Effect of High Temperature on the Aqueous H₂S Corrosion of Mild Steel

As the drilling and exploitation environments in oil and gas industry are becoming harsher and more aggressive, high-temperature hydrogen sulfide (H₂S) corrosion is more frequently encountered.^1-3  High temperatures and high pressures in association with H₂S corrosion may lead to many engineering challenges such as pipeline integrity, materials selection, and corrosion mitigation.

H₂S corrosion at low temperatures (<80°C) has been more widely studied,^4-6  and significant progress has been achieved to reveal the related general corrosion mechanisms. At elevated temperatures, however, H₂S corrosion has not been thoroughly studied and the understanding of the associated corrosion mechanisms is very limited. Overall, high temperature is fully expected to have a significant effect on the corrosion rate and formation/transformation of iron sulfide polymorphs.⁷  Different corrosion product layers may also have different effects that lead to surface heterogeneity, the initiation of galvanic corrosion, and localized attack.

Therefore, in order to better understand, predict, and mitigate H₂S corrosion in oil and gas production at elevated temperatures, further experimental investigations are of great necessity. In this work, H₂S corrosion tests were conducted at 80°C, 120°C, 160°C, and 200°C to identify the effect of high temperature on the kinetics of corrosion and layer formation on mild steel in sour environments.

Experiments were performed in a 7 L Hastelloy^† autoclave. A conventional three-electrode setup was used to conduct linear polarization resistance (LPR) measurements using a potentiostat. The working electrode was a cylindrical sample made from UNS K03014⁽¹⁾ (API 5L X65) carbon steel. A Pt-coated Nb counter electrode and a commercial Zr/ZrO₂ high-temperature, high-pressure pH probe were used. The pH probe reference also served as a reference electrode (exact potential still unknown) as long as its potential was stable at the desired test conditions.⁸  Flat samples were also attached to a stabilized shaft using a polytetrafluoroethylene (PTFE)-coated Type 304 stainless steel (UNS S30400) wire. Before each experiment, the specimens were polished up to 600 grit sandpaper and rinsed with deionized water and isopropanol. A centrally located impeller with a rotation speed of 1,000 rpm was used to keep the solution fully mixed during each test.

The speciation at different temperatures was calculated according to an in-house water chemistry model,⁹  as summarized in Table 1. The dissolved H₂S concentration [H₂S]_aq was kept constant at 0.00385 mol/L, which corresponds to 0.10 bar H₂S at 80°C. After each test, the corroded samples were retrieved and characterized by x-ray diffraction (XRD), scanning electron microscopy/energy dispersive x-ray spectroscopy (SEM/EDS), and surface profilometry. Other experimental procedures and details can be found elsewhere.⁹ 

Figure 1 shows the corrosion rates over time at 80°C, 120°C, 160°C, and 200°C as measured by LPR. It can be seen that the initial corrosion rates increased with increasing temperature, and then quickly decreased to stable corrosion rates of 4.1, 3.8, 1.8, and 2.5 mm/y, respectively, from lowest to highest temperature. Overall, the final steady-state corrosion rate decreased with temperature except at 200°C.

The corrosion products on the steel surface were characterized by XRD as shown in Figure 2. While mackinawite (FeS) was the main corrosion product detected at 80°C, troilite (FeS), pyrrhotite (Fe_1−xS, x = 0 to 0.17), and pyrrhotite/pyrite (FeS₂) became the dominant species as temperature was increased. With increasing temperature, the corrosion product became richer in sulfur; this is an indication of enhanced reaction kinetics for phase transformations.

The morphologies of the formed corrosion products were also characterized by SEM, as shown in Figure 3. The SEM for the 80°C specimen shows a mackinawite layer of 15-μm thickness. From the EDS line scan, the outer layer was identified to likely be an iron sulfide but an inner layer, which consisted mostly of iron and oxygen was assumed to be an iron oxide. At 120°C, the SEM shows troilite-like crystals on the surface and a much thicker layer (61 μm to 73 μm). At 160°C, pyrrhotite crystals were clearly observed on the surface. The corrosion products changed to planar flaky crystals at 200°C. All of the cross sections show a two-layer structure at every temperature tested: an inner layer comprising an iron oxide and an outer layer comprising an iron sulfide. However, the iron oxide was not detected by XRD as a result of the top layer being too thick and compact for XRD penetration.

Until now, iron oxide has not been given much attention as a corrosion product in H₂S corrosion environments. It is hypothesized that this iron oxide is magnetite (Fe₃O₄) because it was already identified at high temperatures in CO₂ environments.¹⁰  The kinetics of Fe₃O₄ formation is very fast, making the scaling tendency (ST, the ratio of precipitation rate to corrosion rate) very high at high temperature. Fe₃O₄ can form on the metal surface according to the following reaction:

However, Fe₃O₄ is thermodynamically less stable than any of the iron sulfides. No Fe₃O₄ is thermodynamically expected in an Fe-H₂S-H₂O system.⁹  Here, for an Fe-CO₂-H₂O system, from the Pourbaix diagram shown in Figure 4, it is seen that Fe₃O₄ is stable in a very limited narrow area at potentials more positive than those for FeCO₃ at 80°C. When the temperature increases to 200°C, the stability region for Fe₃O₄ is greatly increased. New model for high-temperature corrosion of mild steel in aqueous H₂S environment at high temperature should also include Fe₃O₄ formation.

After removal of the corrosion products, the metal surface was characterized by profilometry, as shown in Figure 3. No obvious localized corrosion was observed at 80°C and 120°C. The surface was relatively smooth and the corrosion could be considered as uniform. However, at 160°C some small pits could be observed with approximately a 1.2 pitting ratio (ratio of maximum pit penetration rate to general corrosion rate) and 1.5 mm/y pit penetration rate. This does not constitute localized corrosion but may indicate initiation. At 200°C, many large pits appeared with a 3.2 pitting ratio and 8.2 mm/y pit penetration rate. The pitting ratio may not be a very accurate measure as pitting corrosion actually overwhelmed general corrosion. Because of severe localized corrosion at this temperature, the stable LPR corrosion rate appears to be a little higher than at 160°C (Figure 1). These results fit with Ning, et al.’s previous work¹¹  where it was found that pyrite formation triggers the occurrence of localized attack.

• •

  Sour corrosion experiments were conducted successfully at 80°C, 120°C, 160°C, and 200°C. Initial corrosion rates increased with increasing temperature. Final corrosion rates, after 4 d of exposure, remained high at between 2 mm/y and 4 mm/y. The corrosion product comprised two distinct layers: an inner layer, close to the steel substrate, and an outer layer, on the solution’s side. The inner layer was identified as an iron oxide (postulated to be Fe₃O₄) even though it is not thermodynamically stable in the conditions tested. The outer layer was mainly composed of iron sulfide: mackinawite, troilite, pyrrhotite, and pyrite at 80°C, 120°C, 160°C, and 200°C, respectively.

The authors would like to express sincere appreciation to the following industrial sponsors for their financial support and direction: Anadarko, Baker Hughes, BP, Chevron, China National Offshore Oil Corporation, ConocoPhillips, DNV GL, ExxonMobil, M-I SWACO, Occidental Oil Company, Petroleum Institute, PTT, Saudi Aramco, Shell Global Solutions, SINOPEC, TOTAL, TransCanada, and WGK.