The Synergistic Influence of Trace Impurities and Temperature on the Corrosion Behavior of Tubing in Supercritical CO₂ Environment

Abstract

Carbon dioxide capture, utilization, and storage for enhanced oil recovery (CCUS-EOR) represents an effective strategy for reducing CO₂ emissions while improving oil recovery efficiency. However, harsh environmental conditions during the process can induce a supercritical state in captured CO₂, which may undermine the structural integrity of tubular components through corrosion. This study systematically investigated the corrosion behaviors of two tubing steels (P110 and Super 13Cr) in 20 MPa supercritical CO₂ containing trace H₂S/O₂ impurities at 60–120 °C using weight loss tests and surface analysis. The results demonstrate that in water-unsaturated supercritical CO₂ with ≤500 ppmv H₂S, both steels exhibited low general corrosion rates (P110: 0.03 mm/y; S13Cr: 0.01 mm/y), with incomplete surface films partially covering grinding traces. However, S13Cr suffered pitting corrosion at >500 ppmv H₂S. Oxygen introduction triggered severe general/localized corrosion characterized by cracked, non-protective surface films. Reducing O₂ to 500 ppm yielded thin, continuous protective films, eliminating pitting. Temperature critically influenced S13Cr corrosion: decreasing from 120 °C to 60 °C increased the corrosion rates from 0.0031 mm/y to 0.08 mm/y due to enhanced water precipitation and impurity gas dissolution. These findings establish impurity thresholds to ensure acceptable corrosion performance.

1. Introduction

Carbon neutrality has garnered significant global attention since the 2015 Paris Agreement [1]. Carbon dioxide (CO₂) capture, utilization, and storage (CCUS) are widely recognized as effective emission-reduction strategies crucial for mitigating global warming. Injecting CO₂ into geological formations is a common storage method [2], enabling permanent emission reduction via integration into production processes for carbon recycling. For instance, CO₂ injection enhances fluid properties, reduces interfacial tension, optimizes oil flow, and improves sweep efficiency in reservoirs, thereby promoting oil recovery through miscible displacement (i.e., CCUS for Enhanced Oil Recovery, CCUS-EOR) [2,3,4]. CCUS-EOR delivers dual environmental and economic benefits, driving its widespread application.

Cost minimization in CCUS-EOR necessitates compression of captured CO₂ to a supercritical state (temperature > 31.1 °C, pressure > 7.38 MPa) [5] to avoid biphasic flow and enhance efficiency. However, supercritical CO₂ becomes highly corrosive upon aqueous contact [6,7], threatening infrastructure integrity. Although pure supercritical CO₂ exhibits high water solubility (yielding ~1 μm/y corrosion rates when water is fully dissolved [8]), trace impurities (at the ppm level), such as hydrogen sulfide (H₂S), oxygen (O₂), nitrogen dioxide (NO₂), and sulfur dioxide (SO₂), accelerate corrosion by altering corrosion product solubility. Morland et al. [9] hypothesized that minimal water may form surface films where dissolved impurities generate acids, though their work focused on NO₂ in gaseous CO₂ [9,10]. Extensive studies address O₂/NO₂/SO₂ impacts on carbon steel corrosion in wet supercritical CO₂ [11], particularly pipeline steels [12,13,14], as such impurities originate from capture sources. In contrast, H₂S predominantly exists as a downhole-associated gas in high-sulfur oilfields, primarily influencing injection processes [15]. Crucially, H₂S corrosion research on tubing steels remains scarce for supercritical CO₂ environments. Sun et al. [16] noted that N80 tubing steel suffers severe corrosion (>4 mm/y) in supercritical CO₂, exacerbated by trace H₂S due to degraded film protectiveness. Oxygen introduction during injection further complicates corrosion mechanisms [17]. Although both H₂S and O₂ are present in downhole CCUS-EOR environments, their combined effects on corrosion in supercritical CO₂ have not been systematically studied, particularly regarding how their interaction modifies the corrosion mechanisms of tubing steels.

In addition to water contents and trace impurities, temperature critically influences corrosion by modulating water solubility in supercritical CO₂. Xiang et al. [18] observed that X70 steel corrosion rates in high-pressure (10 MPa) CO₂ systems containing SO₂, O₂, and H₂O peaked at 348 K (75 °C) before declining at 366 K (93 °C), which was attributed to changes in corrosion products. At 366 K, the formation of dense, spherical corrosion products offered enhanced protection to the steel surface. Hua et al. [19] demonstrated that temperature reduction in water-unsaturated systems enhances corrosion, with localized attack being particularly critical. However, the publications above primarily focused on pipeline steels. Other existing studies focus exclusively on pipeline steels, leaving temperature-modulated corrosion of tubing steels in supercritical CO₂ with H₂S/O₂ impurities entirely uninvestigated.

This study aims to investigate the combined effects of trace impurities and temperature on the corrosion behavior of tubing steels in supercritical CO₂ environments containing H₂S and O₂. The corrosion performance of two commonly used tubing steels, P110 and Super 13Cr (S13Cr), was evaluated under supercritical CO₂ conditions with varying H₂S concentrations (10–1400 ppm) and O₂ levels (10–13,000 ppm) across a temperature range of 60 °C to 120 °C. The effects of H₂S, O₂, and temperature on corrosion rates, surface morphology, and corrosion product characteristics were assessed using weight-loss measurements and surface analysis techniques. The findings are expected to inform the development of operational thresholds for impurity levels and temperature conditions that ensure acceptable corrosion resistance in tubing steels.

2. Materials and Methods

2.1. Materials and Sample Preparation

Commercial P110 and S13Cr tubing steels were machined into specimens for corrosion tests, with dimensions of 50 mm × 10 mm × 3 mm. Figure 1 shows their tempered lath martensite microstructure before testing.

Weight loss tests were conducted in a 10 L Cortest-C276 autoclave to evaluate the corrosion rate and corrosion morphology of P110 and S13Cr steel in a supercritical CO₂ with trace impurities, and the schematic diagram is displayed in Figure 2. For each material type, three specimens were analyzed in parallel to ensure reproducibility of the experimental results. The specimens used for the corrosion test were polished with SiC paper of 240, 400, 800, and 1200 grit and then cleaned with deionized water and alcohol, sequentially. The experimental solution consisted of deionized water, which was fully deaerated using 99.999% CO₂ for 3 h prior to being transferred to the autoclave. It is also worth noting that most previous studies have primarily focused on water-saturated supercritical CO₂ systems. In contrast, in practical industrial applications, the EOR projects in the United States permit up to 600 ppm (mole fraction) of water [20]; the DYNAMIS project (EU’s 6th Framework Programme) allows for 500 ppm of water; and the EOR projects in Canada require nearly complete water removal (20 ppm). Therefore, establishing stringent impurity-level specifications is more critical to support EOR projects with significant cost reductions. As such, all the lines connected to the autoclave were thoroughly purged with CO₂ to ensure the complete elimination of O₂ from the system. The required CO₂/H₂S/O₂ mixture was introduced into the autoclave, which was then heated and pressurized to the target conditions. The concentrations of trace impurity gases were precisely controlled using a gas flow controller with a resolution of 0.01 sccm (standard cubic centimeters per minute), as shown in Figure 2, ensuring accurate quantification of individual impurity levels in the gas mixture prior to the experiment. After immersion testing, the corrosion products were removed using a cleaning solution (3.5% methenamine in 10% HCl). The sample was then cleaned, dried, and weighed to calculate the weight loss and corrosion rate using the following equation [14,21].

$$V_{CR}={\displaystyle \frac{8.76\times {10}^4∆W}{S\rho t}}$$

(1)

where V_CR is the corrosion rate, mm/a; $∆W$ is the weight loss, g; S is the exposed surface area of specimen, cm²; $\rho$ is the density specimen, g/cm³; t is the immersion time, h; and $8.76\times {10}^4$ is the unit conversion constant.

As for the pitting rate, it was determined based on the depth of the deepest pit selected from the cleaned surface of the specimen, which was measured using an RH-2000 3D super-depth microscope (Hirox-japan Co., Ltd., Fukuoka, Japan) in accordance with ultra-depth morphological analysis, following the procedure outlined in Equation (2), in alignment with NACE Standard SP0775 [22].

$$V_{PR}={\displaystyle \frac{depth of deepest pit \left(mm\right)\times 365}{exposure time \left(days\right)}}$$

(2)

2.2. Characterization of Corrosion Products

All the corrosion product samples were obtained from high-temperature and high-pressure autoclave immersion experiments. The surface morphologies were analyzed using SEM (EVO MA15, Carl Zeiss, Oberkochen, Germany), and elemental distribution was determined by EDS (X-MaxN, Oxford Instruments Co., Ltd., Abingdon, UK). Furthermore, the phase composition of the corrosion products was identified by X-ray diffraction (XRD) and X-ray photoelectron spectroscopy (XPS).

3. Results and Discussions

3.1. The Influence of H₂S

Figure 3a presents the average corrosion rates, based on mass loss measurements, for P110 and S13Cr steels exposed to supercritical CO₂ (120 °C, 20 MPa) with different H₂S concentrations. Both steels showed similar trends in corrosion rate variations with changing H₂S concentrations. The highest corrosion rates were recorded at 1400 ppm H₂S (0.0434 mm/y for P110 and 0.0313 mm/y for S13Cr). As the H₂S concentration decreased, the rates fluctuated between 0.020 and 0.034 mm/y for P110 and 0.0099 and 0.0116 mm/y for S13Cr. Notably, although P110 consistently exhibited higher average corrosion rates than S13Cr across all the H₂S concentrations, no pitting corrosion was observed in P110, even at 1400 ppm H₂S. In contrast, S13Cr experienced significant pitting corrosion at H₂S concentrations above 1000 ppm, as shown in Figure 3b. The pitting corrosion rates for S13Cr were 2.3846 mm/y at 1400 ppm H₂S and 1.2775 mm/y at 1000 ppm H₂S. These values were derived from the dimensions of the deepest pits, calculated using Equation (2) and illustrated in Figure 3c,d.

As shown in Figure 3a, 1000 ppm H₂S appears to mark a transition point, below which the average corrosion rates stabilized. To further investigate this behavior, the secondary electron (SE) images in Figure 4 show the surface morphologies of the P110 and S13Cr specimens exposed to 1400 ppm and 500 ppm H₂S, respectively. For both steels, the surfaces under both conditions displayed visible grinding and polishing marks, as seen in all the SE images. At 1400 ppm H₂S, small spherical corrosion products were observed on the P110 surface (Figure 4a). These became less prominent at 500 ppm H₂S (Figure 4b). A similar trend was observed for S13Cr (Figure 4c,d). At 1400 ppm H₂S, only a few products formed on the S13Cr surface, which appeared almost clean at 500 ppm, with virtually no corrosion products visible.

To elucidate the chemical composition of these corrosion products, XRD spectral analysis was performed. Distinct FeS peaks were observed for the P110 sample at 1400 ppm H₂S (Figure 5a), whereas only the matrix peaks (i.e., Fe) were detected when the H₂S concentration decreased to 500 ppm (Figure 5b). This observation is consistent with the reduction in the number of corrosion products on the specimen surface, as depicted in Figure 4a,b. Given that certain components of corrosion products cannot be accurately classified using the XRD method due to its resolution limitations, the XPS technique was employed in this study to further identify potential corrosion products formed on the P110 surface at 1400 ppm H₂S. Considering the chemical composition of P110 and the supercritical CO₂-H₂S testing environment, Fe, C, and S were the primary elements detected in the XPS analysis. However, the presence of O was also observed, as shown in Figure 5c, which likely corresponded to the formation of oxides upon exposure of the sample to air, rather than originating from the corrosion process. Figure 5d exhibits the Fe 2p3/2 peak at a binding energy of approximately 711.2 eV, which indicates that Fe existed in an oxidized state, such as Fe²⁺ [23,24]. The S 2p peak centered around 163.2 eV (Figure 5e) was attributed to sulfides [25], consistent with the identification of FeS products from XRD analysis. Additionally, a C 1s peak at a binding energy of approximately 289.0 eV was also detected, corresponding to carbonate species (Figure 5f). Given the presence of oxidized Fe, it is reasonable to infer that FeCO₃, as one of the corrosion products, is also present in supercritical CO₂-H₂S environments. However, its low concentration and weak crystallinity may have rendered it undetectable by XRD. Therefore, based on the XRD and XPS results, it is inferred that at 1400 ppm H₂S, the corrosion process follows the reaction and precipitation mechanisms outlined in Equations (3) and (4):

$$\mathrm{F}\mathrm{e}+ {\mathrm{H}}_2\mathrm{S}\to \mathrm{F}\mathrm{e}\mathrm{S}+2{\mathrm{H}}^{+}$$

(3)

$${\mathrm{F}\mathrm{e}}^{2+}+ {\mathrm{H}\mathrm{S}}^{-}\to \mathrm{F}\mathrm{e}\mathrm{S}+ {\mathrm{H}}^{+}$$

(4)

Notably, H₂S has higher solubility in water than CO₂ [26], and when its concentration exceeds 1000 ppm, the thermodynamic driving force for FeS formation is generally greater than that for FeCO₃ [27]. As a result, H₂S preferentially dissolves in the surface water film and reacts with the steel substrate to form FeS, which precipitates onto the corrosion product layer. When the H₂S concentration was reduced to 500 ppm, these reaction and precipitation mechanisms were likely suppressed. This is supported by SE imaging (Figure 4a,b) and XRD analysis (Figure 5a,b), both of which show minimal FeS deposition on the specimen surfaces.

For S13Cr, although the average corrosion rates leveled off below 500 ppm H₂S, notable pitting corrosion was still observed at 1000 ppm and 1400 ppm (Figure 3). This indicates that, with respect to pitting behavior, the transition threshold for S13Cr should be considered at 1000 ppm H₂S rather than 500 ppm. Therefore, the corrosion products formed on the S13Cr surface at 1000 ppm H₂S were systematically analyzed using XRD and XPS techniques. The XRD analysis revealed only Fe peaks (Figure 6a), corresponding to the metallic matrix. This is consistent with prior surface observations (Figure 4c), where minimal corrosion products were present at 1400 ppm H₂S and even fewer at 1000 ppm. The XPS analysis confirmed the presence of FeS through Fe and S peaks (Figure 6b,c), and detected Cr 2p peaks centered around 577.4 eV (Figure 6d), corresponding to Cr(OH)₃ (Figure 6e) [28,29]. Cr(OH)₃ is notable for its thermodynamic stability and low solubility, which can contribute to protective barrier formation against further corrosion. However, Fierro et al. [30] proposed that Cr enrichment of about 25 at% in the corrosion products formed on 13Cr steels in CO₂ environments, while the Cr concentration reduced to 11 at% of the same material tested in H₂S-containing environments. Accordingly, Ueda concluded that Cr(OH)₃ alone cannot remain stable in CO₂-H₂S environments and emphasized the synergetic effect of Mo sulfides (outer layer) and Cr(OH)₃ (inner layer) in resisting further H₂S corrosion due to their high thermodynamic stability in H₂S-containing environments [27]. In this study, the Mo 3d peaks centered around 231.2 and 229.2 eV were observed (Figure 6f), corresponding to Mo⁴⁺ 3d3/2 and Mo⁴⁺ 3d5/2, respectively [31], typical of MoS₂. Although the Mo 3d peak centered around 234.2 eV corresponds to Mo⁶⁺ 3d3/2, its presence should be attributed to the oxidation of MoS₂ [31]. Despite the coexistence of Cr(OH)₃ and MoS₂ products, severe pitting still occurred in S13Cr specimens under 1000 and 1400 ppm H₂S-containing environments. A similar phenomenon was also observed in a recent study by Wang et al. [26]. In their tests at 180 °C and 27.5 MPa, S13Cr exhibited a low general corrosion rate (0.019 mm/y) but showed deep pitting (~7 μm). These results align with the present study, where S13Cr had a general corrosion rate of 0.013 mm/y and pit depths of ~10 μm at 1000 ppm H₂S (Figure 3). Wang et al. also reported that the corrosion product layer on S13Cr was thin, with FeS, Cr(OH)₃, and other species forming a single, mixed layer rather than a distinct multilayer structure. These findings suggest that neither 1000 ppm nor 1400 ppm H₂S provided enough MoS₂ to form a dense, protective outer layer (Figure 4c). As a result, the Cr(OH)₃ layer alone was insufficiently protective, becoming unstable and vulnerable to H₂S attack, leading to deep pitting. When the H₂S concentrations dropped below 500 ppm, the formation of MoS₂ was suppressed, as was overall H₂S-induced corrosion, and pitting no longer occurred.

Based on the summarized results of general and pitting corrosion rates for both tested materials across varying H₂S concentrations, in conjunction with SE imaging and XRD/XPS analyses, it can be concluded that 500 ppm H₂S represents a critical threshold. Below this concentration, both general corrosion and pitting corrosion remained within acceptable limits. Therefore, 500 ppm H₂S was selected as the representative concentration for subsequent investigations.

3.2. The Coupled Effect of H₂S and O₂

Figure 7a,b present the general and pitting corrosion rates of P110 and S13Cr steels across various O₂ concentrations. A clear trend emerges: corrosion rates increased significantly, and severe pitting occurred (Figure 7c–f) when the O₂ concentrations exceeded 1000 ppm, highlighting the corrosive impact of elevated O₂ levels. In contrast, when the O₂ concentrations were reduced to 100 ppm or below, the general corrosion rates declined to values comparable to those observed in supercritical CO₂ with 500 ppm H₂S (Figure 3), and pitting corrosion was no longer observed. Therefore, 100 ppm O₂ can be considered a critical threshold for both general and localized corrosion. Notably, S13Cr exhibited lower general corrosion rates than P110 across all the conditions, but significantly higher pitting rates at 1000 and 13,000 ppm O₂.

Figure 8 illustrates the variation in corrosion product morphology on the specimen surfaces as the O₂ concentration decreases. For the P110 steel tested in the CO₂-H₂S-O₂ (1000 ppm) environment, a relatively thick corrosion product film was evident, as indicated by the full coverage of grinding and polishing marks (Figure 8a). At higher magnification (Figure 8b), the film exhibited pronounced cracking, likely caused by internal stresses resulting from volume changes during dehydration [32]. When the O₂ concentration was reduced to 100 ppm, corrosion was significantly mitigated (Figure 7a,b), and the film thickness decreased accordingly, exposing the underlying grinding marks (Figure 8c). A similar trend was observed for the S13Cr specimens. At 1000 ppm O₂, a thick, cracked product film was present (Figure 8d,e), while the surface appeared clean and flat with minimal corrosion at 100 ppm O₂ (Figure 8f). Additionally, spherical corrosion products (~10 μm in diameter) were observed on the cracked film surfaces. To further identify the chemical composition of the corrosion products formed under 1000 ppm O₂, XPS analysis was conducted on both P110 and S13Cr specimens.

For P110, the XPS spectra primarily detected Fe 2p, C 1s, O 1s, and S 2p peaks (Figure 9a). The C 1s peak at a binding energy of approximately 289.0 eV (Figure 9c) was attributed to FeCO₃. The S 2p peak centered at approximately 163.2 eV (Figure 9d) was assigned to FeS. Additionally, the Fe 2p3/2 peak around 711.6 eV and the O 1s peak at approximately 529.7 eV indicated the presence of Fe₂O₃. In summary, three corrosion products were identified: FeCO₃, FeS, and Fe₂O₃. For S13Cr, the XPS-detected elements are shown in Figure 10a. Except for the presence of FeS (S 2p peak around 163.2 eV, Figure 9f) and FeCO₃ (C 1s peak around 289.0 eV, Figure 9e), the Cr 2p peak around 577.4 eV (Figure 9d) and Mo 3d peaks around 231.2 and 229.2 eV (Figure 9c) revealed the presence of Cr(OH)₃ and MoS₂. According to the recent publications of Naixin et al. [33] and Wang et al. [26], the inner layer was primarily composed of Cr(OH)₃. During the dehydration process, the Cr-enriched product film transformed Cr(OH)₃ into Cr₂O₃, which led to the cracking of the product film, as shown in Figure 8d,e. The outer layer was likely composed of iron oxides, consistent with the spherical products observed on the surface of the cracked film (Figure 8d). In a supercritical CO₂-H₂O environment, FeCO₃, a commonly observed corrosion product has been extensively studied and is considered to form via either a one-step or two-step reaction mechanism (Equations (5)–(7)) [34]. Meanwhile, Barker et al. reported that impurities in a supercritical CO₂ environment influence the deposition kinetics, morphological characteristics, and protective performance of FeCO₃ [35].

$${\mathrm{F}\mathrm{e}}^{2+}+\mathrm{C}{\mathrm{O}}_3^{2-}\to \mathrm{F}\mathrm{e}\mathrm{C}{\mathrm{O}}_3$$

(5)

$${\mathrm{F}\mathrm{e}}^{2+}+2\mathrm{H}\mathrm{C}{\mathrm{O}}_3^{-}\to \mathrm{F}\mathrm{e}(\mathrm{H}\mathrm{C}{\mathrm{O}}_3{)}_2$$

(6)

$${\mathrm{F}\mathrm{e}}^{2+}+2\mathrm{H}\mathrm{C}{\mathrm{O}}_3^{-}\to \mathrm{F}\mathrm{e}(\mathrm{H}\mathrm{C}{\mathrm{O}}_3{)}_2$$

(7)

The introduction of oxygen impurities promotes additional cathodic reactions (Equation (8)) and inhibits FeCO₃ formation by oxidizing ferrous ions to ferric ions, thereby facilitating the formation of non-protective Fe₂O₃ (Equations (9)–(11)).

$${\mathrm{O}}_2+4{\mathrm{H}}^{+}+4{\mathrm{e}}^{-}\to 2{\mathrm{H}}_2\mathrm{O}$$

(8)

$${\mathrm{F}\mathrm{e}}^{2+}+{2\mathrm{O}\mathrm{H}}^{-}\to \mathrm{F}\mathrm{e}(\mathrm{O}{\mathrm{H})}_2$$

(9)

$${4\mathrm{F}\mathrm{e}\left(\mathrm{O}\mathrm{H}\right)}_2+{\mathrm{O}}_2+2{\mathrm{H}}_2\mathrm{O}\to 4\mathrm{F}\mathrm{e}{\left(\mathrm{O}\mathrm{H}\right)}_3$$

(10)

$$2\mathrm{F}\mathrm{e}{\left(\mathrm{O}\mathrm{H}\right)}_3\to {\mathrm{F}\mathrm{e}}_2{\mathrm{O}}_3+{3\mathrm{H}}_2\mathrm{O}$$

(11)

Generally, at low H₂S concentrations, the uniform corrosion rate can be significantly reduced due to the rapid formation of a relatively thin FeS corrosion product film on the P110 surface, as described by the following formulation (Equation (12)) [36]. However, in the case of S13Cr steel, the addition of Mo and Cr leads to substantial changes in the composition of the product film. Ueda et al. reported that MoS₂ demonstrates considerably higher stability compared to FeS, which favors its preferential formation during the film formation process [27]. Moreover, according to the spontaneous passivation mechanism, MoS₂ is initially formed on the steel surface, followed by the subsequent growth of Cr oxide beneath these sulfides, which acts as a protective barrier against further H₂S corrosion. Additionally, Mo promotes the formation of a stable Cr oxide passive film in the inner layer.

$${\mathrm{F}\mathrm{e}}_{\left(\mathrm{s}\right)}+{\mathrm{H}}_2\mathrm{S}\to \mathrm{F}\mathrm{e}\mathrm{S}+{\mathrm{H}}_2$$

(12)

Based on the observed trends in general and the pitting corrosion rates for both tested materials under CO₂-H₂S-O₂ environments (Figure 7a,b), as well as the variations in corrosion product composition revealed through the SEM observations and XPS analyses, it can be concluded that the combination of 500 ppm H₂S and 100 ppm O₂ serves as a critical threshold condition. Below this threshold, both general and localized corrosion remain within acceptable limits. Therefore, in the following section, the influence of temperature on the corrosion behavior of P110 and S13Cr steels will be investigated under this defined threshold environment.

3.3. The Influence of Temperature

To evaluate the influence of temperature on the corrosion behavior of P110 and S13Cr steels in a supercritical CO₂-H₂S (500 ppm)-O₂ (100 ppm) environment, corrosion tests were conducted at three temperatures: 60 °C, 90 °C, and 120 °C. The average corrosion rates are summarized in Figure 11. A decrease in temperature led to an increase in corrosion severity for both materials. This effect was particularly pronounced for S13Cr, whose average corrosion rate rose sharply from 0.0031 mm/y at 120 °C to 0.08 mm/y at 60 °C. The surface morphologies of the corrosion products are shown in Figure 12. On the P110 specimen (Figure 12a), a protective film was observed to reform, covering the grinding marks and appearing more intact than the film formed under 1000 ppm O₂ at 120 °C (Figure 8a,b). However, shallow cracks were still present, discernible only at high magnification, as shown in the inset of Figure 12a. Similar shallow cracking was also observed on the corrosion product film of the S13Cr specimen (Figure 12b). A plausible explanation for the observed increase in the average corrosion rates of both tested materials is that the reduction in testing temperature affected the solubility of H₂O in supercritical CO₂. Spycher et al. [37] and Choi et al. [5] developed a thermodynamic model to describe the mutual solubility of CO₂ and H₂O and demonstrated that the solubility of H₂O in supercritical CO₂ decreases with decreasing temperature. Consequently, it can be inferred that a reduction in operating temperature may induce aqueous phase precipitation, resulting in the formation of a free aqueous layer above the specimen surface. Subsequently, both supercritical CO₂ and impurity gases (H₂S and O₂) dissolve into the aqueous phase, thereby promoting corrosion of the tested tubing steels.

4. Conclusions

In this study, the effects of impurities and temperature on the corrosion behavior of two commonly used tubing steels, P110 and S13Cr, were systematically investigated through corrosion tests and SEM observations. The chemical compositions of the corrosion products were further characterized using XRD and XPS analyses. Based on the experimental findings, the following conclusions can be drawn:

• In water-unsaturated (500 ppmv H₂O) supercritical CO₂-H₂S environments, both P110 and S13Cr exhibited relatively low and stable general corrosion rates (approximately 0.03 mm/y and 0.01 mm/y, respectively). The corrosion product films were poorly developed and only partially obscured the grinding marks. However, S13Cr showed susceptibility to pitting corrosion when the H₂S concentration exceeded 500 ppm. Therefore, a supercritical CO₂ environment containing 500 ppm H₂S was identified as the threshold condition, below which the corrosion behavior of both materials remained within acceptable limits.
• Upon the introduction of O₂, both steels exhibited markedly accelerated corrosion, characterized by elevated general corrosion rates and the presence of deep pits. Although relatively thick product films formed on the specimen surfaces, these films displayed severe cracking, compromising their protective ability and exacerbating localized corrosion. When the O₂ concentration was reduced to 100 ppm, the corrosion product films became thinner yet continuous, correlating with stable corrosion rates and the complete suppression of pitting in both materials.
• The corrosion behavior of S13Cr was found to be highly sensitive to temperature. As the test temperature decreased from 120 °C to 60 °C, the general corrosion rate of S13Cr increased substantially from 0.0031 mm/y to 0.08 mm/y. This trend is attributed to the influence of temperature on water solubility in supercritical CO₂. At lower temperatures, water is more likely to precipitate on the steel surface, enhancing the dissolution of H₂S into the aqueous phase and thus intensifying its corrosive effect.