Cooling Rate and Graphite Structure of Final Eutectic

Solidification Part in Cast Iron*

Hirokazu Kawashima1, Katutoshi Sigeno1, Masaki Kusubae1 and Kazuhiro Tachibana2

1Matsubara Co. Ltd., Seki 501-3924, Japan2Graduate School of Science and Engineering, Yamagata University, Yonezawa 992-8510, Japan

A CE meter cup was modeled using a casting simulation software and cooling curves at different points of the casting model werecalculated. In setting temp-solid fraction parameters, an Fe-C-Si ternary equilibrium phase diagram was used with a substantive temperaturerange in the equilibrium phases so that the �-Fe+G+L phases can be considered. In studying the calculated cooling curve, a rapid temperaturedrop was found in the cooling curve in the final solidification period. This rapid cooling occurred during eutectic solidification. The structureformed by the rapid cooling was found as chill or D-type graphite because the liquid phase is cooled rapidly to below the eutectic temperature.This rapid cooling may be one of the causes of D-type graphite near the final solidification part of thick iron castings.[doi:10.2320/matertrans.F-M2011829]

(Received January 25, 2011; Accepted September 7, 2011; Published November 25, 2011)

Keywords: computer aided engineering (CAE), cast iron, eutectic reaction, inverse chill, simulation, solidification

1. Introduction

In iron castings, the chilled iron structure called the inversechill can be formed at deep inner regions of thick castings.Although obvious inverse chills have rarely been seen inmodern castings because of the improvement of the inocu-lation technology, the microstructures that seem to be formedby rapid cooling are often observed at the final solidificationpart of the casting, for example, the D-type graphite.

The various causes of the inverse chill have been suggestedas follows:1)

(1) Carbide stabilizing elements segregate at the center ofthick castings,

(2) The inoculation is insufficient,(3) And/or the cooling rate is particularly fast at the center

of the casting.However, these suggestions are not enough to explainsufficiently the reasons why the inverse chill is formed onlyat the thick parts of the casting.

It is widely known that the rapid temperature drop canbe found at the point just after the end of the eutecticsolidification when primary differentiation is conducted ona cooling curve recorded by a CE meter. However thereasonable cause of the rapid temperature drop has not beenreported.

On the other hand, the conventional cooling curvemeasured with a CE cup does not show the real coolingcurve because the temperature is measured through an outersilica tube (�4 mm, 1 mm-thick) for a thermocouple. There-fore, it is difficult to measure the temperature with the CEcup during the eutectic solidification.

We have simulated the cooling curve in the CE cup bynumerical analysis software, in order to see whether ornot the rapid temperature drop can be found at the finalsolidification part just like the way it appears on theconventional cooling curves with the actual measurement.Also, we have allocated certain points in the simulated CE

cup for cooling curve measurement in order to understandthe possibilities of the formations of the inverse chills andD-type graphite.

2. Experimental Procedures

2.1 Actual measurementThe molten iron was poured from a cupola into the

forehearth, and held for ten minutes. Then, the molten ironwith 0.05% amount of carbon inoculation was transferred to aladle. After 30 s from the inoculating, the molten iron waspoured into a CE cup connected with the digital recorder,and the cooling curves were recorded until the end of thesolidification. The CE cup for this experiment was not thetellurium-added type which is commonly used for measure-ment of the cementite eutectic temperature, but the typewithout any additives in order to record a cooling curve ofthe practical cast iron. The temperatures were recorded atintervals of 1/2 s for 300 s just after the pouring in order toobtain the cooling curve until the post-solidification period.Figure 1 shows the shape of the CE cup and the measuringpoints. Table 1 shows the chemical composition of the castiron. Figure 2 shows the result of the actual measurementand the curve calculated by primary differentiation.

2.2 Setting latent heat emission patternIn order to find the appropriate pattern of the latent heat

emission, the temperatures at the thermocouple in the silicatube just like a real CE cup were calculated. The latentheat emission patterns were determined by the relationshipbetween temperature and solid fraction by using simulationsoftware. The definition of the transformation temperatureswere based on the result of the actual measurement and theparameters were set as follows:(1) The liquidus temperature (TL) was assumed to be

1468 K measured by the actual measurement.(2) The eutectic solidification start temperature (TES) was

assumed to be 1418 K measured as the recalescencetemperature.*This Paper was Originally Published in Japanese in J. JFS 83 (2011) 3–6.

Materials Transactions, Vol. 52, No. 12 (2011) pp. 2184 to 2188#2011 Japan Foundry Engineering Society

As to the eutectic solidification end temperature (TEE), threedifferent parameters were set to vary the eutectic temperaturerange (�TE) as follows:(a) Based on the Fe-C binary equilibrium phase diagram,

TEE was assumed to be 1417 K because the eutectictemperature is constant. Therefore, the minimumtemperature difference was defined by subtracting TEE

from TES, i.e., �TE: TES � TEE ¼ 1 K,(b) Based on the Fe-C-Si ternary equilibrium phase

diagram,2) TEE was assumed to be 1408 K becauseSilicon content is taken into account on the eutectictemperature range, i.e., �TE: TES � TEE ¼ 10 K,

(c) Based on the multiple equilibrium phase diagram, TEE

was assumed to be 1403 K because contents ofphosphorus, manganese, sulfur, etc. are also taken intoaccount on the eutectic temperature range, i.e., �TE:TES � TEE ¼ 15 K.

ADSTEFAN Ver.11, a casting simulation software, was usedfor the calculation of molten metal flow and the temperatureduring solidification. Table 2 shows the properties of the cast

iron, the CE cup and the silica tube used for this experiment.Table 3 shows the parameters for the relationship betweentemperature and solid fraction. The results of the calculationare shown in Fig. 3 (a), (b) and (c).

2.3 Cooling curves at each measuring pointAs shown in Fig. 1, nine measuring points were set in the

following order:(1) On the top of the silica tube (Point 1),(2) At the final solidification part, which is 4 mm above

Point 1 (Point 3),(3) On the same level of Point 3, at seven points from

Point 3 to the casting surface with 2 mm intervals(Point 4–10).

The cooling curves at each point were calculated individu-ally. Figure 4 shows the cooling curves at each point.

2.4 Observation of the microstructuresThe graphite structure and the matrix structure of the cast

irons were observed in order to estimate the solidificationprocess and the cooling rate.

3. Results

3.1 Actual measurementAs shown in Fig. 2, the rapid temperature drop was

measured in the final period of the eutectic solidification, aswell as the common case of cast iron cooling. The range ofthe rapid temperature drop was from about 1393 K (1120�C)to about 1360 K (1087�C). The cooling rate during the rapidtemperature drop was about 3 K/s.

3.2 Comparison of latent heat emission patternsComparisons between the cooling curves calculated by the

three different parameters and the measured cooling curveare shown in Fig. 3. The cooling curves of �TE ¼ 10 K(Fig. 3(b)) and �TE ¼ 15 K (Fig. 3(c)) obviously agreedwith the measured cooling curve rather than the cooling

Table 1 Chemical composition of cast iron (mass%).

C Si Mn P S

3.32 2.05 0.739 0.051 0.087

−3.0

−2.0

−1.0

0.0

1.0

(900)

(1000)

(1100)

(1200)

(1300)

0 50 100 150 200 250 300

Coo

ling

rate

,K/s

Tem

p,K

(°C

)

Time,s

Measurement

Differentiation

1573

1473

1373

1273

1173

Fig. 2 Cooling curve and differentiation curve for measured.

Table 2 Material properties.

MaterialDensity

(g/cm3)

Thermal

conductivity

(cal/cm�S�deg)

Specific

heat

(cal/g�deg)

Latent

heat

(cal/g)

Casting FC250 6.4 0.05 0.17 50

CE-cup Shell mold 1.7 0.0015 0.2 —

Quartz glass Silica grass 2.21 0.0032 0.17 —

Table 3 Temp-solid fraction parameters.

TL TES TEE

(a) �TE ¼ 1 KTemp, K 1486 1418 1417

Solid fraction 0.00 0.33 1.00

(b) �TE ¼ 10 KTemp, K 1486 1418 1408

Solid fraction 0.00 0.33 1.00

(c) �TE ¼ 15 KTemp, K 1486 1418 1403

Solid fraction 0.00 0.33 1.00

Sect. A−A

P1

P3

P4

P5

P7P8

P9

P10

P6

A

A

Fig. 1 Shape of CE-cup.

Cooling Rate and Graphite Structure of Final Eutectic Solidification Part in Cast Iron 2185

curve of �TE ¼ 1 K (Fig. 3(a)), especially on the finalsolidification period. Therefore, it is clear that there is atemperature range of the eutectic solidification. Whencomparing the temperature ranges of �TE ¼ 10 K and�TE ¼ 15 K, the temperature ranges are almost the same.Considering all the information, we decided to use thecooling curve of �TE ¼ 10 K for further analysis.

The rapid temperature drop after the eutectic solidificationwas well simulated by the numerical analysis for thisexperiment although the recalescence at the graphite eutecticperiod was not simulated. The rapid temperature drop canbe commonly observed on the cooling curves with smallundercooling by excessive inoculation or with solidificationinto white iron. Therefore, we have concluded that thecalculation of the rapid temperature drop would not beaffected by the nonexistence of the undercooling phenom-enon.

3.3 Analyzing cooling curves for measuring pointsFigure 4 shows that the most outer point (Point 10) took a

steady cooling rate in the eutectic solidification period whilethe temperatures at the more inner points were kept for thelonger time at 1418 K (1145�C) before the rapid temperaturedrop. In the final solidification period in the simulatedcooling curve, the rapid temperature drop of about 30 Koccurred as well as the measured cooling curve.

Figure 5 shows the cooling curve and the solid fractioncurve for the final solidification part (Point 3) during thesolidification period. As shown in Fig. 5, it is clear that the

(1100)

(1140)

(1180)

(1220)

(1260)

0 40 80 120 160

Tem

p, K

(°C

)

Time,s

1413

1453

1493

1533

1373

0 40 80 120 160Time,s

SimulationMeasurementDifferentiation(Simulation)Differentiation(Measurement)

−3.0

−2.0

−1.0

0.0

1.0

0 40 80 120 160

Coo

ling

rate

,K/s

Time,s

(a) ΔTE= 1K (b) ΔTE = 10K (c) ΔTE = 15K

Fig. 3 Cooling curves and the first differentiation curves for measured and calculated values.

(1100)

(1120)

(1140)

(1160)

(1180)

(1200)

(1220)

0 20 40 60 80 100 120 140 160 180

Tem

p,K

(°C

)

Time,s

P3

P4

P5

P6

P7

P8

P9

P10

1373

1413

1433

1453

1473

1493

P7P3

P101393

Fig. 4 Calculated cooling curves of each point (P3�P10).

0 0.2 0.4 0.6 0.8 1.0

(1120)

(1140)

(1160)

(1180)

(1200)

20 40 60 80 100 120 140 160 180

Solid fraction

Tem

p,K

(°C

)

Time,s

P3

Solidus ratio

1393

1413

1433

1453

1473

A type graphite

D type graphite

Primary-γ

TL

TES

TEE

Fig. 5 The cooling curve of final solidification part and solid fraction.

2186 H. Kawashima, K. Sigeno, M. Kusubae and K. Tachibana

rapid temperature drop in the final solidification periodoccurred even at the time when liquid phase remained insidethe castings.

3.4 Structure analysisFigure 6 shows the graphite structures on the final solid-

ification part (Point 3), the most outer point (Point 10) andthe point in the middle of them (Point 7). On Point 3,moderately-cooled structure (i.e., A-type graphite) andrapidly-cooled structure (i.e., D-type graphite) appearedtogether in mottled structure. Figure 7 shows the image ofthe mottled structure by a low magnification. While Point 3had some D-type graphite, Point 7 only had the A-typegraphite. Even on the most outer point (Point 10), its mainstructure was the A-type graphite with the B-type graphiterather than the D-type graphite. The matrix structure of theD-type graphite area on the final solidification part waspearlite structure as shown in Fig. 8.

4. Analysis

The rapid temperature drops detected both on themeasured and the simulated cooling curves for Point 1 wereidentified as the temperature drop during the eutecticsolidification in the final solidification period. By calculatingthe cooling curves of the cast iron based on the temperaturerange of 10 K in the eutectic solidification, it has becomeclear that the temperature drop actually started not after butduring the solidification when the molten iron partiallyremained in the phase of �-Fe+G+L.

As the result of the graphite structure observations at thefinal solidification part, the mottled structure coexists withthe A-type and the D-type graphite. It is considered that themottled structure is an evidence to from the chilled structureby the rapid temperature drop in the liquid/solid phase. Thecooling rate can be assumed to be faster than the generalcooling rate for thin castings due to the evidence that the

Point 3 Point 7 Point 10

200 μμ m/div

Fig. 6 Graphite shape on Point 3, 7 and 10.

Fig. 7 Graphite shape of Point 3. Fig. 8 D-type graphite matrix structure.

Cooling Rate and Graphite Structure of Final Eutectic Solidification Part in Cast Iron 2187

matrix structure around the D-type graphite had pearlitestructure without the transformation to be ferrite.

5. Conclusions

According to the cooling rate of both the measured coolingcurve and the simulated cooling curve of cast iron, andmicrostructure observations, the following conclusions havebeen obtained:

(1) In analyzing the solidification process of cast iron, it isrecommended to use the model that has a certain temperaturerange (approx. of 10 K) in the eutectic solidification inaccordance with the Fe-C-Si ternary equilibrium phasediagram.

(2) The above-mentioned model enables us to confirm thatthe rapid temperature drop on the actual measurement occursduring the eutectic solidification.

(3) The matrix structure around the D-type graphite in thefinal solidification part shows pearlite structure. This factsupports the above-mentioned rapid cooling theory.

REFERENCES

1) Japan Foundry Engineering Society: Casting Defects and Their

Remedies, (Japan Foundry Engineering Society, Tokyo, 2007) p. 171.

2) American Society for Metals: Metals Handbook —Properties and

Selection: Iron and Steels—, 9th Ed. Vol. 1, (American Society of

Metals, Metals Park, 1978) p. 4.

2188 H. Kawashima, K. Sigeno, M. Kusubae and K. Tachibana