Corrosion Testing of Weldable 13%Cr Supermartensitic Steel

TWI Ltd

Per Egil Kvaale
Technip Norge AS

Rolf E. Lye and Stig Lyder Støme
Lundin Norway AS

Presented at CORROSION 2012 Conference, NACE International, March 11-15, 2012, Salt Lake City, Utah, USA

 

Abstract

The use of weldable 13%Cr supermartensitic stainless steel (SMSS) for flowlines is becoming more widespread, largely due to its attractive price, high strength and good corrosion resistance in oxygen-free, high CO₂ brine environments. However, the sensitivity of this material to stress corrosion cracking (SCC) when H₂S is present precludes its use in all but very mild sour environments. 

This paper presents results from the corrosion qualification testing of butt welded DNV SML 13CR 2.5MO grade 13%Cr steel pipes, for use in a subsea flowline project. Welds were manufactured using a manual GTAW process and a superduplex stainless steel consumable, which were then subjected to a plastic straining cycle to simulate reeling, the preferred method of pipe laying. Stress corrosion cracking (SCC), and electrochemical polarization tests were carried out on the welds, in simulated formation water service environments with 178,000mg/l and 235,000mg/l chloride at pH4.5. The results showed no failures in four-point bend SCC tests stressed to 550MPa (100% of the SMYS of the parent material), when tested at 130°C in two simulated formation water environments. Supplementary tests using electrochemical cyclic polarization scans were also conducted on the parent material and HAZ of the welds in the same environments.

Introduction

Lundin Norway AS( ) is operator of the Brynhild field in the southern part of the Norwegian North Sea. The field is located approximately 20km from the Norwegian-UK border at an approximate latitude of 58°N. It will be developed with subsea wells and will be tied back to the Pierce field in the UK sector through an approximately 38km long flowline.

The Brynhild produced fluids are unsaturated oil and formation water which has a very high salt content. See Table 1 for reservoir and fluid data.

Table 1 Brynhild reservoir and fluid data

Parameter	Value	Dimension	Comments
Reservoir pressure	633	bar
Reservoir temperature	145	°C
Bubble point pressure	72	bar	At reservoir temperature
Wellhead shut-in pressure	390	bar
Pipeline operating pressure	50-100	bar
Pipeline operating temperature	90-135	°C
Design life	15	years
Produced fluid CO2 content	3.8	vol %	In the gas phase at the bubble point pressure
Produced fluid H2S content	13	ppm	This value may increase slightly through the lifetime due to water injection
Produced water salinity	276 000	mg/L	Total dissolved salt (TDS)
Produced water chloride content	178 000	mg/L
Produced water pH	>4.6		Not measured, estimated

Due to the CO₂ content and its high salinity the fluid is corrosive. Nevertheless, a production pipeline made from C-Mn steel with continuous corrosion inhibitor injection has previously been qualified. An average corrosion rate of less than 0.2mm/year was achieved, giving a required corrosion allowance of approximately 3mm.

Corrosion inhibitors are surface active chemicals which may affect the separation of the water from the oil. The effect on the mixed Pierce/Brynhild fluid is unknown, and therefore corrosion inhibition of the Brynhild fluids was not accepted from an operational point of view. A pipeline made from a corrosion resistant material was therefore required. The Brynhild fluid composition was outside previously qualified conditions with regard to chloride and temperature for welded ‘super’ 13% chromium supermartensitic stainless steel (SMSS), hence it was decided to test a welded SMSS pipeline material under simulated Brynhild conditions. The HAZ was considered an area of possible concern in terms of material performance.

This paper presents the results of the qualification assessment via stress corrosion cracking (SCC) testing, both for sulfide stress cracking and chloride induced stress corrosion cracking without the presence of H₂S. Technip Norway(1) also asked TWI(1) to conduct supplementary corrosion testing and evaluation of weldable 13%Cr SMSS. The tests were required for welding procedure qualifications in support of the planned developments. The stress corrosion cracking tests were conducted in simulated formation water at 130°C with various combinations of H₂S and CO₂ partial pressures in accordance with ISO15156-3¹ and EFC17² requirements, which were the acceptance criteria. Supplementary potentiodynamic polarization curve measurements were conducted at ambient and elevated temperature and pressure, to determine the pitting corrosion and repassivation potentials of the parent metal and HAZ. The resistance of the materials to environmental cracking and localized corrosion were evaluated to the above standards and current oil and gas industry practice.

Experimental procedure

Material and welding

A steel pipe in grade DNV SML 13CR 2.5MO (seamless, quenched and tempered) with 168.3mm OD and 8.5mm wall thickness was supplied by JFE for the qualification programme. Six girth welds were made using a manual GTAW procedure. All welding passes were deposited using a 2.4mm Zeron 100X( ) superduplex stainless steel (SDSS) filler wire.  After welding, post-weld heat treatment (PWHT) was applied for 5 minutes at 630°C(-0/+20ºC) followed by water cooling, representing industry practice.

Six sections of the welded 13%Cr steel pipes were supplied for testing, with at least 150mm of parent material either side of the centrally located weldment in each case, in the PWHT, but unstrained condition.

Simulated reeling and post-strain aging

Two, full length segments with an approximate weld length of 65mm were cut from each weld. These were subjected to a compressive and tensile plastic strain cycle to simulate reeling conditions (intrados location), using a servo hydraulic tensile test machine. A strain cycle of 0%, -1.01%, 0%, -0.82%, +0.42% was used (Figure 1).

Following the plastic strain cycle, the weld segments were subjected to artificial aging for a period of 1 hour at 250ºC.

Weld characterization

Tensile tests at 130ºC

Tensile tests were conducted at 130°C to determine the tensile properties of the as-received 13%Cr parent and weld metal. For the parent, duplicate standard subsize round specimens (4mm in diameter and 16mm gauge length) were used. For the weld metal specimens, duplicate, all weld metal, standard subsized round specimens (2.5mm diameter and 10mm gauge length) were taken longitudinally from the weld metal.

Hardness measurements

Hardness measurements were conducted on sections taken from the as-received and plastic strained welds. Vickers hardness traverse measurements were conducted at 1mm from the weld cap and weld root face and at mid-section of the weld using a 5kg load. The measurements were taken at an interval of 1mm in the parent and weld metal and at 0.5mm in the HAZ.

Metallography

Metallographic sections were taken from as-received welds and plastic strained welds. These were prepared using standard metallographic techniques and examined using a light microscope. Photographs were taken at regions of interest in the un-etched and etched surface conditions, with special attention paid to the weld root regions.

Corrosion tests

Simulated formation water test environments

A total of three test environments were selected for the test programme, based upon the analysis of the formation water from the two developments concerned. The composition and test temperatures were:

FW1a: 178,000mg/l Chloride (Brynhild), with 1mbara H₂S and 2.3bara CO₂, acidity adjusted to pH4.5 with NaHCO₃, 130ºC.
FW1b: 178,000mg/l Chloride (Brynhild), with 2mbara H₂S and 2.3bara CO₂, acidity adjusted to pH4.5 with NaHCO₃, 130ºC.
FW2: 235,000mg/l Chloride (Krabbe), with no H₂S and 3.1bara CO₂, acidity adjusted to pH4.5 with NaHCO₃, 130ºC.

FW1 test solutions were made using NaCl and deionized water and the acidity adjusted to pH4.5 by the addition of NaHCO₃, whilst being purged with 1bara CO₂ at ambient temperature.

FW2 test solution exceeded the solubility of NaCl at ambient temperatures so a mixture of NaCl and KCl (ratio NaCl:KCl of 200:239.1 g/l) was dissolved in the deionized water at 90ºC to achieve the required chloride concentration. The acidity of FW2 solution was also adjusted to pH4.5 by the addition of NaHCO₃, whilst being purged with 1bara CO₂, at ambient temperature.

For the SCC tests the pressure was set at 22.7bara, and the test gas composition used for each condition was calculated to provide the required partial pressure of H₂S and CO₂ assuming the contribution of steam (water vapor) pressure at 130ºC was 2.7bara.

For the polarization scan tests, the total pressure used was less, 8.7bara, due to the limitations of the test equipment. The test gas composition used for each condition was again calculated to provide the required partial pressure of H₂S and CO₂ taking account of the contribution of steam (water vapor) pressure at 130ºC. In addition to the above, for comparison, polarization scans were also conducted in FW1 solution at 25ºC, in 1bara CO₂ (FW1c).

The acidity of the test solution was also measured at the end of each test undertaken. pH measurements were conducted under 1bara CO₂, at ambient temperature to facilitate comparison with the measurements made before testing.

Stress corrosion cracking (SCC) tests

Three SCC tests were carried out according to conditions shown in Table 2. Specimens measuring 210x15x8mm were machined from the welds which had been plastically strained. These were taken transverse to the weld direction with the weld located at the centre of the specimen. The weld cap faces of the specimens were machined flat and the weld root was left in the as-welded condition.

The specimens were placed in constant deflection 4-point bend jigs with the weld root side (the test face) placed in tension and loaded to a stress of 550MPa (100% of the SYMS of the parent material, as determined by project requirements). Strain gauges were placed at the HAZ of the weld cap face, at either side of the weld, to measure the strain during loading and the stress was calculated from the strain imparted on the specimens assuming a Young’s modulus of 200GPa. The strain gauges were removed once the desired loads were achieved.

Triplicate specimens were exposed in each test environment in an autoclave for a period of 30 days. Purging of the solution with test gas was maintained throughout the duration of the test. At the end of the exposure period, the specimens were removed, washed and dried. They were examined for indications of corrosion and cracking both visually and by the use of a binocular microscope at approximately x20 magnification.

Metallographic sections were taken from one specimen from each test, either taken through areas of suspected pitting or cracking, or through the centre line in specimens showing no indication of pitting or cracking.

Polarization curve measurement

Three polarization curve measurement (PCM) tests were carried out according to conditions shown in Table 2 on duplicate specimens. Specimens of the parent material and HAZ were prepared from the as-received welds.

Table 2 Corrosion test matrix for the assessment of the 13%Cr SMSS welds

 

Test No (test type)	Material	Solution	Gas partial pressure, mbara		pH (*)	Temp., °C
			ρH2S	ρCO2
1 (SCC)	Welded 13%Cr steel	FW1a	1	2500	4.5	130
2 (SCC)	Welded 13%Cr steel	FW1b	2	2500	4.5	130
3 (SCC)	Welded 13%Cr steel	FW2	0	3100	4.5	130
5-1 (PCM)	13%Cr parent	FW1c	0	1000	4.5	25
5-2(PCM)	HAZ of the welded 13%Cr	FW1c	0	1000	4.5	25
6-1 (PCM)	13%Cr parent	FW2	0	3100	4.5	130
6-2 (PCM)	HAZ of the welded 13%Cr	FW2	0	3100	4.5	130
7-1 (PCM)	13%Cr parent	FW1a	1	2500	4.5	130
7-2 (PCM)	HAZ of the welded 13%Cr	FW1a	1	2500	4.5	130

 

SCC = Stress corrosion cracking
PCM = Polarization curve measurements
FW1 = 178000mg/l Cl-, pH 4.5
FW2 = 235,000mg/l Cl-, pH 4.5
(*) Initial test solution pH adjusted by adding NaHCO₃ under 1bar CO₂

The parent specimens were prepared from a region sufficiently remote from the weld to ensure freedom from any effects of the welding process. The HAZ specimens were prepared by cutting through weld metal and parent (etched to reveal the location), parallel to the fusion line either side of the HAZ and grinding back the parent material to expose the HAZ. The resultant HAZ specimen test face was approximately 1mm from the weld fusion line. The area of the exposed test faces of the HAZ specimens ranged from 0.65 to 0.87cm² and for the parent specimens from 1.83 to 2.08cm².

The test surface of each specimen was ground to 320grit finish prior to exposure to the test environment. A standard three electrode set up was used for the PCM tests, consisting of a working (test) electrode, platinized titanium counter electrode and a reference electrode. In the case of the 25ºC test (PCM 5), a glass cell and saturated calomel electrode (SCE) reference electrode were used, and in 130ºC (PCM 6, 7 and 8) tests, an alloy C276 autoclave vessel and a pressure-balanced Ag/AgCl reference electrode was used.

A computer controlled potentiostat was used for conducting the electrochemical polarization curve measurements. The rest potential, E_corr, of each specimen was monitored for a period of at least 1 hour prior to the polarization scan. The scans were conducted at a scan rate of 10mV/min from 50mV below (more negative than) the E_corr to a potential with a corresponding current density value 1mA/cm², at which point, the direction of the scans was reversed.

Results

Tensile tests at 130ºC

The data from the tensile tests are plotted in Figure 2. The average 0.2% proof stress (0.2%PS) of parent material at 130ºC was 642MPa and the UTS was 731MPa. When compared to the material supplier data for ambient temperature properties of 697MPa for the yield strength and 811MPa for the tensile strength a reduction of 55MPa for the 0.2%PS and 80MPa for the UTS was observed. Nevertheless, the values obtained at 130ºC were still above the specified minimum tensile properties according to information provided (550MPa Yield Strength and 700MPa Tensile Strength).

For the weld metal, both the 0.2%PS (479MPa) and UTS (678MPa) were significantly lower than the 13%Cr parent material tested at the same temperature and below the specifications for the 13%Cr parent material. This reduction in strength of the weld metal at 130ºC should be noted and taken into account in the design and qualification for operation of any pipeline made with this type of weldment. 

Tensile data from tests at room temperature, Table 3, show that the 0.2% proof stress of the SDSS weld metal was comparable to the parent SMYS for the SMSS. This shows that the parent and weld metal had matching mechanical strength under the conditions when reeling takes place.

Table 3 Room temperature all weld metal tensile data

Test spec.No.	Material condition	Dim. (mm)	Rp0.2 (MPa)	Rt0.5 (MPa)	Rm (MPa)	AGT (%)	A5 (%)
PA1	Strain- aged	4.02	801	768	981	18	32
PA2		4.03	797	770	978	18	35
AWT 1	As delivered	5.02	720	702	958	20	34
AWT 2		5.03	712	702	950	18	34
AWT 2		5.03	709	698	954	21	34

Hardness measurements

The results from the hardness measurement of the plastically strained and as-received welds are summarized in Table 4. A slight increase in the hardness of the weld root in the strained welds was observed compared to the as-received welds whilst there was a less noticeable increase for the corresponding weld cap measurements.

Figure 3 presents a plot of the weld root hardness values for Weld BW01, and shows that the increase in the average hardness values observed was mainly due to an increase in the parent hardness in the strained weld, probably due to strain hardening. Little or no increase in hardness was observed in the weld metal and HAZ regions when comparing strained and as-received welds.

Table 4 Summary of hardness measurements

Weld	Root		Cap
	As-received	Strained	As-received	Strained
	Min - Max Average	Min - Max Average	Min - Max Average	Min - Max Average
BW01	267-336 296	271-353 305	252-312 279	250-315 286
BW02	254-334 296	269-336 296	224-310 274	239-294 299
BW03	244-308 269	277-327 299	244-308 269	253-310 278

Simulated reeling

An example stress-strain plot of the plastic straining to simulate reeling of the six weld segments is shown in Figure 4.

Metallography

From metallographic examination of the welds no anomalies in the microstructure were observed in the parent, HAZ, or the weld metal in any of the welds examined in both the as-received and post-strained conditions.

Minor weld root lack of fusion features were observed in some of the welds examined. The morphology of the weld root features ranged in shape and size and Figure 5 shows one of the larger typical features observed, however, there were also a number of welds examined which did not exhibit root features (Figure 6). The root features observed were considered to be typical of those found in welds made with this combination of parent and welding consumable and was not considered to have a significant effect on the corrosion performance of the weldment.

Figure 5: An example of lack of fusion weld root feature from Weld BW03 after straining

Figure 6: An example of weld root showing no features from Weld BW01 after straining

SCC tests

The results for the three SCC tests are shown in Table 5 and Figures 7-10 show specimens from Test 1. No cracking was observed in any of the specimens after exposure. Therefore all specimens passed the acceptance criteria for material selection, that no cracking be present. The minor lack of fusion feature shown in Figures 9 and 10 were very similar to those observed in sections of welds which had not undergone SCC testing, Figure 5. Therefore they were considered to be present prior to testing and not as a result of the SCC test.

No pitting was observed in the 13%Cr parent or HAZ in all SCC specimens, and in all but one specimen in the weld metal, when examined after testing. A number of very small pits, of <0.25mm dia., were observed in the compression, non-test (machined cap) face in the SDSS weld metal in one out of the three specimens from SCC Test No. 3, Specimen BW2-B-PS-1. This region was sectioned and examined in more detail. The pits appeared to be hemispherical in shape, and since no corrosion was observed in the 13%Cr material in the same specimen, or in the other two specimens from the same set, they were thought to have resulted from inclusions or porosity which were not observed when the specimens were prepared prior to testing. These were therefore not considered to be detrimental.

No pitting or obvious indications of general corrosion, apart from very small localized surface discoloration, was observed in the 13%Cr parent or HAZ in all SCC specimens when examined after tests. Apart from the minor pits in the weld metal of one specimen, which may have been attributed to inclusions or porosity, all the specimens passed the material qualification criteria that no pitting be present.

Figure 9: Left side weld root of specimen BW01-B-PS-1 after exposure in Test 1.

Figure 10: Right side weld root of specimen BW01-B-PS-1 after exposure in Test 1.

Table 5 Summary results of SCC tests

Test No.	Specimen ID	Cracking	Pits	pH
				At start of test	At end of test
1	BW1-A-PS-1	no	no	4.5	4.8
	BW1-B-PS-1	no	no
	BW1-B-PS-2	no	no
2	BW3-A-PS-1	no	no	4.5	5.0
	BW3-B-PS-1	no	no
	BW3-B-PS-2	no	no
3	BW2-A-PS-1	no	no	4.5	5.7
	BW2-A-PS-2	no	no
	BW2-B-PS-1	no	Yes (*)

(*) pits in weld metal of non-test face <0.25mm dia.

Polarization curve measurement

The polarization scans of the parent and HAZ conducted in the three selected conditions are shown in Figures 11-14. The potential data from the high temperature tests, where a silver/silver chloride (Ag/AgCl (0.1M KCl)) reference electrode was used, has been adjusted to saturated calomel electrode (SCE) reference scale for ease of comparison. The polarization scans show, in general, the pitting potential of the HAZ was higher (less negative) than the parent for specimens tested in the same environment. This indicates that the HAZs were more resistant to pitting compared to the parent.

A possible explanation of this observation could be due to differences in the amount of Cr in solid solution in the HAZ, compared to the parent. The corrosion resistance in Cr steels is related to the volume fraction of chromium in solution in the steel matrix, as this is available to contribute to the passive layer. Supermartensitic steels are generally supplied tempered, to control strength, hardness and toughness. Although the steel has low carbon content, tempering will undoubtedly lower the chromium in solution via precipitation of Cr-rich phases, such as Cr₂₃C₆.

Also, stable austenite forms on tempering, which is likely to have lower chromium than the martensite from which it forms3. Compared to as-formed martensite, where all chromium is in solution and there is no stable austenite. These two microstructural changes in tempered parent steel will lead to (i) general and (ii) local reduction of chromium in solution, and hence corrosion resistance below the maximum that can be achieved for the given chromium content.

In the HAZ, where temperature exceeds Ac1 (about 550-650°C), chromium-rich phases are expected to dissolve and beyond about Ac1+100°C, stable austenite does not reform on cooling. Therefore, material heated in the HAZ to above about 600-700°C and cooled to room temperature is expected to consist mostly of martensite, with few precipitates (some may form during multipass welding but this will be less than during tempering). Hence this part of the HAZ may have higher intrinsic corrosion resistance than the original tempered parent steel. The width of this HAZ zone will depend upon the welding thermal cycle but is likely to be in the order of 1-2mm in a typical pipe girth weld.

Improved HAZ corrosion resistance is not typically observed in practice, as local oxidation from welding tends to reduce the surface corrosion resistance. In the present tests, the as-welded inner pipe surface was not tested, as the corrosion coupons were cut through thickness to provide specimens which were entirely HAZ material.

Overall the polarization data show that the materials exhibited acceptable corrosion behavior in the absence of polarization in the conditions assessed. For the less aggressive environment, the scans from the PCM 5 and PCM 7 tests show repassivation of the parent and HAZ after pitting when anodic polarization was reversed.

In the environment with increased chloride concentration the PCM 6 scans did not exhibit repassivation (E_repass<E_corr). The PCM 6 tests were repeated in PCM 8 to verify the results. The anodic curves of the scans for the HAZ specimens in PCM 8 tests showed initial passive behavior but no repassivation on reversal after breakdown (pitting). However, the hysteresis loops were less pronounced (smaller) compared to the PCM 6 test.

The anodic curve from the polarization scans of the parent material specimens in PCM 8 did not exhibit passive behavior, showing a continuous increase in the current with anodic polarization. Crevice corrosion of the specimens was confirmed as the cause upon examination after testing. Therefore the results from parent material specimens in this test were not considered to be valid, but the environment may be borderline for crevice corrosion when material is subject to polarization

Differences in the repassivation behavior observed between PCM 6 and 8 could be due to the degree of breakdown prior to reversal of polarization. Since the current density is a sum of the current from all active surface pits/regions of loss of passivity, a small number of larger locations will have the same total current as a larger number of smaller locations. Even though the current limit at which the reversal of anodic polarization was the same in both tests, this may explain the slight difference in the repassivation loops between these two tests. Nevertheless, the repeat PCM 8 tests confirmed that the material did not repassivate in this particular test environment.

The valid results from PCM 6 and 8 tests indicate that in FW2 solutions with 3.1bara CO₂ at 130ºC, if sufficient polarization is imparted to the system, e.g. by upset conditions, then this will cause corrosion and once initiated is unlikely to passivate. In PCM 5 and PCM 7 conditions, repassivation of stable pits was shown to occur once the driving force for the polarization was removed.

The difference between the rest potential and pitting potential indicates the degree of polarization, or the amount of driving force that is needed in the system, to cause pitting of 13%Cr parent and HAZ. In service this can result from upset conditions, where the temperature or pressure is increased or if there is ingress of oxygen into the system. The data show the highest differences obtained were from the 25ºC ambient temperature tests (PCM 5) with values ranging from 300 to 343mV, followed by those obtained in PCM 7 tests with values from 211 to 252mV, PCM 6 with values from 117 to 180mV, and the repeat test PCM 8 were 157 and 172mV. This indicates, in PCM 6 (FW2) conditions, the 13%Cr parent and HAZ will be more susceptible to pitting, or are likely to be more sensitive to upset conditions which may cause pitting, when compared to the other two service environments.

Conclusions

Materials characterization

The tensile properties of parent and weld metal were lower at 130ºC, compared with ambient temperature tensile properties data. For the parent, the tensile properties at 130ºC were still above the specifications for the parent material. For the SDSS weld metal, this was less than the parent material specifications and this should be taken into consideration in design and qualification for operation when using this type of weld.

No anomalies were found in the microstructure of the parent, weld metal or HAZ in the as-received and plastically strained welds examined.

There was a slight increase in the hardness of the parent material in the strained welds compared to the as-received welds. This difference was small, and was more pronounced in the weld root compared to the weld cap.

Corrosion testing

No cracking was observed in any of the SCC specimens in the three test environments; therefore all specimens passed the material qualification acceptance criteria to be free from cracking.

No pitting was observed in the 13%Cr parent or HAZ in any of the SCC specimens when examined after tests. Apart from the minor pits in the SDSS weld metal of one specimen, which may have been attributed to inclusions or porosity, all the specimens passed the material qualification acceptance criteria to be free from pitting.

In most of the polarization scans, the data indicated better corrosion resistance in the HAZ compared to the parent; this may be due to a slightly higher amount of Cr in solid solution in the HAZ compared to the parent resulting from differences in their thermal histories.

The degree of polarization required to cause pitting (E_p - E_corr), i.e. the susceptibility to pitting of the 13%Cr parent and HAZ in a particular test environment, was found to decrease in the following order PCM 5 > PCM 7 > PCM 6 and PCM8.

No repassivation of the test specimens was observed in tests carried out in PCM 6 test environment (FW2). If sufficient polarization is imparted to the system by upset conditions, this will cause pitting and once initiated, any stable pits formed are unlikely to passivate, even after the driving force for the polarization has been removed.    

Acknowledgements

The authors would like to thank Lundin Norway AS and Technip Norge AS for their contribution and kind permission to publish this work. The authors would also like to thank colleagues who have assisted in the test programme.

References

1.   NACE MR0175/ISO 15156 (2009) “Petroleum and natural gas industries—Materials for use in H₂S-containing environments in oil and gas production”, (Houston, TX: NACE).

2.   EFC 17 (2nd Edition) “Corrosion Resistant Alloys for Oil and Gas Production - Guidance on General Requirements and Test Methods for H₂S Service”: (EFC 17 - 2nd Edition, 2003).

3.  T.G Gooch, P. Woollin and A.G. Haynes, “Welding Metallurgy of Low Carbon 13% Chromium Martensitic Steels”, Supermartensitic Stainless Steels ’99, (Brussels, Belgium, Belgium Welding Institute, 1999), p.188.