Corrosion fatigue crack growth inhibition of
duplex stainless steel
M. Ahsan & Z. Gasem
King Fahd University of Petroleum and Minerals, Dhahran,
Saudi Arabia
Abstract
The main objective of this study is to examine the inhibition effect of chromate
on corrosion fatigue crack growth in a super duplex stainless steel alloy.
Corrosion fatigue crack propagation tests were performed for Zeron 100 in air,
3.5% NaCl, and in 3.5% NaCl inhibited with sodium chromate (Na2CrO4) using
two chromate concentrations. Specimens were tested in the TL orientation.
Results show that crack growth rates are enhanced in 3.5% NaCl solution
relative to air over the entire range of ∆K tested. Addition of 0.02 M Na2CrO4 to
3.5% NaCl solution results in notable decreases in crack growth rates up to a
specific value of ∆K above which crack growth rates approach values measured
in pure 3.5% NaCl solution. When the chromate concentration is increased by
ten fold, corrosion fatigue crack growth rates are markedly reduced and approach
measurements obtained in air for the entire range of ∆K. Examination of fracture
surfaces by the scanning electron microscope reveals the presence of ductile
striations in austenite and brittle striations in ferrite. The basic striations
characteristics in the austenitic phase are apparently the same in all the
environments. However, ferrite striations appear to be ∆K and environment
dependent. Scratch tests were carried out to examine the effect of chromate
concentration on repassivation kinetics of the duplex alloy. Chromate inhibition
effectiveness during corrosion fatigue of duplex stainless steel is proposed to
correlate with chromate repassivation kinetics and the crack tip strain rate.
1
Introduction
Duplex stainless steels (DSS) are comprised of approximately equal phase
fractions of austenite and ferrite. The objective in designing a duplex structure is
Damage and Fracture Mechanics VIII, C. A. Brebbia & A. Varvani-Farahani (Editors)
© 2004 WIT Press, www.witpress.com, ISBN 1-85312-707-8
60 Damage and Fracture Mechanics VIII
to combine the attractive properties offered by each phase individually and the
commonly high corrosion resistance of both phases due to high chromium
content “stainless”. The ferritic phase is superior in strength and show high
resistance to chloride stress corrosion cracking but it is susceptible to agehardening embrittlement. The austenitic phase exhibits high ductility but it
suffers from chloride stress corrosion cracking. DSS alloys, therefore, have
higher strength and higher resistance to stress corrosion cracking than austenitic
stainless steels and they are less susceptible to embrittlement when compared to
ferritic stainless steels. This has led to the widespread use of duplex stainless
steels in various industries which demand high strength and high resistance to
general as well as localized corrosion.
Due to increasing usage of DSS in applications involving dynamic loading in
aggressive environments, such as offshore gas and oil flowlines in the petroleum
industry, characterizing and understanding corrosion fatigue crack growth
kinetics of DSS alloys is of great importance. Numerous research studies have
characterized fatigue crack growth in air and in chloride containing
environments. Marrow and King [1] reported that fatigue crack growth rates in a
super duplex alloy (Zeron 100) tested in air varied with specimen orientation and
attributed this behavior to crack growth retardation effects due to ferrite/austenite
phase boundaries. Furthermore, they proposed that fatigue crack growth rates
enhancement in 3.5% NaCl was due to hydrogen assisted transgranular cyclic
cleavage of ferrite. Other researchers [2,3] interpreted the corrosion fatigue
behavior of Zeron 100 super duplex stainless steel with general agreement to that
of Marrow and King’s findings and revealed that the austenite phase exhibited
ductile tearing at low ∆K levels and ductile striation fatigue crack growth
mechanism at high ∆K levels.
There are very few studies pertaining to corrosion fatigue crack propagation
inhibition of DSS alloys. The main objective of this work, therefore, is to
examine the fatigue crack growth behaviour of a super duplex stainless steel
immersed in 3.5% NaCl solution containing various concentrations of sodium
chromate (Na2CrO4). The effect of chromate inhibition on corrosion fatigue
crack propagation rates and fracture mechanisms of Zeron 100 DSS in 3.5%
NaCl will be investigated.
2 Experimental procedure
Standard compact tension specimens were machined by EDM (electrical
discharge machining) from 8 mm thick plate of Zeron 100 duplex stainless steel
alloy in the as received condition (solution treated at 1120 ºC for 20 minutes and
water quenched, hardness = 23 HRc). TL specimen orientation was employed in
which the crack propagation direction is parallel to the rolling direction. The
composition of Zeron 100 DSS is given in table 1. Constant load fatigue crack
growth testing was carried out on precracked specimens at a load ratio of R=0.1
and sinusoidal cyclic frequency of 10 Hz in air and 1 Hz in aqueous solutions.
An environmental cell was employed to continuously expose the crack region to
Damage and Fracture Mechanics VIII, C. A. Brebbia & A. Varvani-Farahani (Editors)
© 2004 WIT Press, www.witpress.com, ISBN 1-85312-707-8
Damage and Fracture Mechanics VIII
61
the corrosive environment. Crack length was measured using the compliance
technique following ASTM guidelines E647.
Table 1: Composition of Zeron 100 super duplex stainless steel alloy.
C
Si
Mn
S
P
Cr
Ni
Mo
W
Cu
N
Fe
0.025
0.45
0.60
0.001
0.022
25.25
6.94
3.70
0.67
0.57
0.22
Bal
The seven-point incremental data reduction method was used to obtain da/dN∆K curves. The environments investigated were air and 3.5% NaCl containing
0 M, 0.02 M, and 0.2 M Na2CrO4. Scanning Electron Microscope was used at an
accelerating voltage of 20 kV to examine fracture surfaces.
1.0x10
-2
Zeron 100: TL
R=0.1
Env: Air, F=10 Hz
Env: 3.5% NaCl, F=1 Hz
Env: 3.5% NaCl + 0.02M Na2CrO4, F=1 Hz
Env: 3.5% NaCl + 0.2M Na2CrO4, F=1 Hz
da/dN (mm/cycle)
1.0x10-3
1.0x10
-4
1.0x10-5
10
20
30
40
50
60
∆K (MPa√m)
70
80
90
100
Figure 1: Comparison of fatigue crack growth rate data in air, 3.5% NaCl,
0.02M and 0.2M Na2CrO4 inhibited 3.5% NaCl.
3 Results
3.1 Fatigue crack growth
Fatigue crack propagation results in all environments are shown in figure 1. The
crack growth rates in 3.5% NaCl are enhanced by a factor of 2.2 at lower ∆K
values and by a factor of 1.5 for higher ∆K values relative to those in air at the
same ∆K. Addition of 0.02M Na2CrO4 decreases the crack growth rates at low
∆K levels. The crack growth rates approach those in 3.5% NaCl at high values of
∆K. When the concentration of the inhibitor is increased from 0.02M to 0.2M,
the crack growth rates approach those in air over the entire ∆K tested.
Damage and Fracture Mechanics VIII, C. A. Brebbia & A. Varvani-Farahani (Editors)
© 2004 WIT Press, www.witpress.com, ISBN 1-85312-707-8
62 Damage and Fracture Mechanics VIII
3.2 Fractography
The fracture surface of the specimen tested in air reveals that the dominant crack
growth mode in ferrite as well as in austenite appears to be ductile at low ∆K
levels and shows fatigue striations only at high ∆K (figure 2, a and b). In 3.5%
NaCl, the fracture surface exhibits ductile features at low ∆K levels (figure 3, a
and b).
At intermediate values of ∆K (40-45 MPa√m), distinct features in both phases
are observed where ferrite exhibits flat brittle features and austenite exhibits
ductile fracture (figure 3, c and d). At ∆K=51 MPa√m, ductile striations are
clearly observed in austenite while brittle striations are evident in ferrite, as can
be seen in figure 3e and figure 3f, respectively. The da/dN value at 51 MPa√m
could not be obtained in the da/dN-∆K plot but extrapolation in the Paris regime
gives da/dN value of 1x10-3 mm/cycle which is comparable to the striation
spacing in the austenite phase. The fracture surface of the specimen tested in
0.02M Na2CrO4 inhibited 3.5% NaCl (figure 4), suggests that the cracking mode
at low ∆K level is comprised of flat areas probably in ferrite grains and ductile
features associated with the austenitic phase. At intermediate ∆K level, the
fracture surface shows flat areas and ductile striations similar to that produced in
pure 3.5% NaCl. The fracture features of the specimen tested in 0.2M inhibited
solution are shown in figure 5. At low ∆K level, the fracture mode appears to be
similar to that in air. At intermediate and high ∆K values, ductile fatigue
striations are clearly visible.
3.3 Scratch tests
Scratch tests were performed in all three aqueous environments at the open
circuit potential. The current transient is recorded when the specimen surface is
scratched. The transient charge (Q) is estimated as the area under the currenttime curve.
(a)∆K≅24MPa√m (1000x)
(b) ∆K≅48MPa√m (2000x)
Figure 2: Fractographs of specimen tested in air at R=0.1, f=10 Hz (crack
growth direction: left to right).
Damage and Fracture Mechanics VIII, C. A. Brebbia & A. Varvani-Farahani (Editors)
© 2004 WIT Press, www.witpress.com, ISBN 1-85312-707-8
Damage and Fracture Mechanics VIII
(a) ∆K≅27 MPa√m (2000x)
63
(b) ∆K≅32 MPa√m (2000x)
(c) ∆K≅40 MPa√m (2000x)
(d) ∆K≅45 MPa√m (2000x)
(e) ∆K≅51 Mpa√m (3500x)
(f) ∆K≅51 Mpa√m (2000x)
Figure 3: Fractographs of specimen tested in 3.5% NaCl at R=0.1, f=1Hz
(crack propagation direction: left to right).
Damage and Fracture Mechanics VIII, C. A. Brebbia & A. Varvani-Farahani (Editors)
© 2004 WIT Press, www.witpress.com, ISBN 1-85312-707-8
64 Damage and Fracture Mechanics VIII
(a) ∆K≅29 MPa√m (2000x)
(b) ∆K≅40 MPa√m (3500x)
Figure 4: Fractographs of specimen tested in 0.02M Na2CrO4 inhibited 3.5%
NaCl at R=0.1, f=1 Hz (crack propagation direction: left to right).
(a) ∆K≅26 MPa√m (1500x)
(b) ∆K≅40 MPa√m (2000x)
(c) ∆K≅47 MPa√m (3500x)
Figure 5: Fractographs of specimen tested in 0.2M Na2CrO4 inhibited 3.5%
NaCl at R=0.1, f=1 Hz (Crack propagation direction: Left to right).
Damage and Fracture Mechanics VIII, C. A. Brebbia & A. Varvani-Farahani (Editors)
© 2004 WIT Press, www.witpress.com, ISBN 1-85312-707-8
Damage and Fracture Mechanics VIII
65
Table 2: Charge produced during scratch tests.
Transient Charge (Q)
(µCoulomb/mm2)
Environment
Test 1
Test 2
Average
3.5% NaCl
5.238
5.714
5.476
4.583
4.712
4.648
2.990
2.900
2.945
3.5% NaCl + 0.02M
Na2CrO4
3.5% NaCl + 0.2M
Na2CrO4
Table 2 lists the average values of Q accumulated during scratching at
constant potential in each environment. The average charge produced in 3.5%
NaCl is 5.476x10-6 Coulomb/mm2. When the specimen surface was scratched in
0.02M inhibited 3.5% NaCl environment, the average Q decreased to 4.648 x10-6
Coulomb/mm2 (15% drop) which suggests that chromate ions accelerated film
reformation and reduced the transient charge. When the concentration of the
chromate was increased from 0.02M to 0.2M, the average Q decreased by
approximately 37%.
4 Discussion
Since duplex stainless steel is composed of phases of different strength and
ductility, crack propagation must be a localized process depending upon crack
growth resistance offered by each phase individually. In 3.5% NaCl, the
enhancement mechanism in crack growth rates relative to air can be attributed to:
anodic dissolution, hydrogen embrittlement, or a combination of both. McIntyre
[4] showed that the transgranular faceted growth was a result of hydrogen
embrittlement and the ferrite phase was preferentially attacked by hydrogen.
Hydrogen embrittlement as the main enhancement mechanism in corrosion
fatigue crack growth of Zeron 100 DSS alloys is supported by the work of
Marrow et al [5] who conducted fatigue crack growth tests in pure hydrogen gas
and observed brittle striations even at very low ∆K. Thus, the chromate effect on
fatigue crack growth will be discussed assuming that the major embrittling
mechanism is due to hydrogen absorption and diffusion.
The mechanism by which chromate passivates steel has been studied
extensively, and it appears likely that protection is afforded by adsorption and
oxide formation on the steel surface [6]. Adsorption helps to polarize the anode
to sufficient potentials to form very thin hydrated passive layer which protects
the steel. The oxide film is a mixture of ferric and chromic oxides and is kept in
protective condition by adsorption and oxidation with very little loss of metal as
long as sufficient chromate remains in the solution [6]. The passive oxide film is
conductive and cathodic to steel and passive steels are considered to consist
entirely of cathodic areas. When the passive film is damaged by scratching or by
dissolution, and when insufficient chromate is present to repair the film, the
Damage and Fracture Mechanics VIII, C. A. Brebbia & A. Varvani-Farahani (Editors)
© 2004 WIT Press, www.witpress.com, ISBN 1-85312-707-8
66 Damage and Fracture Mechanics VIII
exposed steel becomes anodic with high current density and severe localized
corrosion attack [6].
Rafaey et al. [7] investigated the concentration effect of different inhibitors
like chromate, molybedate, nitrite, and phosphate on the inhibition of corrosion
of mild steel. They observed that increasing the concentration of chromate
shifted the pitting potential, Ep, to more positive values (more noble). They
explained the inhibition effect of chromate by proposing that CrO42- species plug
the pores in the passive film formed on the steel surface and increased the pitting
corrosion resistance. The effect of chromate inhibition increased with increased
chromate concentration. For the same concentration, chromate was found to be
more effective than nitrite and molybedate. This was attributed to faster
repassivation kinetics of chromate than the other inhibitors.
Dong et al. [8] reported a significant effect of chromate concentration
dependence of pits nucleation and growth in carbon steel using electrochemical
noise analysis. Molybedate was also used to compare the effectiveness of
inhibitors. They found that the pit nucleation rate increased with increased CrO42concentration below a critical value, but decreased quickly when the CrO42concentration was above a critical level. The decrease in nucleation rate in the
case of molybedate was less rapid than in chromate. The enhanced efficiency of
chromate was attributed to its adsorption tendency which was thought to be
stronger than that of molybedate.
McCafferty [9] showed that there is a critical concentration of chromate
required to inhibit the crevice corrosion of iron in the presence of a given
concentration of chloride ion. The critical concentration of chromate required
increased with increased Cl- ion concentration. There was a linear log/log
relationship between the activity of aggressive Cl- ion and the minimum activity
of CrO42- that could provide protection.
Review of previous work shows that chromate inhibition of localized
corrosion processes is strongly dependent on the concentration used and this
effect could be related to repassivation kinetics. Furthermore, scratch tests results
of the present study (table 2) indicate that the transient charges produced are
chromate concentration dependent. Increasing the concentration result in
accelerated repassivation kinetics and limited charge flow. During corrosion
fatigue of DSS in 3.5% NaCl, the passive film is ruptured due to cyclic loading
and unprotected metal is exposed to the environment at the crack tip. The
exposed surface undergoes high density anodic dissolution before repassivation
takes place. The passive region near the crack tip surface is cathodic. Cathodic
reactions generate hydrogen which enters the crack tip region and lead to
embrittlement. The Cl- ions slow the repassivation rate of the protective film and
allow more hydrogen entry into the metal, in accordance with the finding of
McCafferty [9] regarding the chloride effect. When Na2CrO4 is added to 3.5%
NaCl, chromate ions compete with Cl- ions and enhance repassivation kinetics
which reduces the extent of anodic and cathodic reactions and as a result
decrease hydrogen generation and entry.
In general, the effectiveness of chromate inhibition of corrosion fatigue crack
growth depends primarily on Cl- ion concentration, chromate ion concentration,
Damage and Fracture Mechanics VIII, C. A. Brebbia & A. Varvani-Farahani (Editors)
© 2004 WIT Press, www.witpress.com, ISBN 1-85312-707-8
Damage and Fracture Mechanics VIII
67
and the crack tip strain rate έ (proportional to ∆K2 and the loading frequency).
The Cl- ion concentration is constant in all the corrosion fatigue tests performed
in the present study. Higher crack tip strain rates increase the extent of damage in
the passive film at the crack tip region and expose more unprotected metal to the
aggressive environment. Chromate ions repassivate the broken film and
consequently reduce the environmental damage that may take place at the crack
tip. Therefore, it is suggested that either the chromate concentration or the crack
tip strain rate will play the dominant effect on corrosion fatigue crack growth
rates. During corrosion fatigue, if the repassivation kinetics in a solution of a
given chromate concentration is fast enough to repair the mechanically ruptured
film at the crack tip region then effective inhibition is expected. Otherwise, there
will be less effect of chromate and crack growth rates will approach those in
uninhibited solution.
Hence, the corrosion fatigue crack growth results can be explained as follows.
Below ∆K≅33 MPa√m, the repassivation kinetics of 0.02M chromate is fast
enough to repair the limited film damage and results in effective inhibition and
reduced crack growth rates. Above ∆K≅33 MPa√m, the exposed bare metal area
is large such that the repassivation kinetics of 0.02M chromate ions cannot repair
the broken film and the unprotected metal is exposed to the aggressive
environment for longer time and thus the crack growth rates approach those in
3.5% NaCl. When chromate concentration increased from 0.02M to 0.2M,
corrosion fatigue crack growth rates are reduced throughout the Paris region and
approach those in air. This concentration of chromate is sufficient to provide
faster repassivation kinetics which can repair the broken film for all ∆K values
which lead to reduced crack growth rates similar to da/dN measured in air.
5
Conclusions
1.
Inhibiting the 3.5% NaCl solution with chromate with a concentration of
0.02M reduces the corrosion fatigue crack propagation rates of the duplex
stainless steel up to a specific value of ∆K above which da/dN of both the
uninhibited and the inhibited solution converge.
Increasing the concentration of inhibitor to 0.2M significantly reduces crack
growth rates throughout the whole range of ∆K tested and the crack growth
rates approach those measured in air.
Scratch tests performed at constant open circuit potential indicate that
repassivation kinetics largely depend on chromate concentration.
Effective inhibition of corrosion fatigue crack propagation is proposed to
correlate with repassivation kinetics and the crack tip strain rates.
2.
3.
4.
Acknowledgements
Mohammed Ahsan deeply acknowledges King Fahd University of Petroleum and
Minerals (KFUPM) for the M.S. scholarship. Both authors acknowledge
KFUPM for supporting the experimental work.
Damage and Fracture Mechanics VIII, C. A. Brebbia & A. Varvani-Farahani (Editors)
© 2004 WIT Press, www.witpress.com, ISBN 1-85312-707-8
68 Damage and Fracture Mechanics VIII
References
[1] Marrow, T.J., and King, J.E., Microstructural and environmental effects on
fatigue crack propagation in duplex stainless steels. Fatigue and Fracture of
Engineering Materials and Structures, 17(7), pp. 761-771, 1994.
[2] Krishnan, K.N., Knott, J.F., Strangewood., M., Hydrogen Embrittlement
During Corrosion Fatigue of Duplex Stainless Steel. Proceeding of
Hydrogen Effects in Materials, Eds. A.W. Thompson & N.R. Moody, The
Minerals, Metals and Materials Society, pp. 689-696, 1996.
[3] Makhlouf, K., Sidhom, H., Triguia, I. and Braham, C., Corrosion Fatigue
Crack Propagation of a Duplex Stainless Steel X6 Cr Ni Mo Cu 25-6 in Air
and in Artificial Sea Water. International Journal of Fatigue, 25, pp. 167179, 2003.
[4] McIntyre, P., Interactions between hydrogen and steels during cyclic
loading. Corrosion Fatigue, eds. R. N. Parkins and Ya. M. Kolotyrkin, The
Metals Society, pp. 62-72, 1983.
[5] Marrow, T.J., King, J.E. and Charles, J.A., Hydrogen effects on fatigue in a
duplex stainless steel, Proceeding of the Fourth International Conference on
Fatigue and Fatigue Thresholds (Fatigue 90), ed. Kitagawa, H., Hawaii,
USA, pp.1807-1812, 1990.
[6] Hackerman, N. and Snavely, E.S., Inhibitors, Corrosion Basics: An
introduction, NACE, pp.17 – 148, 1984.
[7] Rafaey, S.A.M., Abd El-Rahim, S.S., Taha, F., Saleh, M.B. and Ahmed,
R.A., Inhibition of chloride localized corrosion of mild steel by PO43-,
CrO42-, MoO42- and NO2-, Applied Surface Science, 158, pp. 190-196, 2000.
[8] Dong, Z.H., Guo, X.P., Zheng, J.X. and Xu, L.M., Investigation on
inhibition of CrO42- and MoO42- ions on carbon steel pitting corrosion by
electrochemical noise analysis, Journal of Applied Electrochemistry, 32, pp.
395-400, 2002.
[9] McCafferty, E., Inhibition of the crevice corrosion of iron in chloride
solutions by chromate, Journal of Electrochemical Society, 126 (3), pp. 385390, 1979.
Damage and Fracture Mechanics VIII, C. A. Brebbia & A. Varvani-Farahani (Editors)
© 2004 WIT Press, www.witpress.com, ISBN 1-85312-707-8