Comparison of waterline corrosion of a weathering steel in static and dynamic simulated seawater

2.1 Material and medium

The experimental material was a weathering steel, and its main composition is presented in Table 1. Wire beam specimen with dimensions of 3 mm × 3 mm × 30 mm was machined from a steel plate. A total of 100 wire beam specimens were prepared to form a wire beam electrode (WBE), as depicted in Figure 1, showcasing surface morphology of the WBE and the schematic diagram of WBE corrosion system. The WBE consisted of an array of small electrodes with a side length of approximately 3 mm, which were insulated from each other and sealed with epoxy resin to form a 4 × 25 matrix WBE. The distance between adjacent small electrodes was approximately 1 mm. The small electrodes were numbered from top to bottom in rows (1–25) and left to right in columns (1–4), as shown in Figure 2. Prior to each experiment, the surface of the WBE was sequentially polished using sandpaper with grit sizes of #60, #400, #800, and #1600, respectively, and then degreased with ethanol and rinsed with deionized water. During the experiment, the concave surface of water intersected with the WBE surface at the middle position of the small electrodes in rows 3 and 4, as marked by a line in Figure 2. That is, the rows 1–3 above the waterline were in the air, and the rows 4–25 below the waterline were in the water solution. During the experiments, the ambient temperature was kept at 29 °C, and the relative humidity was maintained at 70 % through air conditioning control, and distilled water was added to the beaker twice a day in the morning and evening to keep the water level unchanged.

Figure 1:

Surface morphology and schematic diagram of WBE in corrosion system. (A) Surface morphology and (B) schematic diagram.

Figure 2:

Schematic diagram of experimental apparatus (RE was SCE).

The experimental medium was a simulated seawater, prepared according to the standard ASTM D1141, and the composition is shown in Table 2. In the experiment, WBE was immersed in a 2,000 mL beaker filled with the simulated seawater, as shown in Figures 1 and 2. The surface of WBE was perpendicular to the water surface of the solution. The dynamic experiment was conducted in a constant temperature shaker (produced by Shanghai Precision Experimental Equipment Company, type THZ-320) with a rotating speed of 45 r/min. The experimental temperature was 29 °C.

2.2 Electrochemical measurements

Electrochemical tests were performed on a Coster CST520 array wire beam electrode potential current scanner, with a saturated calomel electrode (SCE) as reference. The software WBEScan was used to control the test automatically.

In the default state of the CST520 ((including unpowered state), all the 100 small electrodes on WBE were short circuit connected to each other and connected to the ground end, so the 100 small electrodes could be regarded as a complete metal plane. Potential measurement adopted high input impedance mode, while current measurement adopted zero resistance current mode. When conducting current measurement, the single electrode n_x to be measured was connected to the reverse end of the zero-resistance galvanometer, while the remaining 99 electrodes were short circuit connected to the ground, and the galvanic current between the single electrode n_x and the remaining 99 electrodes was measured. By sequentially switching the current measurement circuit consisting of small electrodes n_x (x = 1–100) and the other 99 small electrodes, the current distribution on the entire WBE surface can be scanned. When conducting potential measurement, the single small electrode n_x to be measured was short connected to the ground, while the remaining 99 small electrodes were suspended in the air, and the potential between n_x and the reference (saturated calomel electrode, SCE) was measured. By sequentially switching the small electrode n_x (x = 1–100) and forming a circuit with the reference, the potential distribution on the entire WBE surface can be scanned. All potentials in this paper were relative to SCE.

2.3 Surface analysis

Take out yellow and black corrosion product specimens from the surface of the test piece after 30 days, and analyze them after drying. The composition of corrosion products was analyzed by Fourier transform infrared (FTIR) spectroscopy (Spectrum two). The morphology of the corrosion products was analyzed by a scanning electron microscope (SEM) (SU-1500, Hitachi Ltd, Tokyo, Japan).

3.1 Waterline corrosion behavior of steel in static simulated seawater

Firstly, the change of current density distribution on the surface of WBE in static simulated seawater with time was measured, as shown in Figure 3. The anodic current was set to negative and the cathodic current to positive in figures. That is, the blue represents the cathode region where oxygen reduction reaction mainly occurs, and the red represents the anode region where metal oxidation reaction mainly occurs. The darker the color, the greater the current density.

Figure 3:

Current density distribution of the WBE after immersing in static simulated seawater for (A) 1 h, (B) 1 d, (C) 3 d, (D) 6 d, (E) 12 d, (F) 18 d, (G) 24 d, and (H) 30 d.

Figure 3 shows that during the 30 d experiment, the small electrodes above the waterline exhibited cathodic current, while the small electrodes below the waterline were divided into two regions: the electrodes in rows 4–10 (below the waterline by 0.5 mm–27.5 mm) near the waterline were mainly the cathodic region where cathodic current mainly occurred, while the electrodes below row 10 (below the waterline by 27.5 mm) were mainly the anodic region where anodic current mainly occurs.

From Figure 3A, it can be observed that stable cathode and anode regions have not yet formed on the surface of WBE in the early stage of the experiment (immersing for 1 h), which was similar to the results mentioned by Jeffrey (Jeffrey et al. 2009). The region in rows 4–6 (below the waterline by 0.5 mm–11.5 mm) displayed dark blue, indicating a region with strong cathode reaction. At the initial stage of corrosion, the distribution of dissolved oxygen in the water solution was relatively uniform, and the distribution of anode and cathode on the surface of WBE was random. The region with strong cathode reaction disappeared at 1 d, and the cathode region gradually extended to rows 1–3 (above the waterline by 0.5–11.5 mm), indicating the formation of thin liquid film in the upper region of waterline (Tan 2005). Although the cathode area increased at this time, the cathode was mainly borne by the underwater area. The color of the blue and red areas became significantly darker at 6 d, indicating an increase in the galvanic current. This may be related to the generation of some specific corrosion products on the steel surface, such as corrosion product γ-FeOOH, which can undergo a reduction reaction and increase the cathodic current density (Chen et al. 2020). The cathodic reaction rate in the upper region of the waterline increased at 12 d, and the cathodic current in row 1 was greater than that in row 3. After 12 d, the change in current density distribution was small, and the distribution in cathode and anode regions was relatively fixed.

Figure 4 shows the single-row current density and total anode current density of the WBE in simulated seawater. The single-row current density was obtained by summing the current density vectors of four small electrodes in the same row and then calculating the average value, while the total anode current density was obtained by summing the anode current density of all the small anodes in different time. The single-row current density can more intuitively display the variation of metal corrosion at different positions with time (Tan 2005). The maximum cathodic and anodic current densities appeared at 1 h, with the maximum cathodic current density located in rows 4 and 5 (below the waterline by 0.5 mm–7.5 mm), and the maximum anodic current density located in rows 12 and 24 (below the waterline by 35.5 mm and 83.5 mm, respectively). However, in the first 3 days, there was a rapid decrease in cathodic and anodic current densities with time, which may be related to the quick formation and thickening of corrosion products on the steel surface, leading to an increase in the resistance of metal dissolution reaction. The main cathodic reaction occurred in rows 5–9 (below the waterline by 7.5 mm–23.5 mm) at 6 d, and the maximum single-row current density (reaching 102.5 μA/cm²) was located at row 7 (below the waterline by 15.5 mm). The maximum single-row current density on the upper of the waterline was located in the row 3 (above the waterline by 0.5 mm), reaching 28.6 μA/cm². At 12 d, the single-row current density of the row 1 (above the waterline by 8.5 mm) was 86.2 μA/cm², which exceeded the maximum cathodic single-row current density of 72.1 μA/cm² below the waterline. The distribution of cathodic and anodic current density tended to be stable after 12 d. Figure 4B shows that the total anodic current density did not change significantly after 6 d, fluctuating around 2.5 mA/cm².

Figure 4:

Single-row current density and total anodic current density of the WBE in static simulation of seawater. (A) Single-row current density and (B) total anodic current density.

Figure 5 shows the surface morphology of WBE after corrosion at different times. It can be observed that the corrosion products gradually increased with the extension of time. At 6 d, all small electrodes from row 6 (below the waterline by 11.5 mm) to the bottom of the WBE were obviously covered with corrosion products. Figure 5I shows the surface of WBE after ultrasonic treatment to remove loose corrosion products. It could be observed that there was almost no corrosion on the row 4 and above, and the metal surface presented the original metallic luster, while regions located on row 6 and below were corroded to varying degrees. Some black corrosion products were found on the WBE surface in rows 6–9 (below the waterline by 8.5 mm–23.5 mm), and it is generally believed that these relatively dense black residues have a certain protective effect on the metal (Chen et al. 2020). The metal surfaces in the remaining anode region have undergone varying degrees of corrosion, with more severe corrosion occurring downward, especially on the surfaces in rows 23–24 (below the waterline by 76.5 mm–83.5 mm) at the bottom.

Figure 5:

Surface corrosion morphology of the steel after immersing in static simulated seawater for (A–H) 1 h, 1 d, 3 d, 6 d, 12 d, 18 d, 24 d, 30 d, and (I) after removal of loose surface corrosion products.

It can also be observed from Figure 5 that at the end of the experiment, except for the small electrodes in row 4, the surface of all the other small electrodes below the waterline was covered with a layer of yellowish-brown corrosion products. The small electrodes in row 4 was close to the waterline, where the surface solution could quickly replenish the consumed dissolved oxygen, making it a fixed cathode region that underwent oxygen reduction reaction, as shown in equation (1).

Therefore, the surface solution at row 4 has a high concentration of OH⁻ (Gallagher 2016; Wang et al. 2006), which can cause passivation of the steel surface (Bellanger 1995; Cui et al. 2017; Lee and Rasaiah 2013). The passivated surface had a higher potential, so the steel surface here hardly underwent anodic dissolution. For example, at 1 h, the average potentials in row 4 and row 5 were −0.663 V and −0.669 V, respectively, while the potential in row 13 (below the waterline by 36.5 mm) was −0.699 V. The measured potential difference between the cathode and anode was relatively small, as these small electrodes were short-circuit connected for most of the experimental process and only separated when measuring the potential and galvanic current. Figure 6A shows a schematic diagram of pH distribution on the surface of WBE under static condition. The OH⁻ generated by oxygen reduction reaction continuously diffused from the cathode region at the waterline to the anode region below the waterline due to the concentration difference (Gallagher 2016), gradually forming a pH concentration gradient (Sainis et al. 2020; Yu et al. 2015). Taking the “critical point” of the pH gradient as the boundary (the pH value of the “critical point” is the pH value of the original solution), the following experimental phenomena can be observed: the upper part of the “critical point” was mainly the cathode region, and the lower part was the large anode region. Starting from 12 d, the cathode region was located above the row 11 (below the waterline by 28.5 mm) and the anode region was located below the row 11 in the WBE, until the end of the experiment. At the end of the 30 d experiment, there were fewer or even no corrosion products near the waterline, while the black corrosion products were thicker in the underwater areas far from the waterline, which was consistent with the results in the literature (Chen et al. 2014; Laleh et al. 2023a,b; Li et al. 2020; Pathirana et al. 2023).

Figure 6:

Schematic diagram of pH distribution on WBE surface under (A) static and (B) dynamic condition.

Three stages of waterline corrosion could be observed in the process of static waterline corrosion: initial stage, development stage, and stable stage. The time period of 0∼1 d was the initial stage, during which there was sufficient dissolved oxygen in the water. The oxygen concentration difference at different positions in the WBE was not obvious, and due to the lack of corrosion products on the WBE surface, the resistance to oxygen and electron transfer was small, resulting in a high total anodic current density at this stage. The time period of 1 d–6 d was the development stage, during which the dissolved oxygen in the surface solution of the WBE below the waterline was depleted and cannot be replenished in time, resulting in the formation of an oxygen concentration cell. At this stage, the total anodic current density was greatly reduced because the metal surface was gradually covered by corrosion products. In addition, a thin water film appeared on the upper region of the waterline, but due to the thin thickness, rapid cathodic reaction could not be achieved (Pang et al. 2020). The time period of 6 d–30 d was the stable stage, during which the total anodic current density increased a little and remained stable, possibly due to the formation of a stable water film on the upper of the WBE and the generation of stable corrosion products.

3.2 Waterline corrosion in dynamic seawater

Under dynamic condition, waterline corrosion behavior will undergo change due to the acceleration of mass transfer. Figure 7 shows the variation of the current density distribution on the surface of WBE in dynamic simulated seawater with time, while Figure 8 shows the single-row current density of WBE in simulated seawater, indicating that the waterline corrosion rate under dynamic condition was significantly higher than that under static condition. The accelerated mass transfer of substances related to corrosion reactions, such as oxygen and corrosion products in a flowing state, will promote the reaction of cathode and anode in the corrosion cell.

Figure 7:

Current density distribution of the WBE after immersing in dynamic simulated seawater solution for different times: (A) 1 h, (B) 6 h, (C) 18 h, (D) 1 d, (E) 3 d, (F) 6 d, (G) 12 d, and (H) 30 d.

Figure 8:

Single-row current density and total anodic current density of the WBE in dynamic simulated seawater. (A) Single-row current density and (B) total anodic current density.

It can be observed in Figure 7 that the distribution of anodes and cathodes on the WBE surface was relatively random at 1 h (Figure 7A), which was consistent with the initial waterline corrosion behavior in static simulated seawater. The strongest anode region was located on the small electrode numbered #5-1 below the waterline, with an anodic current density of −1,035.6 μA/cm². At 6 h of corrosion, the anodic current density decreased significantly, and the anode region gradually moved downward to rows 6–23 (below the waterline by 8.5 mm–79.5 mm), forming a large anode structure (Figure 7B). After 3 d, the anode region split into two or more smaller anode regions, and the anode area was significantly smaller at 6 d. It may be that under dynamic condition, the waterline corrosion rate was high in the initial stage, resulting in the rapid formation of relatively dense corrosion products on the anode surface, leading to an increase in reaction resistance (Chen et al. 2020; Raman et al. 1991), even causing electrode reversal in some anode regions due to an increase in potential. For example, from 3 d to 6 d, the potential of the small electrode numbered #11-4 shifted positively from −0.701 V to −0.690 V, and the potential of the small electrode numbered #12-4 shifted positively from −0.710 V to −0.707 V. At 30 d, the anode area further reduced and scattered below 13th row.

Figure 8 shows the single-row current density of WBE in dynamic simulated seawater. The single-row cathodic current density reached its maximum value 719.7 μA/cm² at 1 h, which was located at the 4th row below the waterline. This value was significantly higher than 485.0 μA/cm² under static condition and should be related to the mass transfer acceleration of dissolved oxygen in water. The maximum single-row anodic current density appeared in the 5th row, with a value of −391.8 μA/cm², which was the average of the current density vectors in this row, so it differed significantly from that of the small electrode numbered #5-1 (−1,035.6 μA/cm²). At 6 h, the single-row current density of the 5th row was reversed from anodic current to cathodic current. The maximum single-row anodic current density at 3 d was located in the 18th row (below the waterline by 56.5 mm) with a value of −195.7 μA/cm². After 6 d, the corrosion region was further moved down to row 11 (below the waterline by 28.5 mm) and below. The reason for the reversal of the small electrodes in rows 8–10 (below the waterline by 16.5 mm–27.5 mm) from anode to cathode may be mainly due to the accumulation of relatively dense corrosion products on the steel surface, which thereby increased the anode reaction resistance of the metal and led to a positive shift in the potential. For example, the potential of the small electrode at #10-4 shifted from −0.723 V at 1 d to −0.693 V at 6 d, which was close to the potential −0.692 V of the small cathode #4-4 at 6 d. The electrode reversal phenomenon also appeared in rows 20–22 at 3 d. At 1 d, there was anode current density of −157.22 μA cm⁻², −168.06 μA cm⁻², −143.56 μA cm⁻², respectively, in rows 20–22. While it changed to cathode current density at 3 d, which was 98.61 μA cm⁻², 30.06 μA cm⁻², and 223.33 μA cm⁻², respectively. From the perspective of potential distribution, the average potential of the small electrodes in rows 20–22 at 1 d was −0.725 V, −0.726 V, and −0.724 V, respectively, while that at 3 d was −0.703 V, −0.705 V, and −0.700 V, respectively, indicating a positive shift in the potential. Meanwhile, the average potential of the 23rd row, which was still in the anode region, was more negative (−0.710 V). The anode area further decreased at 30 d. At this time, rows 1–14 (above the waterline by 12.5 mm ∼ below the waterline by 43.5 mm) acted as cathode, rows 15–17 (below the waterline by 44.5 mm–55.5 mm) acted as anode, and rows 18–25 (below the waterline by 56.5 mm–87.5 mm) acted as a cross distribution of anode and cathode.

Figure 9 shows the surface morphology of WBE in dynamic simulated seawater at different times. It could be found that black or brown corrosion products appeared on the surface of small electrodes in rows 3–25 (above the waterline by 3.5 mm ∼ below the waterline by 87.5 mm), which was different from the results under static condition.

Figure 9:

Surface corrosion morphology of the WBE immersing in dynamic simulated seawater for (A–G) 6 h, 18 h, 1 d, 3 d, 6 d, 12 d, and 30 d and (H) after removal of surface loose corrosion products.

From Figures 7–9, it can be observed that under dynamic conditions, column 1 (rows 6–15) shows a cathodic reaction. This phenomenon did not occur in the static state and only occurred under dynamic conditions, which should be related to the shaking of the shaker in the experiment. The shaker caused the beaker to shake clockwise in the experiment, which may cause some changes in the distribution of solution components in the beaker. Column 1 was at the leading edge of rotation and may come into contact with more oxygen molecules.

The development process of waterline corrosion under dynamic condition can be mainly divided into three stages: initial stage (<6 h), development stage (1 d–3 d), and deceleration stage (3 d–30 d). The characteristics of the initial stage and the development stage were the same as those under static condition, distinguished by whether the oxygen concentration difference was formed. However, due to the high metal corrosion rate, the time of these two stages was shortened under dynamic condition. In the deceleration stage, the total anodic current density decreased continuously with time. The formation of this stage should also be related to the high waterline corrosion rate in the early stage under dynamic condition, and the relatively dense corrosion products covering on the anode surface in a short period of time. This first raised the potential of part of the anode surface and reversed to the cathode, resulting in a decrease in the total anode area and an increase in the local anodic current density (such as at 6 d). However, further thickening of the corrosion products led to a continuous increase in the anodic dissolution resistance, and the local anodic current density also decreased with time after 12 d. In addition, under dynamic condition, the pH gradient on the cathode surface was difficult to form and the pH value of the liquid layer on the metal surface near the waterline should be significantly lower than that under static condition. Therefore, the metal surface at rows 3–4 near the waterline was not easily passivated. With the positive shift of the anode potential caused by the coverage of corrosion products, the potential difference between the cathode and the anode decreased. The potential of some original anode underwater even moved positively above the potential of the small electrodes at rows 3–4, causing the corrosion of some small electrodes at rows 3–4 (above the waterline by 3.5 mm ∼ below the waterline by 3.5 mm).

By comparing potentials of the small electrodes in WBE, it was found that the potential difference between the cathode and the anode in WBE under dynamic condition was generally smaller than that at static condition, and decreased continuously with time, and even electrode reversal occurred in the later stage of the experiment. However, the anodic current density under dynamic condition was much higher than that under static condition, indicating that the potential difference was not the main factor accelerating the corrosion of WBE. Mass transfer processes such as dissolved oxygen may have a greater impact on it.

The above results indicate that for the steel studied in this paper, under dynamic conditions, the corroded anode area had shifted from the area far away from the waterline to the waterline position, which was different from the results under static conditions. Therefore, in dynamic corrosion systems, attention should be paid to corrosion control at the waterline.

3.3 Analysis of corrosion products

Figure 10 shows the morphology of corrosion products under SEM. Figures 10A and B shows the morphology of the outer yellow corrosion product and the inner black corrosion product, respectively. It can be observed that the yellow corrosion product was relatively loose and porous, and the black corrosion product was denser. Figure 11 shows the FTIR spectra of black and yellow corrosion products under static and dynamic condition. It can be observed that for yellow corrosion products, there are peaks of 1,017 cm⁻¹ and 737 cm⁻¹ under both static and dynamic condition, corresponding to γ-FeOOH (Raman et al. 1991). A peak at 467 cm⁻¹ corresponding to α-FeOOH was also observed in the static specimen and a peak at 550 cm⁻¹ corresponding to γ-Fe2O3 was observed in the dynamic specimen. For the black corrosion products, broad and strong absorption peaks at 1,059 cm⁻¹ and 1,068 cm⁻¹ indicated the presence of σ-FeOOH under both static and dynamic condition, and the peaks at 1,246 cm⁻¹, 1,399 cm⁻¹, and 1,260 cm⁻¹ correspond to α-FeOOH. However, a broad and weak absorption peak at 566 cm⁻¹ was also observed under dynamic condition, indicating the presence of Fe₃O₄.

Figure 10:

Morphology of (A) the yellow corrosion product and (B) the black corrosion product.

Figure 11:

Infrared analysis of corrosion products: (A) under static condition and (B) under dynamic condition.

The formation process of the yellow corrosion product γ-FeOOH is shown in equations (2) to (4). As a thermodynamically metastable phase, γ-FeOOH can be converted into Fe₃O₄ by cathodic reduction reaction [equation (5)] (Zhang et al. 2023). Fe₃O₄, as a thermodynamically stable phase (Gao et al. 2018), can form a relatively dense corrosion product film on the steel surface, which increased the reaction resistance of the corrosion system. This may be the reason for the polarity reversal in some anode regions and the decrease of total anode area and total anodic current density after dynamic experiment for 3 days.