A Limited Evaluation of the Applicability of NACE MR0175/ISO 15156 H₂S Limits to Supercritical CO₂ Pipelines

In an earlier paper one of the authors raised the question of whether the H₂S partial pressure threshold of 0.05 psia (0.3 kPa) for cracking of steels from sulfide stress cracking (SSC) is pertinent to dense phase CO₂ systems in which there is no free gas phase.¹  The basis for the 0.05 psia (0.3 kPa) threshold is aimed entirely at upstream oil and gas production and transportation for which a separate gas phase is often present. The standard that stipulates this limit is NACE MR0175/ISO 15156²  which recently included the option to evaluate those systems that are at higher pressures and not characterized as ideal gases using fugacity. This also holds for environments that are all liquid with no separate gas phase. Therefore, thermodynamically the correct parameters to define and calculate the contribution of various phases in Supercritical CO₂ (SC-CO₂) are the fugacity and the concentration in the water phase. A limited test program was designed to examine whether SSC would occur in X80 steel at H₂S partial pressures (pH₂S) of 0.05 psia (0.3 kPa) and 0.20 psia (1.38 kPa). It is common practice when considering the possible role of H₂S in SC-CO₂ pipelines to apply the NACE MR0175/ISO 15156 threshold of 0.05 psia (0.3 kPa) without understanding that there is no free gas phase in SC-CO₂ and therefore, partial pressure is not applicable.

Paul³  performed tests similar to the tests in this program using API 5L Grade X65 welded pipe samples stressed to 90% specified minimum yield stress (SMYS) for 30 d. The tests were conducted at 14.5 psia (100 kPa) H₂S and a CO₂ pressure of 1,450 psi (10 MPa) at 40°C. He observed both pitting and crack-like features at high magnification which were not visible at 10× magnification. This pH₂S is far beyond the currently considered H₂S limits for most planned SC-CO₂ pipeline applications and does not represent realistic concentrations from CO₂ emitters at this time.

Li, et al.,⁴  tested API 5L Grade X65 pipe specimens in 1,450 psi (10 MPa) CO₂ at 45°C with 50 mole-ppm and 100 mole-ppm H₂S using four-point bent beams stressed up to 100% yield strength for 137 h and observed no cracking of the specimens. However, these authors acknowledged the test period was likely too short to be certain that cracks would not form after longer testing.

Testing was performed in accordance with NACE TM0316-2023⁵  with modifications in the test environment as explained below. A sample of 20 in OD, API 5L Grade X80 line pipe was used for the tests. The chemical composition of the pipe was 0.052% C, 1.7% Mn, 0.003% S, 0.009% P, 0.45% Cr, 0.25% Cu, < 0.010% Mo, and 0.096% Nb. Triplicate four-point bend beam specimens were machined from the sample in the longitudinal direction and polished to 600-grit finish. All specimens were scribed with a unique identification number. Specimens 1 through 3 were used for test no. 1 (25 mole-ppm H₂S) and specimens 4 through 6 were used for test no. 2 (100 mole-ppm H₂S). The dimensions of four-point bend beam specimens were 4-in (length) by 0.7-in (width) by 0.06-in (thick). Specimens were loaded on Alloy C-276 jigs to 90% (73.0 ksi) of the actual yield strength (AYS), 81.2 ksi, which was applied on all specimens via deflection. The average hardness of the samples is 227 HVN.

TM0316 provides the option to use a test solution that represents a fit-for-purpose environment that represents the service application. Therefore, the conditions shown in Table 1 represent typical SC-CO₂ pipeline conditions except for the chloride content of 1,000 (ppm). Currently, there are no data available on existing CO₂ pipelines to define a typical chloride content in the water phase. In theory, the water condensing out of the CO₂ dense phase would be expected to be fresh with little or no ions present, however, there is a potential for carryover of ions, especially chlorides, from upstream treating processes. Therefore, 1,000 ppm chloride was included to represent some possibility of such events.

Table 1.

Typical Sc-CO2 Pipeline Conditions used for Testing

View large

View Large

Modeling was performed using OLI^†’s Proprietary software (Studio) for the test conditions provided in Table 1 to determine the equilibrium pH, dissolved CO₂ and H₂S in the brine solution, and the fugacity of H₂S, fH₂S. Calculated values were provided in Table 2.

Table 2.

Results of OLI Modeling using the Inputs from Table 1^(A)

View large

View Large

All four-point bend beam specimens were placed inside the respective autoclaves, the autoclaves were subsequently sealed and purged with ultrahigh purity (99.9999%) nitrogen to remove the air present in the autoclave and then transfer it the de-aerated test solution in the autoclaves. The test specimens were completely immersed in the brine test solution and the test fluid volume-to-specimen-surface-area ratio was maintained at >20 mL/cm². A leak test using ultrahigh purity (99.9999%) nitrogen was performed and any gas leaks were fixed. The test solution with the gas at ambient conditions was then heated to the test temperature within ±2°F. The autoclaves were then pressurized to the test pressure within ±25 psig with the test gas. As the mixed gas tank’s maximum pressure is around 580 psig, a CO₂ pressure booster was used to increase test pressures to the target pressures. The test gas was replenished in the autoclaves once, midway, and through the test period. This was accomplished by releasing approximately 500 psi gas from the test vessel into an H₂S scrubber, replenishing it with a fresh gas mixture, and then stabilizing it at the test pressure. The temperature was not adjusted and kept close to the test temperature during the replenishment process. The test temperature and pressure were manually logged throughout the exposure. The pressure was within ±50 psig and the temperature was within ±3°F throughout the test.

After 30-d testing, the heating elements temperature controllers were turned off and the autoclaves were allowed to cool down to ambient temperature. Once the temperature reached ambient temperature the test gas from the autoclaves was released into the H₂S scrubber and then the lids were opened to pump out the test fluids. Finally, the test specimens were removed, cleaned with demineralized water and acetone, and then air dried. Visual examination of all samples was performed at magnifications between 10× and 40×, and corroded areas were documented photographically, including localized corrosion. Wet-magnetic particle testing was performed for all four-point bend beam specimens. Stereoscopic and scanning electron microscopic (SEM) examinations of pitting were performed and photo-documented. Metallography was performed on four-point bend beam specimens through the highest stress area and examined at 100× through 1,000× magnifications using an optical metallographic microscope.

The results of this study indicate that the 0.05 psia (0.3 kPa) threshold for SSC in NACE MR0175/ISO 15156 is likely not applicable to SC-CO₂ systems. While it is recognized that this focused study is not sufficiently comprehensive to unequivocally state the pH₂S threshold is not pertinent it does indicate that further study of the limits of H₂S in SC-CO₂ is important for future CO₂ pipeline applications. Moreover, the application of partial pressures in dense phase systems is incorrect and to establish a justifiable H₂S threshold, if there is one, further studies should use the fugacity of H₂S as the criterion for justifying any such limit. Not only is this question of technical importance it can also have a profound effect on the economics of SC-CO₂ pipelines. At present, the industry applies the NACE/ISO15156 limit for SC-CO₂ pipelines which can lead to significant added costs for a project that may not be necessary.

Furthermore, Caldwell, et al.,⁶  examined the origins of the 0.05 psia rule and found there was no theoretical basis for this limit which was established simply from empirical experience and data. Huizinga⁷  has explained the divergence of pH₂S vs. fH₂S for high-pressure oil and gas systems and the necessity of properly applying fH₂S when there is no gas phase and when pressures are high enough that ideal gas behavior is not correct.

• The application of NACE MR0175/ISO 15156 0.05 psia partial pressure threshold to SC-CO₂ systems is neither theoretically correct nor practically useful and can lead to significant conservatism for SC-CO₂ pipelines. Using H₂S partial pressure was demonstrated not to adhere to the 0.05 psia limit for SSC. The fugacity of H₂S was also considered but there was insufficient testing to predict what, if any, fH₂S would produce a threshold for SSC.

The financial support from the Carbon Capture, Utilization, and Storage (CUUS) Program under the Office of Energy R&D, Natural Resources Canada, and CanmetMATERIALS ia acknowledged.