Tomoyoshi IRIE*
Hiroshi YAKUWA*
Masahiro SAKAI**
*
Technologies, R&D and Intellectual Property Division
**
Muroran Institute of Technology
Exposure tests of the C-ring specimen in ammoniacal environments were carried out to investigate stress corrosion cracking of a pure copper tube. As a result, intergranular corrosion or intergranular cracking was occurring on the surface of phosphorous deoxidized 1/2H-tempered and O-tempered copper tubes under gas-phase ammonia derived from 1 % and 3 % ammonia water, while intergranular corrosion was not observed in the oxygen-free copper and phosphorous deoxidized H-tempered copper tubes. Tempers of these tubes were H, 1/2H and O defined in JIS H0500. It was found that intergranular corrosion did not occur on the surface of phosphorous deoxidized 1/2H copper tubes under gas-phase ammonia derived from 0.1 % ammonia water. Furthermore, the SCC of the phosphorous deoxidized copper tube was suppressed by maintaining a low humidity or a low oxygen environment, and by reducing the stress which was applied to the specimens.
Keywords: Stress corrosion cracking, Pure copper, Phosphorous deoxidized copper, Oxygen-free copper, Ammonia, C-ring test, Exposure test, Crystal grain, Intergranular cracking，Tempered
Stress corrosion cracking (SCC) is a phenomenon in which alloy metals crack when exposed to a specific corrosive environment under static tensile stress within the allowable stress1
）. It has been reported that SCC occurs in phosphorous deoxidized copper exposed to an ammoniacal environment, and that when the phosphorous concentration in the material is 0.02 % or less, the susceptibility to SCC decreases as the phosphorous concentration decreases2
）.
Phosphorous deoxidized copper tubes are used as heat transfer tubes in absorption chillers, which are widely used as heat source equipment for district heating and cooling and building air conditioning. Phosphorous deoxidized copper tubes are classified into JIS H3300 C1201 (P: 0.004 % or more or less than 0.015 %, hereinafter referred to as a low phosphorous copper), which has a low phosphorous concentration, and JIS H3300 C1220 (P: 0.015 % or more or less than 0.040 %, hereinafter referred to as a high phosphorous copper), which has a high phosphorous concentration. Low phosphorous copper tubes are used for the heat transfer tubes of absorption chillers. Low phosphorous copper tubes have the advantage of lower SCC susceptibility, but the market volume is smaller than that of high phosphorous copper tubes. In the future, the use of high phosphorous copper tubes will be desirable to reduce the manufacturing lead time of absorption chillers and to promote global procurement.
The authors have investigated the occurrence of SCC in phosphorous deoxidized copper tubes by exposing C-ring specimens, which can be stressed by applying constant strain, to an ammoniacal environment. In this report, we added newly obtained experimental results to the contents 3) to 5) that have been reported so far, and describe the material factors (phosphorous concentrations, temper, and load stress) and environmental factors (ammonia concentrations, humidity, and oxygen concentrations) on SCC occurrence in phosphorous deoxidized copper tubes.
Three types of copper tube materials were used: high phosphorous (HP), low phosphorous (LP) and phosphorous-free JIS H3300 C1020 (P: 0 %, hereinafter referred to as oxygen-free copper, OF), each with a different degree of workability: H-tempered (work hardened, tensile strength: 415-447 MPa), 1/2H-tempered (work hardened, tensile strength: 248-260 MPa) and O-tempered (annealed, tensile strength: 234-240 MPa).
The specimens used for the exposure tests were copper tubes of nominal diameter 15 A (reference outer diameter 15.88 mm, thickness 0.8 mm) cut to 20 mm length and processed into C-ring specimens according to ASTM G386
）. Fig. 1 shows the schematic diagram of C-ring specimens.
Fig. 1 Schematic diagram of C-ring specimen
Fig. 2 shows the method of stressing the specimens. In the C-ring specimen, stress can be applied to the outer periphery of the specimen in the diagram by tightening the bolt and nut, and the largest tensile stress is applied to the ＊part. The maximum stress applied to the specimen was determined from the correlation between the circumferential stress (calculated from the results of the circumferential and tube axial strain measurements) at the position of ＊relative to the diameter of the copper tube before the bolt and nut were tightened minus the diameter after the bolt and nut were tightened.
Fig. 2 Schematic diagram of stress loading method
The test water to be placed in the sealed container was prepared by dissolving a predetermined amount of 23 % ammonia water in deionized water. Fig. 3 shows the outline of the exposure test device using C-ring specimens in an ammoniacal environment.
Fig. 3 Outline of exposure test
A test tube filled with 15 mL of test water and two C-ring specimens stressed by tightening a bolt and nut were placed in a 500 mL polypropylene container, the container lid was closed and the container was left at room temperature (23 °C) and exposed to an ammonia atmosphere volatilized from the test water. The specimen was pulled out after the predetermined time. The specimen was then washed with deionized water, dried naturally, and the bolt and nut were removed. The specimen was then ultrasonically cleaned in dilute sulfuric acid to remove corrosion products, and embedded in epoxy resin. After the epoxy cured, the polished section of the specimen was observed with an optical microscope.
To reduce the humidity inside the container, a desiccant was put in the container. 1 g of calcium oxide was used as a desiccant, which was placed in a non-woven filter and placed on the bottom of the closed container shown in Fig. 3. The relative humidity (%RH) in the container was measured periodically with a thermo-hygrometer during the test period. When the moisture absorbing ability of the desiccant decreased during the test period, the desiccant was replaced as needed to maintain the humidity in the container at 90 to 95 %RH as much as possible. As a comparative test, a test without a desiccant in the container was also conducted. In the comparative test, the humidity in the container remained at almost 100 %RH.
Fig. 4 shows the schematic diagram of the C-ring test under low oxygen conditions. The basic configuration is the same as in Fig. 3, but a low oxygen environment has been created by replacing the air in the test container with nitrogen. The air in the container was replaced with nitrogen by opening a gas outlet on the lid and injecting high purity nitrogen (99.999 %N2) through the gas inlet on the lid. During nitrogen degassing, the tip of the tube attached to the outlet was placed in a beaker of water to confirm that air bubbles continued to be produced. After replacing the air in the container with nitrogen, the tube attached to the inlet and outlet was closed with a roller clamp to seal the container and maintain a low oxygen state.
Fig. 4 Schematic diagram of the test under a low oxygen concentration environment
To investigate the effects of phosphorous concentration on the material and temper of the material on the occurrence of SCC, we used three types of specimens (HP copper, LP copper, and OF copper), and H-tempered copper, 1/2H-tempered copper, and O-tempered copper for each sample, for a total of nine types of C-ring specimens for the experiment. The stresses applied to the specimens were about 300 MPa for H-tempered copper, about 200 MPa for 1/2H-tempered copper, and about 75 MPa for O-tempered copper, so that all of the applied stresses were within the elastic range. Two types of test water, 1 % and 3 % ammonia water, were used for the exposure environment, and the exposure period was 3 weeks.
Fig. 5 shows the sectional view of the sample exposed to an ammonia atmosphere volatilized from1 % ammonia water. The sections shown in Figs. 5 and 6 below are all at the locations marked with * in Fig. 2. In HP copper and LP copper, intergranular corrosion was observed on the surface of 1/2H-tempered copper and O-tempered copper, and slight rough surface corrosion was observed on the surface of H-tempered copper. On the other hand, in OF copper, no intergranular corrosion was observed, except for slight rough surface corrosion in H-tempered copper, 1/2H-tempered copper, and O-tempered copper.
Fig. 5 Sectional micrographs of each sample exposed to the gas-phase of 1 % ammonia water for 3 weeks (load stress: H-tempered copper = about 300 MPa, 1/2H-tempered copper = about 200 MPa, O-tempered copper = about 75 MPa)
Fig. 6 shows the sectional observation of specimen exposed to an ammonia atmosphere volatilized from 3 % ammonia water. The state of corrosion occurrence was the same as in the case of 1 % ammonia water, as shown in Fig. 5. That is, intergranular corrosion occurred in the 1/2H-tempered and O-tempered HP copper and LP copper, and only rough surface corrosion was observed in the remaining conditions. In Fig. 6, clear cracking in the depth direction was observed in the section of the 1/2H-tempered HP copper (middle of the upper row). It can also be seen that this cracking propagates along the crystal grain boundary. It was found that the stress corrosion cracking of pure copper that occurs under the test conditions of this study is Intergranular Stress Corrosion Cracking (IGSCC).
Fig. 6 Sectional micrographs of each sample exposed to the gas-phase of 3 % ammonia water for 3 weeks (load stress: H-tempered copper = about 300 MPa, 1/2H-tempered copper = about 200 MPa, O-tempered copper = about 75 MPa)
The sections shown in Fig. 5 and 6 are both observed at the point of maximum tensile stress on the outer surface of the specimen in the C-ring specimen, but the intergranular corrosion observed in 1/2H-tempered and O-tempered HP copper and LP copper also occurred at other locations. However, the intergranular corrosion in the sections shown in Figs. 5 and 6 progressed in the thickness direction of the tube more than the intergranular corrosion occurring at other locations. Therefore, the results suggest that the intergranular corrosion propagates in the thickness direction of the tube when stress is applied, and this intergranular corrosion progresses to SCC.
The sectional observation results shown in Figs. 5 and 6 indicate that intergranular corrosion occurred in the 1/2H-tempered and O-tempered HP copper and LP copper containing phosphorous, while no intergranular corrosion occurred in all tempers of the OF copper containing no phosphorous. This suggests that phosphorous contained in pure copper is involved in intergranular corrosion.
On the other hand, intergranular corrosion was not observed in the H-tempered specimen containing phosphorous. In order to investigate the state of the crystal grains of each temper, microscopic observation of the etched HP was carried out. The sectional observations of etched HP copper are shown in Fig. 7. In the 1/2H-tempered copper and O-tempered copper, the crystal grain size is about ten μm, whereas in the H-tempered copper, the crystal grains are flattened and the grain size is about several μm, which is smaller than that of the 1/2H-tempered copper and O-tempered copper. The reason why intergranular corrosion did not occur in the H-tempered copper may be because the crystal grains were finer than in the 1/2H-tempered copper and O-tempered copper. S. Sato et al7
）. also reported that the fracture time for SCC shortens as the grain size increases. The results in this study also showed that grain size affects intergranular corrosion and SCC susceptibility.
Fig. 7 Micrograph of etched HP copper
To investigate the effect of the stress applied to the specimens on the occurrence of SCC, we conducted 3-week exposure tests on two types of C-ring specimens, one made of 1/2H-tempered HP copper tube and 1/2H-tempered LP copper tube, with the stress applied to the specimens at 240 MPa or more which exceeds the yield strength (hereafter referred to as high stress condition) and about 70 MPa (hereafter referred to as low stress condition). The ammonia concentration in the test water was 1 % and 3 %. Fig. 8 shows the sectional observation of the sample after the test with 1 % ammonia water. Under both high and low stress conditions, intergranular corrosion was observed, albeit slightly, over the entire surface of both HP copper and LP copper. However, no intergranular corrosion progression in the thickness direction was observed in any of the samples.
Fig. 8 Sectional micrograph of 1/2H-tempered copper specimen exposed to the gas-phase of 1 % ammonia water for 3 weeks (load stress: high stress condition = 240 MPa or more, low stress condition = about 70 MPa)
Fig. 9 shows the sectional observation of the sample after the test with 3 % ammonia water. In all stress conditions and samples, distinct intergranular corrosion occurs near the surface. In particular, in the case of high stress HP copper, cracks developed from the tube surface in the thickness direction, and indicating that SCC occurred. Under low stress conditions, cracks did not propagate in the thickness direction, and it was confirmed that intergranular corrosion occurred severely near the surface. In particular, it was confirmed that some of the crystal grains near the surface had fallen off due to intergranular corrosion in HP copper.
Fig. 9 Sectional micrograph of 1/2H-tempered copper specimen exposed to the gas-phase of 3 % ammonia water for 3 weeks (load stress: high stress condition = 240 MPa or more, low stress condition = about 70 MPa)
In this experiment, SCC occurred only in the 1/2H-tempered HP copper with 3 % ammonia water under high stress conditions. Exposure of 1/2H-tempered HP copper to the same atmosphere did not result in SCC under low stress conditions, indicating that a loading stress above a certain value is required for SCC to occur. It was also found that SCC occurs when stress is applied above yield stress.
To investigate the effects of humidity and oxygen concentration in the corrosive environment on the occurrence of SCC, exposure tests were conducted in an environment where the humidity and oxygen concentration in the container were not controlled (hereafter referred to as high humidity), an environment where only the humidity in the container was reduced (hereafter referred to as low humidity), and an environment where only the oxygen concentration was reduced (hereafter referred to as low oxygen). The specimens were made of 1/2H-tempered HP copper, and the specimens were stressed under the high stress conditions. The test water was used 5% ammonia water. After the start of the exposure test, the surface of the specimens was visually observed on a daily basis. When a clear crack was visually confirmed, the test was terminated at that point and the specimen was removed from the container. If no cracks were found, the test was terminated after one week and the specimen was removed from the container.
Table 1 shows the time required to complete the test under the three conditions of high humidity, low humidity, and low oxygen, as well as the appearance of the specimen, humidity, oxygen concentration, and gas-phase ammonia concentration at the end of the test. Under the high humidity conditions, the surface of the specimen began to turn dark brown about 1 hour after the start of the test, and after 24 hours, the dark brown color became darker and clear cracks were observed on the surface of the specimen. At low humidity, the color changed to brown about 3 hours after the start of the test, and the color became darker as the test progressed. However, no cracks were visible on the surface of the specimen after 7 days, so the test was terminated after 7 days. Under the low oxygen conditions, discoloration took longer than under the high humidity conditions, but shorter than under the low humidity conditions. After 48 hours, cracks were visually observed on the surface of the specimen, and the test was terminated. The humidity inside the container at the end of the test was 100 %RH in both the high humidity and low oxygen conditions, and 96 %RH in the low humidity condition, which was lower than the other two conditions. It was confirmed that the oxygen concentration under both the high and low humidity conditions was about 20 %, which is the same as the normal atmospheric oxygen concentration, while the low oxygen concentration was maintained at a low value of 4.6 %. The ammonia concentration in the gas-phase was about 2 % under all three conditions, with no significant difference.
Table 1 Appearance photo of the specimen at the end of the test (material: HP, temper: 1/2H, stress: high stress condition = 240 MPa or more)
Fig. 10 shows the sectional observation of the specimens after the test under three conditions: high humidity, low humidity, and low oxygen. The specimens under high humidity and low oxygen conditions had clear cracks that propagated in the thickness direction of the tube. In addition, slight intergranular corrosion was also observed on the surface of the specimens under the two conditions. From the magnified observation of the cracks in the lower row of Fig. 10, cracks propagated along crystal grain boundaries under both high humidity and low oxygen conditions. The cracks that occurred in this study can also be identified as caused by IGSCC.
Fig. 10 Sectional micrograph of the specimen exposed to the gas-phase of 5 % ammonia water (humidity and oxygen concentration controlled) (material: HP, temper: 1/2H,load stress: high stress condition = 240 MPa or more)
On the other hand, no intergranular corrosion, or even cracking due to SCC, was observed in the low humidity conditions. From the sectional observation of Fig. 10, it was found that SCC occurs under high humidity and low oxygen conditions in the three conditions tested. However, cracks were observed on the surface after 24 hours under high humidity conditions, while cracks were observed after 48 hours under low oxygen conditions, and the time to crack initiation was longer under low oxygen conditions. Table 1 shows that humidity and ammonia concentration, which are environmental factors other than oxygen, are almost the same between high humidity and low oxygen, indicating that the longer time for crack initiation was due to the reduction in oxygen concentration.
The results of this experiment suggest that SCC will occur when the oxygen concentration in the container is reduced to 4.6 %, but it will take longer to occur. Regarding the effect of oxygen on SCC in copper materials, a report 8) on brass condenser tubes states that even if NH3 is contained in the condensate boiler water, SCC does not occur, if the O2 concentration is extremely low. Although the material used in this test was not brass but pure copper, the reduction in oxygen concentration delayed the occurrence of SCC, suggesting that, as in the case of brass, oxygen reduction has a suppressing effect on SCC. However, since SCC ultimately occurred at 4.6 % of oxygen concentration under low oxygen conditions in this test, as shown in Figure 10, it can be said that the oxygen concentration at which SCC does not occur is at least less than 4.6 %.
Under low humidity conditions, surface discoloration was slowed and, as shown in Fig. 10, no cracking or intergranular corrosion was observed even after one week. In this experiment, the humidity at the end of the test in low humidity conditions was 96 %RH. During the exposure period, the desiccant was replaced as needed to keep the humidity in the container at 90 to 95 %RH as much as possible, which means that the humidity was reduced by 5 to 10 % compared to the high humidity conditions. Table 1 shows that environmental factors other than humidity, such as oxygen concentration and gas-phase ammonia concentration, are almost the same at high and low humidity, indicating that humidity reduction is effective in reducing the occurrence of SCC. S. Sato et al7
）. conducted SCC tests on phosphorous deoxidized copper sheets in an ammoniacal environment adjusted to 70, 95, and 98 % relative humidity and reported that the time to cracking increased as the relative humidity decreased. Unlike S. Sato et al., the specimen used in this study was a C-ring test, but similar to the S. Sato et al. report, we obtained the result that humidity reduction is highly effective in suppressing SCC.
From the sectional observation shown in Figs. 5 and 8, no SCC was observed in the specimens exposed to 1 % ammonia water for 3 weeks, even when the materials, tempers, and load conditions were changed. To investigate the effect of exposure time on SCC, an 8-week exposure test was conducted in the same environment using 1 % ammonia water. H-tempered and 1/2H-tempered HP copper and LP copper were used as specimens, and the stresses applied to the specimens were high stress conditions. Here, under the high stress condition of H-tempered copper, the specimen was subjected to a stress equal to or greater than 390 MPa, which is the yield stress of H-tempered copper tube.
Fig. 11 shows the sectional observation of the specimen exposed for 8 weeks. Slight rough surface corrosion was observed on the surface of the H-tempered HP copper and LP copper specimens. This corrosion condition was similar to the corrosion condition after 3 weeks of exposure shown in Figs. 5. On the other hand, intergranular corrosion was observed on the surface of the 1/2H-tempered HP copper and LP copper specimens, which was similar to the corrosion conditions shown in Figs. 5 and 8 after 3 weeks of exposure.
Fig. 11 Sectional micrograph of the specimen exposed to the gas-phase of 1 % ammonia water for 8 weeks (load stress: H-tempered copper high stress condition = 390 MPa or more, 1/2H-tempered copper high stress condition = about 240 MPa)
However, as shown in Fig. 12, cracks were observed in the thickness direction of the tube for the 1/2H-tempered HP copper. The depth of this crack was about 0.45 mm for a tube thickness of 0.8 mm. Since SCC was observed in 1/2H-tempered HP copper after 8 weeks of exposure to an atmosphere with 1 % ammonia water, we conducted a 24-week exposure test in an environment with 0.1 % ammonia water, using H-tempered and 1/2H-tempered HP copper and LP copper as specimens under high stress conditions. Fig. 13 shows the sectional observation of the exposed specimen. Regardless of the temper, only slight rough surface corrosion was observed on the surface of HP copper and LP copper specimens, and no intergranular corrosion was observed in the 1/2-tempered HP copper and LP copper specimens, although intergranular corrosion was observed in Fig. 11.
Fig. 12 Sectional micrograph of 1/2H-tempered HP copper specimen exposed to the gas-phase of 1 % ammonia water for 8 weeks (load stress: 1/2H-tempered copper high stress condition = 240 MPa or more)
Fig. 13 Sectional micrograph of the specimen exposed to the gas-phase of 0.1 % ammonia water for 24 weeks (load stress: H-tempered copper high stress condition = 390 MPa or more, 1/2H-tempered copper high stress condition = about 240 MPa)
As shown in Figs. 5, 6, and 11, when the concentration of ammonia water was 1 % and 3 %, only slight rough surface corrosion was observed on the surface of H-tempered HP copper and LP copper. Similarly, when the concentration of ammonia water was 0.1 %, only slight rough surface corrosion was observed on the exposed surfaces, even when the exposure period was 24 weeks. From these results, it suggests that the corrosion of the material surface remains slight, even after exposure of H-tempered HP copper and LP copper to an environment with less than 0.1 % ammonia water for more than 24 weeks.
As shown in Figs. 5, 6, 8, 9, 11, and 12, for 1/2H-tempered HP copper and LP copper, intergranular corrosion and SCC were observed on the surface of the specimens after exposure when the ammonia water concentration was 1 % and 3 %. On the other hand, when the concentration of ammonia water was 0.1 %, intergranular corrosion was not observed on the surface after the exposure, as was the case with H-tempered copper, and only slight rough surface corrosion was observed even when the exposure period was 24 weeks. From these results, it suggests that the corrosion of the material surface remains slight, even after exposure of 1/2H-tempered HP copper and LP copper to an exposure environment with less than 0.1 % ammonia water for more than 24 weeks.
In order to use HP copper for heat transfer tubes of absorption chillers, it was found necessary to reduce the concentration of ammonia in the water, which is the refrigerant injected into the machine, and in the lithium bromide solution, which is the absorbent. In particular, the occurrence of stress corrosion cracking can be suppressed by reducing the ammonia concentration to less than 0.1 %.
For the degree of workability of copper tubes, it is preferable to use H-tempered copper, which showed only rough surface corrosion under the same exposure conditions, rather than 1/2H-tempered copper or O-tempered copper, which showed intergranular corrosion and SCC. It was also found to be effective in keeping the residual stress of the copper tube low, as SCC was suppressed at lower stress levels in the same atmosphere.
Since the inside of a chiller is filled with saturated vapor, it is not possible to reduce the humidity below 100 %RH to suppress the occurrence of SCC. However, since the inside of a chiller is usually kept in a vacuum (oxygen is released to the outside), it is an environment where SCC is less likely to occur due to a decrease in oxygen concentration.
1） For example, Y. Kitamura and T. Suzuki: Boshoku Gijutsu, 2nd edition, Chijin Shokan (2009), p. 44.
2） K. Nagata and S. Sato, Sumitomo light metal technical reports, 24 (1983), p. 97-107
3） M. Sakai, M. Kiya, T. Irie, H. Yakuwa, Zairyo-to-Kankyo, 65 (2016), p. 138-142.
4） M. Sakai, M. Kiya, T. Irie, H. Yakuwa, Zairyo-to-Kankyo, 65 (2016), p. 494-497.
5） M. Sakai, M. Kumagai, A. Osaka, T. Irie, H. Yakuwa, Zairyo-to-Kankyo, 67 (2018), p. 168-171.
6） ASTM G38, “Standard Practice for Making and Using C-Ring Stress-Corrosion Test Specimens”, (1995).
7） S. Sato and K. Nagata, Journal of the JCBRA, 17 (1978), p. 202.
8） H. H. Uhlig and R. W. Revie, Corrosion and Corrosion Control 3rd (Jpn.), Sangyotosho (1989), p. 346.