Anda di halaman 1dari 8

See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/233640258

Ferrite/pearlite band formation in hot rolled


medium carbon steel

Article in Materials Science and Technology · March 2002


DOI: 10.1179/026708301225000752

CITATIONS READS

44 4,051

5 authors, including:

Sven Erik Offerman N. H. van Dijk


Delft University of Technology Delft University of Technology
68 PUBLICATIONS 1,271 CITATIONS 187 PUBLICATIONS 2,775 CITATIONS

SEE PROFILE SEE PROFILE

M. Theo Rekveldt Jilt Sietsma


Delft University of Technology Delft University of Technology
276 PUBLICATIONS 2,843 CITATIONS 414 PUBLICATIONS 6,030 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Expanding the functionality of piezo-particulate composites View project

SESANS View project

All content following this page was uploaded by N. H. van Dijk on 22 May 2014.

The user has requested enhancement of the downloaded file.


Ferrite/pearlite band formation in hot rolled
medium carbon steel
S. E. Offerman, N. H. van Dijk, M. Th. Rekveldt, J. Sietsma, and S. van der Zwaag

The in¯ uence of the microsegregation of Mn, Si, and Cr on the austenite decomposition during isothermal
transformation s in hot rolled medium carbon steel has been studied by neutron depolarisation, electron probe micro-
analysis (EPMA), and optical microscopy. Eight specimens of the same alloy were held at 1173 K for 30 min and
were rapidly cooled to different isothermal transformatio n temperatures. Two-dimensional EPMA maps of the
specimen annealed at 1013 K showed that microsegregation of alloying elements in hot rolled steel is strongly
related to the ferrite/pearlite band formation. The local variations in alloying element concentration lead to
variations in local transition temperatures, which were calculated with the thermodynami c database MTDATA.
Similar EPMA maps for the specimen transformed at 953 K demonstrate the presence of microchemical bands,
while optical microscopy reveals the absence of microstructural bands. It is shown that the formation of micro-
chemical bands is a prerequisite for band formation, but that the kinetics of the phase transformatio n determines
the actual formation of microstructural bands. A quantitative model has been developed, which describes the
observations in terms of the relative difference between ferrite nucleation rates in regions with a high and low
local undercooling and the subsequent growth of the ferrite. The isothermal transformatio n experiments have led
to generalised nucleation and growth criteria for the formation of microstructural bands. MST/5150

Ir Offerman (S.E.Offerman@iri.tudelft.nl), Dr van Dijk, and Dr Rekveldt are in the Interfaculty Reactor Institute, Delft
University of Technology, Mekelweg 15, 2629 JB Delft, The Netherlands and Dr Sietsma and Professor van der Zwaag are
in the Department of Materials Science and Engineering, Delft University of Technology, Rotterdamseweg 137, 2628 AL Delft,
The Netherlands. Manuscript received 9 July 2001; accepted 2 August 2001.
# 2002 IoM Communication s Ltd.

nucleation process, which predominantly takes place at the


Introduction austenite grain boundaries. In steels with large austenite
grains only a limited number of ferrite nucleation sites are
Ferrite/pearlite banding is a common phenomenon in hot available, which is not suf® cient to form ferrite bands.
rolled low alloy steels, which has been known for more than The sequence of events that results in the formation of
forty years.1,2 Recent studies reveal a renewed scienti® c and bands starts with the solidi® cation process. The alloying
technological interest in the origin of microstructural elements having partition coef® cients less than unity (e.g.
banding.3 ± 6 The technological importance of ferrite/pearlite Mn, Si, S, and P) are rejected from the ® rst formed d ferrite
banding is related to the reduced resistance to hydrogen dendrites, resulting in interdendritic regions of high solute
induced cracking and the reduced impact toughness after content, which is retained during the transformation from
an intercritical heat treatment and subsequent rapid cooling d ferrite to austenite. Subsequent hot rolling in the aus-
of these steels.3 The scienti® c interest arises from the aniso- tenitic condition leads to elongated high solute regions. The
tropic behaviour of steels with microstructural bands. A variation in alloying element concentrations may cause an
recent study showed the anisotropic dilatation of steels with inhomogeneous distribution of the carbon in the austenite.2
ferrite/pearlite bands,6 which signi® cantly alters the inter- Certain elements (e.g. Mn and Cr) effectively attract car-
pretation of dilatometry experiments. It can be expected bon, while other elements reject carbon (e.g. Si). Owing to
that microstructural bands also in¯ uence the interpretation its relatively high diffusivity, the carbon in the austenite can
of neutron depolarisation (ND) experiments on the phase be distributed according to thermodynamic equilibrium.
transformation in steel.7 Calculations of Verhoeven 8 have shown that variations in
Microstructural banding is the term used to describe a Mn concentration DCMn result in variations in carbon
microstructure consisting of alternating bands of pro- concentration DCC, which can signi® cantly alter the local A3
eutectoid ferrite and pearlite (or martensite for dual phase transition temperature. The variation in alloying element
steels). Banding mainly depends on three factors: the micro- concentration by itself also raises (e.g. Si) or lowers (e.g. Mn
segregation of alloying elements, the cooling rate (or in and Cr) the A3 transition temperature.
general the imposed temperature pro® le) during the The ferrite starts to nucleate in the regions with a high
transformation, and the austenite grain size. Grossterlinden A3 transition temperature, which causes the carbon to
et al.3 and Thompson and Howell 4 have shown by means of redistribute. Entire grains of austenite transform to pro-
electron probe microanalysis (EPMA) that the ferrite/ eutectoid ferrite, which grows across austenite grain
pearlite bands are strongly related to the microsegregation boundaries,4 resulting in the `bamboo’ structure. The
of alloying elements. Removal of the microsegregation of carbon is piled up in regions with a low A3 transition tem-
alloying elements by annealing for several hours at a high perature. The increase in carbon content will lower the
austenisation temperature eliminates the microstructural local A3 transition temperature even further. Eventually,
bands. Both studies demonstrate that the occurrence of the composition in these regions becomes eutectoidic and
microstructural bands also depends on the cooling rate. A pearlite can form if the temperature is below the A1
low cooling rate results in strong banding, while a high transition temperature.
cooling rate results in the absence of bands. Microstructural Despite the broad interest, mainly qualitative descrip-
banding does not occur when the austenite grain size is tions on the formation of bands can be found in literature.
larger than two or three times the microchemical band The ® rst quantitative description of the formation of bands
spacing.4 This result indicates the importance of the ferrite was given by Kirkaldy et al.,2 who gave a semi-empirical

DOI 10.1179/026708301225000752 Materials Science and Technology March 2002 Vol. 18 297
298 Offerman et al. Ferrite/pearlite band formation in hot rolled medium carbon steel

expression for the critical cooling rate. Thirty years later


Grossterlinden et al.3 developed a ® nite difference analysis
to predict the distance a planar austenite/ferrite interface
can move as a function of the cooling rate, and was able to
predict a critical cooling rate for the steels studied. Little
research has been done on the effect of isothermal trans-
formations on the formation of microstructural bands, and
hence there are no models that predict the formation of
bands during annealing at a constant temperature.9 The
present research aims to provide such a model by quan-
tifying nucleation and growth criteria for the isothermal
austenite to ferrite phase transformation in steels showing
microsegregation of alloying elements.

Experimental

The hot rolled medium carbon steel used had the following
chemical composition: Fe ± 0.364C ± 0 .305Si ± 0 .656Mn ±
0.014P ± 0.021S ± 0.177Cr ± 0.016Mo ± 0.092Ni ± 0.226Cu ± 0.017Sn
(wt-%). Eight specimens of the steel were heated in a nitro-
gen atmosphere to 1173 K and held for 30 min, in order to
form a single phase austenite structure. The specimens were
cooled at a rate of 20 K s­ 1 to 953, 961, 975, 987, 1000,
1013, 1023, and 1031 K respectively, subsequently held 1 Imposed temperature T pro® le and ferrite fraction f
there for ~ 4 .5 h, and then furnace cooled to room tem- as function of time t for specimen annealed at 1013 K:
perature. The specimen dimensions are 100 615 60 .4 mm, measured with neutron depolarisation
which is suitable for ND during isothermal transformation.
The rolling direction is along the long axis of the specimen.
The specimens were studied with optical microscopy and in Fig. 1 for the specimen that was annealed at 1013 K. The
ND. ® gure shows that only pro-eutectoid ferrite forms during the
As part of a larger programme on the kinetics of ferrite isothermal annealing stage. Pearlite starts to form below
formation the isothermal transformation kinetics of this the A1 transition temperature during continuous cooling to
steel was investigated using ND. This ND technique has the room temperature. Similar transformation characteristics
intrinsic advantage over other techniques such as dilato- were observed for the specimens that were annealed at 975,
metry that it not only provides the total ferrite fraction 987, 1000, 1023, and 1031 K. Figure 2 shows the distribu-
transformed, but also the average ferrite grain size.10 This tion of Mn, Si, and Cr as measured with EPMA, compared
work only makes use of the fractional transformation data with the ® nal microstructure at room temperature. The
only. The ND measurements have been performed at the ® gure shows that the pearlite bands coincide with the solute
PANDA instrument at the nuclear reactor of the Inter- rich regions, while the ferrite bands are present in the solute
faculty Reactor Institute. During the heat treatment, the lean regions. The average wavelength of the chemical bands
transmission of a polarised monochromatic neutron beam, is ~ 96 mm. The variation in alloy element concentration
with a wavelength of l ~1 .240 + 0 .012 AÊ , is directed along across several ferrite/pearlite bands of the specimen that was
the short axis of the specimen and an external magnetic ® eld annealed at 1013 K is shown in Fig. 3. The minimum and
of m oH~5 .16 mT is applied along the long axis of the speci- maximum alloying element concentration that was mea-
men. The average rotation of the polarisation vector of the sured with EPMA is given in Table 1. These concentra-
neutron beam was measured during annealing, which yields tion differences originated from the (partial) rejection of
the ferrite fraction in the specimen as a function of time.7,11 alloying elements during the early stages of d ferrite solidi-
The full results of the ND investigation on the ferrite ® cation.
formation kinetics will be published in a subsequent paper. The imposed temperature pro® le and the ferrite fraction
The microsegregation of Mn, Si, and Cr in this steel has f as a function of time, which was measured with ND, are
been studied with EPMA. Only the specimens that were shown in Fig. 4 for the specimen that was annealed at
annealed at 953 and 1013 K were investigated with this 953 K. The ® gure shows that the equilibrium fraction pro-
technique. The EPMA measurements have been performed eutectoid ferrite is formed within the ® rst 2 min of the
on a Jeol JXA 8900R microprobe with a wavelength dis- isothermal annealing stage. The complete transformation
persive spectrometer. Two-dimensional scans were made to pearlite takes place in the next 3 min. The only other
with a step size of 2 mm over an area of 512 6512 mm 2, specimen that showed similar transformation characteristics
which gives qualitative information about the local con- was the one annealed at 961 K. The equilibrium fraction of
centrations of Mn, Si, and Cr. Two additional sets of ® ve pro-eutectoid ferrite of this specimen was reached after
line scans were taken with a step size of 3 mm over a distance 7 min of isothermal annealing. The 2D EPMA scans in
of 300 mm, which give quantitative information about the
local concentrations of Mn, Si, and Cr. The specimens were
prepared for optical microscopy by grinding, polishing, and Table 1 Minimum and maximum concentration of alloy-
etching in 2 vol.-% nital for 20 s. Optical micrographs were ing elements as measured with electron probe
taken of the areas studied with EPMA. microanalysis (EPMA) and calculated with
MTDATA, wt-%
Element MTDATA EPMA
Results
Mn 0.50 ± 0.88 0.57 ± 0.87
Si 0.19 ± 0.50 0.25 ± 0.48
The imposed temperature pro® le and the ferrite fraction f as Cr 0.16 ± 0.20 0.10 ± 0.22
a function of time, which was measured with ND, are shown

Materials Science and Technology March 2002 Vol. 18


Offerman et al. Ferrite/pearlite band formation in hot rolled medium carbon steel 299

a optical micrograph of area (5126512 mm) at which 2D electron probe microanalysis scans were taken; b Mn distribution; c Si distri-
bution; d Cr distribution
2 Microstructure and Mn, Si, and Cr distribution (same area) of specimen annealed at 1013 K: light regions indicate
high concentration

Fig. 5 demonstrate the presence of chemical bands in the


specimen annealed at 953 K, while the optical micrograph
reveals the absence of ferrite/pearlite bands. Although it
is hard to quantify the degree of banding,12,13 it was con-
Concentration (wt-%)

cluded from the micrographs that for isothermal transfor-


mation temperatures of 961 K and higher, microstructural
Temperature (K)

3 Optical micrograph with electron probe microanalysis


line scan (white line, 300 m m) related to Mn, Si, and Cr
distribution and calculated local variation in A3, A +1 , 4 Imposed temperature T pro® le and ferrite fraction f as
and A ­1 temperatures with MTDATA for specimen function of time t for specimen annealed at 953 K:
annealed at Ttrans ~1013 K measured with neutron depolarisation

Materials Science and Technology March 2002 Vol. 18


300 Offerman et al. Ferrite/pearlite band formation in hot rolled medium carbon steel

a optical micrograph of area (5126512 mm) at which 2D electron probe microanalysis scans were taken; b Mn distribution; c Si distri-
bution; d Cr distribution
5 Microstructure and the Mn, Si, and Cr distribution (same area) of specimen annealed at 953 K: light regions indicate
high concentration

bands were observed. In Fig. 6 the concentration variations


of the alloying elements across several ferrite/pearlite bands
are shown for the specimen annealed at 953 K, which is
essentially the same as the other specimen investigated by
Concentration (wt-%)

EPMA.

Discussion

DISTRIBUTION OF CARBON IN AUSTENITE


The minimum and maximum concentrations of Mn, Si, and
Cr that result from the solidi® cation process were calculated
with MTDATA* and are given in Table 1. The calculated
values (except the minimum concentration of Cr) agree well
with the concentrations that were measured with EPMA.
Temperature (K)

Based on the method developed by Brown and Kirkaldy 14


and the observed concentration ranges (Table 1), the varia-
tion in carbon concentration in the austenite phase DCC was
calculated as a result of the Mn, Si and Cr segregation. The
differences between the maximum and minimum in
manganese DCMn, silicon DCSi, and chromium DCCr
concentration result in
DCC ~0:061 DCM n {0 :138 DCSi + 0:156DCC r
&0:005 wt-% : : : : : : : : : : : (1)
for a nominal carbon concentration of 0 .364 wt-%. The
in¯ uence of the Mn and Cr segregation on the variation in
carbon concentration in the austenite is partly cancelled by
6 Optical micrograph with electron probe microanalysis the opposite effect of the Si segregation for the given
line scan (white line 300 m m) related to Mn, Si, and Cr composition. Since DCC is small in comparison with CC,
distribution and calculated local variation in A3, A +1 , effectively, a homogeneous distribution of carbon in the
and A ­1 temperatures with MTDATA for specimen
annealed at Ttrans ~953 K *MTDATA is a registered trademark.

Materials Science and Technology March 2002 Vol. 18


Offerman et al. Ferrite/pearlite band formation in hot rolled medium carbon steel 301

NUCLEATION AND GROWTH MODEL


The experimental data of Figs. 1 ± 6 show that the
formation of ferrite/pearlite bands depends on the presence
of both chemical bands and the isothermal transformation
temperature. The fundamental processes that govern the
formation of bands are the ferrite nucleation and growth
processes during the phase transformation. In this section
the nucleation and growth criteria are derived that have to
be ful® lled for the formation of bands.
In the classical nucleation theory, the rate of ferrite
nucleation is given by15,16
dN kT DG ¤ l QD
~Nn exp { exp { : : : (5)
dt h kT kT
where N is the number of nuclei, Nn the number of potential
nucleation sites, k~1 .38 610 ­ 23 J K ­ 1 (the Boltzmann
constant), h~6 .626 610 ­ 34 J s (Planck’ s constant), l ~
10­ 4 a scaling factor,17 and QD the activation energy for
self-diffusion. The ferrite nucleation, which takes place at
7 Relative difference of ferrite nucleation rate r as func-
tion of isothermal transformation temperature T in
the austenite/austenite grain boundaries, is an important
case that A3min ~1051 K in solute rich regions and parameter in describing the formation of bands. An
A3max ~1059 K in solute lean regions: experiments indi- insuf® cient density of ferrite nucleation sites can prevent
cated by squares; critical undercooling, which marks the formation of microstructural bands. It is of great
transition between band formation and no band for- importance that the shape of the austenite grains is
mation, indicated by broken line; photos show micro- modelled accurately, since it determines the number of
structures of specimens annealed at 953 K (area of available nucleation sites. Van Leeuwen et al.17 showed that
5126512 m m) and 961 K (area of 4206550 m m) the geometry of the austenite grain could be approximated
by a tetrakaidecahedron. The energy barrier DG* for
austenite is present in the specimens, since the Mn, Si, and nucleation on austenite/austenite grain boundaries in a
Cr are accumulated in the same regions. tetrakaidecahedron is given by18
4(z2 c {z1 c c c )3
a c
LOCAL VARIATION IN TRANSITION DG¤ ~ : : : : : : : : : : (6)
27 z3 DGV2
TEMPERATURE
The local variation in the transition temperatures as a where c a c is the surface free energy of the austenite/ferrite
function of the Mn, Si, and Cr concentration was calculated interface (0.6 J m ­ 2),18 c c c the surface energy of an austenite
with the thermodynamic database MTDATA. Since the grain boundary (0 .85 J m ­ 2),18 and z1, z2, and z3 are
expected concentration variations of the other alloying geometrical parameters that depend on the type of
elements are too small to signi® cantly alter the transition nucleation site in the austenite grain (boundary, edge, or
temperatures for the given composition, they have not been corner). For nucleation at the edge of a tetrakaidecahedron
measured and their nominal values were taken. The local z1 ~0 .72, z2 ~1.3, and z3 ~0 .096. 18 The driving force for
transition temperature from austenite to ferrite A3 as nucleation DGV is determined via the parallel tangent
calculated with MTDATA for a carbon concentration of construction for a Fe ± C ± Mn system with standardised
0 .364 wt-% has been ® tted to a linear equation which is data from the Scienti® c Group Thermodata Europe
given by (SGTE). The effect of the silicon and chromium concentra-
tions on DGV is taken into account by using an effective
A3 ~{23 :56 CM n + 26 :58CS i {6 :66CC r + 1065 : : : (2)
manganese concentration, which accounts for the effect of
where the temperature is in K and the concentrations in silicon and chromium on the A3 transition temperature
wt-%. Manganese and silicon have an opposite effect on the according to equation (2). The effective manganese con-
local A3 transition temperature, with a magnitude that is of centration at the minimum and maximum A3 transition
the same order. The local transition temperatures above temperature then becomes CMn,e ~0 .59 and 0 .25 wt-%,
A +1 and below A ¡ 1 the three phase region are for a carbon respectively.
concentration of 0. 364 wt-% given by The relative difference r between the ferrite nucleation
rates in regions with the maximum and the minimum A3
A+1 ~{15 :45 CM n + 19 :64CS i + 26 :72 CC r + 995 (3)
transition temperature is de® ned as
and
(d N=d t)A 3max {(d N=dt)A 3min
A{ : : : r~
1 ~{28 34 CM n + 13 06CS i + 13 26 CC r + 995 (4)
(d N=d t) A 3max
The variation in the local transition temperatures as shown l ¤
in Figs. 3 and 6 for the specimens that were annealed at ~1 {exp (DGA {DGA¤ 3m in ) : : : : : (7)
T~1013 and 953 K respectively, were calculated with equa- kT 3m ax

tions (2) ± (4). The two specimens show, as expected, similar Figure 7 shows how the relative difference between the
¯ uctuations in A3 temperature ranging from 1051 to ferrite nucleation rates changes as a function of the isother-
1059 K. The former specimen has microstructural bands, mal transformation temperature in the case that A3min~
while the latter specimen has not. This leads to the con- 1051 K in the solute rich regions and A3max ~1059 K in
clusion that the formation of ferrite/pearlite bands depends the solute lean regions. In the ® gure the experiments are
on both the presence of chemical bands and the isothermal indicated by squares and the critical undercooling, which
transformation temperature. The formation of strong bands marks the transition between band formation and no band
can already occur for relatively small ¯ uctuations of 8 K in formation is indicated by a broken line. The critical
the local A3 temperature, as compared with ¯ uctuations of undercooling for this material is ~ 100 K.
50 K that were found for the steels that were studied by The ferrite nucleation behaviour largely differs for the
Grossterlinden et al.3 two specimens investigated by EPMA. At 1013 K, the

Materials Science and Technology March 2002 Vol. 18


302 Offerman et al. Ferrite/pearlite band formation in hot rolled medium carbon steel

respectively, which is the period of time between the onset of


the ferrite formation and the onset of the pearlite formation.
As can be seen from this ® gure, there is a steep decrease
in the distance over which the carbon diffuses, below the
minimum A1 transition temperature.

NUCLEATION AND GROWTH CRITERIA


The growth criterion can be established by considering that
the ferrite nucleation predominantly takes place in the
regions with a high A3 temperature. The width of these
regions is denoted by w, which does not have to be equal to
half the wavelength of the chemical bands. The maximum
distance dC over which the carbon has to diffuse in order to
form microstructural bands is then given by
dC ~w=2 : : : : : : : : : : : : : : (11)
The width w can be estimated from Fig. 2 (w» 60 mm). As
can be seen from Fig. 8, the growth criterion was ful® lled
8 Distance d over which carbon diffuses in austenite for all the specimens that were annealed above the minimum
before pearlite formation takes place for eight speci-
A1 transition temperature. This criterion was not ful® lled
mens studied: T temperature
for the specimens that were annealed at T~961 and 953 K.
However, the former specimen does exhibit microstructural
relative difference of the ferrite nucleation rate is r~0 .80. banding ( see Fig. 7), since the relative difference in ferrite
This means that 80% less nucleation events take place in the nucleation rate is still signi® cant at this undercooling.
regions with the lowest A3 temperature compared with The overall nucleation rate in this specimen is so large that
regions with the highest A3 temperature. The redistribu- the equilibrium fraction pro-eutectoid ferrite is reached
tion of the carbon (towards the regions with a low local after 7 min, after which the pearlite formation starts. At
transition temperature) during the phase transformation annealing temperatures below the minimum A1 transition
enhances the difference in nucleation rate even further. At temperature, the in¯ uence of the ferrite growth on the
953 K the relative difference of the ferrite nucleation rate formation of microstructural bands can be neglected.
is only 6 .4%. This indicates that the nucleation rate was The specimen that was annealed at T~953 K does not
approximately homogeneous in the specimen, and hence no have microstructural bands (Fig. 5), since the relative
ferrite/pearlite bands are formed. difference in ferrite nucleation rate is too small at this
The formation of microstructural bands does not only temperature. The overall nucleation rate in this specimen is
depend on the relative difference r of the ferrite nucleation so large that the equilibrium fraction pro-eutectoid ferrite is
rates between regions with the maximum and minimum reached within 2 min, after which the pearlite formation
A3 transition temperature, but also on the distance d over starts. This results in a critical value rc for the relative
which the carbon can diffuse in the austenite before the difference of the ferrite nucleation rate, which is in the order
pearlite formation takes place. If the relative difference of of
the ferrite nucleation rates would be large enough to form
microstructural bands, but the ferrite grains could not rc &6 {8 % : : : : : : : : : : : : : : (12)
become large enough to ® ll the regions with a high A3 tem- If the austenite grain size is larger than 2 ± 3 times the
perature, microstructural bands could not (or only partly) chemical band spacing, there will be too few nucleation sites
form. In these medium carbon steels the ferrite growth rate available to result in a signi® cant difference in ferrite
is dominated by the diffusion rate of carbon in the austenite. nucleation rate between regions with a high and a low A3
The distance over which the carbon can diffuse in the transition temperature. In hot rolled steel this is usually not
austenite can be approximated by the case. With the model developed above and the criteria
established in this section, the formation of microstructural
d&( DcC t)1=2 : : : : : : : : : : : : : : (8)
bands can be predicted for a given heat treatment and a
where t is the time in which the carbon can diffuse in the given hot rolled medium carbon steel. The width w, which is
austenite before the pearlite formation takes place and D cC needed as an input parameter for the model, depends on the
is the diffusion coef® cient of the carbon in the austenite, solidi® cation conditions and is found to depend on the
which depends on the temperature and the nominal carbon primary dendrite arm spacing, 8 which could in principle be
concentration. D cC is given by19 obtained from a solidi® cation model. Another method is to
slowly cool a piece of the steel under investigation from the
DcC ~4:53 |10 {7 (1 + YC (1 {YC )8339 :9 T {1 ) austenite state to room temperature and obtain w from the
(9)
|exp ‰{( T {1 {2:221|10 {4 )(17767 {26436 YC ) Š width of the ferrite bands that have formed.
Since the criteria are expressed in nucleation and growth
where D cC is in m 2 s­ 1 and the temperature in K. The site processes, they can in principle also be used for continuous
fraction YC of carbon on the interstitial sublattice is given cooling experiments. For a continuous cooling experiment
by an integrated nucleation rate and a growth have to be used.
YC ~xC =(1 {xC ) : : : : : : : : : : : (10)
where xC ( ~0 .016704) is the atom fraction of carbon in the
alloy. The distance over which the carbon diffuses in the Conclusions
austenite before the pearlite formation takes place is shown
in Fig. 8 for the eight specimens that were studied. For the Two-dimensional electron probe microanalysis scans across
specimens with an annealing temperature above (or equal several ferrite/pearlite bands show that microstructural
to) the minimum A1 transition temperature the time was banding in hot rolled medium carbon steel is related to the
taken to be 4 .5 h. For annealing temperatures of T~953 microchemical banding of manganese and silicon. The
and 961 K, the time was taken to be t~2 and 7 min experiments show that the degree of banding decreases as

Materials Science and Technology March 2002 Vol. 18


Offerman et al. Ferrite/pearlite band formation in hot rolled medium carbon steel 303

the isothermal transformation temperature decreases, since


the relative difference between the ferrite nucleation rates in References
regions with a low and a high A3 transition temperature
decreases. It is shown that the formation of microchemical 1. p. g. bastien: J. Iron Steel Inst., 1957, 187, 281.
bands is a prerequisite for band formation, but the kinetics 2. j. s. kirkaldy, j. von destinon-forstmann, and r. j. brigham:
of the phase transformation determines the actual existence Can. Metall. Q., 1962, 1, 59.
3. r. grossterlinden, r. kawalla, u. lotter, and h. pircher:
of microstructural bands. A quantitative model has been
Steel Res., 1992, 63 , 331.
developed, with which the formation of microstructural 4. s. w. thompson and p. r. howell: Mater. Sci. Technol., 1992, 8,
bands can be predicted. The nucleation and growth criteria 777.
that have to be satis® ed in order to form microstructural 5. f. a. khalid, m. farooque, a. ul haq, and q. khan: Mater. Sci.
bands are: Technol., 1999, 15, 1209.
1. The relative difference between the ferrite nucleation 6. t. a. kop, j. sietsma, and s. van der zwaag: Mater. Sci.
rates in regions with the maximum and the minimum A3 Technol., 2001, 17, 1569.
transition temperature should be larger than 6 ± 8%. 7. s. g. e. te velthuis, m. th. rekveldt, and s. van der zwaag:
2. In the case that the annealing temperature is above the Ironmaking Steelmaking, 1995, 1, 81.
8. j. d. verhoeven: J. Mater. Eng. Perf., 2000, 9, 286.
minimum A1 transition temperature, the carbon has to
9. h. g. kim and p. r. howell: Proc. Int. Conf. on `Solid-to-solid
diffuse over a critical distance larger than half the width of phase transformations’ , Farmington, PA, USA, July 1994,
a region with a high A3 temperature. At annealing tem- Minerals, Metals, and Materials Society of AIME, 183.
peratures below the minimum A1 transition temperature, 10. g. p. krielaart, m. th. rekveldt, c. m. brakman, and s.
the in¯ uence of the ferrite growth on the formation of van der zwaag: Z. Metallkd., 1994, 85, 709.
microstructural bands can be neglected. 11. s. e. offerman, n. h. van dijk, m. th. rekveldt, j. sietsma , and
s. van der zwaag: Physica B: Phys. Condens. Matter, 2000,
276B ± 278B , 868.
12. j. komenda and r. sandstr OÈ m: Mater. Charact., 1993, 31 , 143.
Acknowledgements 13. a. from and r. sandstr OÈ m: Mater. Charact., 1998, 41 , 11.
14. l. c. brown and j. s. kirkaldy: Trans. AIME, 1964, 230, 223.
15. m. volmer and a. weber: Z. Phys. Chem., 1925, 119 , 277.
The authors thank C. Kwakernaak and W. G. Sloof for 16. r. becker and w. dOÈ ring: Ann. Phys. , 1935, 24 , 719.
performing the electron probe microanalysis measurements 17. y. van leeuwen, s. i. vooijs, j. sietsma, and s. van der zwaag:
and E. R. Peekstok for performing the optical microscopy. Metall. Mater. Trans. A, 1998, 29A, 2925.
This work was ® nancially supported by the Dutch 18. p. j. clemm and j. c. fischer: Acta Metall., 1955, 3, 70.
Technology Foundation (STW). 19. j. AÊ G REN : Scr. Metall., 1986, 20, 841.

Materials Science and Technology March 2002 Vol. 18

View publication stats