Geothermal energy is a clean, reliable source of energy with low life cycle emissions. Its applicability for both power generation and direct district or industrial heating makes it an attractive source of energy. However, the operating environments in geothermal wells pose significant challenges for material integrity due to corrosion and mineral scaling. This article focuses on corrosion; for more information on mineral scaling, see here and here.
Elevated temperatures, high chloride concentrations, and the presence of acid gases, such as CO2 and H2S, make geothermal systems particularly aggressive. To mitigate the risk of general corrosion, large parts of geothermal assets are constructed using corrosion-resistant alloys (CRAs), such as stainless steels, Ni-base alloys, and Ti-base alloys, in place of carbon steel. Though CRAs are mostly resistant to general corrosion, localized corrosion remains a significant threat. It is a key contributor to operational downtime and maintenance costs, which can account for up to 30% of the levelized cost of electricity (LCOE) in geothermal plants.
Unlike general corrosion, which affects large surfaces uniformly, localized corrosion occurs at discrete sites, often in the form of pits or in crevices, and can progress undetected until catastrophic failure. Being stochastic in nature, when and where pitting or crevice corrosion will occur are hard to predict. To proactively manage asset integrity, the ability to predict localized corrosion behavior under real operating conditions is essential. This article outlines recent advancements in OLI’s localized corrosion model, developed as part of US Department of Energy funded Small Business Innovation Research (SBIR) project in collaboration with DNV and MC Consult.
Basics of the OLI Localized Corrosion Model
The localized corrosion model in the OLI Corrosion Analyzer predicts the conditions under which localized corrosion, such as pitting or crevice corrosion, may occur. The basic concept behind the model is depicted in the figure below, which illustrates various stages of localized corrosion:
- Figure 1a: A corrosion-resistant alloy has a passive film, protecting the metal surface.
- Figure 1b: A local breakdown of the passive film occurs, such as from mechanical damage or chemical attack.
- Figure 1c: Metastable pitting begins at an anodic site, where metal dissolution starts. The cathodic reduction reaction occurs outside the pit.
- Figure 1d: The metastable pits either continue to grow, forming stable pits, or depending on the electrochemical conditions –
- Figure 1e: An incipient passive layer forms, preventing further localized corrosion.
Figure 1. Schematic showing the basic concept behind the OLI localized corrosion model
The model predicts conditions for localized corrosion by comparing two key electrochemical parameters:
- Corrosion Potential (Ecorr): The potential at which a metal surface corrodes freely. It is determined by a mixed potential model, which takes into account anodic reactions, such as metal dissolution, and cathodic reactions, such as the reduction of species like oxygen, water, hydronium ions, etc. Ecorr is calculated using a mixed-potential model as the potential where the anodic and cathodic currents are equal (see here and here for a detailed description).
- Repassivation Potential (Erp): The potential below which a growing pit or crevice will repassivate and a protective oxide layer reforms on the metal surface. Erp is influenced by environmental factors such as the concentration of different species (e.g., chlorides) and temperature. Erp is calculated using a repassivation potential model as the potential at the point of repassivation when an incipient passive layer forms (see here and here for a detailed description).
Whether the pit continues to grow or repassivates depends on the relationship between the corrosion potential and the repassivation potential. The model predicts that localized corrosion may occur when:
Ecorr > Erp.
If Ecorr remains above Erp, the pit may grow deeper or another pit may appear, increasing the risk of equipment failure. However, if Ecorr falls below Erp, the pit repassivates, halting further corrosion.
The OLI Corrosion Analyzer calculates these critical potentials to access the likelihood of localized corrosion. By understanding how different components in the geothermal brine as well as temperature and pH affect localized corrosion, users can better assess corrosion risk and take action before failures occur. The sections below highlight how the model captures these effects and supports proactive decision-making.
Effect of high temperature and chlorides
Temperature affects both Ecorr and Erp but is much more significant for Erp, as shown in figure 2. Model predictions for Erp as a function of chloride activity, shown for super duplex alloy 2507 (UNS S32750) in the figure below, exhibit the experimentally observed trends:
- Decreasing Erp with increasing chloride concentration i.e. more aggressive conditions reduce the ability of the alloy to repassivate.
- Decreasing Erp with increase in temperature, leading to higher localized corrosion risk at higher temperature.
- Double-slope behavior in Erp vs. Cl⁻ plots highlighting the competing roles of halides and water in film stability.
Figure 2. Erp for alloy 2507 varying with Cl– activity. Symbols are experimental results37-40 and curves are model predictions.
Effect of inhibitors such as sulfates
Inhibitors act by mainly aiding the formation of a passive film and thus predominantly affect Erp. Figure 3 shows the effect of both aggressive (Cl⁻) and inhibitive (SO₄²⁻) species on the Erp of alloys 2507. At low concentrations of SO42- ions, Erp changes almost negligibly with increasing SO42- concentrations. However, at a critical concentration, Erp increases rapidly reaching values that make localized corrosion impossible, even at high corrosion potentials. The exact location of this transition depends on the temperature and chloride concentration. At low chloride concentrations, the inhibition occurs at lower sulfate concentrations, as shown by the red and blue curves (0.4 and 4M NaCl concentration, respectively, at 60°C). The 0.4M NaCl curve shows a transition at around 0.03M SO42-, whereas the 4M NaCl reaches the solubility limit before inhibition occurs. Higher temperatures require higher sulfate concentrations for effective inhibition, highlighting the importance of the high-temperature data measured at DNV as part of the SBIR effort.
Figure 3. Effect of sulfate ion concentration on Erp for alloy 2507. Symbols are experimental results and curves are model predictions.
Effect of pH
The primary driving force for localized corrosion of an alloy in the passive state is Ecorr which is very sensitive to pH. In contrast, Erp is relatively unaffected by pH. Figure 4 shows the strong dependence of Ecorr on pH in a de-aerated environment for alloy 625 compared with experimental data. In the absence of oxygen, the dominant cathodic reaction is the reduction of water in neutral conditions and reduction of hydronium ions in acidic conditions. This trend is consistent across nickel-base alloys and stainless-steel under de-aerated conditions.
Figure 4. Corrosion potential of alloy 625 varying with pH. Symbols are experimental data and the curve is the model prediction.
Effect of CO2 and H2S
CO2 and H2S, common acid gases present in geothermal systems, impact corrosion differently. CO2 primarily influences Ecorr by lowering the pH of the environment, while H2S affects both Ecorr and Erp. Figure 5 shows the corrosion rate of alloy S13Cr varying with temperature for a system with NaCl and CO2. The addition of CO2 lowers the pH but it generally remains above the depassivation pH and thus in the passive state.
Figure 5. Corrosion rate of alloy S13Cr varying with temperature in a system with NaCl and CO2.
Figure 6 shows the effect of H2S on Ecorr of alloy S13Cr. The presence of H2S increases the Ecorr by about 200mV above the non-H2S baseline. H2S has a compound effect on the cathodic process since it lowers the pH as well as participates via direct reduction. It also affects the anodic passive current density via the dissolution of the sulfide layer. The model incorporates these different effects mechanistically which results in high predictive power even in regions where little to no data is available.
Figure 6. Corrosion potential of alloy S13Cr as a function of temperature in 0.3m NaCl solution in the presence of N2 and N2 + H2S mixtures. Symbols are experimental data and curves are model prediction.
Figure 7 shows the complex effect of H2S on Erp for alloy S13Cr. When no H2S is present in the system, the Erp vs activity of chloride ion follows the standard dual slope curve as shown by the blue curve. At high H2S concentrations (100% in vapor phase) shown by the green curve, the curve is parallel to 0% H2S curve with Erp value shifted to lower potentials by about 200 mV. The presence of H2S accelerates the anodic dissolution in the localized environment which increases the tendency of the alloy to undergo localized corrosion at all chloride concentrations. In contrast, the effect of H2S at 1% strongly depends on the chloride concentration. At high chloride concentrations, the 1% H2S curve behaves similar to the 100% curve with the shift in Erp curve being weaker. However, a drastically different behavior is observed at low chloride concentrations which indicates a change in mechanism. At low concentrations of chloride, Erp increases above the 0% H2S curve. This increase in Erp is due to a metal sulfide formation which competes with the metal oxide formation and has an inhibitive effect. The competition between accelerated anodic dissolution from the adsorption of H2S and inhibition from metal sulfide formation leads to the complex behavior seen at 1% H2S. At low H2S and low chloride activities, the effect of metal sulfide formation dominates while at higher chlorides the acceleration of anodic dissolution dominates. The mechanistic nature of the model makes quantitative prediction of this complex phenomenon possible.
Figure 7. Repassivation potential for alloy S13Cr at 85°C in Cl−+H2S systems as a function of chloride ion activity at various concentrations of H2S in the gas phase (0, 1, and 100 wt%). Symbols are experimental data and curves are model predictions.
Effect of dissolved oxygen
Dissolved oxygen (DO) can significantly raise the risk of localized corrosion by increasing Ecorr even at ppb concentrations. At such low concentrations, mass transfer limitations become important and mixing conditions have a significant influence on Ecorr. This is shown in Figure 8 for alloy 2205 at different temperatures and NaCl concentrations. In the absence of oxygen, Ecorr is very low (~ -0.43V SHE) but dramatically increases with increasing DO. The key partial cathodic reactions are the reduction of water and oxygen. When oxygen is present, even in low quantities, the reduction of oxygen is often the main cathodic reaction. In the absence of oxygen, the dominant cathodic reaction is the reduction of water. Since the reduction of oxygen is subject to mass transfer limitations (due to the diffusion of oxygen to the surface), the current density corresponding to the reduction of oxygen has contributions from an activation-controlled term and a diffusion term. For the activation-controlled term, the reaction orders with respect to O2 and H3O+ ion are specific to the surface, whereas the limiting current density for the diffusion term is independent of the surface and depends only on mixing conditions, diffusivity of O2, and density and viscosity of the solution. This leads to a complex dependence of Ecorr on mixing conditions as shown in Figure 8b.
Figure 8. (a) Comparison of experimental Ecorr vs. OLI model outputs for alloy 2205 in NaCl solutions at 45°C and 60°C as a function of DO concentration (in ppb). Three different NaCl concentrations were used at each temperature: a) 45°C, 0.1 M NaCl, b) 45°C, 0.6 M NaCl, c) 45°C, 3.0 M NaCl, d) 60°C, 0.1 M NaCl, e) 60°C, 0.6 M NaCl, and f) 60°C, 3 M NaCl. (b) Ecorr vs. square root of the DO concentration (in ppb) 60°C and 0.6 M NaCl. Symbols are experimental data195 and the curves are model predictions. Model predictions are shown for different mixing conditions: static systems, rotating disk at ultra-low (5 rpm) and low (30 rpm) rotation speed and complete agitation.
Figure 8b elucidates the effect of mixing to different extents. Flow effects with varying degree of mixing are illustrated using static condition (solid curve), a rotating disk electrode (as an example of well-defined flow conditions) with two very small rotation speeds – 5 rpm (dashed curves) and 30 rpm (dash-dotted curves), and the limiting case of complete agitation (dotted curve) when there is no mass transport limitation. At low DO, Ecorr is strongly flow-dependent because the limiting current density for O2 reduction is very low, as it is related to DO through well-known mass-transfer coefficient relationships. When the limiting current density for O2 reduction is lower than the passive current density, then Ecorr drops abruptly to low values controlled by the reduction of H2O and not by the reduction of O2 as shown in the polarization curves for static conditions at 60°C, 0.6M NaCl, and 8ppb DO in Figure 9a. Even at the ultra-low rotation speed of 5 rpm, there is an effect on Ecorr because mass transport by flow is important to bring oxygen to the surface at low DO concentration when the mean diffusion path is large as flow reduces the thickness of the diffusion zone. In the case of complete agitation, where there are no mass transfer limitations, the O2 reduction is the dominant cathodic reaction and the Ecorr is correspondingly higher. Any realistic corrosion scenario must lie between the solid line (static conditions) and the dotted line (complete agitation) in Figure 8b. Thus, at low DO values (below ca. 100 ppb), modelling can provide a realistic range of Ecorr even if the actual flow conditions cannot be rigorously modelled.
Figure 9. Predicted polarization curves for 60°C and 0.6 M NaCl with (a) static mixing and (b) complete agitation. The partial electrochemical processes are defined in the legend at the bottom.
Critical crevice temperature
To evaluate the risk of localized corrosion, predictions of Ecorr and Erp must be considered together. A strong test of the integrated model is its ability to predict the critical crevice temperature (CCT), defined as the temperature threshold above which crevice corrosion is observed. At temperatures below CCT, Ecorr lies below Erp whereas it exceeds Erp above CCT. Thus, the temperature at which Ecorr and Erp curves intersect should provide an estimate of CCT. Figure 10 shows this approach for alloy 625 in 6% FeCl3 solution. As shown in the figure, Erp and Ecorr curves intersect at around 44°C, predicting a CCT that is in good agreement with experimental values which are scattered from 30 to 50°C.
Figure 10. Prediction of the critical crevice temperature for alloy 625 in a 6% FeCl3 solution.
Conclusion
OLI localized corrosion model enables quantitative prediction of corrosion risk by combining corrosion potential and repassivation potential calculations. By accounting for the effects of temperature, pH, and brine composition, the model provides a powerful tool for assessing alloy performance and defining integrity operating windows. As geothermal deployment expands, such predictive capabilities will be essential for optimizing materials selection and minimizing unplanned downtime.