Presented at Duplex 2000, 6th World Duplex Conference and Expo, Venice, Italy, 17-20
October 2000, published by Associazione Italiana di Metallurgia, pp 733-741 (ISBN 88-85298-37-0)
Abstract
Slow strain rate tests were performed on superduplex stainless steel weld metal containing intermetallic phase, to examine susceptibility to environmental cracking in sour media at 100°C. A range of intermetallic phase content and H2S partial pressure was studied. Results indicated 'moderate susceptibility' to cracking in four specimens out of a total of 15, whilst remaining specimens were 'immune' or 'practically immune'. Susceptibility was associated with the presence of cleavage fracture on the faces of the fractured tensile test specimens. No correlation was found between the volume fraction of intermetallic phase over the range studied (local maximum values of 0-20%) and the susceptibility to cracking, suggesting that particle size might be more important in determining susceptibility.
Introduction
Superduplex stainless steels are employed by the oil and gas industry for their resistance to fluids containing water with CO
2S, H
2S and chloride ions at elevated temperature, combined with high strength and toughness. Superduplex steels have a propensity for formation of intermetallic phases (notably sigma phase) during welding, which is typically controlled to an acceptable level via heat input and interpass temperature control. However, complete avoidance of intermetallic phases in superduplex welds is not always practical. Considerable effort has been expended in the study of intermetallic phase formation in duplex and superduplex steels, and its effect on toughness and corrosion resistance, as reviewed by Karlsson
[1] . Nevertheless, there is little published information on the effect of intermetallic phase on the performance of duplex and superduplex steels in sour media of relevance to the oil and gas industry, in particular with respect to environmental cracking in sour service
[2] .
A study was undertaken to examine the environmental cracking susceptibility of superduplex weld metal with intermetallic phase in sour media via slow strain rate tensile (SSRT) testing.
Experimental
Welding
Welds were produced in UNS S32974 superduplex pipe with an outer diameter of 219mm and a wall thickness of 13.2mm using the TIG (GTAW) process throughout. Filler wire designed to weld UNS S32760 was employed (2.4 and 3.2mm) with Ar shielding and backing gases. The allowable composition ranges of the pipe and filler wire are given in Table 1. Typical analyses of the materials are given in Table 1. All welding was vertical up, in the 5G position, i.e. with the pipe horizontal and flat. Arc energies were in the range 0.8 to 1.8 kJ/mm.
Table 1 Composition ranges of materials used
| | Element, wt% |
|---|
| C | Cr | Ni | Mo | Cu | W | N |
|---|
| UNS S32974 pipe |
0.03 |
24.0-26.0 |
6.0-8.0 |
2.5-3.5 |
0.2-0.8 |
1.5-2.5 |
0.24-0.32 |
| Filler wire for UNS S32760 |
0.03 |
24.5-26.0 |
9.0-10.0 |
3.5-4.0 |
0.5-1.0 |
0.5-1.0 |
0.20-0.30 |
| Figures quoted are maxima except where a range is given. |
Specimen production
The welds were sectioned to allow metallographic examination specifically for intermetallic phases. It was found that intermetallic phase formation was concentrated in the root pass, with lower levels in second and third passes. There was no precipitate in subsequent passes, representing the cap half of the weld. Consequently, 16 cylindrical all-weld metal tensile specimens were machined from as close as possible to the root of the weld, for SSRT testing, to sample as much intermetallic phase as possible. Initial tests showed that cross-weld specimens were inadequate for studying weld metal behaviour as plastic strain concentrated in the parent steel. A further series of three all-weld metal specimens was machined from samples heat treated at 750°C for 11.5 minutes to increase the level of intermetallic phase in the weld metal. This treatment was selected after trials had shown that this gave the desired level of precipitate.
Slow strain rate testing
The specimens were pulled to failure at a strain rate of 1 x 10
-6s
-1, in deoxygenated 25%NaCl solution with 34mg/l NaHCO
3 (calculated using CORMED
TM [3] software to give the desired test pH of 4.0 at the test temperature and pressure) at 100°C. Deoxygenating was by nitrogen purging until the measured oxygen content was <10ppb. The solution was charged by continuous purging with a pre-mixed gas containing CO
2, H
2S and N
2, which was also bubbled continuously for 12 hours prior to testing, to allow saturation. Three levels of H
2S were examined, namely 0.02 psi (0.0014 bar), 0.04 psi(0.0028 bar) and 0.25 psi (0.017 bar) and a CO
2 partial pressure of 35 psi (2.4 bar) was used in each case. Testing was generally in accordance with the recommendations of EFC 17
[4] .
After test, specimens were examined in a scanning electron microscope for evidence of environmentally assisted cracking on the gauge length and on the fracture face. One half of each failed specimen was sectioned longitudinally through the fracture face and the volume fraction of intermetallic phase was point counted adjacent to the fracture face. Intermetallic phase was found to be distributed non-uniformly, with patches showing fairly dense precipitation whilst the bulk of the weld metal had little or no intermetallic phase. Point counting was performed according to ASTM E562 in the area identified as having the locally highest volume fraction of intermetallic phase within the region of interest. Sixteen fields of 100 points were used, at a magnification of x2800. This is equivalent to a measurement area of 0.001mm2 per field with the 16 fields being located in an area of around 0.1mm2. The average volume fraction of intermetallic phase over the whole of the weld area of interest, i.e. the root and second pass; equivalent to 20-25mm2 was estimated also, although insufficient point counting was performed to quantify the latter accurately. The strain to failure of each specimen tested in the sour media was recorded and normalised with respect to data from triplicate similar specimens, both as-welded and heat treated, tested in air at 100°C. The normalised values provide an indication of the susceptibility to environmental cracking of the individual specimens [5] .
Results
Table 2 lists the results of the various SSRT tests, together with the measured volume fractions of intermetallic phase. The approximate range of width of intermetallic particles was around 1-3µm for the as-welded specimens and slightly greater, around 1-5µm for the heat-treated specimens. Examples of the intermetallic phases present in as welded and heat-treated weld metals are shown in Figs. 1 and 2. Figure 3 plots strain to failure against maximum local volume fraction of intermetallic phase for all specimens tested in air and the sour media. A well-defined trend of decreasing elongation with fraction of intermetallic phase was observed, although associated with a considerable degree of scatter. Figure 4 plots the normalised strain to failure for each specimen against the peak volume fraction of intermetallic phase measured adjacent to the fracture face. The plot again shows considerable scatter with respect to the normalised strain to failure. Although the scatter would mask any trend, it is apparent that there is no strong influence of the volume fraction of intermetallic phase on environmental cracking susceptibility, in terms of normalised strain to failure, in this instance. However, it should be noted that four specimens showed evidence of cleavage in the ferrite phase on the fracture face, presumably environmentally-induced, Fig.5. Of these four specimens (marked by an asterisk in Figs.3 and 4), three gave the lowest normalised strains to failure of all specimens, in the range 0.59 to 0.68, and the fourth gave a value of 0.78, which was the joint fourth lowest value. Even allowing for the scatter in the data, there was a trend of low normalised strain to failure for specimens with environmental cracking on the fracture face.
Table 2 Summary of SSRT results from tests on superduplex stainless steel all-weld metal specimens
a) sour media (25%NaCl, pH=4.0, 100°C)
| Specimen | Partial pressure of H2S (psi) | Vol fraction of intermetallic (%) | Strain to failure (%) | Normalised strain to failure | Evidence of environmental cracking |
|---|
| Local maximum* | Average in root | Fracture face | Gauge length |
|---|
| W1 |
0.25 |
3.0 |
<1.0 |
24 |
0.92 |
ductile |
shallow etching |
| W2 |
0.25 |
4.0 |
<1.0 |
27 |
1.04 |
ductile |
shallow etching |
| W3 |
0.25 |
4.0 |
<1.0 |
28 |
1.08 |
ductile |
no corrosion |
| W4 |
0.25 |
4.0 |
<1.0 |
26 |
0.96 |
ductile |
shallow etching |
| W5 |
0.25 |
1.0 |
<0.5 |
16 |
0.59 |
some cleavage |
shallow etching |
| W6 |
0.25 |
1.0 |
<0.5 |
21 |
0.78 |
some cleavage |
shallow etching |
| W7 |
0.02 |
1.0 |
<0.5 |
18 |
0.67 |
some cleavage |
no attack |
| W8 |
0.02 |
0.5 |
<0.5 |
25 |
0.93 |
ductile |
no attack |
| W9 |
0.02 |
0.5 |
<0.5 |
22 |
0.81 |
ductile |
no attack |
| W10 |
0.04 |
3.0 |
<1.0 |
21 |
0.82 |
ductile |
no attack |
| W11 |
0.04 |
0.5 |
<0.5 |
23 |
0.90 |
ductile |
no attack |
| W12 |
0.04 |
2.0 |
<0.5 |
20 |
0.78 |
ductile |
no attack |
| W13 |
0.04 |
4.5 |
<1.0 |
20 |
0.78 |
ductile |
no attack |
| W14 |
0.04 |
2.0 |
<0.5 |
23 |
0.90 |
ductile |
no attack |
| W15 |
0.04 |
7.0 |
1.5 |
21 |
0.82 |
ductile |
no attack |
| HT1 |
0.04 |
19.5 |
12.5 |
9 |
0.68 |
some cleavage |
no attack |
| HT2 |
0.04 |
13.0 |
9.0 |
14 |
0.91 |
ductile |
no attack |
| HT3 |
0.04 |
15.5 |
9.5 |
17 |
1.10 |
ductile |
no attack |
| * measured over approximately 0.1mm2 |
b) air at 100°C
| Specimen | Vol fraction of intermetallic (%) | Strain to failure (%) |
|---|
| Local maximum* | Average in root |
|---|
| W16A |
3.0 |
<1.0 |
24 |
| W17A |
3.0 |
<1.0 |
28 |
| W18A |
3.0 |
<1.0 |
26 |
| W19A |
2.0 |
<1.0 |
26 |
| W20A |
1.5 |
<0.5 |
27 |
| W21A |
1.0 |
<0.5 |
27 |
| W22A |
3.5 |
<1.0 |
27 |
| W23A |
4.0 |
<1.0 |
24 |
| HT4A |
13.5 |
9.5 |
7 |
| HT5A |
15.0 |
12.0 |
13 |
| HT6A |
16.5 |
13.0 |
15 |
| * measured over approximately 0.1mm 2 |
The specimens showing areas of cleavage fracture spanned the range of H2S partial pressure and the range of maximum volume fraction of intermetallic phase: (i) 0.25psi and 1% (two examples) (ii) 0.02psi and 1%, and (iii) 0.04psi and 19.5%. Example (ii) corresponded to a specimen containing a small subsurface weld flaw, adjacent to which the area of cleavage had developed. The other examples represent (i) a specimen at the low end of the intermetallic range studied, tested at the highest partial pressure of H2S and (ii) the specimen with the highest level of intermetallic phase, tested at a low partial pressure of H2S. In each case, the cleavage was localised and at or close to the surface, being subsurface in two instances. There was no evidence of secondary cracking on any of the specimens but some very shallow surface etching was observed on the gauge lengths of specimens tested at the highest H2S level.
Metallographic sections revealed that in the heat treated specimens in particular, which contained the greatest amount of intermetallic phase, numerous particles had fractured during test but typically without extending into the surrounding material ( Fig.6 and, to a lesser extent, Figs.1 and 2).
Discussion
Mechanism of cracking
The results indicate no link between the point counted volume fraction of intermetallic phase and the sensitivity to environmental cracking for the conditions tested. The size of the intermetallic phase particles was similar in each specimen, being of the order of 1-5µm, and largely defined by the width of the ferrite units in the TIG root weld metal examined. Although some specimens had much lower levels of intermetallic phases than others, all showed a tendency to contain clusters of particles. Other authors have found that the intermetallic particle size is more important than the volume fraction in terms of influence on measured properties, notably pitting corrosion resistance and toughness
[6] . It is suggested that a similar relationship may apply in the case of environmental cracking in sour environments also, although particle size was not varied significantly in the present study.
The observations made suggest that environmental cracking in the SSRT test was hydrogen related, giving rise to patches of cleavage in the ferrite phase. This cleavage presumably initiated at a fractured intermetallic particle or a group of particles, with the hydrogen originating from corrosion. However, no significant corrosion was observed on the specimens, just some shallow surface etching. This might have provided sufficient hydrogen for small areas of cleavage fracture but it is also possible that localised crevice attack at cracked intermetallic particles on the surface may have been involved. The location and shape of the patches of cleavage presumably reflected the distribution of intermetallic phase within the individual specimens. Such a cracking mechanism would be favoured by the plastic strain induced in an SSRT test. However, the test is of fairly short duration, typically 2-4 days, and for extended exposure periods, hydrogen generation and uptake may be much higher. Consequently, it is possible that a similar mechanism may operate at much lower strains over longer time periods.
Practical significance of results
It may be noted that the H
2S levels employed in the tests were well below the limit of 0.1 to 0.2 bar for superduplex stainless steels in concentrated chloride solutions, which are listed in NACE MR0175-99. Hence, environmental cracking would not be anticipated for parent material in the absence of intermetallic phases. It is not known whether weld metal without intermetallic phases has significantly lower resistance than parent steel, although it is normally assumed that this is not the case. Hence, it seems that the intermetallic phase has increased sensitivity to environmental cracking.
The SSRT test is acknowledged as a severe test due to the forced plastic straining of the specimen. McIntyre et al [5] proposed a system of classifying SSRT test results and those specimens showing sensitivity here, with normalised strains to failure of 0.59 - 0.78, may be described on 'moderately susceptible' using this system. McIntyre et al stated that environmentally induced brittle fracture on the final fracture surface occurs in moderately susceptible materials, with ductility ratios between 0.5 and 0.75. Remaining specimens were 'immune' or 'practically immune' using the same system. It may be noted that Cooling et al [7] adopted a minimum normalised strain to failure of 0.5 for acceptability for service of 13%Cr steel for downhole applications.
Conclusions
Slow strain rate tensile tests were performed on all-weld metal superduplex stainless steel specimens, to examine the effect of intermetallic phases on susceptibility to environmental cracking in sour media. As-welded and heat treated materials were examined. The distribution of intermetallic phase was inhomogeneous and local maximum volume fractions of 0-20% were present in the specimens tested.
- The strain to failure of all-weld metal superduplex stainless steel specimens decreased with increasing local intermetallic phase content in the range 0-20%.
- The normalised strain to failure in SSRT tests of all-weld metal superduplex stainless steel specimens tested in sour environments with 0.02 to 0.25psi H2S, 25% NaCl and pH = 4.0 at 100°C showed no systematic variation with the volume fraction of intermetallic phase.
- Some SSRT specimens showed evidence of environmental cracking in the form of small areas of cleavage fracture on the faces. These specimens gave the lowest normalised strains to failure of 0.59-0.78.
- It is suggested that the cracking observed was hydrogen-assisted and initiated at fractured intermetallic particles. As such, the intermetallic particle size was probably more significant than the volume fraction.
References
| N° | Author | Title |
|---|
| 1 |
Karlsson L |
'Duplex stainless steel weld metals - effects of secondary phases', Prof Conf 'Duplex Stainless Steels 97', Maastricht, KCI Publishing 1997, paper 204, 43-58. |
| 2 |
Francis R, Byrne G and Warburton G, |
'The role of environmental and metallurgical variables on the resistance of duplex stainless steels to sulphide SCC', Prof Conf 'Corrosion 97', NACE International 1997, paper 12. |
| 3 |
|
Societe Nationale Elf-Aquitaine (production) CORMED TM software, version 1.1, copyright SNEA, 1990. |
| 4 |
|
EFC Publication No.17 'A working party report on corrosion resistant alloys for oil and gas production: Guidance on general requirements and test methods for H2S service', Institute of Materials, 1996. |
| 5 |
McIntyre D R, Kane R D and Wilhelm S M |
'Slow strain rate testing for materials evaluation in high-pressure H2S environments' Corrosion 44 (12) 920-926. |
| 6 |
Francis R and Warburton G R, |
'A model for corrosion of the depleted zones around sigma precipitates produced during welding of superduplex stainless steel' Proc Conf 'Stainless Steel World 99', KCI Publishing 1999, paper 6, 711-720 |
| 7 |
Cooling P J, Kermani M B, Martin J W and Nice P I, |
'The application limits of alloyed 13%Cr tubular steels for downhole duties', Prof Conf 'Corrosion 98', NACE International 1998, Paper 94. |