Table 1 shows the dimensions and residual carbon amount of test specimens. Phosphorus-deoxidized copper tubes (JIS H 3300 C1220, OL) were used in this study. Copper tubes with and without carbon film were prepared. The amount of residual carbon was measured using the rapid filling test [15]. As a result of the measurement, the residual carbon amount was 5.5 mg/m². The copper tubes were cut parallel to the tube axis and used as test specimens. The copper tube with the carbon film was designated as Copper tube A. Copper tube A was cleaned with 25 mass % nitric acid to remove the film, and the resulting specimen was designated as Copper tube B. during cleaning with 25 mass % nitric acid, black material was observed to detach from the surface, suggesting that the carbon film had been removed. Therefore, the amount of residual carbon on Copper tube B was considered to be 0 mg/m². Before the testing, Copper tube A and Copper tube B were acid-cleaned with 5 mass % sulfuric acid to remove the oxide film formed on their surface.

Table 1. Sample size and residual carbon.

The test solutions used Tsukuba City water treated with anion exchange treatment (treated water). Table 2 shows the water quality of the test solutions. The pH and electrical conductivity of all test solutions were measured before the test using a glass electrode and two platinum electrodes coated with platinum black, respectively. The treated water corresponded to the 0 mg/L Cl⁻ solution, and its alkalinity (pH 4.8), SiO₂, calcium hardness, S O 4 2 − , Cl⁻, and N O 3 − concentrations were measured. The alkalinity (pH 4.8) was determined by titration using a bromocresol green-methyl red ethanol solution and sulfuric acid. The concentration of SiO₂ was quantitatively determined by the molybdenum yellow spectrophotometric method. The concentrations of calcium hardness, S O 4 2 − , Cl⁻, and N O 3 − were determined by ion chromatography. The test solutions containing 10, 30, 50, and 100 mg/L Cl⁻ were prepared by adding NaCl, 10 mg/L HEDP and 5 mg/L BTA to the treated water. Hereafter, Cl⁻ concentrations of 0, 10, 30, 50, and 100 mg/L are referred to as Cl⁻ 0 mg/L, Cl⁻ 10 mg/L, Cl⁻ 30 mg/L, Cl⁻ 50 mg/L, and Cl⁻ 100 mg/L, respectively.

Table 2.Water quality of the test solutions.

A schematic drawing of the test apparatus is illustrated in Figure 1. A 500 mL beaker was used as the test vessel. The test solution was added to a volume of 500 mL and stirred at 300 rpm under room temperature. Furthermore, this test was conducted under open-air conditions to evaluate changes in the corrosion potential of copper in the test environment. The immersion period was 5 days, and the corrosion potential was measured over time. The corrosion potential was measured using a digital multimeter (Keysight: 34070A). As the reference electrode used saturated Ag/AgCl (SSE). The SSE was immersed only during measurement to avoid the effects of Cl⁻ leaching from the SSE. The test area was set to 3 cm². The measurements were performed triplicate. After testing, the specimens surface were observed using a digital microscope (KEYENCE: VHX-5000), and their an composition was analyzed using SEM equipped with EDX (Hitachi High-Technologies: S-4300).

Figure 1. Schematic of corrosion potential measurement apparatus.

A schematic drawing of the test apparatus is shown in Figure 2. A 500 mL beaker was used as the test vessel. The test solution was added to a volume of 500 mL and stirred at 300 rpm and under room temperature. The measurement was conducted by deoxidizing the test solution using nitrogen gas. The purpose of this was to remove the effects of oxygen and evaluate the reaction of copper under the test environment. The polarization curve was measured using a potentiostat (HZ-7000; Hokuto Denko, Tokyo, Japan). The measurement was conducted by deoxidizing the test solution using nitrogen gas. The sweep speed was set to 20 mV/min. The specimens were used as the working electrode, a saturated Ag/AgCl electrode (SSE) as the reference electrode, and platinum as the counter electrode. The test area was set to 1 cm². Measurements were performed in triplicate.

Figure 2. Schematic of anodic polarization curve measurement apparatus.

Results of the corrosion potential measurement are shown in Figure 3 and Figure 4. This shows the average of three measurements. About the Copper tube A (residual carbon: 5.5 mg/m²), the corrosion potential increased to +0.07 V vs. SSE after immersion and then tended to stabilize. However, in the Cl⁻ 100 mg/L solution, the potential continued to decrease even after one day of immersion. The final measurement potential increased in the following order: Cl⁻ 100 mg/L < Cl⁻ 10 mg/L < Cl⁻ 30 mg/L < Cl⁻ 0 mg/L < Cl⁻ 50 mg/L. About the Copper tube B (residual carbon: 0 mg/m²) the corrosion potential increased to +0.07 V vs. SSE after immersion and continued to increase thereafter. After two days of immersion, the potential tended to stabilize for Cl⁻ 0 mg/L and 10 mg/L, whereas it tended to decrease slightly for Cl⁻ 30 mg/L, 50 mg/L, and 100 mg/L. The final measurement potential increased in the following order: Cl⁻ 30 mg/L < Cl⁻ 50 mg/L < Cl⁻ 100 mg/L < Cl⁻ 10 mg/L < Cl⁻ 0 mg/L. At each Cl⁻ concentration, the corrosion potential of Copper tube B tended to be higher than that of Copper tube A.

Copper tube A had a carbon film on its inner surface. Yamada et al. reported that carbon film form unevenly on the inner surface of copper tubes [1]. It has also been reported that the carbon film-covered areas act as cathodic sites, whereas exposed copper substrate areas act as the anodic sites [3]. Based on these findings, the initial increases in corrosion potential observed for Copper tube A is attributed to the promotion of the cathodic reaction by the carbon film. As the cathodic reaction proceed, an oxide film forms on the surface, causing the change in corrosion potential became gradual. In the Cl⁻ 100 mg/L, Cl⁻ is considered to damage the film, leading to corrosion and a decrease in the corrosion potential. In contrast, because copper tube B, has no carbon film, the corrosion reaction proceeds slowly and the change in potential was also more gradual.

Figure 3. Time dependent changes in the corrosion potential of Copper tube A in test solutions. (stirred at 300 rpm, under room temperature, open-air conditions).

Figure 4. Time dependent changes in the corrosion potential of Copper tube B in test solutions. (stirred at 300rpm, under room temperature, open-air conditions).

Figure 5 shows the results of the observation and analysis of the test surface after corrosion potential measurement. Corrosion was only observed on Copper tube A at a Cl⁻ 100 mg/L. Figure 6 shows SEM images and the EDX analysis results for areas with and without corrosion products. In the corrosion products, 2.8 mass % Cl was detected, Cl is known to promote pitting corrosion of copper and Yamada et al. and Kano et al. reported that Cl⁻ was detected in pitted areas [12][13]. Based on these results, the corrosion observed on copper tube A in the Cl⁻ 100 mg/L solution is considered to be the first stage of pitting corrosion.

Figure 5.The surface images of Copper tubes A and B after measurement the corrosion potential. ○: The area where corrosion was observed.

Figure 6. SEM image and the EDX analysis of Copper tube A at a Cl⁻ 100 mg/L after measurement of corrosion potential with test water.

The anodic polarization curves are shown in Figure 7 and Figure 8. For Copper tube A, the current density increased sharply from −0.1 V vs. SSE and then changed gradually under all test conditions. Subsequently, the current density increased sharply again at approximately +0.1 to +0.2 V vs. SSE. The potential at which the second sharp increase was observed was defined as the break down potential (BDP). The BDP was defined as the potential at the intersection of the tangent to the increasing section where the current density began to increase again and the line extended from the immediately preceding section of gradual current density increase [7]. Figure 9 shows the BDP under each test condition. The BDP tended to shift to the cathodic side as the Cl⁻ concentration increased. For Copper tube B, the current density increased sharply from around 0 V vs. SSE under all test conditions, following by gradual change. The current density then increased sharply again approximately +0.1 to +0.3 V vs. SSE. Similar Copper tube A, the BDP tended to shift to less noble potential with increasing the Cl⁻ concentration. Comparison of the BDP values of Copper tube A and Copper tube B shows that, Copper tube B tended to exhibit higher values than Copper tube A, irrespective of the Cl⁻ concentration. These results suggest that ‘in environments with high

Figure 7.Anodic polarization curves of copper tube A in test solutions (sweep speed 20 mV/min, stirred at 300 rpm, under room temperature, deoxidizing by nitrogen gas).

Figure 8. Anodic polarization curves of Copper tube B in test solutions (sweep speed 20 mV/min, stirred at 300 rpm, under room temperature, deoxidizing by nitrogen gas).

Figure 9. Breakdown potential of copper tubes with different amounts of residual carbon at various chloride ion concentrations.

Cl⁻ concentrations and the presence of a carbon film, the oxide film is prone to degradation’.

In the corrosion potential measurements, the corrosion potential of Copper tube A increased sharply immediately after immersion, after which the change in potential became gradually. However, at a Cl⁻ concentration 100 mg/L, the potential continued to decrease. After testing, corrosion was observed on the surface of Copper tube A at a Cl⁻ concentration 100 mg/L. In the anodic polarization curve measurements, the BDPs of both Copper tubes A and B shifted to less noble potential as the Cl⁻ concentration increased. These results suggest that, in treated water containing water treatment chemicals, the higher the Cl⁻ concentration, the more likely copper corrosion is to occur. It is thought that the presence of a carbon film makes corrosion even more likely to occur.

Takahashi etal. reported effect of Cl⁻ on pitting corrosion of copper in artificial water [16]. Table 3 summarizes their results and the findings obtained in this study. In artificial water, pitting corrosion was observed at a Cl⁻ cocentrations10 mg/L. In contrast, in treated water with high concentrations of bicarbonate ions, no pitting corrosion was observed up to a Cl⁻ concentration of 100 mg/L for copper tubes without carbon film. Even in the presence of a carbon film, pitting corrosion was suppressed up to a Cl⁻ concentrations of 50 mg/L. Based on these results, it is considered that in treated water containing water treatment chemicals, corrosion is suppressed in the presence of carbon film until the Cl⁻ concentration reaches to 50 mg/L. In the absence of carbon film, corrosion is suppressed at a Cl⁻ concentration of 100 mg/L. Furthermore, the mechanism of pitting corrosion in treated water containing chloride ions and water treatment chemicals was investigated. In the presence of a carbon film, cathodic reactions are promoted in the area surrounding the carbon film. This reaction forms copper oxide films and composite films of HEDP and BTA [5]. The carbon film exists unevenly, resulting in uneven areas where cathodic reactions occur. As a result, the copper oxide films and composite films may not form uniformly, leading to an unstable film. This unstable film is considered to be more easily broken down by Cl⁻. In contrast, on copper tubes without a carbon film, uniform copper oxide and HEDP/BTA

Table 3. Effect of Cl⁻ concentration on copper corrosion resistance in the previous study by Takahashi et al. and in the present study.

○ indicates that pitting corrosion was not observed; × indicates that pitting corrosion was observed.

composite films are likely to form. In addition, treated water contains a high concentration of H C O 3 − , which provides pH buffering capacity. These conditions are considered to promote the formation and maintenance of a stable protective film. Consequently, the film is not easily damaged and corrosion is suppressed, even in environments containing Cl⁻. In addition to differences in residual carbon, changes in surface condition resulting from nitric acid pretreatment might also have influenced these results. The effects of nitric-acid pretreatment on the surface condition remain to be investigated.