CHALLENGES with CORROSION
(see also text on ‘solutions’)
The degradation of metallic components through redox reaction, corrosion, occurs to variable degree when the metals interact with the formation brine. Corrosion can cause installations to fail and the corrosion products may lead to clogging of e.g., the injection wells. Such blockage is a general challenge in operating geothermal plants and may lead to unintentional pressure build-up in the injection wells. Because iron-rich alloys are typically used in the infrastructure, corrosion products comprise iron compounds (oxides, hydroxides, sulphides, carbonates). Furthermore, if dissolved metal ions more noble that Fe, such as lead and copper, occur in the thermal waters, galvanic corrosion can cause their reduction to form metallic particles. Corrosion is commonly related to oxygen ingress, or occurrence of gasses such as CO2 and H2S in the formation waters. In many cases, corrosion processes accelerate with increasing salinity of the formation brines and increased temperature.
Operational challenges and types of corrosion
The interaction between brine and metallic elements in the installations at sites commonly results in corrosion, whose reaction products may cause operational difficulties. Corrosion can occur with a variety of oxidants. Corrosion of Fe(0), a major constituent of many alloys, is particularly rapid through redox reaction with oxygen, which often cause the formation of Fe(III) oxides (rust), e.g.:
4 Fe0 + 3 O2 + n H2O= Fe2O3 * n H2O
In the absence of O2, corrosion occurs anaerobically at much slower rate using water as the oxidant:
Fe0 + H2O = Fe2+ + H2 + 2 OH–
This reaction produces hydrogen gas (H2). It also increases pH locally so that the solubility of ferrous hydroxide may be exceeded. At such places, the net reaction becomes:
Fe0 + 2H2O = Fe(OH)2 + H2
Ferrous hydroxide is not very stable thermodynamically and usually dissolves if pH is not above 7.5 or so. Given that the reported pH of waters in our database typically is lower than 7.5, it is unlikely that ferrous hydroxide can migrate from the site of formation at temperatures above ~75 °C. However, ferrous hydroxide can itself reduce the protons in water at appreciable rates, leading to the formation of the mixed valent iron oxide magnetite, Fe(II)Fe(III)2O4, which is thermodynamically very stable.
Thus, formation of smaller amounts of ferrous, mixed valent, or ferric oxides, hydroxides, or oxyhydroxides from corrosion of infrastructure with metallic iron is unavoidable. If temperatures are sufficient or oxidation occurs with O2, the Fe oxides are likely to be thermodynamically stable. The formed Fe oxides are typically nanoparticulate, aggregated to variable degree, and they may migrate with the flow, which could cause clogging of the injection well. Typically, however, they form surface coatings that protect the metallic installations from further corrosion. Thus, the slow anaerobic corrosion with water as the oxidant generally causes little harm. The ingress of oxygen, however, can dramatically increase corrosion rates. Hence, the geothermal operations are typically operated at significant overpressure to avoid influx of O2 (and minimize formation of CO2). Figure 1 presents an overview of corrosion types that can occur for both unalloyed steel and corrosion resistant alloys (CRA). A brief description of each corrosion type is outlined below.
Uniform corrosion can occur for carbon steel in contact with water, Figure 1a. Several oxidants can participate in the reaction, including oxygen and water. The corrosion rate may be increased when water flow or temperature is high, resulting in variable distribution of passivating corrosion products and patches that are poorly protected.
Pitting and Crevice corrosion
Corrosion resistant alloys are protected by a passive layer of metal oxides, which greatly lowers corrosion rates. However, defects in the layer can cause pitting corrosion in the presence of small amounts of oxygen. Within the pit, the anodic reaction locally generates acidic conditions that prevent formation of passivating layers and keeps the metal reactive. In contrast, the cathodic reaction occurs over large areas of the metal, even when passivated, Figure 1b. Crevice corrosion is largely similar to pitting corrosion but occurs in minute cavities in the infrastructure, such as joints (Figure 1c).
At two sites in our database, analyses of sampled solids display elevated concentrations of Cu or Pb and X-ray diffraction shows the presence of metallic Cu (Cu0) or Pb (Pb0). The Cu0 and Pb0 form from corrosion through galvanic corrosion, a redox reaction between dissolved metal ions and the metallic iron in the infrastructure, for example:
Fe0 + Pb2+ = Pb0 + Fe2+
This mechanism is illustrated in Figure 1e.
Based on the aqueous concentrations in our database, high concentrations of dissolved Pb and Cu occur in solutions with very high concentration of Cl– (an example for Pb is shown in Figure 2). Given that i) chloride is an aggressive ion that generally increases corrosion rates, and ii) rates of reactions typically depend on the concentration of the reactants (in this case the dissolved metal and Fe0), the positive correlation of concentrations of the two metals ions and Cl– supports a hypothesis that the extend of corrosion using dissolved metals ions as oxidants might systematically depend on the brine composition. At the two sites, the most substantial amounts of Cu0 and Pb0 were found on infrastructure in contact with hot water, where reaction kinetics would be generally faster. This effect is supported by the results of corrosion testing of steel in artificial brine, added small amounts of Pb2+. Thus, we propose that the producing well with its hot water is particularly prone to this type of corrosion, whereas the injection well, where cold water is transmitted, should experience the corrosion to smaller extent.
Dissolved carbon dioxide (CO2) is corrosive to iron (steel) because it acts as a weak acid. CO2 corrosion typically results in localised corrosion, often with formation of iron carbonate (FeCO3) as one of the corrosion products (Figure 1f).
More corrosion types are microbial and H2S induced corrosion. These two types are not illustrated in Figure 1, but briefly described below.
Microbial induced corrosion
Sulphate reducing bacteria greatly increase anaerobic corrosions rates if present in substantial amounts because the sulphide production destabilizes the passivating Fe oxide coatings (Hamilton, 1985). Thus, their growth can be deleterious to much infrastructure.
Samples taken in the surface installations at the Sønderborg site contained ZnS and PbS, suggesting that sulphate reducing bacteria generate sulphide.
Consistent with this, sequencing of DNA from water samples indicates the presence of microorganisms carrying the dissimilatory sulphite reductase gene, albeit they exist in small amounts and might not be active. Only one type of gene sequence was identified, which is highly related to that of the halophile bacterium Desulfotomaculum halophilum. This identification of sulphate reducing bacteria in the infrastructure of geothermal operation echoes findings at other sites, including that at Neubrandenburg (Würdemann et al., 2014), where waters also contains high sulphate concentration. The resulting type of corrosion observed would usually be pitting like that depicted in Figure 1b.
H2S induced corrosion
Hydrogen sulphide (H2S) is toxic and corrosive and the presence of hydrogen sulphide gas in the geothermal water causes challenges at several geothermal sites.
FORCE Technology (2019): GEOTERM Project. Best practice guide for operation of geothermal plants to avoid corrosion. Report prepared by FORCE Technology for the GEOTHERM research and development project supported by the Danish Innovation Fund. Link: Best-practice-guide-for-the-operation-of-geothermal-plants-to-avoid-corrosion.pdf (geus.dk)
Hamilton, W. A. (1985): Sulfate-reducing bacteria and anaerobic corrosion. Annu. Rev. Microbiol. v. 39, pp 195-217.
Würdemann, H. et al. (2014): Influence of microbial processes on the operational reliability in a geothermal heat store – Results of long-term monitoring at a full-scale plant and first studies in a bypass system. Energy Procedia v. 59, pp 412-417.