Materials Research, Vol. 3, No. 2, 31-35, 2000.

© 2000

Influence of the Microstructure of Duplex Stainless Steels on their

Failure Characteristics During Hot Deformation

G.S. Reis, A.M. Jorge Jr., O. Balancin*

Departamento de Engenharia de Materiais, Universidade Federal de São Carlos,
13565-905 São Carlos - SP, Brazil

Received: July 20, 1999 Revised: February 11, 2000

Two types of duplex stainless steels were deformed by torsion at a temperature range of 900 to
1200 °C and strain rate of 1.0 s-1 and their final microstructures were observed. The austenite volume
fraction of steel A (26.5Cr - 4.9Ni - 1.6Mo) is approximately 25% at room temperature, after
conventional annealing, while that of steel B (24Cr - 7.5Ni - 2.3Mo) is around 55%. Experimental
data show that steel A is ductile at high temperatures and displays low ductility at low temperatures,
while steel B has low ductility in the entire range of temperatures studied. At high temperatures,
steel A is essentially ferritic and shows dynamic recrystallized grains after deformation. When steel
A is strained at low temperatures and displays low austenite volume fraction, microstructural
observations indicate that failure is triggered by grain boundary sliding due to the formation of an
austenite net structure at the ferrite grain boundaries. At intermediate volume fraction, when
austenite forms a dispersed second-phase in steels A and B, failure begins at the ferrite/ferrite
boundaries since some of the new ferrite grains may become immobilized by the austenite particles.
When steel B is strained at volume fraction of around 50% of austenite and both phases percolate
the microstructure, failure occurs after low straining as a consequence of the different plastic
behaviors of each of the phases. The failure characteristics of both steels are correlated not only with
the volume fraction of austenite but also with its distribution within the ferrite matrix, which limits
attainable strain without failure.

Keywords: duplex stainless steel, hot working, hot ductility

single phase ferritic stainless steels soften by intense dy-

1. Introduction
namic recovery while austenitic stainless steels, with rela-
Duplex stainless steels have become more attractive tively low stacking fault energy, soften by dynamic
than single-phase austenitic and ferritic grades in many recrystallization6,7. When the two phases are jointly de-
industrial applications owing to their high strength and formed, the distribution of strains among the phases is no
corrosion resistance in chloride-containing media1,2. Al- longer uniform. When straining begins in duplex stainless
though advantageous, the processing of these materials steels, strain concentrations appear in the softer ferrite
requires special attention due to the existence of a low phase. As deformation increases, strain gradients may de-
ductility region under hot working conditions3. Generally, crease as a result of accommodation mechanisms such as
the presence of a massive second phase during processing recovery and recrystallization, cracking formation, as well
makes deformation more complex and the microstructure as interphase and grain boundary sliding.
is a decisive limiting factor on plastic behavior. In addition In spite of the complex deformation behavior of duplex
to the work hardening and softening mechanisms required stainless steels, it has been found that the ferrite phase
to deform each phase, the boundaries play an important role continues to show intense dynamic recovery while the
in duplex microstructures; the accommodation of macro- austenite phase undergoes dynamic recrystallization8-10.
scopic deformation depends on the plastic characteristics of Thus, ferrite and austenite display dissimilar behaviors and
both phases and interfaces4,5. microstructures during high temperature deformation, lim-
It is a well known fact that during high temperature iting attainable strain without failure. Recent research has
deformation, after a certain amount of work hardening, shown that the ductility of this kind of steel depends on the
* balancin@power.ufscar.br
32 Reis et al. Materials Research

deformation conditions, the behavior of the constituent rite/austenite ratio changes with temperature, steel B al-
phases, and the volume fraction of austenite in the ferrite ways contains more austenite than steel A under similar
matrix9,11,12. Also, it has been suggested that the ductility reheating and deformation conditions.
of duplex steels depends on phase distribution and on the Mechanical tests were carried out on a computerized hot
nature of the interface between these phases5,9. Although torsion machine described previously14. The samples were
attention has been paid to the softening mechanisms of the 10-mm in length and 10-mm in diameter in the reduced
ferrite and austenite phases and on deformation conditions, central gage section. These were heated by means of an
very few systematic studies have focussed on the identifi- induction furnace assembled on the testing machine.
cation of the relationship between the microstructure and Chromel-alumel thermocouples were used to control the
failure characteristics of these steels during high tempera- power supply to the furnace and to monitor the specimen
ture deformation. temperature during testing. To retain the high temperature
The purpose of this work was to observe the failure microstructures for further observations, water was injected
characteristics during hot deformation of two-phase alloys into a quartz tube surrounding the sample, immediately
relating then to the different microstructures which were upon reaching the fracture strain. After standard metal-
present. In order to obtain a large number of initial micro- lographic procedures, a deposition color etch composed of
structures, two kinds of duplex stainless steels were strained 100 mL distilled H2O, 20 mL HCl and 2 g potassium
by torsion tests under various deformation conditions after metabissulfite was applied, in order to observe the micro-
two initial thermal treatments. structure.
Hot torsion tests were carried out over a 900 to 1200 °C
2. Materials and Experimental Procedures temperature range, at an equivalent strain rate of 1 s-1. Some
of the tests were conducted after heating to 1250 °C and
Two types of two-phase alloys with different Cr/Ni
then cooling to deformation temperature, and others after
equivalent ratios were investigated in this work. The
heating to test temperature only, in order to produce a
austenite volume fraction of steel A (26.5Cr - 4.9Ni -
variety of microstructures before deformation.
1.6Mo) is around 25% at room temperature, after conven-
The temperatures required for complete dissolution of
tional annealing and cooling, while that of steel B (24Cr -
the austenite were estimated at around 1250 °C for steel A
7.5Ni - 2.3Mo) is approximately 55%. Figure 1 illustrates
and above 1300 °C for steel B (see Fig. 1). It was, therefore,
the cross section of the Fe-Cr-Ni system with 70% Fe and
expected that all the austenite was dissolved during the
indicates the constituent phases expected for these steels
treatment at 1250 °C in steel A and somewhat less in steel
after equilibrium is achieved during reheating at high tem-
B. In tests carried out on heating, the austenite volume
peratures before deformation. Although the fer-
fraction decreased as the temperature rose, until the mate-
rial was almost entirely ferritic at solution temperature. As
the temperature was reduced to below the reheat tempera-
ture on cooling, the phase transformation began with the
nucleation and growth of the austenite phase in the grains
and at the ferrite grain boundaries.
3. Results
Ductility was evaluated measuring the strain to fracture
from flow curves obtained from torsion tests. Figure 2
shows the dependence of the strain to fracture on deforma-
tion temperature for steel A, samples tested on cooling and
on heating, and for steel B, samples tested on cooling only.
Steel A is very ductile at high temperatures, but its ductility
decreases with decreasing test temperature. As the volume
fraction of austenite in the ferrite matrix increases, ductility
also decreases; that is the reason steel B exhibits such low
ductility over the entire temperature range.
Figure 2 shows three different behaviors in the depend-
ence of ductility on test temperatures: (i) steel A is very
ductile at high temperatures; (ii) steel A has low ductility
Figure 1. Cross section of the Fe-Cr-Ni system with 70% Fe13. The at low temperatures, regardless of the previous thermal
relative locations of steels A and B were determined taking into account treatment; and (iii) steel B has low ductility in the entire
the Cr/Ni equivalent ratios. range of temperatures studied.
Vol. 3, No. 2, 1999 Duplex Stainless Steels on their Failure Characteristics During Hot Deformation 33

Figure 4. Steel A tested on cooling to 950 °C, displaying formation of

Figure 2. Dependence of strain to fracture on deformation temperature cracks by grain boundary sliding after straining to ε = 2.5.
for steels A and B.

containing austenite particles within the grains and at grain

At high temperatures, as illustrated in the diagram of
boundaries. The volume fraction of austenite is small (Fv ~
Fig. 1, steel A is essentially ferritic. An example of this kind
5%) since the sample was heated to 1250 °C, which is
of microstructure is given in Fig. 3. In this case, the sample
higher than the austenite solution temperature, cooled to
was tested on cooling to 1100 °C and strained to ε = 1.2. test temperature and strained before the equilibrium volume
The microstructure consisted of the small new grains in the fraction of austenite was attained. In this case, deformation
old ferrite grains; initial grains of 400 µm were reduced to conditions represent the behavior (ii) described above. This
about 20 µm. This leads to the conclusion that dynamic figure clearly shows the formation of cracks at ferrite grain
recrystallization occurred, although this material tends to boundaries, indicating that the failure process is triggered
soften by intense dynamic recovery. Previous studies on by grain boundary sliding.
ferrite stainless steel15 and α - iron16 show that continuous Figure 5 presents the microstructure of a sample with
dynamic recrystallization takes place in these materials intermediate austenite volume fraction. This was obtained
after large deformation, with the formation of new grains from steel A tested on heating to 950 °C and strained to
deriving from the growth of subgrains and a gradual in- ε = 6.0. The microstructure consists of a ferrite matrix
crease in sub-boundary misorientation. Also, it is worth containing a volume fraction of around 20% of austenite.
observing in Fig. 3 that the old grain boundaries were Ferrite grains, formed as a result of dynamic recrystalliza-
bulged by the grains formed during deformation. tion, are small and equiaxial, while austenite particles are
Figure 4 shows the microstructure of a steel A sample elongated and aligned in the direction of deformation. This
tested on cooling to 950 °C and strained until it fractured micrography also shows the formation of some cracks at
(ε = 2.5). The microstructure consists of a ferrite matrix interphase boundaries, although most of them are located
at the ferrite/ferrite boundaries.
When the volume fraction of austenite is increased, the
structure of the material tends to become duplex. Figure 6

Figure 3. Microstructure of steel A tested on cooling to 1100 °C,

displaying new ferrite grains formed by continuous dynamic recristalliza- Figure 5. Microstructure of steel A strained at 950 °C on heating, showing
tion after straining to ε = 1.2. cracking at ferrite/ferrite boundaries.
34 Reis et al. Materials Research

The sample tested on cooling was heated to above the

austenite dissolution temperature and then cooled to test
temperature. Although at 950 °C the equilibrium volume
fraction of austenite is relatively high, the time spent on
cooling was insufficient for the alfa-gamma transformation
to be completed, allowing for the observation of only a
small volume fraction of austenite. Figure 2 shows that
most of the austenite phase is located at the ferrite grain
boundaries, forming a thin film that characterizes a net
structure18. The failure process occurs by grain boundary
sliding, which is independent of the austenite volume frac-
tion in the ferrite matrix.
Figure 6. Microstructure of steel B tested on cooling to 1100 °C, showing When tests are carried out with the on heating samples,
the formation of cracks after straining to ε = 1.3. part of the austenite that is present at room temperature is
dissolved before deformation. As diffusion is easier at the
shows the microstructure of a steel B sample obtained after boundaries than inside the grains, the austenite dissolution
straining to ε = 1.3 and tested on cooling to 1100 °C. The begins on the film at the grain boundaries. After some time
volume fraction of austenite is close to 50% and the micro- of reheating, the ferrite grain boundaries are no longer
structure consists of alternating austenite and ferrite phases covered by a continuous film of austenite. This inhibits
aligned in the direction of deformation. The austenite re- grain boundary sliding and continuous dynamic recrystal-
veals coarse elongated grains, typical of the work hardening lization proceeds to increase ductility. Thus, the ferrite
regime, while the ferrite displays fine equiaxial grains that matrix flows around the harder austenite phase, which acts
characterize its dynamic recrystallization. This figure also as a dispersed second phase. After large deformations,
shows the formation of cracks at ferrite/austenite interfaces,
austenite particles constrain the motion of the new grains
and on ferrite/ferrite and austenite/austenite grains.
and failure begins with cracking at the ferrite/ferrite
4. Discussion boundaries (see Fig. 5).
As the volume fraction of austenite is increased further
Two steels were used in this investigation, two different
and the microstructure is characterized by ferrite and
initial thermal treatments were applied, and various defor-
austenite percolation18, the plastic behavior depends on
mation conditions were used. As a result of the variety of
both the phases. Because the diffusion rates in ferrite are
conditions employed, a large number of initial microstruc-
about one order of magnitude higher than in austenite and
tures was obtained. These ranged from single phase ferrite
to mixtures of ferrite and austenite with several volume the dislocation movement in ferrite is much faster due to its
fractions, distributions and austenite particle shapes. high stacking fault energy, ferrite and austenite display
When steel A is tested at high temperatures (above dissimilar behaviors during high temperature deformation.
1100 °C), it is essentially ferritic and highly ductile. It When straining starts, the softer ferrite phase deforms plas-
would appear that the high ductility obtained in these tically to a larger extent than the harder austenite. After
experiments may be attributed to two causes: i) one is the some work hardening, the ferrite softens while the austenite
ease of dislocation annihilation and of sub-boundary for- begins working hardening. Thus, the ferrite matrix flows
mation, which decrease the material’s strength; ii) the sec- alongside the harder austenite phase, forming cavities at
ond cause is the bulging and motion of high angle grain triple junctions and at ferrite/austenite interfaces parallel to
boundaries, which have the beneficial effect of isolating the the direction of shear (see Fig. 6), which greatly decreases
cracks formed at original boundaries and of inhibiting grain ductility.
boundary sliding17. To confirm that this sharp decrease in ductility is asso-
As the temperature is decreased and particles of ciated with the microstructure, a complementary experi-
austenite are observed in the matrix, so ductility decreases. ment was carried out using steel B, in which the austenite
Figure 2 shows that, at 950 °C, steel A can be strained to particles are no longer contiguous (Fig. 7). This sample was
2.5 on cooling and 6.0 on heating without failure as com- tested on cooling to 1200 °C after reheating to 1350 °C for
pared to strains of about 40 at 1100 °C. From Figs. 4 and 5 30 min. The micrography displays cracking at ferrite/ferrite
it can be observed that the volume fraction of austenite is boundaries, indicating that the austenite particles constrain
larger in the heat tested sample than in the cooling tested the motion of the recrystallized ferrite grains, which is
one. Hence, one can infer that the level of ductility is typical of dispersed second-phase behavior. The strain to
correlated not only to the volume fraction of austenite but fracture was 4.8, i.e., 3.4 times the value obtained at the
also to its distribution within the ferrite matrix. same temperature when both phases percolate.
Vol. 3, No. 2, 1999 Duplex Stainless Steels on their Failure Characteristics During Hot Deformation 35

References
1. Solomon, H.D.; Devine Jr, T.M. Duplex Stainless
Steels. ASM, Metals Park, Ohio (R.A. Lula, ed.), p.
693, 1982.
2. Nilsson, J.O. Mater. Sci. Tech., v. 8, p. 685, 1992.
3. Kawasaki, T.; Takada, I.; Ohtsubo, H.; Suzuki, S.
Kawasaki Steel Technical Report n. 14, p. 50, 1986.
4. Akdut, N.; Foct, J. ISIJ International, v. 36, p. 883,
1996.
5. Iza-Mendia, A.; Pinõl-Juez, A.; Urcola, J.J.; Gutiérrez,
I. Metall. Mat. Trans., v. 29A, p. 2975, 1998.
6. Richards, P.; Sheppard, T. Mat. Sci. Tech., v. 2, p. 836,
Figure 7. Microstructure of steel B tested on cooling to 1200 °C after 1986.
reheating to 1350 °C. 7. McQueen, H.J.; Ryan, N.D.; Evangelista, E. Mat. Sci.
Forum, v. 113-115, p. 435, 1993.
5. Conclusions 8. Wang, R.; Lei, T.C. Mat. Sci. Eng., A165, p. 19, 1993.
- The ductility of duplex stainless steels depends on the 9. Arboledas, J.B.; Tirado, J.L.M.; Rodriguez, R.S. Proc.
deformation temperature and volume fraction of austenite, Int. Congress Stainless Steels’96, p. 116. Dussel-
and it is also sensitive to austenite particle distribution. dorf/Neuss, June, 1996.
- These materials are very ductile at high temperatures 10. Cizek, P.; Wynne, B.P. Mat. Sci. Eng., A230, p.88,
if very low austenite volume fraction is present. 1997.
- When the second phase forms a net structure, material 11. Hoffmann, W.A.M.; Balancin, O. Informacion Tec-
failure occurs by grain boundary sliding after little strain- nologica, v. 9, p. 11, 1998.
ing, regardless of the volume fraction of austenite. 12. Mecozzi, M.G.; Capotosti, L. Proc. Int. Congress
- When the austenite is distributed as a dispersed second Stainless Steels’96, p. 317. Dusseldorf/Neuss, June,
phase, ductility increases and failure is caused by austenite 1996.
particles that immobilize the new ferrite grains formed by 13. Pugh, J.W.; Niebest, J.O. Trans. AIME, v. 188, p. 268,
dynamic recrystallization. 1950.
- As the volume fraction of austenite increases and the 14. Jorge-Jr., A.M.; Balancin, O. Revista de Engenharia
microstructure is characterized by percolation of both e Ciências Aplicadas, v. 2, p. 133, 1994/1995.
phases, failure occurs after low straining as a result of the 15. Belyakov, A.; Kaibyshev, R.; Zaripova, R. Mat. Sci.
different plastic behaviors of each phase, regardless of test Forum, v. 113-115, p. 385, 1993.
temperatures. 16. Baczynski, J.; Jonas, J.J. Metall. Mater. Trans., v.
29A, p. 447, 1998.
Acknowledgements 17. McQueen, H.J.; Ryan, N.D.; Konopleva, E.V.; Xia, X.
The financial support of the Brazilian agencies Can. Metall. Quart., v. 34, p. 219, 1995.
FAPESP, CAPES and CNPq are gratefully acknowledged. 18. Hornbogen, E. Acta Metall., v. 32, p. 615, 1984.