1

Growth kinetics of proeutectoid ferrite in Fe-0.1C-1.5Mn-1Si quaternary

and Fe-0.1C-1.5Mn-1Si-0.2Al quinary alloys

G. H. Zhang, R. Wei
a
, M. Enomoto
a
and D. W. Suh

Graduate Institute of Ferrous Technology, Pohang University of Science and Technology,
Pohang 790-784, Republic of Korea
a
Department of Materials Science and Engineering, Ibaraki University, Hitachi 316-8511, Japan

Abstract
The growth kinetics of proeutectoid ferrite in the early stages of transformations were studied
in Fe-0.1C-1.5Mn-1Si (mass pct) quaternary and Fe-0.1C-1.5Mn-1Si-0.2Al quinary alloys. The
observed kinetic transition temperatures from partitioned slow-growth to unpartitioned fast-
growth of ferrite in both alloys are in good agreement with predictions using a local equilibrium
model for multi-component systems. The measured parabolic growth rate constants were smaller
than those calculated assuming paraequilibrium in the unpartitioned growth region, but the
difference between the measured and calculated growth rate constants gradually diminished with
decreasing the reaction temperature. The dissipation of driving force, derived from the diffusion
of the substitutional solute within the transformation interface, possibly constitutes a major part
of the discrepancy between the measured and calculated growth kinetics.

Key words: proeutectoid ferrite; kinetics; local equilibrium; paraequilibrium; solute drag
Metallurgical and Materials Transactions A
43 (2012) 833-842
2

1. Introduction
Driven by the scientific and industrial importance of ferrite transformations, its diffusional
growth in steels has been the focus of considerable research during the past several decades. Of
the variety of morphologies the ferrite can adopt when forming from the austenite, the grain
boundary allotriomorph is the most representative microstructure. In theoretical treatments of
allotriomorphic ferrite growth in the binary Fe-C system, it is often assumed that the free energy
dissipated by the finite mobility of interface is negligible and the problem is treated as a
diffusion-controlled process. Consequently, predictions of ferrite growth kinetics reduce to
solving the flux balance equation for carbon at the austenite/ferrite interface subjected to an
appropriate interfacial condition. A local equilibrium (LE) interfacial condition proposed first by
Zener [1] is usually chosen so that there are no chemical potential gradients for C or Fe across
the migrating interface. Thus, for isothermal transformation in Fe-C alloy, the compositions at
the interface are given by the ends of a tie-lie which passes through the bulk alloy composition in
the equilibrium phase diagram. Some experimental results of the ferrite growth kinetics in the
binary Fe-C system indicate that these assumptions are indeed reasonable [2,3].
The situation immediately becomes complicated when a substitutional alloying element (e.g., X
= Mn, Si, Ni, Al, Cr etc.) is added to the Fe-C system. Because of the large differences in the
diffusivities of C and X in austenite at the temperatures of interest, it is not possible to
simultaneously satisfy the flux balance equations for both C and X with interface compositions
given by the tie-line passing through the bulk alloy composition in two-phase region. In
order to satisfy the flux balance equations simultaneously and yet maintain the local equilibrium
at the interface, another tie-line which deviates from the bulk composition must be considered.
This leads to two cases [4-9]. In the first case, the tie-line is chosen so that the carbon activity
3

gradient in austenite near the / interIace is almost to zero as illustrated in Figs. 1(a), then the
driving force for carbon diffusion in austenite will become very small and the slow diffusion of
alloying element can keep pace with carbon diffusion. Because the concentration of alloying
element in ferrite differs significantly from that in the bulk, it will be partitioned between
austenite and ferrite. This growth mode is, therefore, called a partitioning local equilibrium
(PLE) mode. In the second case, the tie-line is chosen so that the composition ratio (Fe/X) at the
/() terminus oI the tie-line is the same as that in the bulk alloy as illustrated in Figs. 1(b).
Because the supersaturation of alloying element is close to unity, the alloying element diffusion
can keep pace with that of carbon by forming a sharp and narrow spike just ahead of the
interface. The growth of ferrite in this manner is called a negligible partitioning local equilibrium
(NPLE) mode. In general, the partitioning growth takes place at higher temperatures while the
unpartitioned growth occurs at lower temperatures. The transition temperature from partitioned
slow growth to unpartitioned fast growth depends on the alloy system and composition. For
ferrite growth without partitioning of substitutional alloying element, there is an alternative
model assuming that the local equilibrium is achieved just for the carbon and the other elements
(Fe and X) are not affected by the passage of the interface. This is a constrained local
equilibrium known as paraequilibrium (PE) [10,11]. Under the PE condition, substitutional
alloying elements are not partitioned between austenite and ferrite and no concentration spike
exists as illustrated in Figs. 1(c).
The effects of individual alloying element on the growth kinetics of ferrite allotriomorphs have
been studied extensively in Fe-C-X alloys [12-15]. Experimental results indicate that the growth
kinetics of ferrite are in good agreement with theory in alloys containing Al, Si, V, Co and Ni.
However, in alloys containing Mn, Cr and Mo, the measured growth kinetics are significantly
4

slower than the PE and/or the NPLE prediction in the unpartioned growth regime. The
discrepancy has been attributed to a solute drag effect caused by the interaction of the
substitutional solutes with the moving interface [13,16,17].
If more than one substitutional solute is present, the growth behavior of ferrite can be different
from that expected with a simple superposition of the effects of the individual elements because
the interaction between substitutional solutes may not be negligible. Indeed, Tanaka et al. [18]
reported that synergistic effects of Mn and X (=Si, Ni, Co) are likely to exist in Fe-C-Mn-X
quaternary alloys, and they may lead to the deceleration of the thickening kinetics of ferrite. Guo
et al. [19] reported that three growth stages were recognized in the ferrite transformation in a
quaternary Fe-0.04C-3Mn-1.9Si alloy, which starts with a rapid growth stage during which no
partition or interfacial accumulation of Mn or Si could occur, followed by second stage of no
bulk partition but enrichment of Mn and Si at the ferrite/austenite boundaries accompanying a
considerable drop of growth rate and third slow growth stage accompanying the significant
partitioning of alloying element between ferrite and austenite. Recently, Wei et al [20] studied
the partitioning of Mn and Si and growth kinetics of ferrite in a deformed Fe-0.1C-3Mn-1.5Si
alloy. They found that both solutes partition simultaneously above the calculated PLE/NPLE
transition temperature. Although the growth kinetics were considerably faster than the
calculation assuming local equilibrium, fine ferrite particles were readily obtained in the
partitioned growth region. Compared with the comprehensive work on Fe-C-X ternary alloy
systems, however, the studies on the growth behavior in higher order systems are still rare.
The purpose of the present study was to investigate the growth kinetics of proeutectoid ferrite in
the multi-component steels containing more than one substitutional alloying element. Fe-C-Mn-
Si is the basic system of most modern high strength steels, eg. TRansformation-Induced-
5

Plasticity (TRIP) steel and Dual-Phase (DP) steel. Recently, the influence of Al on the
improvement of advanced high strength steel received much interest [21,22]. Thus, the Fe-C-
Mn-Si quaternary and Fe-C-Mn-Si-Al quinary alloys were chosen in this study. The growth
kinetics of proeutectoid ferrite in both alloys were investigated from a time-series of
metallographic observations on samples treated isothermally and the results were examined with
respect to the local equilibrium and paraequilibrium models incorporated with the solute drag
effect.

2. Experimental procedures
High-purity Fe-C-Mn-Si and Fe-C-Mn-Si-Al alloys were prepared by vacuum induction melting.
Their chemical compositions, Ae
3
temperatures calculated using ThermoCalc with the TCFE6
database, are shown in Table . Cast ingot of 50 kg each were hot-rolled into 20 mm thick plates
between 1173~1473 K (900~1200
o
C). In order to homogenize the distribution of alloying
elements, the samples were sealed in a quartz tube and soaked at 1523 K (1250 C) for 48 hours.
Isothermal transformation experiments were performed using a quenching dilatometer.
Cylindrical specimens with 3 mm diameter and 10 mm length were first austenitized at 1373 K
(1100
o
C) for 10 min. They were then rapidly cooled to the programmed isothermal holding
temperature ranging from 1003 to 1108 K (730 to 835
o
C). After isothermal holding for various
times up to 30 min, the specimens were quenched by helium jet down to ambient temperature.
The mean austenite grain size after austenitization was approximately 100 m.
Each specimen was cut at the center along the transverse axis to perform optical microscopy.
Volume fractions of ferrite were deduced from area fraction using image analysis. The half-
thickness of ferrite allotriomorphs was estimated from the average value of ferrite nucleated at
6

the faces of austenite grains upon several fields of optical micrographs [13,18]. The
concentrations of Mn, Si and Al within and near the ferrite particles were assessed by energy
dispersive X-ray (EDX) analysis, using a field-emission scanning electron microscope (FE-
SEM). The data of Si and Al concentration showed a considerable scatter, perhaps because of the
lower sensitivity of the X-ray detector for Si radiation and the low concentration of Al in the
present alloys [19]. In contrast, the Mn concentrations in ferrite are much more reliable and
informative, and these are reported here.

3. Calculation of interfacial compositions and parabolic growth rate constants
The calculations of the local equilibrium interfacial tie-lines and the parabolic growth rate
constants in Fe-C-X
i
alloys usually require a numerical manipulation of a set of transcendental
equations. But, this involves multiple regression analysis of o/ phase boundary compositions,
which takes much time and often causes numerical problems in higher order alloys. Wei et al.
[20] proposed an approximate approach to determine the operative tie-line in any Fe-C-X
i
multi-
component system. This method is used in the present study and is summarized as follows. Since
the difference in diffusivities among substitutional alloying elements is much smaller than that of
carbon, one can reasonably assume that the diffusivities of substitutional elements are
approximately equal suggesting their supersaturations are also identical. In the PLE mode, one
can trace a series of interfacial compositions from (o+)/ terminus of tie-lines which have the
same carbon activity to the austenite with bulk composition. Among the tie-lines, the operative
tie-line can be determined uniquely because it should intersect with the carbon component ray
passing through the bulk composition ensuring the same supersaturation in substitutional
components. In the NPLE mode, the intersection with the o/(o+) phase boundary of the carbon
7

component ray which passes through the bulk composition defines the terminus composition of
the interfacial tie-line on the ferrite side and the other terminus composition on the austenite side
is readily obtained. The process was performed using the ThermoCalc software with the TCFE6
database. Once the operating tie-line is determined, the growth rate constant o under PLE can be
calculated considering the diffusion of substitutional elements, i.e.

|
|
.
|

\
|
=
i
i
D
f
2
o
[1]
where ) /( ) (
0 o
i i i i i
x x x x = is the supersaturation of the ith alloying element and D
i
is the
diffusivity of the ith alloying element in austenite. The form of the function
|
|
.
|

\
|
i
D
f
2
o
depends on
the shape of the precipitate. The simple planar form is used in this study. For planar
precipitates,
|
|
.
|

\
|
i
D
f
2
o
satisfies the following transcendental equation [7],


f (
o
2 D
i
) = t
o
2 D
i
exp
o
2
4D
i
|
\

|
.
| erfc
o
2 D
i
|
\


|
.
|
|
[2]
On the other hand, the growth rate constant in NPLE condition is calculated taking the
PLE/NPLE boundary as the operative /() boundary and the carbon diffusion as rate-
controlling one. The parabolic growth rate constant of ferrite under paraequilibrium is calculated
from the equation

|
|
.
|

\
|
=
C
C
2 D
f
o
[3]

4. Results and discussion
8

4.1 Microstructures and partitioning of alloying elements
Figs. 2a through 2d show optical micrographs of Fe-C-Mn-Si alloy isothermally reacted for 30
min at various temperatures. The white microstructural constituent is ferrite and the dark or grey
one is martensite or bainite formed during quenching. Ferrite grains predominantly nucleated on
the austenite grain boundary and grew into the grains with a morphology that would be described
as allotriomorphic. The allotriomorphic ferrite formed at 1078 K (805
o
C) (Fig. 2a) and 1068 K
(795
o
C) (Fig. 2b) was very thin. But a number of large ferrite grains formed at 1058 K (785
o
C)
(Fig. 2c) and 1048 K (775
o
C) (Fig. 2d). This indicates that growth rate of ferrite increased
dramatically when the temperature is reduced from 1068 to 1058 K (795 to 785
o
C). Figs. 3a
through 3d show optical micrographs of Fe-C-Mn-Si-Al alloy isothermally reacted for 30 min at
various temperatures. Similarly, one can observe a notable increase of ferrite thickness and
fraction when temperature is decreased from 1088 to 1078 K (815 to 805
o
C).
In general, the volume change accompanying the formation of ferrite from austenite can be
considered to be isotropic, and it is very small compared to the initial volume of the sample.
Thus, the ferrite fraction can be thought to be proportional to the relative dilatation during the
isothermal transformation [23,24]. With the volume fraction of ferrite at the end of isothermal
holding determined from the optical micrographs, the time evolution of ferrite fraction can be
evaluated from the measured dilatometric data during the isothermal holding. Fig. 4 shows time
evolution curves of ferrite volume fraction for both alloys reacted at 1048 K (775
o
C), which is
estimated from the corresponding dilatometric data. The measured ferrite fraction by quantitative
metallography is also shown for comparison. The calculated curve agrees well with the
experimental values, suggesting that the evaluation of ferrite volume fraction as described above
is reliable. The time evolutions of ferrite volume fraction at various isothermal temperatures for
9

Fe-C-Mn-Si and Fe-C-Mn-Si-Al alloys are shown in Figs. 5a and 5b, respectively. For Fe-C-Mn-
Si alloy (Fig. 5a), the ferrite transformation is sluggish at temperatures higher than 1068 K (795
o
C), but the ferrite transformation is accelerated subsequently below 1058 K (785
o
C). One can
see a wide gap between 1068 and 1058 K (795 and 785
o
C). For Fe-C-Mn-Si-Al alloy (Fig. 5b),
a similar gap is also observed between 1088 and 1078 K (815 and 805
o
C), which is 20
o
C higher
than that in Fe-C-Mn-Si alloy. It indicates there would be a kinetics transition from slow to fast
transformation at those temperature intervals. The calculated PLE/NPLE transition temperatures
for Fe-C-Mn-Si and Fe-C-Mn-Si-Al alloys in the present study are 1068 K and 1088 K (795
o
C
and 815
o
C), respectively. The transition temperatures of apparent growth behavior of ferrite
accord well with the critical temperatures for ferrite growth without partitioning of alloying
elements. This suggests the change in growth kinetics monitored in the microstructural
observation is related closely with the occurrence of redistribution of alloying elements during
transformation as reported in the earlier study [14].
The partitioning behavior of Mn is examined in terms of its concentration in ferrite depending on
the isothermal temperature as shown in Fig. 6. The o/(o+) terminus compositions of Mn, Si,
and Al calculated assuming local equilibrium are also shown. It is seen that Mn partition occurs
above 1068 K (795
o
C) for Fe-C-Mn-Si alloy, while it does above 1088 K (815
o
C) for Fe-C-Mn-
Si-Al alloy. The critical temperatures for partitioning of Mn are just about at the NPLE/PLE
boundary, in agreement with assumption of local equilibrium at the interface. Although Si and Al
could not be reliably detected in this study, the partition of Si and Al is presumed because the
diffusivity of Si and Al is appreciably faster than that of Mn at a given temperature both in
austenite and ferrite phase [25]. As a ferrite stabilizer, Si and Al would partition to ferrite in the
10

PLE region according to the calculation. Indeed, Si and Al enrichment in ferrite has been
reported in Fe-C-Si and Fe-C-Al ternary alloys [26].

4.2 Comparison of observed growth kinetics with calculation
After the nucleation at the austenite grain boundaries, the thin layers of ferrite thicken
parabolically. The growth can be described by the following equation,


R
2
= o
2
(t t
0
) [4]
where R is the half-thickness of ferrite allotriomorphs, t is the isothermal reaction time and t
0
is
the incubation time for ferrite nucleation. The parabolic growth rate constant can be obtained
from the slope in R
2
vs t plot. In addition, at a large undercooling from Ae
3
temperature, ferrite
growth is so fast that the impingement of carbon diffusion fields associated with different ferrite
allotriomorphs is possibly attained within a rather short reaction time. Only the data at those
short reaction times, when the grain boundary allotriomorphs thickness could be identified
without impingement, were used to obtain the parabolic growth rate constants. Fig. 7 plots the
square of ferrite half-thickness against the isothermal holding time for both alloys isothermally
reacted at 1048 K (775
o
C). Parabolic growth rate constants were obtained from the slopes of the
plots. Figs. 8a and 8b compares the measured and calculated parabolic growth rate constant in
Fe-C-Mn-Si and Fe-C-Mn-Si-Al alloys, respectively. In the PLE region, the possible range of
growth rate constant was evaluated from the diffusivities of Mn, Si and Al. It is seen from the
calculations that the addition of 0.2 mass pct Al significantly raises Ae
3
temperature of Fe-C-
Mn-Si alloy, and thus, the driving force for ferrite transformation is also increased at the same
reaction temperature. As a result, the calculated parabolic growth rate constant of Fe-C-Mn-Si-Al
alloy is larger than that of Fe-C-Mn-Si alloy, which is also in agreement with the experimental
11

observation. In addition, Al raises the NPLE/PLE transition temperatures. As mentioned, the
NPLE/PLE transition temperature is 1068 K and 1088 K (795
o
C and 815
o
C) for Fe-C-Mn-Si
and Fe-C-Mn-Si-Al alloys, respectively. This implies that the growth mode will be different for
both alloys in the temperature range between 1068 K and 1088 K (795
o
C and 815
o
C). It
possibly leads to approximately 3 times larger growth rate constant of Fe-C-Mn-Si-Al alloys
than that of Fe-C-Mn-Si alloy in that temperature interval.
Meanwhile, in both alloys investigated, the measured parabolic growth rate constants fall
between the paraequilibrium and local equilibrium predictions. The measured growth rate
constants at higher temperatures where Mn partition was observed are larger by 1~2 orders of
magnitude than the values calculated from the PLE condition. At lower temperatures where Mn
was unpartitioned, the measured growth rate constants are smaller than the calculated ones under
the PE condition but approach the calculated ones as the reaction temperature decreases. In Fe-
C-Mn ternary alloys, it is consistently reported that the measured growth rate is significantly
higher than predicted one with PLE condition in the partitioned growth region but smaller than
that under PE condition in the unpartitioned growth region [27,28]. In Fe-C-Si and Fe-C-Al
alloys, however, it is reported that the measured growth rate constant agrees well not only with
the paraequilibrium calculation but also with the local equilibrium one because the growth rate
constants calculated under the PE and NPLE are very similar [15]. The temperature dependence
of growth kinetics in the investigated Fe-C-Mn-Si and Fe-C-Mn-Si-Al alloys is comparable to
that in Fe-C-Mn ternary alloy. It indicates that the presence of Mn in both alloys possibly has a
critical effect on the growth kinetics of ferrite. It is reported that the growth of ferrite can be
faster than calculated one in the region of partitioned growth because substitutional alloying
elements diffuse along the grain boundary, the so-called rejector/collector plate mechanism [27].
12

On the other hand, in the region of unpartitioned growth, the observed growth is slower than the
prediction with paraequilibrium possibly due to the drag effect of substitutional alloying
elements [17]. Here, the deceleration of growth kinetics due to the solute drag effect is examined
in more detail by expanding the previous formalism into Fe-C-Mn-Si-Al multi-component
system [17,29].
If the interaction oI the substitutional solute with the moving / interIace is signiIicantly strong,
the substitutional element will tend to accumulate there and to move with the interface. A drag of
alloying element then retards the migration of the / interIace. Solute drag eIIect has been
proposed [13,16,17,29] as a necessary complement to the diffusion-controlled growth theory.
Following the previous studies [17,29], a triangular potential well is assumed for interaction of
solute with / interIace. The chemical potential of solute in the boundary region (in reference to
the chemical potential in ferrite) is given by
) 1 ( ) (ln RT
4
1
o
c
s
E E x x
i i
j
j ij i i
+ A + + + =

=
[5]
Then the solute drag force is calculated using the Hillert-Sundman model [30] from the equation,
ds
ds
d
J
S
V
G
i
i
i
m

}
=

= A
4
2
drag
o
o

[6]
where x
i
is the concentration of the ith solute (i=0, 1, 2, 3 and 4 for Fe, C, Mn, Si, Al,
respectively),
ij
is the Wagners interaction parameter in austenite between the ith and jth solutes,
E
i
is the binding energy oI solute to the / interIace, E
i
is a half of the difference in the
reference free energy of the ith solute between ferrite and austenite, V
m
is the molar volume of Fe
atom, S

is the velocity of interface, J

i
is the diffusion flux of solute through the interface and
i

is the chemical potential of the ith element (i=0, 1, 2, 3 and 4 for Fe, C, Mn, Si, Al, respectively).
13

In the present calculation, the binding energy of Mn, Si and Al to the / interIace was assumed
to be 8,10 and 12 kJ/mol, respectively [29, 31]. Some Wagner interaction parameters between
solutes were taken from Ref. [32] and the others were estimated with the help of ThermoCalc
software*. They were summarized in Table II. The diffusivities of alloying elements across the
*The Wagners interaction parameter in austenite between the ith and jth solutes is defined as
ij
=ln
i
/x
i
, where

i
is the activity coefficient of the ith solute in austenite. ln
i
/x
i
can be calculated by using ThermoCalc software.
Assuming
ij
is inverse proportion to the temperature [32], the temperature dependence of
ij
can be obtained via
regression analysis.

interface are not well known. It is assumed to be equal to the half of the volume diffusivity in
ferrite following recent reports [33,34] that the cross-boundary diffusivity fall between the values
of the volume diffusivity in ferrite and austenite. The volume diffusivity of alloying element in
ferrite was taken from the compilation by Fridberg et al. [25]. From Table II, one can notice
there is an attractive interaction between Mn and Si atoms, while a repulsive interaction exists
between Si and Al atoms. Compared with Si-Al interaction, the repulsive interaction between
Mn and Al atoms is weak. Figs. 9a and 9b show the calculated free energy dissipation by solute
drag (G
drag
) at 1023 K (750
o
C) for Fe-C-Mn-Si and Fe-C-Mn-Si-Al alloys, respectively. The
free energy dissipations by Mn, Si and Al drag pass through maximum values, however, the
growth rate at the maximum G
drag
varies depending on the alloying element because the cross-
boundary diffusivity D
b
(~0.5D

) is different for each alloying element. For Fe-C-Mn-Si alloy,

the sum of the free energy dissipation by Mn and Si also passes through a maximum as if the free
energy dissipation occurred by drag of one solute with a higher binding energy, which indicates
that co-segregation of Mn and Si enhances the drag effects. When Al was added to the Fe-C-Mn-
14

Si alloy, the free energy dissipations by Mn and Si drag were slightly decreased due to the Mn-
Al and Si-Al repulsive interaction. Consequently, the maximum of the overall G
drag
in Fe-C-
Mn-Si-Al alloy is smaller than that in Fe-C-Mn-Si alloy. The peak of the overall G
drag
shows a
plateau in Fe-C-Mn-Si-Al alloy due to the large difference in cross-boundary diffusivities of Mn,
Si and Al.
In Figs. 10a and 10b, the growth rate v
max
at which the total drag force becomes maximum is
compared with the measured and calculated growth rates (one-half of the parabolic growth rate
constant, corresponding to the growth rate at t=1s). In both alloys observed, at the two higher
temperatures (white circles) where the partition of Mn was experimentally observed, the
measured growth rates are much lower than v
max
, which means that the drag force will be
considerably small. Furthermore, the drag force will become even smaller with the elapse of
reaction time because the growth rate will decrease. Therefore, the solute drag force is unlikely
to be significant at these higher temperatures. In contrast, at lower temperatures (black circles),
the measured growth rates are larger than v
max
except for 1068 K (795 C) in the Fe-C-Mn-Si-Al
alloy. In these conditions, the drag force will increase with time and reach to the maximum.
Hence, the solute drag possibly has a substantial influence on the ferrite growth by the
dissipation of the free energy in the moving / interIace. In Table III, the driving forces
attending the Ierrite transIormation assuming the PE, G

, are compared with the largest drag

forces,
M
drag
G A , in Fe-C-Mn-Si and Fe-C-Mn-Si-Al alloys at the temperatures where the solute
drag is thought to be important. Both G

and
M
drag
G A are increased with lowering the reaction
temperature. However, the ratios of
o
A A G /
M
drag
G are decreased with the decrease in the
reaction temperature. This implies the solute drag will become less dominant as the temperature
15

decreases. It accords with the gradual approaches of the measured growth rate constants to the
calculated ones under the PE condition with decreasing the reaction temperature (see Figs. 8a
and 8b).

5. Conclusions
The growth kinetics of proeutectoid ferrite at early stages of transformations were studied in Fe-
0.1C-1.5Mn-1Si quaternary and Fe-0.1C-1.5Mn-1Si-0.2Al quinary alloys.
(1) The kinetic transition temperature from partitioned slow to unpartitioned fast growth of
ferrite was determined through the time evolution of ferrite fraction and the redistribution Mn
during the ferrite formation. The observed transition temperature of the alloys investigated
agreed fairly well with the PLE/NPLE transition temperature predicted from the local
equilibrium model.
(2) The growth rate of ferrite at a given temperature was higher in Fe-C-Mn-Si-Al alloy because
the addition of 0.2 mass pct Al raised significantly the Ae
3
temperature and the kinetic transition
temperature from partitioned to unpartitioned growth of ferrite.
(3) For both alloys, the measured parabolic growth rate constants in unpartitioned temperature
region were smaller than the calculated ones with the PE model, but the discrepancy between the
measured and calculated ones gradually decreased with decreasing the reaction temperature. It
could be understood with the drag effect of substitutional solutes.

Acknowledgments
G.H. Zhang is grateful to professor N. J. Kim for the provision of laboratory facilities at GIFT,
POSTECH. D.W. Suh appreciates the financial support from POSCO through Steel Innovation
16

Program. The authors are grateful to Professor H.K.D.H. Bhadeshia for careful reading of this
manuscript.















17

References
1. C. Zener: Trans. AIME, 1946, vol.167, pp.550-95.
2. G.R. Purdy and J.S. Kirkaldy: Trans TMS-AIME, 1963, vol. 227, pp. 1255-66.
3. S. Crusius, L. Hogland, U. Knoop, G. Inden, and J. Agren: Z. Metallkd., 1992, vol. 83, pp.
729-38.
4. M. Hillert: The Mechanism of Phase Transformations in Crystalline Solids, Institute of
Metals, London, 1969, p.231-47.
5. J.S. Kirkaldy: Can. J. Phys., 1958, vol. 36, pp. 907-16.
6. G.R. Purdy, D.H. Weichert, and J.S. Kirkaldy: Trans. TMS-AIME, 1964, vol. 230, pp.1025-34.
7. D.E. Coates: Metall. Trans., 1972,vol. 3, pp.1203-12.
8. D.E. Coates: Metall. Trans., 1973, vol. 4, pp.1077-86.
9. D.E. Coates: Metall. Trans. 1973, vol. 4, pp. 2313-25.
10. A. Hultgren: Trans. ASM, 1947, vol. 39, pp. 915-1005.
11. M. Hillert: Paraequilibrium, Internal Report, Swedish Institute for Metal Research,
Stockhelm, Sweden, 1953.
12. K.R. Kinsman and H.I. Aaronson: Metall. Trans., 1973, vol. 4, pp. 959-67.
13. J. R. Bradley and H. I. Aaronson: Metall Mater Trans A, 1981, vol. 12A, pp.1729-41.
14. K. Oi, C. Lux, and G. R. Purdy: Acta Mater., 2000, vol. 48, pp. 2142-57.
15. M. Enomoto: Trans. ISIJ, 1988, vol. 28, pp.826-35.
16. G. R. Purdy, and Y. J. M. Brechet, Acta metall. Mater, 1995, vol. 43, 3763-74
17. M. Enomoto: Acta Mater. 1999, vol. 47, 3533-40.
18. T. Tanaka, H.I. Aaronson, and M. Enomoto: Metall. Mater. Trans. A, 1995, vol. 26, pp. 561-
80.
18

19. H. Guo and M. Enomoto, Metall. Mater. Trans. A, 2006, vol. 37A, pp. 1721-29.
20. R. Wei, K. Kanno, and M. Enomoto: Metall. Mater. Trans. A, 2011, vol. 42A, pp. 2189-98.
21. D. W. Suh, S.J. Park, C. S. Oh, and S. J. Kim: Script. Mater., 2007, vol. 57, pp. 1097-1100.
22. D. W. Suh, S.J. Park,T. H. Lee, C. S. Oh, and S. J. Kim: Metall. Mater. Trans. A, 2010, vol.
41A, pp. 397-408.
23. X. Sun, H. Dong, Q. Liu, and Y. Weng: Mater. Sci. Eng. A, 2008, vol.487, pp. 93-98.
24. Y. Liu, D. Wang, F. Sommer, and E. J. Mittemeijer: Acta Mater., 2008, vol. 56, pp. 3833-42.
25. J. Fridberg, L.-E. Torndahl, and M. Hillert: Jernkontorets Ann., 1969, vol. 153, pp. 263-76.
26. H. I. Aaronson and H.A. Domain: Trans. TMS-AIME, 1966, vol.236, pp.781-96.
27. M. Enomoto and H. I. Aaronson: Metall Mater Trans A, 1987, vol. 18A, pp.1547-57.
28. G. H. Zhang, M. Enomoto, N. Hosokawa, M. Kagayama, and Y. Adachi: J. Magn. Magn.
Mater., 2009, vol. 321, pp. 4010-16.
29. H. Guo and M. Enomoto: Metall. Mater. Trans. A, 2007, vol. 38A, pp. 1152-61.
30. M. Hillert and B. Sundman: Acta metal., 1976, vol. 24, pp. 731-43.
31. M. Enomoto, N. Nojiri, and Y. Sato: Mater. Trans. JIM, 1994, vol. 35, pp.859-67.
32. J.S. Kirkaldy, B.A. Thomson, and E.A. Baganis: in Hardenability Concepts with
Applications to Steel, D.V. Doane, and J.S. Kirkaldy, eds., TMS-AIME, Warrendale, PA,
1978, pp. 82125.
33. C.R. Hutchison, A. Fuchsmann and Y. Brechet: Metall. Mater. Trans. A, 2004, vol. 35A, pp.
1211-21.
34. A. Beche, H. S. Zurob, and C. R. Hutchison: Metall. Mater. Trans. A, 2007, vol. 38A, pp.
2950-55.

19






Table . Chemical compositions of the alloys used in the present study (mass pct.)



Table II. Wagner interaction parameters in austenite used in present calculation











Alloy C Mn Si Al Ae
3
(ortho) Ae
3
(para)
Fe-C-Mn-Si 0.1 1.5 0.94 857 832
Fe-C-Mn-Si-Al 0.1 1.5 1.0 0.22 894 861
(at 750
o
C)
C-Mn -5070/T
[29]
-5.0
C-Si 4.84+7370/T
[29]
12
C-Al 10-14185/T -3.8
Mn-Al 2.9+1319/T 4.2
Si-Al 4.5+26928/T 21.8
Mn-Si -11+1215/T -10
20






Table III Comparison of driving force and the largest drag force










Alloys
Reaction
temperature,
o
C
Driving force,
G

, J/mol
The largest
drag force,
M
drag
G A , J/mol

o
A
A
G
G
M
drag

Fe-C-Mn-Si
775 -90 97 1.08
750 -140 104 0.74
730 -188 112 0.60
Fe-C-Mn-Si-Al
795 -88 84 0.95
775 -122 88 0.72
750 -174 94 0.54
21

Figure captions:
Fig.1 Schematic illustration of the construction for the operative interfacial tie-line and the
corresponding C and X proIiles in and during growth oI Ierrite in an FeCX alloy (X is a
austenite stabilizer) under (a) PLE, (b) NPLE and (c) PE conditions. The bulk alloy compositions
are designated by the solid circle in each case.

Fig.2 Optical micrographs of Fe-C-Mn-Si alloy isothermally transformed for 30 min at (a) 1078
K (805
o
C), (b) 1068 K (795
o
C), (c) 1058 K (785
o
C) , and (d) 1048 K (775
o
C).

Fig. 3 Optical micrographs of Fe-C-Mn-Si-Al alloy isothermally transformed for 30 min at (a)
1098 K (825
o
C), (b) 1088 K (815
o
C), (c) 1078 K (805
o
C), and (d) 1068 K (795
o
C).

Fig. 4 Time evolution of the ferrite volume fraction at 1048 K (775
o
C) (Solid lines were
estimated from dilatometric data and the symbols were determined by metallographic
measurements).

Fig. 5 Time evolution of the ferrite volume fraction at various temperatures in (a) Fe-C-Mn-Si
and (b) Fe-C-Mn-Si-Al alloys.

Fig. 6 (a) Mn, (b) Si and Al concentrations in ferrite calculated from local equilibrium model
(Open and solid symbols in (a) represent the measured Mn concentration in ferrite in Fe-C-Mn-
Si and Fe-C-Mn-Si-Al alloys, respectively).

22

Fig. 7 Square of ferrrite half-thickness plotted against the isothermal holding time at 1048 K
(775 C).

Fig. 8 Comparison of measured and calculated parabolic growth rate constants of ferrite in (a)
Fe-C-Mn-Si and (b) Fe-C-Mn-Si-Al alloys.

Fig. 9 Calculated free energy dissipation by solute drag at 1023 K (750
o
C) in (a) Fe-C-Mn-Si
and (b) Fe-C-Mn-Si-Al alloys.

Fig. 10 Comparison of measured and calculated growth rate of ferrite overlapped with v
max
in (a)
Fe-C-Mn-Si and (b) Fe-C-Mn-Si-Al alloys.












23
















(a)







24


















(b)





25
















(c)

Fig.1 Schematic illustration of the construction for the operative interfacial tie-line and the
corresponding C and X proIiles in and during growth oI Ierrite in an FeCX alloy (X is a
austenite stabilizer) under (a) PLE, (b) NPLE and (c) PE conditions. The bulk alloy compositions
are designated by the solid circle in each case.


26






















Fig.2 Optical micrographs of Fe-C-Mn-Si alloy isothermally transformed for 30 min at (a) 805
o
C,
(b) 795
o
C, (c) 785
o
C, and (d) 775
o
C.
100m
(d)
775
o
C
(c)
100m
785
o
C
805
o
C
(a)
100m
795
o
C
(b)
100m
27





















Fig. 3 Optical micrographs of Fe-C-Mn-Si-Al alloy isothermally transformed for 30 min at (a)
825
o
C, (b) 815
o
C, (c)805
o
C, and (d)795
o
C.

(a)
100m
825
o
C
(b)
100m
815
o
C
(c)
100m
805
o
C
795
o
C
100m
(d)
28
















Fig. 4 Time evolution of the ferrite volume fraction at 775
o
C (Solid lines were estimated from
dilatometric data and the symbols were determined by metallographic measurements).








29











(a)











(b)
Fig. 5 Time evolution of the ferrite volume fraction at various temperatures in (a) Fe-C-Mn-Si
and (b) Fe-C-Mn-Si-Al alloys.

30











(a)










(b)
Fig. 6 (a) Mn, (b) Si and Al concentrations in ferrite calculated from local equilibrium model
(Open and solid symbols in (a) represent the measured Mn concentration in ferrite in Fe-C-Mn-
Si and Fe-C-Mn-Si-Al alloys, respectively).
31

















Fig. 7 Square of ferrrite half-thickness plotted against the isothermal holding time at 775C.







32










(a)










(b)
Fig. 8 Comparison of measured and calculated parabolic growth rate constants of ferrite in (a)
Fe-C-Mn-Si and (b) Fe-C-Mn-Si-Al alloys.

33











(a)









(b)
Fig. 9 Calculated free energy dissipation by solute drag at 750
o
C in (a) Fe-C-Mn-Si and (b) Fe-
C-Mn-Si-Al alloys.

34











(a)










(b)
Fig. 10 Comparison of measured and calculated growth rate of ferrite overlapped with v
max
in (a)
Fe-C-Mn-Si and (b) Fe-C-Mn-Si-Al alloys.