2 [%Ni] - 7 [%Cr]
2. The coefficient of various elements for carbon
equivalent which are calculated from the liquidus
Shrinkage Test in hypoeutectic to eutectic composition is as
follows: CEL = [%C] ? 0.23[%Si] - 0.03[%
Figure 11 shows the relationship between carbon equiva- Mn] ? 0.32[%P] ? 0.64[%S] ? 0.02[%Ni] ?
lent and the shrinkage volume.12 Carbon content was 0.06 [%Cr]
changed with 2.1% target of Si content. The shape of the 3. Since CE = [%C] ? (1/3) [%Si] is the equation
test piece was a column and its size as follows: diameter calculated from carbon solubility in hypereutectic
200 9 H200 mm (45 kg). The shrinkage volume formed in composition, it is not applicable for normal iron

780 International Journal of Metalcasting/Volume 14, Issue 3, 2020

 casting. On the other hand, CEL = [%C] ? 0.23 5. H.W. Dietert, On gating iron castings. Am. Foundry
[%Si] is calculated by the equation of solidifica- Soc. Trans. 34, 1038–1048 (1926)
tion temperature drop in hypoeutectic to eutectic 6. L.F. Porter, P.C. Rosenthal, factors affecting fluidity
composition. So it coincides with field experience of cast iron. Am. Foundry Soc. Trans. 60, 725–735
such as carbon flotation or shrinkage defect. (1952)
4. Si coefficient of carbon equivalent within the 7. Special steel committee, Toku 33-4-ji 10, Daido steel
range of hypoeutectic to eutectic composition is (1967) (as cited in ISIJ, Tekko Benran, vol. 1, 3rd edn.
assumed to be 0.23. (Maruzen, Tokyo, 1981), p. 205)
8. T. Kanno, J. JFS 76, 62–66 (2004)
5. Si coefficient a is constant as 0.23 until the Si 9. L. Wang, H. Nakae, J. JFS 91, 213–220 (2019)
content is less than 3.65%, but it increases if Si 10. M. Hansen, K. Anderko, Constitution of Binary
content exceeds 3.65%. Alloys, 2nd edn. (McGraw-Hill Book Co, New York,
1958), p. 354
11. T. Kanno, I. Kang, J. JFS 85, 414–419 (2013)
12. T. Kanno, I. Kang, Y. Fukuda, T. Mizuki, S. Kiguchi,
REFERENCES J. JFS 77, 9–17 (2005)

1. P.A. Heller, H. Jungbluth, Giesserei 42, 255 (1955)

2. F. Neumann, H. Schenck, W. Patterson, Giesserei 47, Publisher’s Note Springer Nature remains neutral with
1217–1246 (1959) regard to jurisdictional claims in published maps and
3. W. Patterson, H. Siepman, H. Paycna, Giesserei 48, institutional affiliations.
239 (1961)
4. T.B. Massalski, H. Okamoto, Binary Alloy Phase
Diagrams, vol. 1, 2nd edn. (ASM International, OH,
1990), p. 843

International Journal of Metalcasting/Volume 14, Issue 3, 2020 781