Materials Science and Engineering A283 (2000) 65 69
www.elsevier.com/locate/msea
Mössbauer study of the retained austenitic phase in
multiphase steels
a, b a a
A. Mijovilovich *, A. Gonçalves Vieira , R. Paniago , H.D. Pfannes ,
b
B. Mendonça Gonzalez
a
Departamento de Física, Uni ersidade Federal de Minas Gerais, C.P. 702, 30123-970 Belo Horizonte, Brazil
b
Escola de Engenharia, Uni ersidade Federal de Minas Gerais, Rua Espírito Santo 35, 30160-030 Belo Horizonte, Brazil
Received 31 May 1999; received in revised form 21 December 1999
Abstract
Samples of steels with composition 0.30%C-1.5%Mn-1.5%Si-0.5%Al-0.5%Mo (wt.%) were subjected to different thermomechan-
ical treatments to produce ferrite/pearlite/bainite (FPB), spheroidized (ESF) and martensite (MAR) microstructures. Subsequently
they underwent a two stage annealing to obtain a final structure comprising of ferrite, bainite, martensite and austenite. The
samples were studied by means of Mössbauer spectroscopy (transmission and conversion electron Mössbauer spectrocopy
(CEMS)), X-ray diffraction (XRD), and metallographic analysis. Austenite contents were found to be the same for all samples
except for the spheroidized sample annealed at 750°C that showed an increase of the austenite with increasing temperature of the
treatment. Mössbauer spectroscopy and quantitative XRD analysis exhibited significant discrepancies ascribed to texture effects.
It is shown that the thermal treatment was successful in retaining significant quantities of the austenite phase for steels of this
composition. © 2000 Elsevier Science S.A. All rights reserved.
Keywords: Multiphase steel; Mössbauer; Austenite; Martensite; Bainite
spite that the result may be strongly influenced by
1. Introduction
texture effects. Mössbauer spectroscopy is a well-known
In order to enhance the ductility in high-strength technique used in the study of Fe-containing alloys [4].
steels it was shown that it is necessary to increase the
The different phases can be distinguished from their
content of their retained austenite. Alloys with high
different signals, and different magnetic behaviors re-
ductility and excellent levels of mechanical strength can
gardless of the state of aggregation of the phases.
be obtained by the transformation of austenite to
Martensite and austenite are easily distinguished from
martensite during plastic deformation (i.e. trip: trans-
their different hyperfine patterns in the Mössbauer
formation induced plasticity effect) [1]. Matsumura et
spectra with better accuracy than by other techniques.
al. [2] increased the content of retained austenite in an
Due to the low solubility of carbon in -Fe in equi-
alloy of C Mn Si by a two stage thermal treatment:
librium, the interstitial solute C can not be detected by
an annealing followed by a quick quenching to the
Mössbauer spectroscopy. In the transmission made sig-
range of temperatures for the bainitic transformation.
57
nals from all the Fe atoms in the sample are obtained
The amount of retained austenite increased with in-
regardless of the state of the aggregation or crystallinity
creasing content of Mn and Si in the alloy [3].
in the material. In the case of conversion electron
It is usual to determine the phases present by metal-
Mössbauer spectrocopy (CEMS) the spectrum stems
lographic analysis as well as X-ray diffraction. The last
from a region of 10 100 nm below the surface of the
method is sometimes used for quantitative analysis in
sample, and thus becomes an efficient tool to analyse
the surface. With Mössbauer spectroscopy the texture
* Corresponding author. Present address: EMBL c/o DESY,
effect does not affect the total area of the subspectra
Notkestrasse 85, Geb. 25A, Notkestrasse 85, 2603, Hamburg, Ger-
corresponding to the different phases.
many. Tel.: [phone]; fax: [phone].
0921-5093/00/$ - see front matter © 2000 Elsevier Science S.A. All rights reserved.
PII: S0921-5093(00)00620-1
66 A. Mijo ilo ich et al. / Materials Science and Engineering A283 (2000) 65 69
We used transmission Mössbauer spectroscopy to
determine the amount of austenitic retained phase after
the thermal treatments. By CEMS we were able to
study the mechanical stability and transformation to
martensite during the laminating process. Quantitative
X-ray diffraction analyses were employed to determine
the relative amounts of retained austenite, and ferritic
phases present.
2. Experimental
By a specific thermal treatment of an alloy of compo-
sition 0.30%C-1.5%Mn-1.5%Si-0.5%Al-0.5%Mo three
initial structures were obtained: ferrite/pearlite/bainite
(FPB), spheroidized (ESF) and martensite (MAR).
Subsequently they underwent a two stage annealing
(Fig. 1) to obtain a final structure of ferrite, bainite,
martensite and austenite. We will keep the acronyms of
Fig. 1. Schematic diagram of two stage annealing. T1=intercritical
the initial phases when we refer to the samples after the
temperature; t1=annealing time; T2=bainitic temperature (425°C),
two-stage annealing. For the metallographic analysis a
t2=bainitic transformation time.
selective etching with Nital 2%, Picral 5% and Na-thio-
sulfate [5] was used.
Fig. 2. Photographs of the microstructures: (a), FPB780; (b), MAR780; (c), ESF750 and (d), ESF840. Ferrite is gray color, bainite in dark gray
and austenite-martensite in light gray.
A. Mijo ilo ich et al. / Materials Science and Engineering A283 (2000) 65 69 67
Table 1
were able to take into account the thickness of the
Hyperfine parameters and relative areas (%) for the identified phases
samples. The sample FPB780 was investigated also by
from transmission Mössbauer spectraa
CEMS using the same source as above. In this case, as
the effective thickness is small, we used a least
Sample H(T) (mm s-1) (mm s-1) Relative area
(%) squared fit with simple Lorentzian lines [11]. The values
2
of to measure the quality of the fit ranged from 1.7
FPB780
to 3.4.
Ferrite 1 33.0 0.047  45.42
Ferrite 2 30.6 0.067  28.02
Ferrite 3 27.9 0.072  6.84
Austenite 1  -0.023  13.67
Austenite 2  0.032 0.6 6.08
MAR780
Ferrite 1 33.0 0.045  45.70
Ferrite 2 30.7 0.064  26.07
Ferrite 3 28.4 0.073  8.34
Austenite 1  -0.027  13.51
Austenite 2  0.033 0.6 6.38
Mar810
Ferrite 1 33.0 0.008  43.80
Ferrite 2 31.0 0.03  22.83
Ferrite 3 29.4 0.033  12.46
Austenite 1  -0.059  13.72
Austenite 2  0.005 0.6 7.20
Esf 750
Ferrite 1 33.0 0.008  48.2
Ferrite 2 31.0 0.031  28.4
Ferrite 3 29.2 0.025  13.0
Austenite 1  -0.053  7.1
Austenite 2  0.051 0.6 3.2
Esf 840
Ferrite 1 33.0 0.005  40.2
Ferrite 2 31.3 0.028  22.8
Ferrite 3 29.5 0.041  17.4
Austenite 1  -0.062  13.6
Austenite 2  -0.004 0.6 6.0
a
H(T) is the hyperfine magnetic field in Tesla, (mm s-1) is the
isomer shift refered to -Fe in mm s-1, and (mm s-1) is the
quadrupole splitting.
Integrated intensities of X-ray diffraction peaks were
used to determine the content of retained austenite [6].
The samples were grounded, embedded in epoxy, pol-
ished with 1 m diamond paste and measured in a
Philips diffractometer (K -Cu radiation, 0.01°-steps of
Fig. 3. Mössbauer spectra for: (a), MAR780 and (b), FP780 samples.
2 ). The reflections used for the quantitative analysis
were: (200), (220) and (311) for the -phase; (200) and
Table 2
(211) for the -phase.
Hyperfine parameters and percentages of phases from CEMS spectra
We measured transmission Mössbauer spectra of all
(surfaces)a
samples. The measuring temperature was room temper-
57
H(T) (mm s-1) (mm s-1) Relative area
ature and the source was Co in Rh matrix. Two
(%)
samples (FPB780 and MAR780) were also measured at
77 K. Since the steels were 50 nm thick foils the lines
FPB780
were broadened because of thickness effect. It is com- Ferrite 1 33.0 0.00  48.12
Ferrite 2 31.0 0.04  39.36
mon to use fits with hyperfine field distributions or
Ferrite 3 27.0 -0.08  3.37
Voigtian line profiles [7] to take this effect into account.
Austenite 1  -0.18  5.38
In fitting these spectra we used the program WOTAN
Austenite 2  0.00 0.6 3.79
[8] which is based on the integral form of the absorp-
a
tion line calculated by Margulies [9,10]. In this way we Symbols as in Table 1.
68 A. Mijo ilo ich et al. / Materials Science and Engineering A283 (2000) 65 69
bainite shows dark gray, and both, the martensite and
the austenite become light gray. The microstructure of
the FPB sample is heterogeneous with fine and coarser
regions (Fig. 2a). The final structure for the sample
obtained from an initial martensite is finer (Fig. 2b).
For the samples obtained from the spheroidized initial
structure an increase of the amount of bainite and
martensite-austenite with increasing intercritical tem-
perature is found (Fig. 2c, d).
The main contributions to the Mössbauer spectra
arise from the ferrite and martensite phases of the steel,
which lead to magnetically split spectra. We used the
designations Ferrite 1, 2 and 3 to denote subspectra
corresponding to Fe atoms in the ferritic/bainitic/
martensitic matrix with three different environments,
following the nomenclature of Uwakweh et al. [12]. But
due to the different nomenclatures used in the literature
Fig. 4. CEMS spectra for sample FP780. it is not possible to give unique assignments for the
different Fe C configurations [13]. The austenitic para-
Table 3
magnetic phase is clearly distinguishable with two con-
XRD results for retained austenite phase (volume%) for different
tributions: austenite 1 and 2 for the singlet and the
two-stage thermal treatments
doublet, respectively. It amounts 21% in all samples.
Since there is insignificant contribution of cementite or
Sample Thermal treatment Volume (%) of
other carbide precipitates seen in the spectra, they have
phase
been neglected in the fitting. The hyperfine parameters
FPB780 780°C 20 min-425°C 10 min 13.2 1.5
and the percentages of each phase as determined from
ESF750 750°C 20 min-425°C 10 min 7.1 1.5
the Mössbauer spectra are listed in Table 1. Typical
ESF840 840°C 20 min-425°C 10 min 16.1 1.5
Mössbauer spectra are shown in Fig. 3.
MAR780 780°C 20 min-425°C 10 min 17.8 1.5
The Mössbauer results indicate similar austenite con-
MAR810 810°C 20 min-425°C 10 min 17.2 1.5
tents for the MAR and FPB samples. For the ESF
samples it is clearly observed that by the treatment at
higher temperature more austenite is retained. By com-
paring the room temperature and liquid nitrogen spec-
tra of FPB780 (ferrite-perlite system) and MAR780
(martensite rich sample), we conclude that both exhibit
similar austenite contents.
We measured the FPB780 sample also with CEMS
and deduced a decrease of the austenitic phase, indicat-
ing a transformation from austenite to martensite in the
surface during the polishing process (see Table 2 and
the corresponding spectrum in Fig. 4). This was simi-
larly observed by [14].
The quantitative determination of phases by XRD is
given in Table 3 and a characteristic pattern is shown in
Fig. 5. The Mössbauer results for the austenite content
differs significantly from the XRD results, except for
the ESF samples where both techniques indicate the
Fig. 5. X-ray diffraction pattern of a multiphased structure obtained
same trend. This is ascribed to a cristallographic texture
from martensitic initial structure-MAR-780 (intercritically annealed
effect that strongly influences the XRD measurements.
at 780°C for 20 min and at 425°C for 10 min).
Mössbauer results concerning the spin texture in these
steels will be published elsewhere [15].
3. Results and discussion
Photographs of optical microscopy are shown in Fig.
4. Conclusions
1. Three phases can be distinguished in the photo-
graphs, namely martensite plus austenite, bainite and
Metallographic analysis showed the presence of
ferrite. With the used etching, ferrite becomes gray,
bainite, ferrite and martensite-austenite in all samples.
A. Mijo ilo ich et al. / Materials Science and Engineering A283 (2000) 65 69 69
The finer microstructure is present in the samples ob- References
tained from the martensitic phase due to its acicular
[1] V.F. Zackey, D.F. Parker, R. Busch, Trans. ASM 60 (1967)
morphology.
252 259.
From the micrographs it is seen that the amount of
[2] O. Matsumura, Y. Sakuma, H. Takechi, Trans. ISIJ 27 (1987)
carbide precipitates is negligible which is in concordance
570 579.
with the Mössbauer results, which do not indicate any
[3] Y. Sakuma, O. Matsumara, H. Takeshi, Metallur. Trans. A 22A
significant contribution from carbides. For the evalua- (1991) 489 498.
[4] F.E. Fujita, in: U. Gonser (Ed.), Mössbauer Spectroscopy,
tion of the retained austenite the Mössbauer spectra
Springer-Verlag, New York, 1975, p. 5.
indicate that the annealing at higher temperature is
[5] S. Bandoh, O. Matsumara, Y. Sakuma, Trans. ISIJ 28 (1988)
effective in stabilizing the austenite phase in the
569 574.
spheroidized sample. Results for other samples under
[6] J. Durnin, K.A. Ridal, J. Iron Steel Inst. (1968) 60 67.
different thermal treatments are similar, in disagreement
[7] J.Y. Ping, D.G. Rancourt, Hyp. Int. 71 (1992) 1433.
with XRD measurements. The difference was attributed
[8] R. Hollatz, Wotan Fitting Programm, Institut für Experimental-
to a texture effect. physik, Universität Hamburg, Germany, 1992.
[9] S. Margulies, P. Debrunner, H. Frauenfelder, Nucl. Inst. Meth.
From CEMS results a significant decrease of the
21 (1963) 217.
austenite content in the surface due to the mechanical
[10] S. Margulies, J.R. Ehrman, Nucl. Inst. Meth 12 (1961) 131.
polishing is deduced.
[11] R.A. Brand, Angewandte Physik, Universität Duisburg, Ger-
many, 1988.
[12] O.N.C. Uwakweh, J.P.H. Bauer, J.M.R. Génin, Metallur. Trans.
21A (1990) 589.
Acknowledgements
[13] M. Ron, in: R.L. Cohen (Ed.), Applications of the Mössabuer
Spectroscopy, vol. 2, Academic Press, New York, 1980, p. 329.
The support of the Brazilian research agencies
[14] R.C. Mercader, J. Desimoni, Hyp. Int. 110 (1997) 101 109.
Fapemig, CAPES and CNPq is greatfully acknowl-
[15] A. Mijovilovich, R. Paniago, H.D. Pfannes, A. Gonçalves
edged. Vieira, B. Mendonça Gonzalez, (in press).
.