ISIJ International, Vol. 52 (2012), No. 11, pp.

2036–2041

Critical Temperature Range in Standard and Ni-bearing Spheroidal

Graphite Cast Irons

Roxana Maria GHERGU,1) Jon SERTUCHA,2) Yannick THEBAULT1) and Jacques LACAZE1)*

1) CIRIMAT, Université de Toulouse, BP 44362, 31030 Toulouse cedex 4, France.

2) Ingeniería y Procesos de Fundición, IK4-AZTERLAN, Aliendalde auzunea 6, E-48200 Durango (Bizkaia), Spain.
(Received on May 15, 2012; accepted on July 17, 2012)

Describing the conditions for reaustenitization of spheroidal graphite cast irons is of interest for their
heat-treatment after casting, e.g. for manufacturing austempered ductile irons. Differential thermal analy-
sis has been used to characterize the direct eutectoid transformation and the reverse transformation, i.e.
the reaustenitization. This has been applied to a standard and a Ni-bearing alloy, with a ferritic matrix for
the former, both a ferritic and a pearlitic matrix for the latter. The results are discussed in relation with the
stable and metastable three phase fields. While earlier description of the direct eutectoid transformation
is confirmed, the one for reverse eutectoid has been found more complex and is amended.

KEY WORDS: nodular cast irons; eutectoid transformation; austenitization; ferrite; pearlite.

could proceed only when the lower temperature of the

1. Introduction
respective three-phase field has been reached. This approach
To control solid-state heat-treatment of cast irons, a precise has been applied successfully to explain that Cu and Mn
knowledge of the effect of alloying elements on the stable favour pearlite because they both decrease the driving force
(austenite-ferrite-graphite) and metastable (austenite-ferrite- for carbon diffusion in ferrite and decrease more Tα than
cementite) three-phase fields is required. The corresponding Tp.9,10)
temperature ranges have to be considered for describing the Using the same hypothesis, this approach was extended to
eutectoid transformation of austenite upon cooling as well as the reverse eutectoid transformation.11) Upon heating, trans-
for controlling full or partial reaustenitization that may be formation of ferrite to austenite forming halos around the
called reverse eutectoid transformation. Recent interest has graphite phase should start at the upper limit of the stable
been focused on this latter for manufacturing austempered three phase field ( Tα0 ) while decomposition of pearlite could
ductile irons (ADI) parts as stressed by a number of works proceed as soon as the lower limit of the metastable three-
dealing with kinetics of austenitization1,2) and its physical phase field (Tp) is reached as it is not related to the graphite
modelling.3,4) A clear understanding of the conditions for phase. Predictions for both direct and reverse eutectoid
nucleation and growth of the new phases is necessary for mod- transformations were confronted against an extensive
elling of direct and reverse eutectoid transformations. review of literature data.7) The agreement was good for finite
In a previous work, the conditions for ferrite and pearlite scanning rates, though some features should be stressed:
growth upon cooling from the austenite field have been
described on the basis of results obtained by differential
thermal analysis (DTA).5,6) Assuming their growth does not
involve long range diffusion of substitutional solutes (e.g.
Si, Cu, Cr, Mn, etc.)7,8) as must be the case for finite cooling
rates, the characteristic temperatures for transformation
could be evaluated on a Fe–C isopleth section of the phase
diagram drawn for the appropriate composition of the cast
iron in substitutional solutes. Such a section for a Fe–C–Si
alloy with 2.5 mass% Si is shown in Fig. 1,7) with the upper
and lower temperatures of the three-phase field denoted Tα0
0
and Tα for the stable system, Tp and Tp for the metastable
system. If ferrite precipitates as halos around graphite nod-
ules and pearlite grows with no connection with graphite
phase, it was thus found that growth of ferrite and pearlite
* Corresponding author: E-mail: [email protected] Fig. 1. Isopleth section of the stable (solid line) and metastable
DOI: http://dx.doi.org/10.2355/isijinternational.52.2036 (dotted lines) Fe–C–Si phase diagrams.7)

© 2012 ISIJ 2036

 ISIJ International, Vol. 52 (2012), No. 11

• for very low scanning rates, in practice lower than 2 K/

min, long range diffusion of substitutional solutes can Tαo = 739 + 31.5wSi − 7.7wCu
take place in which case the stable transformation may −18.7wMn + 3.3wMo − 10.7wCr − 26.0wNi .......... (1)
start within the three-phase field. This would apply
when partial reaustenitization is looked for as is the Tα = 739 + 18.4 wSi + 2.0(wSi )2 − 14.0 wCu
case for manufacturing so-called dual phase ADI.12,13) −45.0 wMn + 2.0 wMo − 24.0 wCr − 27.5wNi ......... (2)
• Upon cooling, the results showed that ferrite growth
can start at very low undercooling with respect to Tα Tpo = 727 + 30.07wSi − 1.98(wSi )2
while pearlite needs some finite undercooling relative
−10.7wCu − 13.7wMn + 9.3wMo + 24.3wCr − 12.0 wNi ..... (3)
to Tp to appear.14) For modelling the eutectoid transfor-
mation, it has been accepted that ferrite appears at the Tp = 727 + 21.6 wSi + 0.023(wSi )2
interface between austenite and graphite15,16) giving
rise to the bull-eye structure. In some cases however, −21.0 wCu − 25.0 wMn + 8.0 wMo + 13.0 wCr − 33.0 wNi .... (4)
intergranular ferrite may precipitate upon cooling These formulas have been calculated using a thermody-
before bull-eye ferrite after reaustenitization.9) Similar- namic software and associated databases as previously
ly, austenite has been observed to appear at the ferrite/ described.7) They are thus valid for the low carbon content
graphite interface for high heating rates17) and at both that prevails at the eutectoid reaction, and account should be
the ferrite/graphite interface and grain boundaries upon made of the presence of graphite that leads to redistribution
reverse eutectoid transformation.17,18) of all alloying elements but carbon in the matrix. Accord-
• For finite cooling rates, microsegregation issued from ingly, the wi values entered in the formulas for evaluating
solidification was considered to affect only the last the characteristic temperatures of an alloy are 1.05 times
stage of the eutectoid transformation though the pres- those listed in Table 1.7)
ence of even low levels of some alloying elements such DTA experiments were performed with a SETARAM
as Cr, Mn and Cu could favour pearlite instead of fer- SETSYS apparatus, with alumina powder as reference mate-
rite.19) This effect has been related to chemical hetero- rial and samples machined out from the cast irons with a
geneities built-up during solidification and seems to diameter of 3.9 mm and a height of 5 mm. The DTA cell is
more dramatically affect reaustenitization and subse- regularly checked against transformation temperature of
quent solid-phase transformations18,20,21) as compared pure elements, namely melting of aluminium, silver, iron
to the direct eutectoid transformation during cooling and nickel. The accuracy on the temperatures reported here
after casting. is estimated to be ± 2°C.
Since very few precise data were in fact available for As received, the standard alloy showed an as-cast ferrito-
checking the reaustenitization process against the former pearlitic matrix and it was guessed that pearlite should start
assumptions,7) it seemed appropriate to test these latter. For to decompose to ferrite and graphite upon heating at a tem-
this, an experimental investigation in which both ferritic and perature far below the metastable three-phase field. It was
pearlitic matrix were processed has been performed. For thus considered more significant to study only the transfor-
such an attempt, it was decided to compare the transforma- mation of ferrite upon heating of this alloy. DTA samples
tion of a Ni-bearing cast iron that is prone to give a pearlitic were thus subjected to a subcritical anneal which consisted
matrix to a standard alloy. Both alloys were first studied in heating at a rate of 10 K/min to 720°C, holding for 60 min
after a ferritization treatment and the Ni-bearing one was and then cooling to 400°C at 2 K/min. The first sample was
also studied after a pearlitizing thermal cycle. DTA was then cooled to room temperature (RT) and it was checked
used for controlling the heating and cooling of the samples. that all pearlite had decomposed. The subsequent samples
were maintained for 1 min at 400°C after the subcritical
anneal and then heated to 950°C, held 5 minutes and then
2. Experimental Details
cooled to RT. The heating and cooling rates used were 2, 5,
The alloys investigated have been prepared for previous 10, 15 and 20 K/min.
studies and all details on melt preparation and casting con- In the as-received condition, the Ni-bearing material had
ditions have already been given.22) The content in the main been ferritized by means of a subcritical anneal.22) A first set
alloying elements for the standard and the Ni-bearing grades of DTA experiments was thus conducted to investigate the
are given in Table 1. The isopleth Fe–C sections for the Fe– effect of heating rate on the transformation of ferrite to auste-
C–Si and Fe–C–Si–Ni alloys under investigation would be nite with upper temperature set at 1 050°C. Upon cooling, the
very similar to that shown in Fig. 1, with only slight changes effect of the scanning rate on the transformation of austenite
in the characteristic temperatures. In this table are also listed to a ferrito-pearlitic matrix was investigated. Finally, for
the characteristic temperatures (°C) estimated according to studying the effect of the heating rate on the transformation of
the following formulas where wi (mass%) is the alloying pearlite to austenite, the samples were first heated to 1 050°C,
content of element “i”: 7) then cooled at the highest possible rate (about 60 K/min in the
eutectoid range) to maximize the amount of pearlite, and final-
Table 1. Composition of the alloys (mass%). ly reheated to 1 050°C and cooled to RT for the same range of
scanning rates as above. Figure 2 shows the nearly fully pearl-
C Si Mn Ni Tα0 Tα Tp0 Tp
itic matrix obtained after rapid cooling from 1 050°C.
standard grade 3.73 2.08 0.13 0.01 805.2 782.6 781.3 770.9
Ni-bearing grade 3.89 1.88 0.13 0.92 773.5 750.4 765.2 734.4

2037 © 2012 ISIJ

 ISIJ International, Vol. 52 (2012), No. 11

Table 2. Characteristic transformation temperatures (°C) evaluated

on the DTA records of the ferritized standard cast iron for
various scanning rates (°C/min).

Heating Cooling
Scanning Ferrite Ferrite Ferrite Pearlite Pearlite
rate start peak start start peak
2 785 828 778 736 724
5 804 835 764 730 712
10 816 845 757 717 698
15 826 853 746 708 686
20 833 856 741 701 678

Fig. 2. Micrograph of the Ni-bearing alloy with nearly fully pearl-

itic matrix obtained after rapid cooling from 1 050°C.

Fig. 4. DTA records upon cooling the standard material from

950°C.

Fig. 3. DTA records upon heating the standard material in the fer- two thermal arrests are clearly distinguished that relate to
ritized state (solid lines) and in the as-received condition the appearance of ferrite for the high temperature and the
(dotted line). formation of pearlite for the low temperature. As expected,
the relative amplitude of the former with respect to the latter
decreases as the cooling rate is increased. As indicated with
3. Results
the open arrows in Fig. 4, the characteristic temperatures
Figure 3 shows the five DTA records obtained upon heat- that could be evaluated, namely the start of ferrite precipi-
ing the ferritized standard alloy up to 950°C at different tation, the start of pearlite precipitation and the peak tem-
rates (solid lines) plus the record corresponding to the as- perature associated to pearlite, are listed in Table 2. As men-
received alloy with a 5 K/min heating rate (dotted line). tioned before, the start of the transformations is evaluated
Note that, in this figure and the following, the records have with a higher uncertainty than the peak temperature.
been shifted along the ordinate direction for clarity. Com- The set of characteristic temperatures have been plotted in
paring the two records at the heating rate of 5 K/min, it is Fig. 5 as function of the square root of the scanning rate during
seen that the as-received material shows three peaks while the DTA run. Looking first at the transformation temperature
the ferritized one shows only two peaks. It is thus easy to from austenite to ferrite upon cooling, it is seen that the data
associate the peak at intermediate temperature with pearlite extrapolates to Tα with the value for the lowest cooling rate
decomposition, while the peak at lower temperature relates slightly above the regression line as expected. The two series
to the magnetic transformation of ferrite and that at higher of data for pearlite formation upon cooling, namely the start
temperature to the transformation of ferrite to austenite. It and peak temperatures, may be extrapolated to nearly the same
could be noticed however that a small bump is still apparent temperature i.e. 20°C below Tp. That pearlite needs such an
at intermediate temperatures on all DTA records with the underccoling to appear is in agreement with previous
ferritized samples, indicating that a small amount of pearlite works.6,9,14) Summing up, the results obtained upon cooling the
was anyway present. standard alloy agree with those expected according to the pro-
The temperature for the beginning of austenite formation posed model for the eutectoid transformation.11)
and the peak temperature of the thermal arrest related to ferrite On the contrary, the data for ferrite transformation to aus-
decomposition were recorded as indicated by the open arrows tenite upon heating are at definite difference with respect to
in Fig. 3. While the latter is easily evaluated, the former is sub- those expected. The temperature associated with the start of
ject to a higher uncertainty. The obtained values for ferrite ferrite decomposition may be extrapolated to a temperature
transformation start and ferrite peak are listed in Table 2. close to Tα at zero scanning rate. This means that the trans-
The DTA records obtained on cooling the samples of the formation proceeds within the three phase field and thus that
standard alloy are shown in Fig. 4. On each of these records, austenite does not form halos around the graphite nodules as

© 2012 ISIJ 2038

 ISIJ International, Vol. 52 (2012), No. 11

Table 3. Characteristic transformation temperatures (°C) evaluated

on the DTA records of the Ni-bearing cast iron for various
scanning rates (°C/min).

Heating Heating Cooling

Ferritized material Pearlitized material Pearlitized material
Scanning Ferrite Ferrite Pearlite Pearlite ferrite Ferrite Pearlite Pearlite
rate start peak start peak peak start start peak
2 785 804 763 769 802 742 707 703
5 791 812 767 778 809 734 699 690
10 801 823 776 789 818 727 687 675
15 781 799 822 721 680 663
20 814 838 788 806 830 714 672 653

Fig. 5. Evolution of the characteristic temperatures of the DTA

records obtained for the standard material as function of the
square root of the scanning rate. Solid symbols are for the
ferritic reaction, open symbols for the pearlitic one.

Fig. 7. DTA records upon heating the Ni-bearing material in the

pearlitized state.

lized by nickel, some of the pearlite transforms to ferrite and

graphite during heating before it decomposes to austenite
when the temperature is high enough. The characteristic
Fig. 6. DTA records upon heating the Ni-bearing material in the temperatures considered here are the start and peak temper-
ferritized state. atures for pearlite transformation and the peak temperature
for ferrite transformation to austenite as it is shown with the
postulated in earlier work.11) Instead, the transformation open arrows in Fig. 7. Note that there are no shoulders on
must proceed by the growth of austenite along grain bound- the high temperature side of the ferrite peak appearing, this
aries as reported by several authors17,18,23) and is controlled will be discussed further below.
by carbon diffusion along these grain boundaries. Finally, Fig. 8 shows the set of DTA records that were
Figure 6 presents the DTA records on heating to 1 050°C obtained from the Ni-bearing alloy upon cooling during the
of the as-received ferritized Ni-bearing material. These second series of experiments. As the initial conditions
records show two main peaks, the Curie peak at low tem- before cooling were the same in the first series of experi-
perature and the peak associated with the ferrite to austenite ments, i.e. a short holding after heating to 1 050°C, the
transformation at higher temperatures. On all records, a records obtained upon cooling during the first and second
shoulder peak at even higher temperature could be observed series were similar. The characteristic temperatures consid-
indicating that the transformation of ferrite to austenite is ered in Fig. 8 are the same as for Fig. 4, i.e. the start of fer-
much longer for this alloy than for the standard one. The rite precipitation, the start of pearlite precipitation and the
temperature for the start of ferrite decomposition and the peak temperature associated to pearlite, they are listed in
peak temperature (open arrows in Fig. 6) have been consid- Table 3.
ered as characteristic temperatures and are listed in Table 3. The characteristic temperatures obtained from Figs. 6, 7
Figure 7 shows the DTA records upon heating after the and 8 have been plotted in Fig. 9. Starting again with the
Ni-bearing material had been heated up to 1 050°C and rap- transformation upon cooling, it is seen that the start of ferrite
idly cooled to RT to get a nearly fully pearlitic matrix. The growth corresponds well to Tα. Similarly, the two character-
records all show three peaks, as for the dotted curve in Fig. istic temperatures for formation of pearlite upon cooling
3, which can be allocated to the Curie transformation, pearl- may be extrapolated to a single temperature that is about
ite decomposition and ferrite transformation at increasing 10°C below Tp. These results agree with predictions11) as
temperature. It is seen that the height of the peak associated was already the case for the standard alloy.
with ferrite increases with respect to that of pearlite as the Concerning the transformation from ferrite to austenite
scanning rate is decreased. This shows that, though stabi- upon heating, it is seen that the temperature for the start of

2039 © 2012 ISIJ

 ISIJ International, Vol. 52 (2012), No. 11

Fig. 8. DTA records upon cooling the Ni-bearing material from

1 050°C.

Fig. 9. Evolution of the characteristic temperatures of the DTA

records obtained for the Ni-bearing material as function of
the square root of the scanning rate. Solid symbols are for the
ferritic reactions, open symbols for the pearlitic one. Solid
diamonds are for the material with initially a pearlitic matrix.
Fig. 10. Optical micrograph after etching with 2% Nital of Ni-
this reaction in an initially fully ferritic matrix may well be bearing samples held 5 minutes (a dnd c) and 15 minutes
extrapolated to Tα0 as expected from the model. This should (b and d) at 800°C (a and b) and 5 and 15 minutes at
825°C (c and d).
mean that for this alloy the transformation proceeds, at least
for a part, by the formation and growth of austenite halos
around graphite nodules, and this is contrary to the result on ferritic or pearlitic. This suggests that this peak relates to the
the standard alloy presented above. This is further discussed same part of the material at the microstructure scale, and
in the next section. hence that the ferrite associated to the shoulder detected in
Decomposition of pearlite during heating shows start and alloys with an initially ferritic matrix (see Fig. 6) gives rise
peak characteristic temperatures that may well be extrapo- to the pearlite that first transforms after pearlitization. For a
lated to the same temperature that appears to be significantly more precise understanding of the origin of the shoulders,
higher than Tp and in fact quite close to Tα. This suggests samples of the ferritized Ni-bearing alloy have been heat-
that decomposition of pearlite to austenite upon heating pro- treated by introducing them in a furnace at either 800°C or
ceeds provided the material temperature is above Tα, i.e. with 825°C for 5 and 15 minutes, and then quenched into brine.
the condition that austenite may be stable at that temperature. Typical micrographs of each sample are shown in Fig. 10
Thus, the description that have been proposed previously11) where it is seen that austenite (appearing as martensite after
must be modified: decomposition of pearlite to austenite will quenching) forms at the graphite/ferrite interface but also at
start at Tp if this temperature is higher than Tα, it will start at cell boundaries away from the graphite nodules (Figs. 10(a)
Tα in the opposite case as for this Ni-bearing alloy. and 10(b)). Growth of austenite then proceeds by thickening
of the austenite shell around the graphite nodules and of the
austenite formed at cell boundaries (Fig. 10(c)) and finishes
4. Discussion
in the regions in between nodules and cell boundaries (Fig.
In Fig. 9 it is noticed that the peak temperature for ferrite 10(d)) which are then associated to the shoulders.
transformation is about the same whatever the initial matrix, Using a sample of the Ni-bearing alloy that had been

© 2012 ISIJ 2040

 ISIJ International, Vol. 52 (2012), No. 11

5. Conclusion
The DTA results obtained in the present study upon cool-
ing from the austenite + graphite field are in full agreement
with previous reports and confirm the model proposed for
the eutectoid transformation. On the other hand, the DTA
results upon heating the materials and the few micrograph
observations performed show that the previously developed
model for reverse eutectoid transformation had to be chal-
lenged with new experiments. The following conclusions
could be drawn for reaustenitization of a ferritic matrix:
• in standard cast irons with very low levels of alloying
additions, segregation of Si is such that the predicted
Fig. 11. Example of composition profile in between two nodules
ferrite to austenite transformation temperature is lower
in the Ni-bearing alloy. at cell boundaries than at the nodule/ferrite interfaces.
Thus, austenite appears at cell boundaries and grows by
Table 4. Estimates of the Tα temperature at locations close to the
carbon diffusion along grain boundaries at a tempera-
graphite nodules and at cell boundaries. ture lower than the Tα0 temperature of the alloy. The ref-
erence temperature for ferrite decomposition is then Tα.
Standard alloy Ni-bearing alloy • Negative segregation of Ni leads to a predicted trans-
Nodule/ferrite interface 774.9 754.0 formation temperature for ferrite lower at the graphite/
Intercellular areas 769.8 758.8 ferrite interface than at cell boundaries, so that austenite
halos may form around the nodules before ferrite starts
transforming at cell boundaries. In that case, the refer-
pearlitized, a few line-scans between two neighbouring ence temperature for ferrite decomposition is Tα0 .
graphite nodules were recorded using an energy dispersive In the case of a pearlitic matrix, decomposition of pearlite
X-ray spectrometry (EDS) system on a scanning electron to austenite needs this latter to be stable which is possible
microscope (SEM). One of these line-scans is shown in Fig. if the Tα temperature of the alloy is lower than its Tp tem-
11 where it is seen that Si and Ni contents decrease as perature. In such a case, the reference temperature for the
expected from the graphite/matrix interface to the inter- start of pearlite decomposition is Tp while it will be Tα in
nodule areas, from 2.2 mass% to 1.9 mass% for Si and from the opposite case. The previous model for pearlite decompo-
1.05 mass% to 0.6 mass% for Ni. Because of the low Mn sition upon heating may be made more precise by stating that
content, this element appears more erratic when an increase the temperature at which it starts is the highest of Tα and Tp.
would have been expected, thus it will be considered as con-
stant at 0.14 mass% in the following. The Si profile in the REFERENCES
standard alloy could be similarly considered to change from
1) R. C. Thomson, J. S. James and D. C. Putman: MST, 16 (2000), 1412.
2.0 mass% to 1.7 mass% while the Mn was again taken as 2) R. Ivanova, W. Sha and S. Malinov: ISIJ Int., 44 (2004), 886.
constant at 0.14 mass%. 3) W. Kapturkiewicz, E. Fras and A. A. Burbelko: MSE A, 413–414
Using the compositions mentioned above, the Tα temper- (2005), 352.
4) Uma Batra, S. Ray and S. R. Prabhakar: JMEPEG, 14 (2005), 574.
ature could be estimated by means of Eq. (2) for locations 5) J. Lacaze, S. Ford, C. Wilson and E. Dubu: Scand. J. Metall., 22
close to the graphite nodules and at cell boundaries. The val- (1993), 300.
6) J. Lacaze, C. Wilson and C. Bak: Scand. J. Metall., 23 (1994), 151.
ues obtained are listed in Table 4. It appears clearly that the 7) V. Gerval and J. Lacaze: ISIJ Int., 40 (2000), 386.
difference in the reaustenitizing process between the two 8) S. N. Lekakh and V. L. Richards: AFS Trans., 118 (2010), 225.
alloys may be directly related to the change in Tα value from 9) J. Lacaze, A. Boudot, V. Gerval, D. Oquab and H. Santos: Metall.
Mater. Trans. A, 28A (1997), 2015.
the graphite/ferrite interface to cell boundaries: this change 10) J. Sertucha, P. Larrañaga, J. Lacaze and M. Insausti: Int. J. Metal-
is negative for the standard alloy while it is positive for the casting, 4 (2010), 51.
11) E. Linares, V. Gerval and J. Lacaze: Scr. Mater., 38 (1998), 279.
Ni-bearing alloy. Thus, for the standard alloy, transforma- 12) A. Basso, R. Martinez and J. Sikora: Key Eng. Mater., 457 (2011), 145.
tion of ferrite to austenite could start at lower temperature 13) A. P. Druschitz, R. E. Aristizabal, E. Druschitz, C. R. Hubbard, T. R.
Watkins, L. Walker and M. Ostrander: Metall. Mater. Trans. A, 43A
in the cell boundaries provided carbon can diffuse from the (2012), 1468.
graphite nodules to these areas. As mentioned above, sever- 14) J. Lacaze: Int. J. Cast Metals Res., 11 (1999), 431.
al authors have observed this for the intermediate heating 15) M. Wessén and I. Svensson: Metall. Mater. Trans. A, 27A (1996),
2209.
rates investigated in the present work. In the case of the Ni- 16) J. Lacaze and V. Gerval: ISIJ Int., 38 (1998), 714.
bearing alloy, the decomposition of ferrite to austenite is not 17) N. Wade and Y. Ueda: Trans. Iron Steel Inst. Jpn., 20 (1980), 857.
18) J. M. Chou, M. H. Hon and J. L. Lee: Scr. Metall. Mater., 27 (1992),
favoured at cell boundaries (see Table 4) and the transfor- 61.
mation should preferentially start with the formation of aus- 19) J. Lacaze: Proc. of Erstarrung Metallischer Schmelzen in Forschung
tenite halos around nodules when the material temperature and Giessereipraxis, ed. by A. Ludwig, Wiley-VCH, Weinheim,
(1999), 25.
has reached Tα0 . The microstructures in Fig. 10 show that 20) J. Sikora and R. Boeri: Int. J. Cast Metals Res., 11 (1999), 395.
both transformation processes do in fact compete for this 21) J. M. Schissler and J. Saverna: J. Heat Treating, 4 (1985), 167.
22) J. Lacaze, P. Larrañaga, I. Asenjo, R. Suárez and J. Sertucha: MST,
alloy. This discussion shows the importance of microsegre- 28 (2012), 603.
gation in the reaustenitization process. 23) J. E. Rehder: AFS Trans., 73 (1965), 473.

2041 © 2012 ISIJ