THE ROLES OF LOCAL HARD ZONES & MATERIAL INCLUSIONS
On the Sulfide Stress Cracking of a Pipe in a Sour Gas Processing Plant
1. IntroductionSulfide Stress Cracking (SSC) from wet H2S damage is an established degradation mechanism with well-documented failure case histories [1]. In this article, an unconventional failure case where SSC was a main culprit is discussed. A leak on a NPS 12” pipe carrying sour natural gas was discovered by a major oil and gas producer. Further investigation found the specific leak source at the six o’clock location (Figure 1). The submitted failed pipe cut-out was 1.5 meters (m) and contained a girth weld approximately 1 m upstream of the cracked location as-received (Figure 1). Notably, the sample had a wall thickness larger than the standard (i.e., Schedule 40) for NPS 12 pipe. This article summarizes the scope, methods, results, and interesting findings from the failure analysis conducted.
2. ScopeThe scope of work followed these steps: (1) visual examinations, (2) analysis of chemical composition of scale, (3) microhardness measurements, (4) fractography, and (5) microscopic examinations. In this article, only select results are presented, but discussions and conclusions are based on findings from all experimental methods.3. Results and Discussion
Material evidence found in this analysis indicated that the failure occurred due to SSC initiating from the external surface (outer diameter, OD) and then propagating towards the internal surface (inner diameter, ID) until a through wall crack had formed resulting in the leak of internal media.
3.1 Sulfide Stress Cracking
SSC is defined as cracking of a susceptible material under the combined action of tensile stress and corrosion in the presence of water and H2S [2]. Carbon steel is a susceptible material and all three primary components necessary for SSC to occur are present here:
• H2S: present in significant amounts (~40%) in gaseous internal media. Only a partial pressure of H2S >0.0003 MPa (0.05 psia) in the gas phase is enough for SSC to occur [2].
• Water: evidence showing staining of bottom surface where cracking was found supports the presence of an aqueous phase which previously pooled there (Figure 2). This is likely due to condensation of gas-phase moisture in the line, as circumferential streaking was found from the top to pooling stains.
• Tensile stress: the normal operating pressure of the line creates a hoop stress, which can initiate and sustain cracking in the longitudinal direction perpendicular to the stress. Additionally, non-uniform residual stress due to pipe manufacturing are present as evidenced by the hardness measurements (Table 2 and Schematic 1).
3.2 Sequence of Events
The evidence found in the analysis suggests that the following sequence events most likely occurred:
• H2S corrosion occurred on the bottom internal surfaces of the sample where a water phase was present. This resulted in a well-adhered black sulfide surface scale. Dissolved H2S dissociates into protons (H+) and sulfide (S2-). The main two electrochemical reactions of H2S corrosion are the anodic Fe dissolution from the steel into ferrous (Fe2+) and the cathodic reduction of ionic hydrogen in the aqueous phase into atomic hydrogen (H). The chemical association of Fe2+ and S2- forms the black FeS-rich scale found on the ID (Table 1).
• H formed on the ID surface during (i) permeate through the scale and into the steel. It is noted that:
• Condensed water from moisture in the gas phase does not contain the buffering capability of produced water due to the absence of bi/carbonate ions, hence the pH of the water phase previously pooled at the bottom of the pipe is expected to be quite low.
• At lower pH, hydrogen evolution by the reduction of H+ is dominant (over water reduction) due to the high concentration of H+.
• H2S is a poison for the hydrogen recombination reaction, which would normally result in the formation and off-gassing of diatomic hydrogen (H2(g)).
• Elevated temperatures promote the dissociation of H2S and increase the diffusion rate of H in metals. Therefore, H ingress to the steel was relatively easy at the 70°C operating temperature. However, the SSC cracking potential is maximized at near-ambient temperature. This distinction is important because the steel may have been charged during operating temperature and subsequently cracked during excursions to lower temperatures (e.g., shutdowns).
• With time, subsurface H diffused to the OD surface. During this process, some atoms are trapped at inclusions and dislocations in the material.
• SSC initiated on the OD surface at multiple locations at similar times. SSC is possible in base metals with localized hardness points. It is expected that hardness points >200 HB will eventually crack as per [4], especially if localized hardness exceeds 237 HB [2]. It is noted that:
• Threshold hardness in standards (e.g., maximum 22 HRC [5]) do not indicate immunity of materials to SSC below them. Rather, they indicate low susceptibility, but the material may still experience SSC with the confluence of other factors.
• Some carbon steels contain residual elements that form high hardness areas that will not temper at normal stress relieving temperatures.
• The microstructures examined from this sample showed evidence of spherical holes and inclusions rich in aluminum (Al), calcium (Ca), and silicon (Si) (Figure 3). Ca concentrations are not specified in ASTM A106 Grade B.
• Based on the outer shell with inner core morphology of the inclusions and the presence of calcium and aluminum, it is probable that during the manufacturing process calcium was injected into the liquid steel to reduce Sulphur levels to specification and change inclusion morphology and composition [6].
• Inclusions are typically less ductile than the surrounding matrix and therefore create localized triaxial stresses where crack initiation is more likely. Furthermore,
inclusions and high-stress areas are locations where diffusing hydrogen is trapped, embrittling the material.
• The density of inclusions near the OD and FF of the failed metallographic cross-section was higher than the intact counterpart. Also, secondary cracks initiating on the OD of the former were seen aligning in the same direction of a string of inclusions (Figure 4b).
• The highest hardness was measured at the intersection of the OD and the FF. Also, the highest drop in hardness between the OD and the adjacent mid-wall was along the FF.
• Once the SSC initiated, the stress intensity factor of the sharp crack tips sustained slow propagation towards the ID. It is noted that:
• H in steel reduces the threshold stress intensity factor (Kth). The following equation describes the dependence of Kth on H concentration [7]:
where α, α”, and β’ are constants, kIG is the intrinsic Griffith toughness for cleavage fracture without H, σYS is the yield strength of the material, and CHσ,T is
the H concentration at the crack tip. CHσ,T depends on subsurface H concentration at the ingress side using [7]:
where σH is the crack-tip hydrostatic stress and VH is the molar volume of H in the steel.
• Triaxial stresses at the crack tip are typically higher in thick-walled pipe versus thinner walled pipe due to the constraining effect of more surrounding material.
• The cracks propagated together at different rates and coalesced where they became close to one another or overlapped. This indicates slow crack growth.
• Leaking to the outside occurred when one crack fully penetrated the wall thickness, corroding areas near the crack on the OD.
4. Summary
Material evidence found in the analysis this case indicates the failure of the pipe was primarily due to SSC initiating from the OD and then propagating towards the ID. The combination of higher hardness on the OD, the presence of inclusions and their higher concentration near the OD surface, and the presence of hydrogen in the susceptible steel
material created a situation where SSC initiation was possible on the OD. Once the SSC initiated, the stress intensity factor of the sharp crack tip was enough to sustain slow
propagation of the crack towards the ID, aided by the lower threshold stress intensity of the material due to hydrogen, high crack-tip stresses due to the constraining effect of the
thick-walled pipe, and the existence of higher stress regions of harder inclusions along the crack direction where embrittlement is augmented by trapped hydrogen.
5. References
[1] R. Pourazizia, M. A.; Mohtadi-Bonabb, J. A.; Szpunara, Engineering Failure Analysis 109 (2020).
[2] API RP 571, 2nd Edition, 2011.
[3] ASTM E140, 1st Edition, 2005.
[4] NACE RP0472, 2005.
[5] ANSI/NACE MR0175/ISO 15156-2, 2015.
[6] W. Cubberly, Metals Handbook: 9th Edition, 1975.
[7] H. M. Ha, I. M. Gadala, A. Alfantazi, Electrochimica Acta 204 (2016).
Contributed by: IBRAHIM M. GADALA, PH.D., P.ENG., RESEARCH ENGINEER, NOVA CHEMICALS CORPORATION
The contents of articles and any opinions expressed therein are those of the authors and do not represent those of MTI.