Influence of Chromium Content in Alloys on Corrosion in Saline Water Saturated with Supercritical CO₂

Abstract

Amid growing global efforts toward carbon capture, utilization, and storage (CCUS), this study investigates the influence of chromium (Cr) content in candidate construction alloys on their corrosion modes and kinetics in supercritical CO₂ (s-CO₂)-saturated saline water at 8 MPa and 50 °C. The results indicate that alloys with a Cr concentration of over approximately 9 wt.%, including P91, 316L, and Alloy 800, exhibit a satisfactory corrosion performance in this environment. During exposure to s-CO₂-saturated saline water, a non-protective FeCO₃ layer forms on all tested alloys. For alloys containing more than 2 wt.% Cr, an inner Cr-enriched layer concurrently grows and acts as a barrier to resist environmental attack. The integrity of the inner and outer corrosion layers becomes more compact and uniform on alloys with at least 9 wt.% Cr. Pitting is unlikely to occur on candidate alloys used for s-CO₂ storage or enhanced oil recovery.

1. Introduction

The adoption of economy-wide decarbonization targets for 2050 has stimulated intensive efforts in the development and deployment of carbon capture, utilization, and storage (CCUS) technology for various industrial applications [1,2]. Among them, a prominent example is the integration of CO₂-enhanced oil recovery (EOR) with CCUS, which involves capturing CO₂ from large emitters, transporting it as supercritical CO₂ (s-CO₂), and utilizing it in direct EOR operations [3,4,5]. In a CCUS chain, long-term, safe CO₂ transportation and EOR application are two crucial concerns to avoid catastrophic risks and significant capital loss due to the existing corrosion and stress corrosion cracking challenges for primary constructional materials. For effective CO₂ pipeline transport, the transported CO₂ streams with corrosive impurities (such as H₂O, O₂, SO₂, NO_x, etc.) are compressed above their critical point (7.4 MPa, 31 °C) to maintain a single-phase flow for optimum efficiency [6,7]. The density and viscosity of s-CO₂ at these conditions highlight its suitability for pipeline transmission. For instance, the density and viscosity of s-CO₂ at 50 °C and 10 MPa are 395 kg/m³ and 0.03 cP, respectively [8,9]. The general process of s-CO₂ transportation, storage, and utilization systems is to collect CO₂ from large industrial emitters, transport it under s-CO₂, and re-use it for specific field EOR operations [10,11].

Based on the dominant reactions, corrosion in these s-CO₂ environments are divided into two categories: the s-CO₂-rich phase in pipeline and upper-level EOR well environments, and the s-CO₂-saturated aqueous phase (i.e., H₂O-rich phase) in localized pipeline condensation areas and lower-level EOR and storage systems [12]. In the past years, considerable studies have been conducted to clarify the effects of aggressive impurities on corrosion in the s-CO₂-rich phase [2,13,14,15]. Corrosion damage is expected to be more severe in the s-CO₂-saturated aqueous phase compared to the s-CO₂-rich phase, and only a few investigations have been performed in this area [16]. Moreover, most studies on the s-CO₂-saturated H₂O-rich phase mainly focused on low-alloyed steels, as shown in Appendix A,

Table A1. Summary of studies for corrosion in the S-CO₂-saturated H₂O-rich phase.

Material	T /°C	P /MPa	NaCl	Time/h	Rate /mm/y	Refs.
X70	80	9.5	3.5 wt.%	0.5	119	Gao 2015 [16]
X70	80	9.5		7	49.5
X70	80	9.5		12	57.4
X70	80	9.5		24	37
X70	80	9.5		96	7
X70	80	9.5		168	5
X70	80	9.5		384	1.5
X65	80	10	3.5 wt.%	240	8.46	Gao 2016 [27]
X65 1	80	10		240	15.48
P110 2	80	10	3.5 wt.%	240	10	Gao 2016 [21]
3Cr 2	80	10		240	2.2
316L 2	80	10		240	0.018
P110	80	9.5	3.5 wt.%	0.5	84.2	Gao 2018 [20]
P110	80	9.5		12	48
P110	80	9.5		50	14.4
P110	80	9.5		96	13.0
P110	80	9.5		168	6.12
P110	80	9.5		384	1.8
X65	60	10	35,249 ppm Cl−	6	20	Hua 2018 [30]
X65	60	10		24	14
X65	60	10		48	7.5
X65	60	10		96	3.8
X65	60	10	1.0 wt.%	6	10.8	Hua 2020 [62]
1Cr	60	10		6	10.5
3Cr	60	10		6	10.1
5Cr	60	10		6	6.8
X65	60	10		192	1.5
1Cr	60	10		192	1.9
3Cr	60	10		192	2.0
5Cr	60	10		192	2.1
N80	80	8	3.0 wt.%	1	45	Zhang 2021 [63]
N80	80	8		12	29
N80	80	8		24	20
N80	80	8		48	14
N80	80	8		96	8
N80	80	8		168	6
X65 3	50	8	3.5 wt.%	1.5	21	Sun 2021 [12]
X65 3	50	8		24	4
X65 3	50	8		72	2.4
X65	50	8	DI	96	5.0	Li 2024 [26]
X65	50	8	3.5 wt.%	96	1.5
X65	75	8		96	1.0
X65	100	8		96	0.9

Note: ¹, ², ³ are gas conditions. ¹, with 50 ppmv H₂S; ², with saturated H₂S; ³, with 10,000 ppmv of O₂, SO₂, and H₂S, respectively.

. Furthermore, except for water and s-CO₂, the constructional alloys used to inject CO₂ into saline aquifer reservoirs encounter saline water (SW) with high concentrations of chloride, possibly experiencing both active general and localized corrosion (pitting) attacks [8,9,17,18]. Unfortunately, previous studies on carbon steels (CSs) and low Cr steels indicated that they are unsuitable for engineering applications due to their high corrosion rates (CRs) in s-CO₂-saturated saline water [11,19].

There are several key factors controlling corrosion in the s-CO₂-saturated H₂O-rich phase: alloy chemistry and structures, temperature, pressure, environmental aggressive agents, etc. Firstly, previous studies imply that the suitable materials of construction (MOC) should have a certain amount of alloying elements (such as Cr and Mo) to effectively resist environmental attacks. For instance, carbon steels exhibit high CRs (7 and 13 mm/y, respectively) in the s-CO₂-saturated 3.5 wt.% NaCl solution at 80 °C and 9.5 MPa [16,20], while austenitic stainless steels have excellent corrosion resistance (less than 0.02 mm/y) under similar conditions [21,22]. Secondly, the influence of temperature on corrosion in the s-CO₂-saturated H₂O-rich phase is different from that observed in condensed acidic solutions [23]. The CRs of X65 are higher at lower temperatures (50–80 °C) than at higher temperatures (110–130 °C) after 168 h of exposure [24,25]. Similarly, Li et al. reported that the CR of X65 decreased from 1.47 mm/y at 50 °C to 0.93 mm/y at 100 °C in 8 MPa s-CO₂-saturated H₂O [26]. Thus, the corrosion in the temperature range of 50–80 °C needs to be carefully addressed to avoid potential underestimations. Thirdly, increasing s-CO₂ pressure leads to a decrease in the pH value of the s-CO₂-saturated H₂O-rich phase, consequently introducing a more acidic condition [26]. Interestingly, at the pressure range from 7.4 to 10 MPa, previous studies indicated that the presence of impurities (such as H₂S) had a more noticeable impact on the CRs of carbon steels compared to pressure change [27]. However, above 10 MPa, the pressure influence cannot be ignored [10,28,29]. Finally, the corrosion performance of alloys in the s-CO₂-saturated H₂O-rich phase is affected by the presence of different cations and anions. For example, accelerated pitting kinetics might be attributed to the co-precipitation of Ca²⁺ and Mg²⁺ ions with FeCO₃ [30]. Compared to Mg²⁺, Ca²⁺ demonstrated a much greater propensity to co-precipitate with FeCO₃, thus resulting in more severe pit propagation. Moreover, Cl⁻ is a notorious agent to trigger pitting and even the stress corrosion cracking of steels in aqueous solutions [26].

In saline water environments, limited information is available to determine the appropriate MOCs for cost-competition construction and long-term safe operation. Herein, this study aims to investigate the effect of the Cr content in alloys on corrosion modes and rates by exposing commercial candidate alloys with various Cr contents in s-CO₂-saturated saline water (SW) to fill the knowledge gaps. High-pressure autoclave testing, weight loss measurement, and microscopic characterization of as-formed corrosion products on the alloys were conducted to reveal the corrosion mechanisms of these alloys in the s-CO₂-saturated H₂O-rich phase, and partially support the industrial applications of CCUS technology.

2. Experimental Procedure

2.1. Testing Materials and Solution Preparation

Various candidate commercial alloys with a Cr content ranging from 0.28 to 19.2 wt.%, including X80 (European equivalent grade #L555MB), P91 (UNS #K91560), and 316L (UNS #S31603) steels made by EVRAZ North America (Regina, SK, Canada), 2Cr (GR2) and 5Cr (GR5) purchased from ArcelorMittal (Contrecoeur, QC, Canada), and Alloy 800 (UNS #N08800) produced by CAMBRIDGE (Cambridge, ON, Canada) respectively, were selected in this study. Their chemical compositions, which were measured using a photoelectric direct-reading spectrometer (Bruker, Billerica, MA, USA) and carbon sulfur analyzer (LECO Corporation, St. Joseph, MI, USA), are listed in

Table 1. Bulk chemical compositions (wt.%) of alloys tested in this study.

Alloys	Chemical Composition (wt.%)
	Cr	Ni	Fe	Si	Mn	C	Al	Others
X80	0.28	-	Bal.	0.17	1.83	0.04	-	0.3Mo
2Cr	2.2	0.1	Bal.	0.21	0.4	0.1	0.02	1.0Mo
5Cr	4.6	0.1	Bal.	0.33	0.3	0.1	0.05	0.5Mo
P91	8.9	0.1	Bal.	0.3	0.4	0.1	0.03	0.9Mo
316L	17.3	8.3	Bal.	0.4	1.0	0.02	-	0.2Mo
Alloy 800	19.2	31.2	Bal.	0.8	0.8	0.06	0.2	-

. All specimens were prepared via the process of grinding the exposed surfaces sequentially with a series of SiC abrasive papers up to 600 grit, cleaning with deionized (DI) water, dehydrating them in alcohol, and finally drying in compressed air. A 3.5 wt.% NaCl solution was utilized as the exposed H₂O-rich phase, which was continuously purged with CO₂ for 2 h before the autoclave experiments to eliminate dissolved oxygen.

2.2. Autoclave Corrosion Experiments and Post-Experiment Characterization

The setup for the autoclave corrosion experiments is schematically shown in Figure 1, which comprises a s-CO₂ supply system, a s-CO₂ pump, an autoclave (316 stainless steel, 1 L capacity), a real-time temperature controller, and a gas disposal line. All experiments were controlled at a total pressure of 8 MPa and 50 °C for 96 h. As displayed in Figure 1, the specimens were immersed in the H₂O-rich phase (i.e., s-CO₂-saturated 3.5 wt.% NaCl solution) to simulate the CO₂ injection environment.

To avoid potential contamination, each type of alloy was evaluated separately. Specifically, four specimens were experimented/tested for each alloy, of which the first three samples were used for corrosion rate calculations, and the remaining one was for the post-experiment characterization of the formed corrosion products. Before an experiment, the initial weight of specimens was measured using an analytical balance with a precision of 0.00001 g. The freshly prepared specimens were accommodated on Teflon holders (Figure 1), and 250 mL of s de-aerated NaCl solution was added to the autoclave. The autoclave was then immediately sealed, and residual air inside was removed by purging with CO₂ for 2 h. Following that, the autoclave was heated up to the target temperature of 50 °C. Liquid CO₂ was subsequently injected into the autoclave to reach the desired total pressure of 8 MPa through a s-CO₂ pump. Note that s-CO₂ was introduced only at the beginning of the experiment in this study and was not renewed during the experiment. All experiments were performed under static conditions. After 96 h, the autoclave was turned off, and the corroded specimens were removed for subsequent weight loss measurements and microscopic examinations. Note that 96 h duration was selected based on our comprehensive literature review, which indicates that an exposure time of at least 96 h for immersion experiments is needed to reach a steady state [10]. The corrosion products formed on the first 3 samples were removed using a pickling solution (composed of 3.5 g hexamethylene tetramine, 500 mL HCl and deionized water to make a 1 L solution) at room temperature, following the cleaning procedures outlined in ASTM Standard G1-03 [31]. The weight of specimens after removing the corrosion products was measured again to determine the weight loss. All weight loss measurements were repeated three times to ensure the result’s accuracy. Assuming a uniform thickness loss across each sample surface, the corrosion rate (CR, mm/y) was calculated using Equations (1) and (2):

$$∆W=W_o-W_f$$

(1)

where $W_f$ is the final weight; $W_o$ is the initial weight in g. According to ASTM Standard G31-21 [32], the corrosion rate, CR, in mm/y, was calculated as:

$$CR=(K\times ∆W)/\left(A\times t\times \mathsf{\rho }\right)$$

(2)

where $t$ is the time of exposure in hours, $A$ is the exposed surface area in cm², $W$ is the mass loss in g, $\mathsf{\rho }$ is the density in g/cm³, and $K$ is a constant 8.76 × 10⁴ for mm/y.

To identify the surface corrosion product grown on tested samples without cleaning, an X-ray diffractometer (XRD, Ultima IV, Rigaku, Tokyo, Japan) with a Cu X-ray source and a scanning speed of 1°/step and a scanning electron microscope (SEM, Tescan Vega3, Tescan, Kohoutovice, Czech Republic) equipped with an energy-dispersive X-ray spectroscopy (EDS) detector were employed. After the surface characterization, the corroded steels with a Cr content ≤ 5 wt.% were embedded in epoxy and cut to expose the cross-sectional areas of corrosion layers for SEM examinations. As shown below, the corrosion rates of P91 and SS316 fall within the transition range. The cross-sections of corrosion layers on the two steels were prepared by using the focused ion beam (FIB) milling technique and then analyzed using transmission electron microscopy (TEM, JEOL JEM-ARM200cF S/TEM, JEOL, Akishima, Tokyo, Japan). Note that before FIB operation, a very thin Pt layer was coated on the tested samples to protect the formed corrosion layers on the steels and to make the layer conductive. The microscope was equipped with a cold field-emission gun (cFEG) and a probe spherical aberration corrector. To ensure the reliability of the results, the EDS analyses obtained through SEM and TEM were repeated five times.

3. Results

3.1. Average Corrosion Rates of Alloys

Figure 2 presents the corrosion rates of various alloys in the s-CO₂-saturated SW at 50 °C and 8 MPa with a 96 h immersion time based on weight loss measurements. As shown, X80 steel suffered the highest corrosion rate of 3.14 mm/y, while 2Cr and 5Cr steels had corrosion rates of 1.75 and 0.91 mm/y, respectively. As reported, X65 shows a corrosion rate of 1.5 mm/y in the same experimental condition after 96 h [26], and X70 shows a corrosion rate of 7 mm/y at 70 °C and 9.5 MPa after 96 h [16]. For these low Cr steels, the introduction of Mo (≤1%) is unlikely to remarkably improve their corrosion resistance in the s-CO₂-saturated NaCl solution. Further increase in Cr concentration in Fe-based steels led to a noticeable drop in the corrosion rate. The average corrosion rate of P91 steel specimens, which have a higher Cr content of 8.9 wt.%. and contents of 0.1 wt.% Ni and 0.9 wt.% Mo, was found to be around 0.17 mm/y, indicating the vital role of Cr content in boosting corrosion resistance in the s-CO₂ environment. For the austenitic stainless steel (SS316L) and Ni-based alloy (Alloy 800) with a Cr content above 17 wt.%, their corrosion rates were less than 0.02 mm/y and consistently decreased with increasing Cr content, further confirming the critical role of Cr content in the alloys on corrosion resistance in the environment. Similarly, Gao et al. found that P110, 3Cr, and 316L have corrosion rates of 10, 2.2, and 0.018 mm/y in the 3.5% NaCl solution at 80 °C and 10 MPa after 240 h [21], which indicates the effect of Cr. As the two alloys (SS316L and Alloy 800) contain a certain amount of Ni, FIB and SEM/TEM characterizations were conducted on tested SS316L samples to clarify whether the presence of Ni in alloys could also contribute to corrosion resistance improvement. Furthermore, as illustrated in Figure 2, based on the standards [33], corrosion rates less than 0.02 mm/y are outlined as outstanding, 0.02–0.1 mm/y as excellent, 0.1–0.5 mm/y as good, 0.5–1 mm/y as fair, 1–5 mm/y as poor, and greater than 5 mm/y as unacceptable. Therefore, the above results indicate the MOC materials suitable for s-CO₂ permanent storage or EOR need to have at least 9 wt.% Cr to meet the long-term application requirements.

3.2. Corrosion Product Characterization

Figure 3 shows the photographic images and XRD spectra of all the alloys after being exposed to s-CO₂-saturated SW at 50 °C and 8 MPa for 96 h. Compared with freshly prepared specimens with a metallic color, the dark color presented on X80, 2Cr, and 5Cr steels suggested the formation of thick corrosion layers on their surface. Unlike them, P91 exhibited a lighter dark color on its surface, implying the presence of a thinner corrosion layer on its surface. For the alloys with a higher Cr content, their surface color was brighter compared to P91. To identify the phase compositions of the formed corrosion products, XRD analyses were conducted, and the collated XRD spectra are shown in Figure 3. The results indicate that corrosion products formed on these alloys are mainly composed of iron carbonate (FeCO₃, ICDD PDF No. 29-0696), along with their respective substrates.

Based on the above results, the corroded surfaces of the samples in the s-CO₂-saturated SW were subjected to further SEM/EDS examinations on elemental distribution mapping. The backscattered electron (BSE) images obtained are presented in Figure 4. X80 formed a dense corrosion layer, while corrosion layer spalling and/or microcrack formation occurred on 2Cr and 5Cr surfaces. With an increase in the Cr content, the surface morphologies of corroded P91 and 316L were different from the low-Cr steels as relatively uniform and small crystalline products were presented on them. For Alloy 800 with a Cr content close to 20%, a very thin and compact layer was grown on its surface after exposure. To analyze the elemental variation and distinguish the corresponding phases as shown in the XRD spectra (see Figure 3), point-scanning EDS analyses were carried out at specific sites marked on the SEM images, and the results are listed in

Table 2. Average EDS elemental composition (at. %; metal cation basis) of the corresponding sites marked in Figure 4.

Alloys	Site #		Average Composition (at. %)				Products
		O	Fe	Cr	Ni	C
X80	1	59.6 ± 2.6	17.6 ± 4.4	-	-	22.4 ± 1.8	FeCO3
2Cr	2	57.9 ± 3.5	19.5 ± 4.9	-	-	22.0 ± 2.2	FeCO3
	3	60.9 ± 1.9	14.5 ± 2.9	2.7 ± 0.4	-	21.2 ± 0.6	-
5Cr	4	61.5 ± 1.0	12.1 ± 1.7	-	-	25.6 ± 1.7	FeCO3
	5	60.9 ± 1.0	9.5 ± 0.7	10.8 ± 0.7	-	18.8 ± 0.5	-
P91	6	58.7 ± 1.3	19.3 ± 3.0	-	-	22.0 ± 1.7	FeCO3
316L	7	63.5 ± 2.4	19.0 ± 4.8	-	-	17.5 ± 2.4	FeCO3
Alloy 800	8	61.6 ± 0.8	18.6 ± 3.7	-	-	19.8 ± 2.9	FeCO3

(Sites #1 to #8). To improve the accuracy of the results, the element composition of each area was collected five times. Figure 4 presents the EDS spectra for sites #4 and #5, illustrating the process of characterizing elemental composition. The uniform layer (site #1) on the surface of X80 is composed of Fe, C, and O, consistent with the XRD results. Similar to X80, the outer corrosion layer on 2Cr (site #2) and 5Cr (site #4) also consisted of Fe, C, and O. But their inner corrosion layer (sites #3 and 5) had Cr, and the concentration of the Cr cation in the layer likely increased with an increase in the Cr content in the bulk steels. A similar two-layer structure was also found on a 3Cr steel after being exposed to a NaCl solution saturated with 0.8 MPa CO₂ for 32 h, where the outer layer was FeCO₃ and the inner layer was a mixture of amorphous FeCO₃ and Cr(OH)₃ [34]. These results suggest that the corrosion process of low-Cr (≤5 wt.%) steels in CO₂ or a high-pressure s-CO₂-saturated NaCl solution might be dominated by the same mechanism. The results in Figure 3 and Figure 4 also indicate that the inner Fe/Cr layer might be much thinner compared to the outer FeCO₃ layer on the steels. For the alloys (P91, 316L, and Alloy 800) with a higher Cr content, the EDS results for sites #6, 7, and 8 show the surface corrosion layers on these alloys are mainly composed of Fe, C, and O, consistent with the XRD results.

As corrosion layer spalling and microcracking occurred on 2Cr and 5 Cr steels, EDS mapping analyses were also conducted on the two steels, and the results are presented in Figure 5. The O element is likely similar for both the outer and inner layers, but Cr occurs in the inner layer. Upon comparing the enlarged SEM images in Figure 5, it was observed that the two steels had a relatively dense Cr-included inner layer, but the outer FeCO₃ layer grown on 5Cr steel was somewhat looser than that on 2Cr steel.

As the above characterization results are insufficient to reveal the role of Cr content in the alloys on corrosion resistance improvement, cross-sectional examinations of corroded samples were performed to reveal the potential multiple-layer depth information. As stated above, the cross-sectional samples of X80, 2Cr, and 5Cr steels, after being exposed to the s-CO₂-saturated SW for 96 h, were prepared using cold mounting with epoxy and the saw-cutting method. Figure 6 illustrates the SEM cross-sectional results of the three steels. A single, dense, FeCO₃ layer with an average thickness of approximately 70 μm was grown on X80 steel. On the other hand, the cross-sectional morphology of 2Cr indicated that two distinct layers were presented with a total thickness of about/approximately 25 μm. The average thickness of the inner Cr-included layer was around 10 μm. In the case of 5Cr, the poor adhesion of the outer-layer FeCO₃ on the surface resulted in only the inner layer being left, and visible with a thickness of ~5 μm. These cross-sectional results are in agreement with the surface analyses shown in Figure 4 and Figure 5. Moreover, these observations as well as the obtained corrosion rates (see Figure 2) suggest that the three low-Cr alloys would experience active corrosion in the s-CO₂-containing environment, and their long-term corrosion rates are likely to remain at high levels, close to those obtained from 96 h of testing.

Based on the SEM examination on cross-sections of alloys (P91, SS316L, and Alloy 800) with a Cr content higher than 9%, only a single corrosion product of FeCO₃ was detected, different from that formed on 2Cr and 5Cr alloys. Consistent with the corrosion rate results, there is a threshold Cr content (~9%) above which the alloys have sufficient resistance to s-CO₂ environmental attacks. Thus, the tested samples of P91 (close to the threshold) and SS316L (above the threshold) were selected for FIB/TEM characterization to further clarify the role of alloying elements, particularly Cr, on corrosion in the s-CO₂-saturated saline water environment. Figure 7 shows the SEM images taken during the FIB operation to produce cross-sectional samples for TEM examinations. A uniform and compact double-layer structure formed on the two steels after exposure. Compared with P91, the outer corrosion layer (FeCO₃) formed on SS 316L was thicker, but the inner Cr-included layer became thinner. Note that microvoids and cracks were artificially introduced and observed on the sectional areas during the FIB operation, which could be attributed to volume expansion-induced stress accumulation within formed corrosion layers and the mismatch of mechanical properties among the two layers and substrate.

Figure 8 shows the cross-sectional TEM analysis results for the P91 sample. The TEM DF image (Figure 8a) of P91 confirms a thin layer is present between the outer FeCO₃ layer and substrate, and the corresponding EDS mapping results (see Figure 8 and

Table 3. Average EDS elemental composition (at. %) from the corresponding regions numbered in Figure 8 and Figure 9.

Region #	Chemical Composition (at. %)
	O	Fe	Cr	C
1	22	18	-	60
2	21	13	8	58
3	22	23	-	55
4	20	15	10	55

) show the existence of Cr cations within the layer. As the thin Cr-included layer was partially peeled off during FIB operation, the Cr content within the layer could be higher than that listed in Table 3. Moreover, the selected-area electron diffraction (SAED) results reveal that the outer layer (Figure 8c) is crystalline FeCO₃ and the Cr-included inner layer is an amorphous structure (Figure 8d). Furthermore, only a trace amount of Ni cation was observed along the formed corrosion layer, suggesting that the alloying element, Ni, in the steel was unlikely involved in the corrosion process in the testing environment. In addition, Mo cations were found to be uniformly distributed within the outer FeCO₃ and inner Cr-rich layers instead of local accumulation within the Cr-rich layer, further confirming that the addition of a small amount of Mo (≤1%) to Fe-based steels could not lead to a remarkable improvement of their corrosion resistance in such a specific s-CO₂-related environment. As former studies reported that the presence of Mo may suppress the chemical dissolution of formed Cr-rich oxides and enhance their resistance to pitting [35,36], further investigation is needed to explore the performance of Fe-based steels with a higher Mo content and clarify the role of Mo in corrosion in s-CO₂ environments.

The dark-phase (DF) image reveals the outer FeCO₃ layer on the surface, consistent with the above surface characterization results. Upon enlarging the interface between the FeCO₃ layer and substrate, a thin secondary layer is visible at the interface (Figure 9b), and corresponding EDS mapping analyses show that this is a Cr-included layer, which can act as a major barrier to resist an attack from s-CO₂-saturated SW. Localized chemical compositions of regions 3 and 4 were collected and are provided in Table 3, in which region 3 consists of Fe, C, and O, and region 4 contains certain amounts of Cr cations. In general, the TEM results are consistent with the corrosion rates identified on P91, SS316L, and Alloy 800, indicating that they are suitable for next-phase long-term corrosion kinetic assessments to determine whether their corrosion rates follow linear, parabolic, or exponential laws with time.

4. Discussion

4.1. Thermodynamic Calculation of the Tested Alloy

In the s-CO₂-saturated SW solution, the dissolution of CO₂ into 3.5% NaCl through reactions (3)–(5) leads to the formation of an acidic environment containing H⁺, HCO₃⁻, and CO₃²⁻.

$${\mathrm{C}\mathrm{O}}_2+{\mathrm{H}}_2\mathrm{O}\leftrightarrow {\mathrm{H}}_2{\mathrm{C}\mathrm{O}}_3$$

(3)

$${\mathrm{H}}_2{\mathrm{C}\mathrm{O}}_3\leftrightarrow {\mathrm{H}}^{+}+\mathrm{H}\mathrm{C}{\mathrm{O}}_3^{-}$$

(4)

$$\mathrm{H}\mathrm{C}{\mathrm{O}}_3^{-}\leftrightarrow {\mathrm{H}}^{+}+\mathrm{C}{\mathrm{O}}_3^{2-}$$

(5)

According to the previous studies [28,37], increasing s-CO₂ pressure leads to higher acidity levels, and the pH value of the SW solution in this study is around 3.1. In the s-CO₂-saturated NaCl solution, the corrosion process of an alloy is dominated by electrochemical reactions as follows [28].

Anodic reactions [38]:

$$\mathrm{M}\to {\mathrm{M}}^{\mathrm{n}+}+\mathrm{n}{\mathrm{e}}^{-}$$

(6)

Cathodic proton reduction [39,40]:

$$2{\mathrm{H}}^{+}+{2\mathrm{e}}^{-}\to {\mathrm{H}}_2(\mathrm{p}\mathrm{H}<4)$$

(7)

As shown in Table 1, four alloying elements, including Fe, Cr, Ni, and Mo, are major contributors to the corrosion performance of alloys tested in this study. Based on the data retrieved from a public thermodynamic database at room temperature and commercial chemical software Hydra-Medusa V. 1, the electrochemical potential of an alloy element to form a specific compound in the s-CO₂-saturated NaCl solution at 50 °C was calculated, and the corresponding Pourbaix diagram of this alloy was constructed. Detailed information about the Pourbaix diagram’s construction can be found in our recent publications [41,42]. Figure 10 presents the established Pourbaix diagrams of the four alloying elements in the testing solution. At a pH of 3.1, only Mo could form a stable oxide (MoO₂) instead of Fe, Ni, or Cr. Thermodynamically, Fe, Ni, or Cr suffers active corrosion, while Mo likely experiences passivation in such an acidic s-CO₂ environment. However, as shown in Figure 2, the corrosion rates of the alloys decrease with increasing Cr content in them, indicating that the corrosion processes are controlled by the formation and chemical dissolution of the carbonates and/or oxides of these alloying elements in the alloys in the s-CO₂ SW at 50 °C. In addition, based on the testing results (see Figure 2 and Figure 8) and thermodynamic prediction (Figure 10d), it is likely that there is a critical Mo content in the alloys, below which the addition of Mo only has a marginal effect on their corrosion resistance improvement in s-CO₂-saturated aqueous solutions.

4.2. Formation and Dissolution of Carbonate and Oxide

The above post-mortem characterization of corroded samples shows that FeCO₃ layers are formed on all the tested alloys. It is thus necessary to clarify FeCO₃ formation. As Fe is unlikely passivated in s-CO₂ SW, FeCO₃ formation likely involves the active dissolution of Fe, followed by FeCO₃ precipitation. The electrochemical reactions involved in FeCO₃ precipitation include [43]:

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

(8)

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

(9)

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

(10)

Supersaturation (S) is a critical driver of FeCO₃ precipitation involving nucleation and particle growth processes [44]. S is defined as $\mathrm{S}=\left({\mathrm{a}}_{{Fe}^{2+}}{·\mathrm{a}}_{{CO}_3^{2-}}\right)/K_{sp}$ [45], where ${\mathrm{a}}_{{Fe}^{2+}}$ and ${\mathrm{a}}_{{CO}_3^{2-}}$ (in mol/L) are the ferrous and carbonate ion activities, respectively, and $K_{sp}$ (in mol²/L²) is the solubility product associated with FeCO₃. At lower S values, particle growth is predominant; higher values favor nucleation, leading to amorphous or nano-crystalline film formation [34]. In this study, the S value is supposed to be high for low-Cr steel (<3 wt.%.) and low for steels/alloys with a Cr value larger than 3 wt.%. Except for those, the pH can also remarkably affect the morphology and integrity of FeCO₃ films. With a decrease in pH, the critical Fe²⁺ concentration required to exceed FeCO₃ solubility significantly increased [46,47]. For example, reducing the pH from 6 to 5 and then from 5 to 4 increases Fe²⁺ solubility by 100 times and 5 times, respectively [44]. A study on X65 steel in a 3.5 wt.% NaCl solution at 50 °C found that decreasing the pH from 7.5 to 3.8 resulted in the replacement of a protective crystalline FeCO₃ layer with non-protective amorphous FeCO₃ film [48]. Different from this report, the results in this study show that a thick, non-protective, but crystalline, FeCO₃ layer was formed on all the tested alloys. The difference could be attributed to the presence of high concentrations of HCO₃⁻ and CO₃²⁻ enhancing the nucleation and growth of FeCO₃ crystals in the s-CO₂ environment.

The formation of a non-protective FeCO₃ layer on the alloys may also be related to environmental temperature. It is suggested that, at low temperatures (<80 °C), controlling the solution’s pH above 6 is necessary for the formation of a protective film [49]. For X70 steel, it was found that a compact and protective FeCO₃ layer could be formed only in the solutions with a pH range of 5.5–6.5 at temperatures ≥ 75 °C [50]. It is worth noting that the solubility of CO₂ in a brine solution exhibits a negative correlation with temperature, indicating the pH value of the solution increases with the temperature [51]. These observations are consistent with our findings that a protective FeCO₃ layer cannot be formed on the alloys with different alloying elements in the s-CO₂-saturated NaCl solution at 50 °C. Therefore, the corrosion-resistance improvement of the alloys, as shown in Figure 2, is due to the formation of an inner (Fe, Cr) oxide layer.

Thermodynamically, Fe, Cr, and Mo could form their oxides in an aqueous solution at 50 °C, as predicted in

Table 4. Calculated Gibbs energy of formation of Fe, Cr, Ni, and Mo oxide with saline water at 50 °C.

Reaction	Gibbs Energy at 50 °C (kJ/mol)
2Fe + 3H2O = Fe2O3 + 3H2	−34.911
2Cr + 3H2O = Cr2O3 + 3H2	−347.030
Ni + H2O = NiO + H2	23.862
Mo + 2H2O = MoO2 + 2H2	−61.229

. Once formed, they would experience chemical dissolution in the acidic s-CO₂ SW. Compared to Fe oxides, the Cr-included layer and Cr oxide exhibit lower dissolution rates in an acidic solution [52,53,54]. With an elongated testing duration, a Cr-included inner layer would inevitably be formed on the alloy containing the Cr element, accompanied by the precipitation of a non-protective FeCO₃ layer. Compared to FeCO₃ and Fe oxide, the (Fe, Cr) oxide layer can act as a better barrier to: (1) reduce outward metal cations and inward oxygen anions mitigation within the corrosion layer [55], and (2) retard the electron transfer from the interface of corrosion layer/substrate to the interface of the corrosion layer/solution [10]. Increasing the Cr³⁺ content in the (Fe, Cr) oxide can lead to the formation of a more protective layer, as expected. Former studies on hot aqueous solutions indicated that only the alloys with a Cr content > 25 wt.% can form and maintain a thin and compact Cr₂O₃ layer [56,57]. These factors result in corrosion rates decreasing with increasing Cr content in the alloys, as shown in Figure 2. Another interesting observation is that the presence of Cr can also influence the adhesion between FeCO₃ and (Fe, Cr) oxide. For the alloys with a Cr content ≤ 5 wt.%, the formed inner (Fe, Cr) oxide layer has poor adhesion with substrate or outer FeCO₃ layer (see SEM results in Figure 4). Microstructural analyses of SEM images in Figure 5 show that 2Cr steel forms a less cohesive amorphous layer topped by a dense FeCO₃ layer, whereas 5Cr steel features the opposite: a dense amorphous layer topped by a less-cohesive FeCO₃ layer. However, further increasing the Cr content above 8.9% can noticeably improve the adhesive property of the inner (Fe, Cr) layer with both the substrate and FeCO₃ layer, as demonstrated by FIB-SEM sample images in Figure 7. Ni cations were not detected within the corrosion scales, suggesting that the improved performance of these alloys is primarily due to their Cr content, accompanied by unrevealed indirect benefits from Ni. This will help achieve the compact and protective integrity of the formed corrosion layer and effectively resist s-CO₂ environmental attacks. Although the selection of these alloys with a Cr content > 9% inevitably leads to a relatively high capital investment, the long-term successful operation of the system core components can significantly reduce the operation cost to an even lower value compared to that of low-Cr alloys with poor corrosion resistance in s-CO₂ environments. Note that the price increased by the partial replacement of Fe with Cr in the constructional alloys will be much lower than the cost to replace core components during the service [58].

In addition, Cl is a notorious agent that triggers and propagates localizing pitting. Many studies have reported that Cl⁻ ion can be preferentially adsorbed and accumulated at oxide defects to replace O²⁻ for the formation of highly soluble metal–anion complexes [59]. However, in this study, pitting did not occur on the tested alloys. This can be due to: (1) the relatively high chemical dissolution of the corrosion layer suppressing the localized accumulation of Cl⁻ with corrosion layers. (2) Cl⁻-induced pitting being inhibited by HCO₃⁻. HCO₃⁻ was found to act as a pitting inhibitor against chloride pitting in an aqueous solution for Fe-based steels and Ni-based alloys [23,60]. (3) The influence of Cl⁻ on s-CO₂ solubility in aqueous solutions. It was found that the solubility of CO₂ decreases with increasing NaCl concentration in the solution [61].

5. Conclusions

This study investigated the corrosion of seven commercial alloys in saline water (3.5 wt.% NaCl) in an 8 MPa s-CO₂ environment at 50 °C over 96 h. The key findings are summarized as follows:

(1)

Corrosion rates consistently decrease with increasing Cr content in Fe-Cr and Fe-Cr-Ni alloys. There is a critical Cr content (~9 wt.%), above which the alloys exhibit a satisfactory corrosion performance for s-CO₂ permanent storage or enhanced oil recovery in saline water reservoirs.

(2)

During exposure to the s-CO₂ saline water environment, a non-protective FeCO₃ layer forms on the constructional alloys, likely via a precipitation mechanism due to environmental acidity and the presence of high amounts of HCO₃⁻ and CO₃²⁻. For alloys with more than 2 wt.% Cr, an inner Cr-enriched layer forms between the outer FeCO₃ layer and the substrate, acting as the main barrier controlling the corrosion process in the s-CO₂ saline water. The properties of the inner Cr-containing layer significantly improve with increasing bulk Cr content above ~9 wt.%.

(3)

In the s-CO₂ saline water environment, pitting is unlikely to occur on candidate construction alloys due to the solution’s acidity and the presence of bicarbonate ions, even though the Cl⁻ ion concentration reaches up to 3.5 wt.%.

(4)

Alloys with a Cr content above 9% are suitable candidates for next-phase long-term corrosion kinetic evaluations. Given the complicate chemistry of s-CO₂ storage environments, it is highly recommended to assess their performance in s-CO₂-saturated solutions with typical ions (such as, Ca²⁺, Mg²⁺, or S²⁻ presented in the reservoirs) at higher temperatures, as the testing temperature (50 °C) in this study is the lower end of the 50–80 °C operation range under s-CO₂ storage.