D. Raabe - Alloy Design, Combinatorial Synthesis, and Microstructure Property Relations for Low-Density Fe-Mn-Al-C Austenitic Steels.pdf

(2478 KB) Pobierz
Author's personal copy
JOM, Vol. 66, No. 9, 2014
DOI: 10.1007/s11837-014-1032-x
Ó
2014 The Minerals, Metals & Materials Society
Alloy Design, Combinatorial Synthesis, and Microstructure–
Property Relations for Low-Density Fe-Mn-Al-C Austenitic
Steels
D. RAABE,
1,3
H. SPRINGER,
1
I. GUTIERREZ-URRUTIA,
1,4
F. ROTERS,
1
M. BAUSCH,
1
J. -B. SEOL,
1
M. KOYAMA,
2
P. -P. CHOI,
1
and K. TSUZAKI
2
1.—Max-Planck-Institut fur Eisenforschung, 40237 Dusseldorf, Germany. 2.—Department of
¨
¨
Mechanical Engineering, Kyushu University, 744 Motooka, Nishi-ku, Fukuoka, Fukuoka 819-0395,
Japan. 3.—e-mail: d.raabe@mpie.de. 4.—e-mail: i.gutierrez@mpie.de
We present recent developments in the eld of austenitic steels with up to 18%
reduced mass density. The alloys are based on the Fe-Mn-Al-C system. Here,
two steel types are addressed. The rst one is a class of low-density twinning-
induced plasticity or single phase austenitic TWIP (SIMPLEX) steels with
25–30 wt.% Mn and
<4–5
wt.% Al or even
<8
wt.% Al when naturally aged.
The second one is a class of
j-carbide
strengthened austenitic steels with even
higher Al content. Here,
j-carbides
form either at 500–600°C or even during
quenching for
>10
wt.% Al. Three topics are addressed in more detail,
namely, the combinatorial bulk high-throughput design of a wide range of
corresponding alloy variants, the development of microstructure–property
relations for such steels, and their susceptibility to hydrogen embrittlement.
INTRODUCTION TO LOW DENSITY STEELS
Reducing energy consumption in conjunction with
improving safety standards is a paramount target in
modern mobility concepts. Hence, the development
of strong, tough, and ductile steels for automo-
tive applications is an essential topic in steel
research.
1–15
In this context, twinning-induced
plasticity (TWIP) steels with up to 30 wt.% Mn and
>0.4
wt.% C content have shown an excellent
combination of ductility and strength.
16–34
Increas-
ingly, the reduction in mass density of TWIP steels
becomes an additional challenge. Two effects enable
such efforts. The rst one is that Mn increases the
face-centered cubic (fcc) lattice parameter. The sec-
ond one is that very high Mn and C alloying stabi-
lizes the austenite, so that it can tolerate Al
additions up to about 10 wt.% without becoming
unstable, i.e., transforming into body-centered cubic
(bcc)-ferrite.
35
Such an alloy concept sustains many advantages
associated with TWIP steels, e.g., mechanical
twinning and very high strain hardening;
16,19,24
yet,
it enables density reductions of up to 18% (Fig.
1).
Hence, alloys based on the quaternary system
Fe-Mn-Al-C are specically promising for the design
of low-density TWIP steels.
36–47
Regarding the excellent mechanical properties of
TWIP steels, which are characterized by the tran-
sition from dislocation and cell hardening to mas-
sive mechanical twinning, it has to be considered
that Al increases the stacking fault energy
(SFE).
17,18
This means that the overall strain-
hardening behavior and the onset of mechanical
twinning in density-reduced TWIP grades
41
may
differ from those observed in conventional TWIP
steels.
17,19,24
However, alloys based on the Fe-Mn-Al-C system
offer an even larger variety in deformation and
strain-hardening mechanisms than those associated
with the TWIP effect alone. This is due to the
characteristic dislocation substructures and the
higher number of phases present in the Fe-Mn-Al-C
system, namely, fcc-austenite, bcc-ferrite, and
ordered structures such as DO
3
and L’1
2
-type car-
bides. Depending on composition, low-density steels
can assume austenitic structure for the composition
regime Mn: 15–30 wt.%, Al: 2–12 wt.%, and C: 0.5–
1.2 wt.%. To combine the advantages of the TWIP
mechanisms with the reduction in specic weight,
(Published online June 24, 2014)
1845
personal copy
1846
Raabe, Springer, Gutierrez-Urrutia, Roters, Bausch, Seol, Koyama, Choi, and Tsuzaki
Fig. 1. Density reduction in Fe-Mn-Al-C TWIP steels as a function of
Al content.
Fig. 2. Specic energy absorption values of selected Fe-Mn-Al-C
steels in comparison with different conventional deep-drawing steels
(Color gure online).
this alloy range is hence the most promising one.
When increasing the Al content to a range
>6–8
wt.%, strain hardening in these steels is less
dominated by the TWIP effect but instead by the
formation of nanosized L’1
2
-type carbides, so-called
j-carbides.
44,45,48
Density-reduced steels with ferritic structure have
compositions in the range Mn
<
8 wt.%, Al: 5–8
wt.%, and C
<
0.3 wt.%. Corresponding complex
grades, consisting of austenite and ferrite, can be
synthesized by using compositions Mn: 5–30 wt.%,
Al: 3–10 wt.%, and C: 0.1–0.7 wt.%.
48
Besides these
compositions, ordered D0
3
structures, i.e., near-fer-
ritic Fe-Al-Cr alloys without Mn, have also been
addressed in the past in the context of density-re-
duced alloy design.
When comparing the synthesis and properties
among the different classes of weight reduced steels,
alloys based on the austenitic Fe-Mn-Al-C system
are most attractive due to their superior strain
hardening, high energy absorption (Fig.
2),
high-
density reduction and robust response to minor
changes in composition and processing.
37–47
Even
thin-strip casting with associated in-line hot rolling
has been successfully conducted in our group as a
pathway for efcient small-scale manufacturing of
such grades.
Recent publications on austenitic Fe-Mn-Al-C
alloys have reported yield strength values of 0.5–
1.0 GPa, elongations to fracture in the range 30–
80%, and ultimate tensile strength in the range of
1.0–1.5 GPa.
37–47
When blended with an Al content below 5 wt.%, a
single austenite phase prevails at room tempera-
ture, showing excellent strain hardening, which was
attributed to the hierarchical evolution of the
deformation substructure.
41
Al also promotes the
formation of nanoprecipitates upon aging with L’1
2
structure and approximate stoichiometry of (Fe,
Mn)
3
AlC.
44,45,48
These phases are referred to as
j-carbides.
They belong to the group of non-oxide
perovskites. Because of their ordered fcc structure,
j-carbides
have a lattice mismatch below 3% with
respect to an austenitic Fe-Mn-C matrix phase and
can hence form cuboidal nanoprecipitates.
44,48
When embedded in a ferritic matrix, the lattice
mismatch can be as large as
$6%,
45,48
which leads
to semicoherent interfaces and, hence, different
precipitate morphologies.
This article provides a concise introduction to
some recent developments in the eld of low-density
Fe-Mn-Al-C TWIP steels placing attention on alloy
design, synthesis routes, and microstructure–prop-
erty relations. We also provide a brief outlook on
pending questions associated with the role of
j-carbides
on strain hardening and hydrogen
embrittlement.
BULK COMBINATORIAL DESIGN
OF LOW-DENSITY AUSTENITIC STEELS
Here, we use a combinatorial approach for rapid
trend screening and alloy maturation of metallur-
gically melted and processed Fe-Mn-Al-C low-den-
sity TWIP and
j-carbide
hardened steels. The
approach is referred to as rapid alloy prototyping
(RAP).
49
We apply it here to one group of Fe-30Mn-
1.2C-xAl (wt.%) TWIP steels and to a second group
of Fe-20Mn-0.4C-xAl TWIP steels, both with vary-
ing Al content (x) and different aging conditions.
In both cases, the samples were synthesized by
melting and casting in a vacuum-induction melting
(VIM) furnace under 400 mbar Ar pressure. The
system was modied to enable synthesis of ve dif-
ferent alloys in one operation for each alloy system.
We used ve Cu molds, which could be moved
stepwise inside the furnace. They were successively
lled with melt from a 4-kg ingot. After each cast,
the remaining melt composition in the ingot was
adjusted by charging Al through an air lock. After
cooling and cutting, the 10
9
50
9
130-mm
3
-sized
blocks were hot rolled at 1100°C into 2
±
0.1-mm-
Author's personal copy
Alloy Design, Combinatorial Synthesis, and Microstructure–Property Relations
for Low-Density Fe-Mn-Al-C Austenitic Steels
1847
thick and
$500-mm-long
sheets. These were
reheated to 1100°C, water quenched, and cut per-
pendicular to the rolling direction into sets of nine
segments with dimensions 2
9
60
9
55 mm
3
for
each alloy composition. Homogenization was per-
formed at 1100°C for 2 h under Ar, followed by
water quenching. Aging was conducted in air at
450°C, 500°C, 550°C, and 600°C for 0 h, 1 h, and
24 h at each temperature, followed by oil quenching.
This results in a matrix of 45 different sample con-
ditions. Scales were removed from the surfaces by
low-pressure, ne-grit sandblasting after the heat
treatments. Samples for mechanical testing and
microstructure investigation were prepared from
the segments by package spark erosion. Tensile
testing was conducted at room temperature with an
initial strain rate of 10
À3
s
À1
. All values plotted
represent averages of three measurements for every
material state. Cross-sectional areas of selected
samples were prepared in the plane perpendicular
to the rolling direction by grinding and polishing
with standard metallographic techniques. X-ray
diffraction (XRD) analysis was performed on the
rolling plane of samples ground to a thickness of
1 mm. Further details of the method are explained
in Ref.
49.
The RAP method enabled us to screen two dif-
ferent sets of ve Fe-Mn-C-based weight-reduced
Al-containing compositions each exposed to nine
respective heat treatments within 35 h. For each
alloy base set, synthesis, processing, mechanical
screening, and phase characterization are included.
The metallographic analysis showed no cracks,
pores, or macrosegregations in the nal materials.
The as-cast samples had a coarse dendritic micro-
structure. Hot rolling and water quenching resulted
in a fully recrystallized microstructure with a grain
size of
$20
lm
with some retained microsegrega-
tions of Mn.
Figure
3
shows the mechanical properties of 45
different material conditions obtained for the Fe-
30Mn-1.2C base composition (i.e., in total 135 ten-
sile tests and hardness measurements) in terms of
the yield strength (YS, Fig.
3a),
ultimate tensile
strength (UTS, Fig.
3b),
total elongation (TE,
Fig.
3c),
and hardness (Fig.
3d).
The results are
plotted according to the systematically varied Al
content. Color-coding reects individual aging con-
ditions. The data are reproduced from an earlier
publication.
49
They show a clear dependence of the
mechanical behavior on both composition and heat
treatment.
For the reference material (no Al addition, i.e.,
ternary Fe-30Mn-1.2C alloy), the best mechanical
behavior is found for the as-homogenized state,
namely, 360 MPa YS, high work hardening
(830 MPa UTS), and high ductility (77% TE). Aging
of the Fe-30Mn-1.2C alloy leaves the YS virtually
unchanged and increases the hardness slightly.
However, it reduces UTS and TE. Embrittlement
becomes most apparent for long aging times (24 h)
and higher temperatures (>500°C). For the alloy
Fe-30Mn-1.2C-8Al, i.e., the material with the high-
est Al content and lowest density, the opposite trend
applies. Without aging, the mechanical response is
similar to that of the Al-free alloy. Only a minor
change in YS (increase), UTS, and TE (decrease) is
found. Aging for 1 h leads to an increase in YS,
UTS, and hardness (increasing with temperature)
and only to a minor reduction in TE. Aging of the
alloy Fe-30Mn-1.2C-8Al for 24 h further increases
YS, UTS, and hardness to levels twice as high as in
the as-homogenized state and drastically reduces
ductility. The mechanical data for the alloys with
intermediate Al contents fall between the two dif-
ferent behaviors described above, i.e., alloys with 2–
6 wt.% Al have properties between the respective
values from materials without Al and with 8 wt.%
Al. Especially the alloys with Al additions of 4 wt.%
and 6 wt.% are only very weakly affected by the
applied aging treatments in terms of their
mechanical data compared with the alloys Fe-30Mn-
1.2C (weakening/embrittlement) and Fe-30Mn-
1.2C-8Al (strengthening). To better extract the
mechanical trends for the investigated compositions
and heat treatments, selected results are summa-
rized in Fig.
4.
By using the RAP approach, we nd for the
30 wt.% Mn system that high amounts of Al
($8 wt.%) result in pronounced strengthening dur-
ing aging, depending on time and temperature
(Fig.
4).
The observed effects are attributed to the
formation of nm-sized
j-carbides
during aging. The
intermediate alloy variants with Al concentrations
of 2–6 wt.% do not show equivalent mechanical
properties compared to the
$8
wt.% Al variant.
Also, a much smaller inuence of aging on the ten-
sile behavior can be observed in the 2–6 wt.% Al
cases. A detailed and systematic screening of the
stoichiometry, interface structure, lattice mist,
and mechanical effects of the nm-sized
j-carbides
is
still pending, but rst results suggest that they can
occur in a wide compositional existence range
(Fig.
5).
Also it is observed that they are thermally
very stable at 600°C and even prevail after 100 h
heat treatment with only modest coarsening
(Fig.
5).
For the second alloy class investigated, namely,
the Fe-20Mn-0.4C-xAl TWIP steels, a more homo-
geneous trend of the mechanical properties as a
function of the Al content is observed (Fig.
6).
While
yield strength and ultimate tensile strength both
increase, the tensile elongation drops as a function
of the Al content in the range between 2 wt.% and
11 wt.% Al. As observed by XRD screening of the
corresponding samples, the increasing ferrite con-
tent seems to be more relevant for this trend than
the presence of
j
carbides or
e-martensite.
Author's personal copy
1848
Raabe, Springer, Gutierrez-Urrutia, Roters, Bausch, Seol, Koyama, Choi, and Tsuzaki
Fig. 3. Overview of the obtained mechanical properties from RAP experiments as a function of alloy composition and applied aging treatment:
(a) yield stress (YS), (b) ultimate tensile stress (UTS), (c) total elongation (TE), (d) hardness. (RAP: Rapid Alloy Prototyping). Data and gures
are reproduced from an earlier publication
49
(Color gure online).
MICROSTRUCTURE–PROPERTY
RELATIONS FOR LOW-DENSITY
AUSTENITIC STEELS
When deriving microstructure–property relation-
ships for austenitic low-density steels, two scenarios
must be considered: For compositions with
25–30 wt.% Mn and up to 4–5 wt.% Al or even
8 wt.% Al, when naturally aged, the TWIP effect
prevails as deformation mechanism.
41,48
Thus,
dislocation-based strain hardening, which is fol-
lowed at higher loads by mechanical twinning, is
the primary strain-hardening mechanism. Such
low-density TWIP alloys are referred to as single
phase austenitic TWIP (SIMPLEX) steels.
48
After
aging at 500–600°C or for alloys with higher Al
content, twinning is reduced and strain hardening
is essentially associated to
j-carbide
and solid-
solution strengthening (Figs.
5
and
6).
If the Al content exceeds 10 wt.%, then
j-carbides
can form during quenching. The exact strain-hard-
ening mechanism for the latter case is still under
investigation. In a recent study, Gutierrez-Urrutia
and Raabe
48
suggested that the prevalent deformation
Fig. 4. Selected mechanical testing results from RAP experiments:
Tensile strength and total elongation for alloys Fe-30Mn-1.2C
through Fe-30Mn-1.2C-8Al. Some data were taken from Ref.
49
(Color gure online).
Author's personal copy
Alloy Design, Combinatorial Synthesis, and Microstructure–Property Relations
for Low-Density Fe-Mn-Al-C Austenitic Steels
1849
Fig. 5. (a) Joint atom probe tomographic and (b) Transmission Electron Microscopy (TEM) analysis of
j-carbides
in a Fe-30Mn-1.2C-8Al low-
density steel after 24 h at 600°C. This sample was synthesized after identifying suited compositions by the RAP process (Color gure online).
mechanisms in austenitic steels that are strengthened
by regularly arranged nanosized
j-carbides,
such as
those shown in Fig.
5,
consist in Orowan bypassing of
longitudinal rods of
j-carbides
and subsequent
expansion of dislocation loops, which is assisted by
dislocation cross-slip and, to a minor extent, shearing
of
j-carbides.
They further suggest that the higher dislocation
densities observed within the dislocation bundles are
caused by minor topological differences in the spac-
ing between the
j-carbide
rods, so that dislocations
follow preferential soft paths within the widest
channels. In certain cases, these dislocation accu-
mulations can lead to stress values that are high
enough to shear
j-carbide
interfaces. Although these
complex nanoscale interactions are still subject to
further analysis, we give below some suggestions
how to incorporate such effects into a mean eld
dislocation density-based model of strain hardening
of
j-carbide
containing Fe-Mn-Al-C steels.
A general form of a temperature-sensitive consti-
tutive model for fcc metals with low stacking fault
energy and mechanical twinning, depending on
chemical composition, deformation rate, and tem-
perature, was introduced by Steinmetz et al.
50
It
extends the three-internal-variable model of Roters
et al.
51
with a physical description of twin nucleation.
The model is based upon experimental observations
performed on Fe-Mn-C TWIP steels (without
j-car-
bides) by electron channeling contrast imaging,
Transmission Electron Microscopy (TEM), and
electron backscatter diffraction (EBSD).
19–34
These
works revealed that the important microstructural
internal state variables in such alloys include dislo-
cations, grain size, mechanical twins, and dislocation
cells. In this regard, this model provides a quantita-
tive description of strain-hardening behavior of low-
density TWIP steels, i.e., SIMPLEX steels. Following
the comments made above, strain hardening in
j-
carbide-containing steels can be described in a
seamless fashion where the activation barrier for
twinning can vary as a function of the Al content and
a carbide-rod-dependent Orowan loop mechanism
can be introduced in the form of a corresponding
stress term in the kinetic equation of state. The latter
formulation is not yet presented here.
In this section, we provide a summarized
description of the temperature-sensitive constitu-
tive model of strain hardening for single-phase low-
density TWIP steels, i.e., SIMPLEX steels. The
model uses three different dislocation densities (q
c
,
q
w
, and
q
d
) and the volume fraction of mechanical
twins (f
tw
) as state variables. The three dislocation
densities are those in the cell interior, in the cell
walls, and dipoles. The evolutions of these state
variables represent the microstructural changes
that occur during plastic deformation. The evolution
of the dislocation densities is given by the following
equations:
_
e
M
1 2d
d
_
q
(1)
À
q
c
¼
b
K
c
n
c
Zgłoś jeśli naruszono regulamin