Low alloy steels combine relatively low cost with exceptional mechanical properties, making them commonplace in oil and gas equipment. However, their strength and hardness are restricted for sour environments to prevent different forms of hydrogen embrittlement. Materials used in sour services are regulated by the ISO 15156-2 standard, which imposes a maximum hardness of 250 HV (22 HRC) and allows up to 1.0 wt% Ni additions due to hydrogen embrittlement concerns. Low alloy steels that exceed the ISO 15156-2 limit have to be qualified for service, lowering their commercial appeal. As a result, high-performing, usually high-nickel, low alloy steels used successfully in other industries are rarely considered for sour service. In this work, the hydrogen stress cracking resistance of the high-nickel (3.41 wt%), quenched and tempered, nuclear-grade ASTM A508 Gr.4N low alloy steel was investigated using slow strain rate testing as a function of applied cathodic potential. Results showed that the yield strength and ultimate tensile strength were unaffected by hydrogen, even at a high negative potential of −2.00 VAg/AgCl. Hydrogen embrittlement effects were observed once the material started necking, manifested by a loss in ductility with increasing applied cathodic potentials. Indeed, A508 Gr.4N was less affected by hydrogen at high cathodic potentials than a low-strength (yield strength = 340 MPa) ferritic-pearlitic low alloy steel of similar nickel content. Additionally, hydrogen diffusivity was measured using the hydrogen permeation test. The calculated hydrogen diffusion coefficient of the ASTM A508 Gr.4N was two orders of magnitude smaller when compared to that of ferritic-pearlitic steels. Hydrogen embrittlement and diffusion results were linked to the microstructure features. The microstructure consisted of a bainitic/martensitic matrix with the presence of Cr23C6 carbides as well as Mo- and V-rich precipitates, which might have played a role in retarding hydrogen diffusion, kept responsible for the improved HE resistance.

Oil and Gas Environments

A large amount of recoverable oil and gas (O&G) reserves pose significant challenges to materials. These challenges can be associated with deep-water offshore fields, reservoirs in arctic locations—where materials need to withstand temperatures as low as −60°C—or high-temperature (>177°C) and high-pressure (>103 MPa) fields (HPHT).1-2  Another critical factor is the likely presence of atomic hydrogen, generated at the surface by corrosion reactions or cathodic protection systems, that can lead to hydrogen stress cracking (HSC). Sulfide stress cracking (SSC) is a particular case of HSC when H2S is present in the production fluids or the environment.2  Both HSC and SSC are considered hydrogen embrittlement (HE) phenomena, where generally ductile—albeit susceptible—materials can fracture in a brittle manner, leading to catastrophic failures.

Low alloy steels (LASs) are the preferred choice for many applications in O&G equipment. In fact, LASs are, by volume, the most widely used alloy systems for critical O&G components.2-3  The popularity of LAS in O&G production equipment can be attributed to their excellent mechanical properties, achievable through appropriate heat treatment and processing, and their relatively low cost.2,4  LASs with high strength and toughness could play an important role in the development of unconventional O&G reserves. It is believed that LASs with improved hardenability, excellent toughness, ductile-to-brittle transition temperatures (DBTTs) below −60°C, and yield strengths above 690 MPa (100 ksi) could be essential to overcome design challenges in extreme O&G environments, e.g., HPHT fields or fields in arctic regions.1  Molybdenum (Mo), chromium (Cr), and nickel (Ni) can be added to the steel composition to achieve higher hardenability and, therefore, uniform properties through-thickness for larger wall thicknesses and heavy forgings. It has been shown that Ni further improves toughness, strength and can lead to grain refinement in LASs.5-6  However, LASs have been found to be susceptible to HE in H2S and hydrogen-bearing environments.7 

Today, the selection of LASs for sour service (i.e., fields containing measurable amounts of H2S) is governed by the ISO(1) 15156-2 Standard.8  ISO 15156-2 restricts LASs to a nickel content <1.0 wt% and imposes a hardness limit of 250 HV (22 HRC) to surfaces directly exposed to H2S, among other requirements. Alloys that do not meet the ISO 15156-2 requirements must undertake a comprehensive qualification testing program. As a result, many technologically advanced LASs with improved strength and toughness are, nowadays, excluded in practice from sour service environments. As summarized by Kappes, et al.,1  and Craig,9  the evidence collected since the early 1960s points to a much more complex interplay between a LAS’s susceptibility to SSC and its microstructure, chemical composition, impurities, and thermomechanical processing. The sole restriction on Ni remains highly controversial. In the data compiled by Kappes, et al.,1  certain microstructures, i.e., tempered martensite and lower bainite, seemed to improve SSC resistance while untempered martensite drastically reduced it. Moreover, LAS with a ferritic-pearlitic microstructure systematically had a lower SSC resistance than quenched and tempered (Q&T) LAS at the same hardness and strength levels. Further clarification is needed on the exact role of Ni on SSC resistance. Nevertheless, it has been proposed that qualifying higher strength, higher toughness LASs with Ni contents > 1.0 wt% for sour service could be a technological enabler to develop the O&G reserves of the future.1,3 

ASTM A508 Gr4N: A High-Strength Low Alloy Steels for Oil and Gas Applications

The nuclear-grade ASTM(2) A508 Gr.4N alloy is a Ni-Cr-Mo high-strength LAS (HSLAS) that has been developed for reactor pressure vessel (RPV) forgings.5  The ASTM A508/A508M standard specifies the chemical composition, mechanical properties, and manufacturing processes.10  The Grade 4N designation is a high-Ni (2.8-3.9 wt%) LAS with a specified minimum yield strength (SMYS) of 690 MPa (100 ksi). Compared to conventional RPV steels, A508 Gr.4N exhibits improved strength, toughness, DBTT, and hardenability.11  Therefore, A508 Gr.4N has been proposed as the future of RPV steels,5  where, among other material degradation concerns, resistance to HE plays an important role.

A thorough investigation by Park, et al.,12  linking microstructure and mechanical properties with various Ni, Cr, and Mo compositions showed that Ni was the most effective alloying element in increasing both strength and toughness. The improvement in mechanical properties was achieved through solid solution strengthening of the matrix, refinement of the microstructure (prior austenite grains and sub-grain structures), and an increase in the martensite volume fraction. Ni does not affect the precipitation behavior, whereas higher additions of Cr suppress cementite formation, enhancing the precipitation of M23C6 and M7C3 carbides and leading to a reduced carbide size and uniform carbide distribution. The combination of these properties, i.e., refined microstructure and small, homogeneously distributed carbides, is believed to be most favorable to resist HSC and SSC.13 

As discussed by Depover, et al.,14  microstructural features influence the HE resistance of LASs. In this regard, Nagao, et al.,15  investigated the influence of carbide formers due to their potential beneficial effect on the HSC resistance of LASs. According to this work, carbides acted as traps reducing the amount of diffusible hydrogen, which is thought to play a detrimental role in hydrogen embrittlement. The trapping capacity of chromium carbides was quantified by Depover and Verbeken16  using thermal desorption spectroscopy (TDS), slow strain rate testing (SSRT), and hydrogen permeation tests. The work indicated the beneficial effect of small M23C6 carbides (less than 100 nm) as they acted as hydrogen traps hindering hydrogen diffusion. The effect of molybdenum additions on HSC performance was studied by Hinotani, et al.,17  on medium carbon steel with 3.9 wt% Mo. The authors used a tempering stage to precipitate Mo2C and evaluated its effect on hydrogen diffusion. It was concluded that after tempering, the fine Mo2C carbides decreased hydrogen diffusivity abruptly. The addition of vanadium as a carbide former was evaluated by Lee, et al.,18  on tempered martensitic steels with and without V. It was concluded that those steels with V additions trapped more hydrogen and were more resistant to HE than the V-free baseline steels. These findings on Mo2C and V carbides were also observed by Depover and Verbeken.19-20 

The ASTM A508 Gr.4N LAS could be a promising candidate to be introduced in the O&G industry for components where improved strength, hardenability, and excellent toughness are critical. However, limited information is present in the literature regarding A508 Gr.4N’s compatibility with typical oil and gas environments. Therefore, in this work, the HSC resistance and hydrogen transport properties of A508 Gr.4N were evaluated under different hydrogen-charging conditions using SSRT and hydrogen permeation testing, respectively. In addition, microstructural features were characterized and their effect on the diffusion properties discussed. Finally, results for the ASTM A508 Gr.4N LAS were compared against a LAS with similar Ni composition but different microstructure to evaluate the role of the microstructure on the hydrogen embrittlement resistance and diffusion properties.

Materials

Forged ASTM A508 Gr.4N plates were fabricated in a lab-scale furnace from 50 kg ingots to the ASTM A508/A508M-1610  specification. The dimensions of the plates were 300 mm × 250 mm × 35 mm (length × width × thickness). The forging reduction ratio was 4.0, and the chemical composition of the produced plates is shown in Table 1. The heat-treatment condition is described in Table 2 and consisted of an austenitizing step at 860°C for 12 h, followed by quenching using fan cooling with a cooling rate of 50°C/min. This corresponds to the cooling rate of water cooling with a plate thickness of 100 mm. Tempering was done at 620°C for 10 h, and the plates were left to cool in the furnace. Table 3 lists the mechanical properties of the ASTM A508 Gr.4N used in this investigation, as provided by the supplier in the materials certificate of compliance.

Table 1.

Chemical Composition ASTM A508 Gr.4N

Chemical Composition ASTM A508 Gr.4N
Chemical Composition ASTM A508 Gr.4N
Table 2.

Heat Treatment Conditions

Heat Treatment Conditions
Heat Treatment Conditions

A 25 mm × 25 mm 99.99% palladium membrane with a 0.775 mm thickness was used to verify the reproducibility of the hydrogen permeation technique. Palladium was selected because it is an inert material, and its diffusion parameters are widely known.

Metallographic Preparation and Analysis

For the metallographic preparation, samples were cut from the forged plates, then mounted in epoxy resin and cured overnight at room temperature before grinding and polishing. The final polishing step was a chemical-mechanical polish with 0.04 μm colloidal silica suspension. Subsequently, samples were etched immediately after polishing for 30 s in a 2.0% Nital solution.

The microstructure characterization was performed using light optical microscopy (LOM), scanning electron microscopy (SEM) coupled with electron backscatter diffraction (EBSD). SEM micrographs and EBSD data were obtained with a field emission Tescan Mira3 SEM using an accelerating voltage of 15 kV, and the step size for the EBSD data collection was 0.11 μm.

High-resolution imaging was performed using transmission electron microscopy (TEM) and scanning transmission electron microscopy (STEM), which was coupled with energy dispersive x-ray spectroscopy (EDS) for elemental composition analysis. TEM samples were extracted and prepared using a high-resolution focused ion beam (FIB) Tescan LYRA3 GM instrument with a Ga+ ion source. The TEM characterization was performed using a FEI Talos FS200X G2 with a field emission gun (FEG) operating at 200 keV. Diffraction patterns were taken for further crystallographic characterization using selected area diffraction (SAD) of 40 μm.21 

Slow Strain Rate Testing

Rectangular bars from the as-received heat-treated plates were first cut in the S-T plane, as shown in Figure 1(a). Cylindrical full-size SSRT samples that met the specifications outlined in the ASTM G129-00 standard,22  were then machined in a lathe with final dimensions as shown in Figure 1(b). The gauge section was further ground using SiC sandpaper with grit sizes of P1200, P2500, and P4000 to achieve the required surface roughness of 0.25 μm or less. The final grinding step with P4000 sandpaper was performed parallel to the gauge section to minimize grinding marks perpendicular to the later applied load. Both end sections of the specimens were painted with Dulux Duramax metal primer and Dulux Duramax enamel topcoat to (i) electrically insulate the specimen from the metallic bottom and top plates of the environmental chamber and (ii) solely expose the gauge section of the test sample to the cathodic polarization. After each coating, the paint layers were cured overnight.

FIGURE 1.

Full-size SSRT test sample. All dimensions are given in mm.

FIGURE 1.

Full-size SSRT test sample. All dimensions are given in mm.

Close modal

The SSRT tests were performed at 20±2°C with a constant strain rate of 1 × 10−6 s−1 in a deaerated 3.5% NaCl solution. Deaeration was performed by N2 bubbling to evacuate any chlorine formation. Due to the presence of an electrolyte, it was impractical to use an extensometer during in situ testing, therefore the machine compliance was included in the measured elongation. However, once yielding starts, most of the recorded displacement is associated with the gauge section plastic deformation.23  SSRT specimens were immersed in the test solution within the environmental chamber and cathodically polarized to three different hydrogen charging conditions to study the effect of hydrogen embrittlement, as follows:

  • Eapp,1 = −1.05 VAg/AgCl

  • Eapp,2 = −1.40 VAg/AgCl

  • Eapp,3 = −2.00 VAg/AgCl

Two tests were performed for each condition. The samples were precharged for 48 h inside the SSRT equipment without any load applied before the tensile test. The precharging time and potentials were selected to compare the A508 G4.4N results with previous work by Husby, et al.,24  on a ferritic pearlitic LAS with similar nickel content. Cathodic charging continued at the same predefined potentials during SSRT straining. On average, tests took 4 d until fracture. Additionally, three SSRT tests were performed in the air to obtain baseline results.

SSRT results are typically plotted as comparative stress-strain curves for the control and the test environments. Further, SSRT results can be quantified calculating several ratios, as follows:25 
formula

In the present study, the evaluation of the SSRT results was performed in accordance with the recommendations given in the ASTM G12922  and NACE(3) TM109826  Standards. The calculated SSRT ratios are outlined in Equations (2) through (5).

Time to Failure Ratio (RTTF):
formula
where TTFe and TTFc are the time to failure in the test and the control environment, respectively.
Reduction of cross-sectional area ratio (RRA):
formula
where RAe and RAc are the reduction of cross-sectional area in the test and control environment, respectively. The areas of the fractured samples were calculated from measurements obtained with a stereo microscope with samples laying horizontally. The measurements correspond to the length across the thinnest part of the sample.
Reduction of plastic elongation ratio (RE):
formula
where Ee and Ec are the reduction of plastic elongation in the test and control environment, respectively. To calculate the strain, the plastic elongation was obtained by the SSRT software.
Reduction of toughness (area under the stress-strain curve) ratio:
formula
where Te and Tc are the reduction of toughness/area under the stress-strain curve in the test and control environment, respectively.

Fractography

SEM analysis of the fracture surface and the gauge section was performed to evaluate the fracture mode and to observe possible secondary cracking as well as the correlation between crack propagation and microstructural features.

Hydrogen Permeation Testing

Coin-shaped samples were cut from the as-received material using wire electrical discharge machining (EDM). The samples had a diameter of 35 mm and a thickness of 1.5 mm. Samples were ground to a final thickness using SiC sandpaper with incremental P200, P600, and P1200 grit sizes. Finally, the samples were rinsed with ethanol, dried with N2, and stored in a desiccator. Due to the physical limitations imposed by the permeation cell, the exposed area had a radius of 8 mm. To ensure that H transport was limited by bulk diffusion, the thickness was set to 1 mm (as suggested in appendix X1.4.3 of the ISO 17081).26  In addition, the standard recommends a radius-to-thickness ratio of 10:1 to enable the analysis of the permeation transients using one-dimensional diffusion assumptions. Instead of using the recommended thickness of 0.8 mm, samples were approximately 1 mm, which would represent an 8:1 ratio. As reported by Hutchings, et al.,27  when using a ratio of 10:1, the error associated with 2D diffusion was 4.9%, whereas for a 5:1 ratio, the error was 10%. With a cubic polynomial fit, the interpolated error corresponding to the 8:1 ratio was 6%.

The need for a coating on the anodic side was stated by Manolatos, et al.28  In the present work, samples were sputtered-coated with Pd by magnetron sputtering using a 208HR High-Resolution Sputter Coater due to irregularities found on electrodeposited coatings in previous tests (not shown). SEM and EDS analysis were performed on both the top and cross sections to verify the Pd layer’s uniformity and thickness. Cross sectioning was done by FIB.

Permeation experiments were performed to evaluate the hydrogen diffusion parameters of A508 Gr.4N at different charging conditions while keeping the rest of the variables fixed (i.e., surface, thickness, and environment). Electrochemical hydrogen permeation studies were conducted as per the Devanathan and Stachurski experimental approach.29  The arrangement consisted of two individual semi-cells, i.e., a detection cell and a charging cell, as shown in Figure 2, with the sample clamped in between the two compartments. The evaluation of the results and setup characteristics followed both ISO 1708126  and ASTM G14830  standards. The cells’ volume was 200 cm3, and the exposed area on both exit and entry sides was 2 cm2. Silicon washers covered with high-vacuum grease were used to reduce crevice formation and improve the seal between the samples and the glass cells. The temperature of the electrolyte was controlled using a recirculating bath and maintained at 15°C to enable the comparison with previous literature.31  Silver-silver chloride (Ag/AgCl in saturated KCl solution) reference electrodes were used with Luggin capillaries to reduce the effect of the solution ohmic resistance in both cells. The Luggin capillaries were filled with 3.5 M KCl. Both detection and charging cells included Pt/Ir 0.25 mm diameter wire and 25 mm × 25 mm 99.9% Pt mesh as counter electrodes. The solution used in the anodic cell was 0.1 M NaOH, in which deaeration was performed by continuous N2 bubbling. The pH measured in the aerated condition was approximately 13. For the charging compartment, the solution was 3.5 wt% NaCl and the measured pH was 6.5. Each test was performed with the following procedure: (i) a 0.1 M Na(OH) solution was purged with N2 in a separate cell before introducing it into the permeation anodic compartment; (ii) the solution was added to the anodic cell and the anodic polarization was started. At this stage, the oxidation current (background current) decreased and stabilized, and N2 flowed through both compartments. N2 purging on the cathodic side minimized corrosion during the stabilization stage. (iii) Once the background current had a value lower than 0.1 μm/cm2, N2 bubbling on the cathodic compartment was stopped, followed by the addition of 3.5 wt% NaCl. (iv) Immediately after, the corresponded cathodic potential was applied.

FIGURE 2.

Hydrogen permeation setup. The anodic compartment (left) was filled with 0.1 M NaOH and the cathodic compartment (right) with 3.5 wt% NaCl.

FIGURE 2.

Hydrogen permeation setup. The anodic compartment (left) was filled with 0.1 M NaOH and the cathodic compartment (right) with 3.5 wt% NaCl.

Close modal

A bi-potentiostat with floating grounds was used in both compartments. The anodic side was polarized to +345 m VAg/AgCl (+550 mVNHE) to ensure hydrogen oxidation. Experiments only varied in the applied potentiostatic charging of the entry surface as follows:

  • Eapp,1 = −1.05 VAg/AgCl

  • Eapp,2 = −1.40 VAg/AgCl

  • Eapp,3 = −2.00 VAg/AgCl

A palladium membrane was tested following the same procedure described above to corroborate the reproducibility of the test setup and the accuracy of the obtained parameters. Similarly, three transients were performed with an applied potential of −1.05 VAg/AgCl.

The flux was calculated as per Equation (6) using the hydrogen oxidation current (IH) from the anodic side.
formula
where F is the Faraday constant (96,485 C/mol), IH the hydrogen oxidation current expressed in A, exposed area in cm2, and the current density (iH) in A/cm2 units.
At steady-state, the hydrogen sub-surface concentration at lattice sites and reversible traps (COR) is proportional to the steady-state hydrogen permeation flux Jss.30  When irreversible trapping is negligible, COR reflects the severity of the environment, and it can be obtained from Fick’s second law as follows:
formula
where Deff is the effective diffusion coefficient and L is the sample thickness. The Deff values were determined from the permeation transients following ASTM G14830  and ISO 17081.26  From the three methods presented in the standards, two of those were implemented in this work to assess Deff. The tlag method is based on the time to reach 63% of the flux at steady-state (J(t)/Jss = 0.63). Deff is, thus, defined as:
formula
The tb method is based on the breakthrough time, i.e., the time that it takes to detect hydrogen on the anodic side. In this case, Deff is defined as:
formula

Cathodic Polarization Curves

Cathodic potentiodynamic polarization curves were performed to correlate the charging current densities to the given applied cathodic potentials. Tests were done on an A508 Gr.4N sample with an exposed area of 2 cm2 and its surface was ground with SiC paper down to 1200 grit. The polarization curves were conducted in aerated 3.5 wt% NaCl solution at 15°C to replicate the hydrogen permeation cathodic side conditions. After recording the open-circuit potential (OCP) for 15,000 s, the polarization started from +25 mVOCP to a final cathodic polarization of −2.10 VAg/AgCl with a scan rate of 0.1667 mV/s.

Hot Extraction

Hot extraction was performed at 700°C to quantify the hydrogen content present in the SSRT samples charged for a specific amount of time. Therefore, cylindrical samples with the SSRT gauge section dimensions (i.e., 6.31 mm diameter and 25.4 mm length) were used. The analysis comprised two steps: charging and desorption. Samples were charged electrochemically for 1 d, 2 d, 5 d, 10 d, and 15 d in a 3.5 wt% NaCl solution. The applied potential was −1.05 VAg/AgCl, and charging was performed at room temperature to replicate the SSRT charging conditions. The −1.05 VAg/AgCl potential was selected due to the lower hydrogen recombination and the better result reproducibility. After charging, samples were rinsed with ethanol followed by deionized water, and dried before introducing the cylinders in the tube furnace preheated at 700°C. The experimental procedure required 30 s from the end of charging to the start of the hot extraction measurement to reduce the amount of hydrogen desorbed prior to the measurement while being consistent through the tested samples. As the sample temperature increased, the hydrogen was desorbed and carried by N2 carrier gas to a thermal conductivity detector. The software then determined the hydrogen concentration based on the difference in thermal conductivity between hydrogen and nitrogen. The hydrogen concentration is presented as wppm alongside the sample mass.

Additionally, the effect of irreversible trapping was evaluated using the same analysis technique. Hydrogen desorbed from samples that were electrochemically charged for 2 d were compared against discharged samples, i.e., samples that were charged for the same time and then introduced in a vacuum chamber for 5 d at room temperature. The aim was to desorb any hydrogen trapped at reversible traps during the vacuum step, leaving only the hydrogen trapped at irreversible traps for the hot extraction analysis.14  In both cases, samples were subjected to hot extraction at 700°C.

Microstructure

Figures 3(a) through (d) show A508 Gr.4N LOM and SEM micrographs. In the literature, the A508 Gr.4N microstructure is described as a complex mixture of lower bainite and tempered martensite.5,12  Due to their similar appearance, these two phases are rather difficult to distinguish.5  The prior austenite grain (PAG) boundaries could be revealed by etching with a 2% Nital solution. Based on a statistical analysis of the production material done by the vendor, the average PAG size is 75 μm, corresponding to an ASTM grain size number of 4.8.

FIGURE 3.

(a) and (b) LOM, and (c) and (d) SEM micrographs of the A508 Gr.4N LAS in the Q&T condition. The microstructure is a mixture of lower bainite and tempered martensite. Finely dispersed carbides can be seen within the laths, at lath and PAG boundaries.

FIGURE 3.

(a) and (b) LOM, and (c) and (d) SEM micrographs of the A508 Gr.4N LAS in the Q&T condition. The microstructure is a mixture of lower bainite and tempered martensite. Finely dispersed carbides can be seen within the laths, at lath and PAG boundaries.

Close modal

The presence of retained austenite in the microstructure of steels is suggested to have a negative impact on HSC and SSC resistance.1  The reasons for this are twofold. First, retained austenite can convert into untempered martensite when a critical stress or strain is applied or if the alloy is improperly tempered. Snape, et al., were able to show that this can drastically reduce a LAS’s resistance to SSC.32  Second, the higher solubility of hydrogen in the austenite phase showing a face-centered cubic (fcc) crystal structure, compared to body-centered cubic (bcc) phases, could lead to an increased release of hydrogen in the crack zone and promote crack growth.1  The EBSD phase map presented in Figure 4(a) exhibited a microstructure entirely consisting of bcc phases, either lower bainite or tempered martensite, as represented by the red color code. No retained austenite could be detected in the EBSD scans, which would have been indicated in a blue color code. The absence of retained austenite contrasted findings by other researchers who identified up to 17% retained austenite in an A508 Gr.4N test alloy.6  The visible black lines represent PAG grain boundaries or lath boundaries within the PAGs. Figure 4(b) illustrates an Euler orientation map, which emphasizes the fine sub-grain microstructure within the PAG boundaries. Different colors represent different orientations within the microstructure.

FIGURE 4.

EBSD maps of the A508 Gr.4N LAS. (a) Phase map showing an exclusively bcc microstructure (red) with no retained austenite. The present bcc structure is suggested to be mainly lower bainite. (b) The Euler orientation map emphasizes the lath structures formed within the prior austenite grain boundaries. Different colors represent different orientations of the grains and sub-grain structures.

FIGURE 4.

EBSD maps of the A508 Gr.4N LAS. (a) Phase map showing an exclusively bcc microstructure (red) with no retained austenite. The present bcc structure is suggested to be mainly lower bainite. (b) The Euler orientation map emphasizes the lath structures formed within the prior austenite grain boundaries. Different colors represent different orientations of the grains and sub-grain structures.

Close modal

STEM micrographs and EDS element composition maps are shown in Figures 5(a) through (f). A variety of suspected carbides can be observed in Figure 5(a). Round-shaped particles presented smaller sizes (10 nm to 50 nm), whereas elongated precipitates were up to 450 nm long. The shape and size were correlated to the chemical composition of the suspected carbides by coupling TEM and EDS analysis (Figures 5[b] through [f]). EDS maps (Figures 5[b], [c], and [d]) revealed that Cr-Mn-C particles had an elongated-type shape and the largest sizes (i.e., 50 nm to 450 nm). For the case of Mo-rich particles, as illustrated in Figure 5(e), suspected carbides also presented an elongated shape but had a smaller size (50 nm to 100 nm). As seen in Figure 5(f), V-rich particles were found in a discrete round-shape form with sizes that ranged between 5 nm and 30 nm. Regardless of their chemical composition, particles were predominantly located along lath boundaries.

FIGURE 5.

(a) High-angle annular dark-field (HAADF) STEM micrograph, (b) through (f) EDS elemental maps of C, Cr, Mn, Mo, and V, respectively, from the A508 Gr.4N sample.

FIGURE 5.

(a) High-angle annular dark-field (HAADF) STEM micrograph, (b) through (f) EDS elemental maps of C, Cr, Mn, Mo, and V, respectively, from the A508 Gr.4N sample.

Close modal

The crystal structure of the carbides of the Cr-rich particles was determined by their diffraction patterns obtained from the SAD, represented by a dash lined in the TEM micrograph in Figure 6(a). The obtained pattern shown in Figure 6(b) matched the M23C6 crystal structure, and the dark fields corresponded to the circled diffraction spot are shown in Figures 6(c) and (d). The crystal structure of Mo- and V-rich particles could not be characterized by diffraction patterns due to their small size and relatively low quantity.

FIGURE 6.

(a) TEM micrograph of ASTM A508 Gr.4N. (b) Diffraction pattern obtained from the selected area indicated on (a) placed on the Cr carbides. (c) and (d) Dark field images obtained from the circled spot on (b).

FIGURE 6.

(a) TEM micrograph of ASTM A508 Gr.4N. (b) Diffraction pattern obtained from the selected area indicated on (a) placed on the Cr carbides. (c) and (d) Dark field images obtained from the circled spot on (b).

Close modal

Hydrogen Stress Cracking Resistance Evaluated by Slow Strain Rate Testing

Representative SSRTs engineering stress-strain curves are presented in Figure 7 for the four different testing conditions, i.e., air, −1.05 VAg/AgCl, −1.40 VAg/AgCl, and −2.00 VAg/AgCl. The largest elongation to fracture was observed by the material in the air with an average value of 23.6%. With an increased applied cathodic potential, i.e., more hydrogen generated, the alloy lost some of its ductility. Nevertheless, for all three in situ test conditions, the fracture occurred only after significant plastic elongation and past the material’s ultimate tensile stress (UTS). The polarized samples still exhibited an average elongation to fracture value of 15.7%, 14.5%, and 14.0% for −1.05 VAg/AgCl, −1.40 VAg/AgCl, and −2.00 VAg/AgCl, respectively. Further, 0.2% YS and UTS were similar for all test conditions and all test samples started necking at approximately 8% elongation.

FIGURE 7.

Representative stress-strain curves, derived from SSRT, of the A508 Gr.4N alloy for the control and the three different hydrogen charging conditions. The alloy behavior was similar up to the UTS, showing a large amount of plastic elongation in all conditions.

FIGURE 7.

Representative stress-strain curves, derived from SSRT, of the A508 Gr.4N alloy for the control and the three different hydrogen charging conditions. The alloy behavior was similar up to the UTS, showing a large amount of plastic elongation in all conditions.

Close modal

It is also noticeable that the stress-strain curves declined sharply shortly before the final fracture for the hydrogen charged samples. This can be attributed to the fact that hydrogen diffuses to areas with high triaxial stresses,33  which in this test scenario was centered around the necking zone where most plastic deformation occurred. With increased cathodic polarization, the sharp drop happened at lower elongations. Hence, crack initiation and propagation could be accelerated with increasing hydrogen concentration once the material started to severely neck.

The SSRT ratios were calculated according to Equations (2) through (5). The obtained average values are given in Table 4. As expected, RTTF, RRA, RE, and RT decreased with increasing hydrogen concentration. Further, it is noteworthy that, due to the 45° angled fracture of the charged samples, the RRA was the most error-prone parameter as accurate measurements of the diagonals/area is difficult.

Table 3.

Mechanical Properties of the A508 Gr.4N Test Material

Mechanical Properties of the A508 Gr.4N Test Material
Mechanical Properties of the A508 Gr.4N Test Material
Table 4.

SSRT Ratios for the Three Hydrogen Charging Conditions

SSRT Ratios for the Three Hydrogen Charging Conditions
SSRT Ratios for the Three Hydrogen Charging Conditions

Fractography

SEM micrographs of the fracture surface, as well as from the gauge section close to the main crack, for the control sample are shown in Figures 8(a) through (c). The SSRT samples tested in air showed the characteristic cup and cone shape, typical of ductile materials (Figure 8[a]). Further, the fracture surface exhibited an entirely ductile rupture fracture appearance, which can be divided in the central crack initiation zone, showing characteristic radial marks, and the outer shear-lip zone (Figure 8[b]). Higher magnification micrographs (Figure 8[c]), taken from the center of the test sample, revealed the presence of dimples that resulted from microvoids nucleation, void coalescence, and the final shear fracture. In contrast, the hydrogen-charged samples showed a shear fracture where the main crack propagated at a 45° angle to the plane in which the load was applied (Figure 9). Secondary cracks developed in the gauge section where necking occurred in the charged samples. Radial machining or grinding marks from the sample manufacturing and preparation, not shown, could act as stress raisers and crack initiation points.

FIGURE 8.

SEM micrographs of a control SSRT sample. The side view (a) shows the typical cup and cone shape, characteristic of ductile rupture. (b) The fracture surface can be divided into the central crack initiation zone, showing radial marks and the outer shear-lip zone. At higher magnifications (c), the dimpled fracture surface can be observed, which results from microvoid nucleation and coalescence.

FIGURE 8.

SEM micrographs of a control SSRT sample. The side view (a) shows the typical cup and cone shape, characteristic of ductile rupture. (b) The fracture surface can be divided into the central crack initiation zone, showing radial marks and the outer shear-lip zone. At higher magnifications (c), the dimpled fracture surface can be observed, which results from microvoid nucleation and coalescence.

Close modal
FIGURE 9.

Side view and fracture surface of an SSRT sample polarized at −1.40 VAg/AgCl. Shear fracture with crack propagation at a 45° angle. Secondary cracks were present and radial machining or grinding marks acted as crack initiation points.

FIGURE 9.

Side view and fracture surface of an SSRT sample polarized at −1.40 VAg/AgCl. Shear fracture with crack propagation at a 45° angle. Secondary cracks were present and radial machining or grinding marks acted as crack initiation points.

Close modal

SEM analysis of the fracture surfaces of the charged samples revealed a hydrogen-induced fracture with three distinct zones (Figure 10). It is suggested that crack initiation occurred at the sample’s surface or at a near-surface defect. The area adjacent to the crack initiation showed a brittle fracture zone with quasi-cleavage and intergranular features. The intergranular crack propagation occurred along the PAG boundaries. The main crack, then, propagated at a 45° angle, exhibiting a mixed fracture mode of transgranular cleavage next to areas of ductile rupture. The final fracture region was characterized by intergranular and quasi-cleavage features, similar to those observed in the crack initiation zone.

FIGURE 10.

SEM micrographs of SSRT samples polarized at −1.05 VAg/AgCl, −1.40 VAg/AgCl, and −2.00 VAg/AgCl presented in (a), (b), and (c), respectively. The fracture surface revealed three distinct regions: (I) crack initiation zone, (II) shear fracture, and (III) final fracture. The white arrows indicate possible crack propagation pathways.

FIGURE 10.

SEM micrographs of SSRT samples polarized at −1.05 VAg/AgCl, −1.40 VAg/AgCl, and −2.00 VAg/AgCl presented in (a), (b), and (c), respectively. The fracture surface revealed three distinct regions: (I) crack initiation zone, (II) shear fracture, and (III) final fracture. The white arrows indicate possible crack propagation pathways.

Close modal

Figures 11(a) through (c) show the fracture surfaces in the shear fracture zone where the crack propagated at a 45° angle to the applied load. This area was characterized by a mixed fracture mode exhibiting transgranular cleavage and ductile rupture. The brittle cleavage regions appeared to be more predominant in the sample with increased hydrogen charging, i.e., −1.40 VAg/AgCl and −2.0 VAg/AgCl.

FIGURE 11.

SEM micrographs of SSRT samples polarized at −1.05 VAg/AgCl, −1.40 VAg/AgCl, and −2.00 VAg/AgCl presented in (a), (b), and (c), respectively. Areas centered in the shear fracture zone, characterized by a mixed fracture mode exhibiting transgranular cleavage and ductile rupture.

FIGURE 11.

SEM micrographs of SSRT samples polarized at −1.05 VAg/AgCl, −1.40 VAg/AgCl, and −2.00 VAg/AgCl presented in (a), (b), and (c), respectively. Areas centered in the shear fracture zone, characterized by a mixed fracture mode exhibiting transgranular cleavage and ductile rupture.

Close modal

Hydrogen Permeation Tests

Hydrogen permeation tests were first performed on a Pd membrane to verify the experimental setup and the reproducibility of the curves. Good reproducibility of the transient was observed, as illustrated by Figure 12. Deff values calculated by the tb method ranged from 4.05×10−8 cm2/s to 4.96×10−8 cm2/s, whereas for the tlag method, the Deff had a minimum and maximum value of 2.72×10−8 cm2/s and 2.98×10−8 cm2/s, respectively. From the Devanathan and Stachurski work made on palladium membranes that were 0.51 mm thick, the reported Deff values were 8.9×10−8 cm2/s and 9.6×10−8 cm2/s for the tb and tlag methods, respectively. Therefore, Deff values obtained in the present work were in good agreement with Devanathan and Stachurski. Discrepancies in the diffusion coefficients could be due to the charging environment, the choice of radius-to-thickness ratio, sample thickness, and preparation.

FIGURE 12.

Reproducibility of transients on palladium membranes.

FIGURE 12.

Reproducibility of transients on palladium membranes.

Close modal

The Pd-sputtered coat layer uniformity and thickness were evaluated by SEM and EDS of the top surface and crosssection using the A508 Gr.4N as substrate (Figure 13). The EDS analysis (Figure 13[b]) of the top surface confirmed the full coverage of the Pd-sputtered layer and revealed high contrast in areas where FIB removed the sputtered layer, making Pd coverage evident. The thickness of the coating measured at the cross section shown in Figure 13(c) ranged from 84 nm to 105 nm.

FIGURE 13.

SEM images of the palladium-sputtered coat layer on the anodic side of the A508 Gr.4N samples. (a) SEM micrograph and (b) EDS analysis of palladium from the top surface; and (c) SEM micrograph of the cross section.

FIGURE 13.

SEM images of the palladium-sputtered coat layer on the anodic side of the A508 Gr.4N samples. (a) SEM micrograph and (b) EDS analysis of palladium from the top surface; and (c) SEM micrograph of the cross section.

Close modal

Three permeation transients with an applied potential of −1.05 VAg/AgCl at 15°C were performed to evaluate the presence of irreversible trapping on A508 Gr.4N. Figure 14 shows the permeation curves plotted as steady-state normalized flux vs. the normalized time. Although the curves did not entirely overlap, no trend on the delay and slope of the curves was found, suggesting the absence of irreversible trapping.26 

FIGURE 14.

Permeation transients of A508 Gr.4N charged at −1.05 VAg/AgCl plotted as steady-state normalized permeation flux vs. normalized time.

FIGURE 14.

Permeation transients of A508 Gr.4N charged at −1.05 VAg/AgCl plotted as steady-state normalized permeation flux vs. normalized time.

Close modal

A508 Gr.4N Permeation Parameters and Comparison Against Ferritic-Pearlitic 3%Ni LAS

Figure 15 shows the first permeation transients for −1.05 VAg/AgCl, −1.40 VAg/AgCl, and −2.00 VAg/AgCl charging conditions at 15°C. It can be observed that an increase in cathodic potential displaced the transients to shorter times. In other words, the more severe the hydrogen charging environment, the faster hydrogen was detected at the anodic side. Table 5 summarizes A508 Gr.4N’s Deff, normalized steady-state flux by thickness (Jss·L), and COR together with the permeation parameters obtained by Husby, et al.,24  for a 3 wt% Ni ferritic-pearlitic LAS. The Jss·L of A508 Gr.4N was highest for the charging corresponding to the −2.00 VAg/AgCl polarization. In this regard, increasing the applied cathodic potential from −1.05 VAg/AgCl to −1.40 VAg/AgCl resulted in a two-fold increase in Jss·L, while an increase from −1.40 VAg/AgCl to −2.00 VAg/AgCl led to a more modest 20% increment. A twofold increase in Deff values was found irrespectively of the method when increasing the applied cathodic potential from −1.05 VAg/AgCl to −1.40 VAg/AgCl. In contrast, when analyzing the effect of potential on the Deff value for the more severe condition, the methods differed in the resulting trend. In the case of tb, an increase of the cathodic potential from −1.40 VAg/AgCl to −2.00 VAg/AgCl resulted in an increase of approximately 17% in Deff, while the tlag method showed a small decrease in Deff of about 2%.

Table 5.

Permeation Parameters Calculated by the tb and tlag Methods for A508 Gr.4N Tested at 15°C in 3.5 wt% NaCl Solution Under Three Hydrogen Charging Conditions(A)

Permeation Parameters Calculated by the tb and tlag Methods for A508 Gr.4N Tested at 15°C in 3.5 wt% NaCl Solution Under Three Hydrogen Charging Conditions(A)
Permeation Parameters Calculated by the tb and tlag Methods for A508 Gr.4N Tested at 15°C in 3.5 wt% NaCl Solution Under Three Hydrogen Charging Conditions(A)
FIGURE 15.

Permeation transients of A508 Gr.4N at 15°C plotted as steady-state normalized permeation flux vs. normalized time for −1.05 VAg/AgCl, −1.40 VAg/AgCl, and −2.00 VAg/AgCl charging conditions.

FIGURE 15.

Permeation transients of A508 Gr.4N at 15°C plotted as steady-state normalized permeation flux vs. normalized time for −1.05 VAg/AgCl, −1.40 VAg/AgCl, and −2.00 VAg/AgCl charging conditions.

Close modal

When comparing A508 Gr.4N against the ferritic-pearlitic 3 wt% Ni LAS studied by Husby, et al.,24  the nuclear-grade steel had Deff values that were two orders of magnitude lower (Eapp= −1.05 VAg/AgCl). Moreover, the Jss·L value reported by Husby, et al., was approximately four times larger than that of A508 Gr.4N at −1.05 VAg/AgCl. Although the ferritic-pearlitic LAS had a higher Jss·L, due to the difference in Deff, the COR values of A508 Gr.4N were an order of magnitude larger regardless of the analysis methodology.

Polarization Curves

The cathodic polarization curve and the actual current densities values of the applied potentials, i.e., −1.05 VAg/AgCl, −1.40 VAg/AgCl, and −2.00 VAg/AgCl, of an A508 Gr.4N sample are displayed in Figure 16. The curve showed a limiting current density around 3 × 10−2 mA/cm2 between the OCP and −900 mV, which corresponds to the oxygen reduction reaction (ORR) limiting current density.34  Hydrogen reduction became the predominant cathodic reaction at more negative potentials, increasing the current density linearly (in the logarithmic scale) until reaching −1.2 VAg/AgCl. For increased negative potentials, the effect of ohmic drop due to the solution resistance became relevant. An increase of the applied cathodic potential from −1.05 VAg/AgCl to −1.40 VAg/AgCl resulted in an order of magnitude increase of the measured current density. A further increase of the applied cathodic potential from −1.40 VAg/AgCl to −2.00 VAg/AgCl resulted in a four-fold increase in current density.

FIGURE 16.

Potentiodynamic cathodic polarization curves of ASTM A508 Gr.4N.

FIGURE 16.

Potentiodynamic cathodic polarization curves of ASTM A508 Gr.4N.

Close modal

Hot Extraction

Figure 17 provides an overview of the hydrogen content as a function of the charging time, with two samples per condition. Results showed an abrupt increase in the hydrogen concentration between two and five charging days. The hydrogen concentration reached an equilibrium hydrogen concentration after 5 d, which can be interpreted as the saturated condition for the given potential.

FIGURE 17.

Hydrogen concentration on ASTM A508 Gr.4N with an applied potential of −1.05 VAg/AgCl plotted against charging time. The dashed line was added to aid trend visualization.

FIGURE 17.

Hydrogen concentration on ASTM A508 Gr.4N with an applied potential of −1.05 VAg/AgCl plotted against charging time. The dashed line was added to aid trend visualization.

Close modal

Figure 18 shows the desorption curves for the charged and discharged condition. From the area under the curve, the calculated amount of desorbed hydrogen for the charged sample was 0.57 wppm, whereas for the discharged condition, the desorbed hydrogen was difficult to measure due to the detection limit of the instrument (i.e., 0.01 wppm).

FIGURE 18.

Desorption curves for a charged sample (solid line) and discharged sample (dashed line).

FIGURE 18.

Desorption curves for a charged sample (solid line) and discharged sample (dashed line).

Close modal

Hydrogen Stress Cracking Resistance

Stress-strain curves obtained from the SSRTs and examination of the fracture surfaces showed that the A508 Gr.4N LAS was influenced by hydrogen. The ductility parameters decreased with more negative potentials, i.e., higher hydrogen concentrations35  and the fracture mode changed from a ductile rupture failure exhibited by the control samples to a hydrogen-induced shear fracture. Further, secondary cracking was observed in the necking zone of the gauge section for the hydrogen charged samples. However, it was also apparent that the 0.2% YS and UTS were unaffected even at a cathodic polarization of −2.00 VAg/AgCl. Signs of HE could be observed from the stress-strain curves only after the steel reached its UTS and after the SSRT specimens started necking. Similar results were reported by Cabrini, et al.,36-38  who evaluated several pipeline steels with YS between 430 MPa (62 ksi) and 770 MPa (110 ksi) with different microstructures. Cabrini, et al., performed constant load (CL), interrupted slow strain rate tests (ISSRTs), and conventional SSRTs under cathodic polarization. Specimens stressed until their YS in the CL tests showed no signs of cracking. Even for the ISSRTs, where the samples were strained to their UTS and held at this stress level for one week, no cracking was evident. Therefore, the materials were considered immune to HE in the tested conditions by the authors. However, SSRTs performed on the pipeline steels revealed loss of ductility and, therefore, susceptibility to HE when continuous straining in the plastic region was applied. These results emphasize that continuous straining was not only critical for the crack initiation but also crack propagation under the test conditions. As a result, SSRTs may be over-conservative to evaluate a material’s susceptibility to HSC and SSC, indicating hydrogen degradation where no failures would occur in service. Constant load testing using, e.g., proof rings,37  could be used as an alternative approach to determine the HSC threshold stress in simulated service conditions, in which the proof load is set to the desired stress level.

Researchers have studied the role of alloy microstructure on the HSC and SSC susceptibility of various LASs. Cabrini, et al.,36-38  also demonstrated that irrespectively of the strength level, bainitic and fully tempered martensitic structures with a fine carbide distribution showed the best resistance to HE. Here, an API X80 (0.2% YS = 587 MPa) pipeline steel with a fine bainitic and acicular ferrite microstructure exhibited improved resistance to HE compared to an API X70 grade (0.2% YS = 457 MPa), characterized by a polygonal ferritic microstructure.36  Likewise, Yoshino38  and Snape13  concluded that quenched and tempered microstructures exhibited the highest resistance to SSC. The A508 Gr.4N steel investigated herein was characterized by a microstructure comprising lower bainite and tempered martensite with finely dispersed carbides. The fine sub-grain structures, i.e., laths and packets, are best observed in the Euler orientation map shown in Figure 4(b). Further, no retained austenite was detected in the microstructure, which has been reported to have a negative effect on the resistance to HE.1,39 

Hydrogen Diffusion and Transport Kinetics in ASTM A508 Gr.4N

The TEM analysis of A508 Gr.4N samples showed carbides located along lath boundaries of the Q&T microstructure. EDS mapping revealed Cr, Mo, Mn, and V to be the predominant elements. The Cr carbides matched the M23C6 crystal structure by diffraction pattern analysis. These observations are in accordance with Kim, et al.5,12  Irreversible trapping was evaluated by analyzing the consecutive permeation transients at −1.05 VAg/AgCl (Figure 14). The observed variations in the curves were considered experimental variability rather than caused by irreversible trapping effect as no clear trend on permeation transients was found. In addition, hydrogen desorption curves obtained by hot extraction exhibited the same behavior because no hydrogen was found in the discharged sample, suggesting the absence of irreversible trapping.

Increasing the applied cathodic potential from −1.05 VAg/AgCl to −1.40 VAg/AgCl resulted in an abrupt increase in steady-state normalized flux from 1.32 × 10−13 mol·cm·s to 7.32 × 10−12 mol·cm·s (Table 5). Nevertheless, when increasing the applied cathodic potential from −1.40 VAg/AgCl to −2.00 VAg/AgCl, the Jss·L value increased only by 20%. As a large amount of hydrogen gas evolution was observed on the cathodic side when charging at −2.00 VAg/AgCl, the relatively small increase in Jss·L could be due to the hydrogen recombination to form H2. As defined by Zafra, et al.,40  the efficiency of the permeation process is the relationship of the applied cathodic charging current and the steady-state oxidation permeation current. According to the author, the efficiency decreases with the increase of the applied cathodic potential due to the hydrogen recombination rate at the entry surface.

Some observations can be drawn from the calculated Deff values. First, as a general trend, Deff increased with applied cathodic potential, which can be explained in terms of trap occupancy. In this regard, Griffith and Turnbull41  reported the sensitivity of the effective diffusivity to hydrogen with the varying trap occupancy. The authors demonstrated that Deff values could vary by more than an order of magnitude when increasing the severity of the environment. Second, an increase of the applied cathodic potential from −1.05 VAg/AgCl to −1.40 VAg/AgCl resulted in a two-fold increase of the Deff, regardless of the implemented method. Conversely, different Deff trends were observed when the charging potential increased from −1.40 VAg/AgCl to −2.00 VAg/AgCl, depending on the analysis methodology. When tb was implemented, the Deff increased by 17%, whereas for tlag, its value decreased slightly with the increasing cathodic potential. The latter behavior could be due to the lack of a clear steady-state obtained when charging at −2.00 VAg/AgCl as the tlag method relies on the steady-state flux.

Comparison Between ASTM A508 Gr.4N and a Low-Strength Ferritic-Pearlitic LAS with a Similar Nickel Content

Comparing the SSRT results with work performed by Husby, et al.,24  on ferritic-pearlitic LASs with varying Ni content indicated that Q&T microstructures exhibited improved or comparable resistance to HE even at higher strength levels, as is the case of the A508 Gr.4N alloy. Husby, et al.,24  studied low-strength 3 wt% Ni ferritic-pearlitic LAS with a YS of ca. 340 MPa and a UTS of approximately 530 MPa. In this regard, SSRT conditions in Husby, et al., were similar to those in the present study, which allowed a direct comparison with our results. RE and RT ratios of the 3 wt% LAS investigated by Husby, et al., and the A508 Gr.4N cathodically polarized are shown in Figure 19. For the lower hydrogen concentration, i.e., specimens polarized at −1.05 VAg/AgCl, A508 Gr.4N exhibited comparable or slightly lower ratios, which would indicate a higher susceptibility to HE, attributed to the amount of hydrogen after the precharging. With an applied potential of −1.05 VAg/AgCl, the ferritic-pearlitic LAS was expected to reach the hydrogen saturation after 48 h,24  whereas the nuclear-grade steel reached the hydrogen saturation level by the fifth day (hot extraction results). In other words, the ferritic-pearlitic LAS would reach saturation in a shorter time, thus, hindering the further hydrogen ingress during the test. Therefore, keeping the precharging time of 2 d as a constant, the obtained results of the ASMT A508 Gr.4N HSC resistance would be conservative. With increased hydrogen charging (–2.00 VAg/AgCl) the SSRT ratios for the A508 Gr.4N LAS remained relatively high and indicated a noticeably better resistance to HE compared to the 3 wt% Ni LAS in Husby’s work (i.e., on average 0.60 for A508 Gr.4N and 0.37 for the ferritic-pearlitic LAS). Assuming that the slope of the SSRT ratios with the increase cathodic potential shown in Figure 19 is an indicator of the susceptibility of the material to HE, the A508 Gr.4N showed a greater resistance (lower slope) than the 3 wt% Ni ferritic-pearlitic LAS (higher slope). These results supported the assumption that finely dispersed carbides and Q&T microstructures improved A508 Gr.4 HSC resistance despite its much higher strength and hardness, which was above the 250 HV limit of the ISO 15156-2 standard. Furthermore, the comparison between the A508 Gr4N and Husby, et al.,24  clearly illustrated that HE resistance depends on the complex interplay between hydrogen and microstructure. Hence, HSC (SSC) resistance cannot be gauged by hardness measurements alone.

FIGURE 19.

SSRT ratios compared for the A508 Gr.4N alloy and a research-grade LAS with 3.0 wt% Ni investigated by Husby, et al.24  The SSRT ratios are comparable for both LASs at low hydrogen charging conditions, i.e., −1.05 VAg/AgCl. For higher hydrogen charging conditions, i.e., −2.00 VAg/AgCl, the A508 Gr.4N exhibits better resistance to HE. The results illustrate the beneficial effects of a lower bainitic microstructure with a fine dispersion of carbides at more severe hydrogen charging.

FIGURE 19.

SSRT ratios compared for the A508 Gr.4N alloy and a research-grade LAS with 3.0 wt% Ni investigated by Husby, et al.24  The SSRT ratios are comparable for both LASs at low hydrogen charging conditions, i.e., −1.05 VAg/AgCl. For higher hydrogen charging conditions, i.e., −2.00 VAg/AgCl, the A508 Gr.4N exhibits better resistance to HE. The results illustrate the beneficial effects of a lower bainitic microstructure with a fine dispersion of carbides at more severe hydrogen charging.

Close modal

Comparing the hydrogen diffusion and transport kinetics results with the ferritic-pearlitic steel tested by Husby, et al., the nuclear-grade steel had Deff values that were two orders of magnitude smaller. This difference was attributed to both chemical composition and heattreatment history. In contrast to the ferritic-pearlitic microstructure obtained by austenitization and slow cooling, the A508 Gr.4N heat treatment consisted of a Q&T process that produced a bainitic/tempered martensitic microstructure. Depover, et al.,42  studied the influence of the bainite/ferrite ratio on the diffusion coefficient. They found that the Deff obtained for the 98.1% bainite-1.9% ferrite was half of that obtained from the steel with 89.2% bainite-10.8% ferrite. Additionally, the two LASs differed in the shape, composition, and distribution of carbides. Whereas the 3 wt% LAS used by Husby, et al., contained FeC3 exclusively, the A508 Gr.4N comprised several precipitates. In this regard, Cr carbides were detected by EDS analysis and their crystal structure was determined to be M23C6 by electron diffraction. The effect of Cr23C6 on HE was evaluated by Depover and Verbeken16  by comparing the as-quenched (as-Q) and Q&T conditions on Fe-C-Cr alloys. The Q&T microstructure showed a decrease in hydrogen diffusivity compared to the as-Q condition, with diffusion coefficients of 5.78 × 10−11 m2/s and 1.71 × 10−10 m2/s, respectively. The difference was attributed to carbides with sizes less than 100 nm that acted as traps hindering hydrogen diffusion. Similarly, Mo-rich particles were found by TEM-EDS. The Mo2C precipitation and its effect on the hydrogen diffusion were studied by Hinotani, et al., on a 3.9 wt% Mo-0.20 wt% C steel.17  Mo2C carbides that form during the tempering stage caused a decrease of two orders of magnitude in the Deff values when the steel was tempered for 1 h at 600°C. In addition, Depover and Verbeken19  studied the role of Mo2C on the HE resistance. The authors reported the beneficial effect of the carbide addition as they act as deeper traps, reducing the amount of mobile hydrogen, thus, increasing the HE resistance. Last, some V-rich particles were also found in A508 Gr.4N. In this regard, Li, et al., studied V-containing carbides and their effect on hydrogen diffusion in an API X80 pipeline steel.43  The authors found a decrease in the hydrogen diffusion coefficient from 4.74 × 10−6 cm2/s to 8.48  × 10−7 cm2/s when adding 0.16% V to the vanadium-free API X80 steel, which could also explain the marked decrease in Deff of A508 Gr.4N. Although the evidence suggests that the nature of the carbides affected Deff substantially, more research is needed to elucidate the mechanisms by which carbides reduced hydrogen transport kinetics. Future work will use thermal desorption spectroscopy to complement the permeation work presented herein.

Despite the promising results presented herein, currently, the A508 Gr.4N LAS is, in practice, excluded from sour service applications. This restriction is due to the 1.0 wt% Ni and the 250 HV hardness limits of the ISO 15156-28  standard. Using this high-Ni grade with its improved strength and toughness would require a costly and extensive qualification program. Krishnan4  showed in a recent study that the Ni content might not influence the SSC susceptibility. On the contrary, the higher Ni grade showed better or equal performance in the H2S-containing test environments compared to the lower Ni grades. The tests were performed on quenched and tempered AISI 4140 (0.11 wt% Ni), 8630Mod (0.81 wt% Ni), and AISI 4340 (1.81 wt% Ni) LASs using the uniaxial tensile test Method A in NACE TM0177.37  Krishnan’s work4  underpins the encouraging SSRT results of the A508 Gr.4N alloy presented here. Indeed, A508 Gr.4N might be suitable for O&G equipment in sour environments if an adequate SSC resistance could be demonstrated with future testing.

In this work, the HSC resistance and the hydrogen diffusion transport kinetics of the nuclear-grade ASTM A508 Gr.4N LAS were studied using the slow strain rate testing technique as a function of applied cathodic potential and electrochemical hydrogen permeation testing under potentiostatic cathodic charging. The following conclusions were drawn based on the evidence presented herein:

  • LOM and SEM micrographs revealed a microstructure comprising lower bainite and tempered martensite with uniform and finely dispersed precipitates within the laths and at interlath and PAG boundaries. EBSD mapping revealed the absence of retained austenite in the microstructure. TEM analysis confirmed the presence of Cr23C6 carbides and precipitates rich in Mo and V. The combination of these microstructural features retarded hydrogen diffusion and can improve HSC resistance.

  • SSRTs results suggested A508 Gr.4N had an adequate HSC resistance at all applied cathodic potentials. YS and UTS remained unaffected even at the cathodic polarization of −2.00 VAg/AgCl. Hydrogen effects were only observed after necking. It is speculated that once necking started, hydrogen diffused to areas with high triaxial stresses (necking zone) and caused premature crack initiation and embrittlement.

  • ASTM A508 Gr.4N had a noticeably better HSC resistance at −2.00 VAg/AgCl than a low-strength 3 wt% Ni ferritic-pearlitic LAS with a YS of 340 MPa, demonstrating that HSC resistance depends on the complex interplay between hydrogen and microstructure. Thus, HSC resistance cannot be gauged by hardness measurements alone.

  • No effect of irreversible trapping was found based on the evaluation of consecutive permeation transients and hot extraction.

  • Deff values obtained for the bainitic/martensitic A508 Gr.4N LAS were two orders of magnitude smaller than those of the ferritic/pearlitic 3% Ni LAS studied by Husby, et al.24  The difference could be attributed to differences in chemical composition and microstructure.

(1)

International Organization for Standardization (ISO), 1214 Vernier, Geneva, Switzerland.

(2)

ASTM International (ASTM), 100 Barr Harbor Drive, West Conshohocken, PA, 19428-2959.

(3)

National Association of Corrosion Engineers (NACE) International, 15835 Park Ten Place, Houston, TX 77084.

Trade name.

The authors acknowledge the financial support of Chevron (Australia Business Unit) and Woodside Ltd. as well as the support from FWO (fellow grant 12ZO420N). We also thank Japan Steel Works for preparing the materials used in this investigation. Furthermore, we acknowledge the support of the John de Laeter Centre at Curtin University, where the characterization took place. Finally, we thank colleagues that were part of this work, namely Yu Long, Elaine Miller, and Dr. Sam Bakhtiar.

1.
Kappes
M.
,
Iannuzzi
M.
,
Rebak
R.B.
,
Carranza
R.M.
,
Corros. Rev.
32
,
3–4
(
2014
):
p
.
101
128
.
2.
Perez
T.E.
,
JOM
65
,
8
(
2013
):
p
.
1033
1042
.
3.
Iannuzzi
M.
,
Barnoush
A.
,
Johnsen
R.
,
npj Mater. Degrad.
1
,
1
(
2017
):
article 2
.
4.
Krishnan
K.
, “
Sulfide Stress Cracking Resistance of High Strength, Low Alloy Steels with Varying Nickel Content
,”
CORROSION 2019, paper no. 12900
(
Houston, TX
:
NACE International
,
2019
).
5.
Kim
M.-C.
,
Park
S.-G.
,
Lee
K.-H.
,
Lee
B.-S.
,
Int. J. Press. Vess. Piping
131
(
2015
):
p
.
60
66
.
6.
Park
S.-G.
,
Lee
K.-H.
,
Min
K.-D.
,
Kim
M.-C.
,
Lee
B.-S.
,
Met. Mater. Int.
19
,
1
(
2013
):
p
.
49
54
.
7.
Wilhelm
S.M.
,
Kane
R.D.
,
J. Petrol. Technol.
38
,
10
(
1986
):
p
.
1051
1061
.
8.
ISO 15156-2
, “
Petroleum and Natural Gas Industries–Materials for Use in H2S-Containing Environments in Oil and Gas Production–Part 2: Cracking-Resistant Carbon and Low-Alloy Steels, and the Use of Cast Irons
” (
Geneva, Switzerland
:
International Organization for Standardization
,
2020
).
9.
Craig
B.
,
Corrosion
44
,
11
(
1988
):
p
.
776
782
.
10.
ASTM A508/A508M-16
, “
Standard Specification for Quenched and Tempered Vacuum-Treated Carbon and Alloy Steel Forgings for Pressure Vessels
” (
West Conshohocken, PA
:
ASTM International
,
2018
).
11.
Lee
K.
,
Park
S.G.
,
Kim
M.C.
,
Lee
B.S.
,
Wee
D.M.
,
Mater. Sci. Eng. A-Struct. Mater. Prop. Microstruct. Process.
529
,
1
(
2011
):
p
.
156
163
.
12.
Park
S.G.
,
Kim
M.C.
,
Lee
B.S.
,
Wee
D.M.
,
J. Nucl. Mater.
407
,
2
(
2010
):
p
.
126
135
.
13.
Snape
E.
,
Corrosion
23
,
6
(
1967
):
p
.
154
172
.
14.
Depover
T.
,
Laureys
A.
,
Perez Escobar
D.
,
Van den Eeckhout
E.
,
Wallaert
E.
,
Verbeken
K.
,
Materials (Basel)
11
,
5
(
2018
):
p
.
698
.
15.
Nagao
A.
,
Martin
M.L.
,
Dadfarnia
M.
,
Sofronis
P.
,
Robertson
I.M.
,
Acta Mater.
74
(
2014
):
p
.
244
254
.
16.
Depover
T.
,
Verbeken
K.
,
Mater. Sci. Eng. A
669
(
2016
):
p
.
134
149
.
17.
Hinotani
S.
,
Ohmori
Y.
,
Terasaki
F.
,
Mater. Sci. Eng.
76
(
1985
):
p
.
57
69
.
18.
Lee
J.
,
Lee
T.
,
Kwon
Y.J.
,
Mun
D.-J.
,
Yoo
J.-Y.
,
Lee
C.S.
,
Met. Mater. Int.
22
,
3
(
2016
):
p
.
364
372
.
19.
Depover
T.
,
Verbeken
K.
,
Int. J. Hydrogen Energy
41
,
32
(
2016
):
p
.
14310
14329
.
20.
Depover
T.
,
Verbeken
K.
,
Mater. Sci. Eng. A
675
(
2016
):
p
.
299
313
.
21.
Williams
D.B.
,
Carter
C.B.
,
The Transmission Electron Microscope
(
New York, NY
:
Springer
,
1996
),
p
.
3
17
.
22.
ASTM G129-00
, “
Standard Practice for Slow Strain Rate Testing to Evaluate the Susceptibility of Metallic Materials to Environmentally Assisted Cracking
” (
West Conshohocken, PA
:
ASTM International
,
2017
).
23.
NACE TM0198-2011
, “
Slow Strain Rate Test Method for Screening Corrosion-Resistant Alloys for Stress Corrosion Cracking in Sour Oilfield Service
” (
Houston, TX
:
NACE
,
2011
),
p
.
27
.
24.
Husby
H.
,
Wagstaff
P.
,
Iannuzzi
M.
,
Johnsen
R.
,
Kappes
M.
,
Corrosion
74
,
7
(
2018
):
p
.
801
818
.
25.
Henthorne
M.
,
Corrosion
72
,
12
(
2016
):
p
.
1488
1518
.
26.
ISO 17081:2014
, “
Method of Measurement of Hydrogen Permeation and Determination of Hydrogen Uptake and Transport in Metals by an Electrochemical Technique
” (
Geneva, Switzerland
:
International Organization for Standardization, 2014
).
27.
Hutchings
R.B.
,
Ferriss
D.H.
,
Turnbull
A.
,
Brit. Corros. J.
28
,
4
(
1993
):
p
.
309
313
.
28.
Manolatos
P.
,
Jerome
M.
,
Galland
J.
,
Electrochim. Acta
40
,
7
(
1995
):
p
.
867
871
.
29.
Devanathan
M.A.V.
,
Stachurski
Z.
,
Proc. R. Soc. London. Ser. A. Math. Phys. Sci.
270
,
1340
(
1962
):
p
.
90
102
.
30.
ASTM G148–97
, “
Standard Practice for Evaluation of Hydrogen Uptake, Permeation, and Transport in Metals by an Electrochemical Technique
” (
West Conshohocken, PA
:
ASTM International
,
2018
).
31.
Husby
H.
,
Iannuzzi
M.
,
Johnsen
R.
,
Kappes
M.
,
Barnoush
A.
,
Int. J. Hydrogen Energy
43
,
7
(
2018
):
p
.
3845
3861
.
32.
Snape
E.
,
Schaller
R.F.
,
Jones
F.
,
Corrosion
25
,
9
(
1968
):
p
.
380
388
.
33.
Kane
R.D.
,
Cayard
M.S.
, “
Roles of H2S in the Behavior of Engineering Alloys: A Review of Literature and Experience
,”
CORROSION 1998, paper no. 98274
(
Houston, TX
:
NACE
,
1998
).
34.
Galvele
J.R.
,
Degradación de materiales I: corrosión, Monografía tecnológica
,
no
.
3
. (
Buenos Aires, Brazil
:
Jorge Baudino Ediciones
,
2006
),
p
.
100
.
35.
Cabrini
M.
,
Lorenzi
S.
,
Marcassoli
P.
,
Pastore
T.
,
Corros. Rev.
29
,
5–6
(
2011
):
p.
261
274
.
36.
Cabrini
M.
,
Sinigaglia
E.
,
Spinelli
C.
,
Tarenzi
M.
,
Testa
C.
,
Bolzoni
F.M.
,
Materials (Basel, Switzerland)
12
,
11
(
2019
):
p.
1843
.
37.
NACE TM0177-2016-SG
, “
Laboratory Testing of Metals for Resistance to Sulfide Stress Cracking and Stress Corrosion Cracking in H2S Environments
” (
Houston, TX
:
NACE
,
2016
).
38.
Yoshino
Y.
,
Corrosion
38
,
3
(
1982
):
p
.
156
167
.
39.
Solheim
K.G.
,
Solberg
J.K.
,
Walmsley
J.
,
Rosenqvist
F.
,
Bjørnå
T.H.
,
Eng. Fail. Anal.
34
(
2013
):
p
.
140
149
.
40.
Zafra
A.
,
Belzunce
J.
,
Rodríguez
C.
,
Mater. Chem. Phys.
255
(
2020
):
p
.
123599
.
41.
Griffiths
A.J.
,
Turnbull
A.
,
Corros. Sci.
37
,
11
(
1995
):
p
.
1879
1881
.
42.
Depover
T.
,
Van den Eeckhout
E.
,
Verbeken
K.
,
Mater. Sci. Technol.
32
,
15
(
2016
):
p
.
1625
1631
.
43.
Li
L.
,
Song
B.
,
Cheng
J.
,
Yang
Z.
,
Cai
Z.
,
Int. J. Hydrogen Energy
43
,
36
(
2018
):
p
.
17353
17363
.