Failure Analysis
Failure Analysis
Failure Analysis
Abstract
The effect of the addition of W instead of Mo on fatigue crack growth (FCG) behavior has been studied for duplex
(ferritic–austenitic) stainless steels. The microstructure and the FCG properties of the homogenized alloys are basically insensitive
to the variation of W content. For the alloys aged at 900°C for 30 min after homogenization, however, the apparent FCG
resistance is significantly increased and is attributed principally to the roughness-induced crack closure due to the existence of
brittle secondary phases precipitated during aging. With an increase in W content, the quantity and size of the brittle s and x
precipitates decrease and the ductility of the aged alloys increases. In accordance, the intrinsic FCG resistance of the aged alloys
is improved with increasing W content. © 1998 Elsevier Science S.A. All rights reserved.
Table 1
Chemical composition (wt.%) of the DSSs
specific heat treatments, the effect of isothermal aging transverse (T) direction (L–T orientation). The geome-
on the microstructure and the FCG behavior is also try of the CT specimens for the FCG tests is shown in
investigated for purposes of comparison with the ho- Fig. 1. The crack length was monitored with a traveling
mogenization treatment. microscope. The precracking was conducted at a fre-
quency of 12 Hz and a load ratio (R) of 0.1 using an
Instron servohydraulic testing machine. All the speci-
mens were precracked to a depth of at least 2 mm from
2. Experimental procedure the crack tip and were subsequently tested at a constant
load amplitude with a sinusoidal waveform. Fractogra-
Four different alloys with the base composition of phy of fatigue failed specimens was performed in a
25.5Cr–6Ni–3.5Mo were used in this study. While the JEOL scanning electron microscope (SEM). To observe
PRE value was fixed at 41, i.e. Mo+0.69W =3.5, Mo the microstructure under the transmission electron mi-
was substituted for W at different contents as 3.5Mo– croscope (TEM), the thinned discs were electropolished
0W, 2.5Mo–1.5W, 1.5Mo – 3W and 1Mo – 3.6W, re- in a twin jet electropolisher using a solution of 75 ml
spectively. A detailed description of chemical perchloric acid and 925 ml acetic acid. Optimum polish-
compositions in different phases of the DSSs studied is ing conditions were obtained between 30 and 40 V at
given in Table 1. room temperature. The thin foils were examined in a
After hot rolling the ingots at 1100°C, 12 mm thick Philips EM400 microscope operated at 200 kV.
plates with microstructures of nearly equal volume
fractions of ferrite and austenite [10] were produced.
Solution treatment was carried out at 1100°C for 10
min to dissolve secondary phases, followed by water
quenching. Island-like austenite grains were found in
the ferritic matrix, while no precipitates could be ob-
served by light microscopy. The volume fractions of
austenite (g) and ferrite (a) were determined with image
analysis application software on representative micro-
graphs. The results are shown in Table 2 for different
compositions. Isothermal aging treatment at 900°C for
30 min, corresponding to the ‘nose’ of the C-curve for
s transformation, was performed to evaluate the mi-
crostructural effects resulting from the alloy composi-
tion. To get good contrast, a color etching technique
was applied to the specimens.
Tensile specimens were oriented with the tensile axis
parallel to the longitudinal (L) direction, and compact
tension (CT) specimens were oriented with the crack
plane normal to the L direction, propagating in the
Table 2
Volume fractions of the a- and g-phases in the DSSs
3.1. Microstructures
Fig. 5. (a) Yield strength (sy) and ultimate tensile strength (UTS), and (b) elongation as a function of aging time for various compositions.
which contrast the fatigue fracture surfaces of the ho- erable areas of intergranular failure and becomes much
mogenized and the aged alloys. It is found that the rougher. The roughness analysis of the fatigue fractured
fracture surface of the homogenized alloy is basically surfaces corresponding to Fig. 7 has been made and the
transgranular but that of the aged alloy contains consid- results are plotted for different profiles, as shown in Fig.
8. As expected, the fracture surface roughness of the
aged alloy is evidently greater than that of the homoge-
nized alloy. It is then inferred that the slight increase in
the apparent FCG resistance of the aged 3.5Mo–0W
alloy is attributed to the roughness-induced crack clo-
sure effect due to the intergranular failure via grain
boundaries weakened by the secondary phase (s) precip-
itation. Intergranular cracking mainly occurs along s or
x interfaces. This type of crack closure effect should
decrease as the brittle precipitates diminish in quantity;
however, the apparent FCG resistance of the aged
1Mo–3.6W alloy containing the smallest amounts of
brittle precipitates is significantly improved. To account
for the fatigue behavior in this alloy, which has smaller
roughness-induced crack closure but exhibits higher
FCG resistance compared with the 3.5Mo–0W alloy,
Fig. 6. The da/dN– DK curves for the homogenized and aged alloys intrinsic and extrinsic contributions have been separately
with various compositions. considered, as stated below.
T.H. Kang et al. / Materials Science and Engineering A251 (1998) 192–199 197
Fig. 7. SEM fractography for fatigue failed specimens of: (a) homogenized, and (b) aged alloys (3.5Mo – 0W composition); CPD, crack
propagation direction.
3.4. Extrinsic 6ersus intrinsic FCG resistance increase in W content, the intrinsic FCG resistance of
the aged alloys increases accordingly. This implies that
To distinguish between extrinsic and intrinsic contri- the intrinsic FCG resistance is remarkably decreased,
butions to the FCG behavior, the crack closure load by the aging treatment, in the 3.5Mo–0W alloy, but is
has been measured in order to calculate the effective not much influenced in the alloys containing W con-
stress intensity range, DKeff ( : Kmax −Kcl), so that the tents. In view of the microstructural change during
intrinsic FCG resistance can be evaluated. Although aging (Figs. 2–4), it has been clarified that the quantity
there are several other extrinsic effects induced by crack and the size of the brittle s-phase become smaller as
branching and crack deflection [15,16], it is assumed the W content increases, and finally (as in the 1Mo–
here that only the crack closure has a major role in the 3.6W alloy) the precipitation of the s-phase is halted
extrinsic effects. The normalized crack closure levels and replaced by that of the tougher x-phase. As a
(Kcl/Kmax), plotted as a function of DK for the homoge- result, the ductility in tensile tests gets improved as the
nized and aged microstructures, are shown in Fig. 9. It W content increases (Fig. 5(b)). It is then inferred that
is observed that the crack closure levels of aged mi- this improvement in ductility contributes to the intrinsic
crostructures are much higher than those of homoge- FCG resistance in the aged alloys (Fig. 10).
nized microstructures, which is in good agreement with Considering the above extrinsic and intrinsic contri-
Figs. 7 and 8. However, a variation of crack closure butions, the apparent FCG behavior of the aged alloys
levels for the alloys with different W contents is not regarding the compositional effect (Fig. 6) can be inter-
significant in either homogenized or aged preted as follows. For the 3.5Mo–0W alloy, its FCG
microstructures. resistance is counterbalanced by two factors, i.e. the
After subtracting this crack closure effect from the large amounts of crack closure mainly due to the rough
apparent FCG rate, the intrinsic FCG rate is plotted surface, and the inferior intrinsic FCG resistance due to
against DKeff to obtain the effect of alloy composition the precipitation of the brittle s-phase upon aging. For
for both the homogenized and the aged alloys (Fig. 10). the 2.5Mo–1.5W alloy, while the crack closure contri-
As indicated in Fig. 10, the intrinsic FCG behavior of bution does not change much (Fig. 9), the intrinsic
the homogenized alloys is nearly insensitive to the (ductility-facilitating) contribution resulting from the
variation of the alloy composition, in agreement with suppression of s precipitation takes effect (Fig. 10). The
the fact that the microstructure in these alloys does not net result is the improvement in the apparent FCG
change much by alloying. It is noted that the intrinsic resistance of the 2.5Mo–1.5W alloy compared with the
FCG resistance of the homogenized alloys is obviously 3.5Mo–0W alloy. For the 1Mo–3.6W alloy, although
superior to (for the 3.5Mo – 0W composition) or no the crack closure effect is minimized due to the disap-
worse than (for the 2.5Mo – 1.5W and 1Mo – 3.6W com- pearance of the brittle s-phase, the intrinsic contribu-
positions) that of the aged alloys. Moreover, with the tion predominates in the overall FCG resistance. As to
198 T.H. Kang et al. / Materials Science and Engineering A251 (1998) 192–199
Fig. 10. The da/dN– DKeff curves for the homogenized and aged
alloys with various compositions.
4. Conclusions
Acknowledgements
References [8] J. Li, T. Wu, Y. Riquiet, Mater. Sci. Eng. A174 (1994)
231.
[9] H. Okamoto, K. Tsuda, Proc. Conf. Duplex Stainless Steels ’94,
[1] J. Charles, Proc. Conf. Duplex Stainless Steels ’91, France, Les
Glasgow, Scotland, 1994, vol. 1, paper 91.
Editions de Physique, Paris, 1991, p. 3.
[10] J.C. Jippold, W. Lin, Proc. Conf. Duplex Stainless Steels ’94,
[2] J. Charles, Proc. Conf. Duplex Stainless Steels ’94, Glasgow,
Glasgow, Scotland, 1994, vol. 1, paper 104.
Scotland, 1994, vol. 1, paper K1. [11] T. Huhtah, J.O. Nilsson, Proc. Conf. Duplex Stainless Steels ’94,
[3] J.O. Nisson, Mater. Sci. Technol. 8 (1992) 685. Glasgow, Scotland, 1994, vol. 1, paper 43.
[4] H. Okamoto, Proc. Conf. Application of Stainless Steels ’92, [12] J.O. Nilsson, P. Liu, Mater. Sci. Technol. 7 (1991) 853.
Stockholm, Sweden, The Institute of Materials, 1992, p. 390. [13] J.O. Nilsson, A. Wilson, B. Josefsson, T. Thorvaldsson, Proc.
[5] T.Y. Marrow, J.E. King, Mater. Sci. Eng. A183 (1994) 91. Conf. Stainless Steels ’92, Stockholm, Sweden, The Institute of
[6] M. Strangwood, S.G. Druce, Mater. Sci. Technol. 6 (1990) 211. Materials, 1992, vol. 1, p. 280.
[7] J. Sakai, J. Matsushima, Y. Kamemura, M. Tanimura, T. Os- [14] J.O. Nilsson, Mater. Sci. Technol. 8 (1992) 685.
aka, in: R.A. Rula (Ed.), Duplex Stainless Steels, ASM, St. [15] S. Suresh, R.O. Ritchie, Metall. Trans. A 13 (1982) 1627.
Louis, MO, 1983, p. 211. [16] K.S. Chan, Metall. Trans. A 23 (1992) 183.
.
.