1. Introduction

A dual-structure disposal canister consisting of inner carbon steel with excellent mechanical rigidity and outer copper with excellent corrosion resistance is being developed in several countries such as Sweden, Finland, and Canada [1] for the disposal of the spent nuclear fuel in a deep geologic place. KAERI (Korea Atomic Energy Research Institute) is also currently considering coppercast iron dual-structure canister as a prominent option [2].

Copper was believed that it proceeds only with uniform corrosion at very low rate without any local corrosion in the deep geologic environment so far, and so it is evaluated that several millimeter-thick copper canisters can withstand its leak-tightness for a long period of disposal [1,3]. However, if local corrosion progresses, the leak-tightness could be broken sooner than expected, because according to previous studies examining crevice corrosion [4], pitting [5,6], and SCC (Stress Corrosion Cracking) [7] of copper, its local corrosion can be possible in a specific environment. Moreover, it is not certain how the deep geologic environment will change over the long period of disposal by unexpected crustal movements. In addition, non-uniform corrosion is also expected at the welded joints of disposal canister. For this reason, when designing the disposal canister, the thickness of the copper layer is generally set about 2 to 5 times thicker in consideration of safety [3,8]. In this aspect, if the concerns about non-uniform corrosion of disposal containers that may occur during the long-term disposal period can be resolved, the long-term reliability of the copper disposal canister can be enhanced, and the thickness of the copper layer can be diminished.

In this paper, even if copper is locally peeled off due to local corrosion or uneven corrosion earlier than expected, a solution of preventing severe galvanic corrosion of carbon steel was suggested by inserting more galvanic inactive metal than copper between copper and cast iron layers. The insertion metal may be selectively applied through coating method on a cast nodular iron surface where local corrosion is concerned or structural stress concentrated, so that the efficiency may be enhanced.

2. Experiment

2.1 Test Specimens

Some copper-insert double layers were formed on a carbon steel plate using a cold spray coating method. As insert, nickel, titanium, and silver with lower galvanic activity than copper were selected [9]. Silver is thermodynamically more stable than copper and so it has a higher corrosion potential than copper. Nickel and titanium have a lower corrosion potential than copper and so those are thermodynamically disadvantageous in corrosion [10]. However, TiO₂ formed by oxidation is an effective passive oxide, so its galvanic activity is lower than that of copper [11]. In addition, nickel has a crystal plane of FCC (Face Center Cubic) structure, which is known to be difficult for oxygen to approach on it, and resulting in strong corrosion resistance [12].

Four types of cold spray coating were formed on a carbon steel (SS275) plate (100 × 100 mm). First, insert coating of silver, nickel, or titanium with 2 mm thick were formed on a carbon steel plate, and then 5 mm thick copper coating was stacked on each insert coating plates and a bare carbon steel plate. Fig. 1 shows the side view of four kinds of cold spray coating plates manufactured.

Fig. 1

Lateral view of 4 kinds of cold spray coating plates; from top Cu- Ti, Cu-Ni, Cu-Ag, and Cu on carbon steel plate.

In order to measure the corrosion potential of each coating layer, a specimen was machined into a coin shape in 15 mm diameter as shown in Fig. 2(a). For the galvanic corrosion test, the central part of coin shaped specimen was excavated in cone shaped hole to reveal the insert in 5 mm diameter as shown in Fig. 2(b). In Fig. 2(b), the disclosed copper area was about 0.971 cm², and the exposed area of the central insert was 0.196 cm², so the area ratio is about 5 vs. 1.

Fig. 2

Electrochemical specimens of (a) single metal surface and (b) twin metal surface.

2.2 Electrochemical Test

In order to measure the corrosion potential of each coating layer, a polarization test was performed using an electrochemical cell. The reference electrode was SCE (Standard Calomel Electrode), the counter electrode was platinum wire, and the test specimen was used as a working electrode. KURT groundwater, which have pH 7.93 and DO 5.31 mg·L⁻¹, was used as a test solution. The exposure area of the working electrode was about 1.0 cm². The measured temperature was 25℃, at which the reference potential of the SCE is about E= +0.244 V [13]. The electrochemical cell was equipped in a potentiostat (SP-300, BioLogic). Before the test, the electrochemical cell was stabilized at OCP (Open Circuit Potential) state for 1 hour, and then scanned at 1 mV·sec⁻¹ scan rate in the range of −400 mV to + 600 mV based on OCP. The corrosion potential of each specimen was derived from a Tafel plot in the measured polarization curve. Polarization tests were done three times at same condition for the reproducibility. The corrosion potential measurement was also performed for the galvanic specimens in Fig. 2(b).

To determine the long-term corrosion behavior of galvanic specimens, electrostatic voltammetric tests were performed on galvanic specimens in the KURT field. To look at apparent corrosion behavior in the short term, a voltage of 1.0 V, which is much higher than the corrosion potential of copper (approx. −0.2 V), but lower than the electrolysis voltage of water (approx. +1.23 V), was applied to the electrochemical cell and the current change was measured for one week. To keep the water quality in the corrosion cell constant during the long-term test, KURT groundwater was continuously flowed into the corrosion cell. A schematic diagram of the electrostatic voltammetry test is shown in Fig. 3.

Fig. 3

Schematic diagram of electrostatic voltammetry test at KURT; 1: Potentiostat, 2: Working electrode, 2-1: Real view of 2, 3: Reference electrode, 4: Counter electrode, 5-1: Inlet of KURT water, 5-2: Outlet of KURT water.

2.3 Surface Observations

The corrosion specimen from the electrostatic voltammetry test was observed with an optical microscope. Wirecut electric discharge machining was applied to sectioning the galvanic specimens. After the sectioning, the corrosion pattern of the interlayer was investigated. In addition, SEM (Scanning Electron Microscope) observations and EDAX (Energy Dispersive X-ray Spectroscopy) analysis were performed to evaluate corrosion products.

3. Experimental Results and Discussions

3.1 Corrosion Potentials

Table 1 shows the OCP, corrosion potential, and corrosion current for the single and galvanic specimens. The values of the corrosion potential for single specimens are in the order of Fe < Ni < Cu < Ti < Ag. Iron (Fe) was the most sensitive to corrosion, and silver (Ag) was the least sensitive to corrosion. But nickel (Ni) was found to be more vulnerable to corrosion than copper (Cu) in a view of corrosion potential. But looking at the corrosion current, it was certain that nickel have a slower corrosion rate than copper since corrosion current of nickel showed lower values than that of copper like silver and titanium (Ti).

Fig. 4 shows the polarization curves of the galvanic specimens. As shown in Fig. 4, the galvanic specimens showed only one corrosion potential within the test range of −400 mV to + 600 mV. The measured corrosion currents and corrosion potentials derived from Tafel plot are listed in Table 1. The corrosion potentials for galvanic specimens except Cu-Fe was higher than that of copper only. Therefore, it was conceived that copper was likely to be corroded when silver, nickel, and titanium were added. It can be seen that the most specimens except Cu-Fe were slightly lower than copper in the corrosion current. From this, it was assumed that corrosion may occur on the copper mainly in these galvanic specimens. On the other hand, Cu-Fe showed much higher corrosion current than copper, so it was considered that iron may more sensitive to corrosion than copper. From the electrochemical tests, it was evaluated that the copper may be more active in corrosion than silver, nickel, and titanium in the galvanic specimens.

Fig. 4

Electrochemical polarization curves for galvanic specimens in KURT groundwater.

3.2 Electrostatic Voltammetry Test

Four galvanic specimens of Cu-Ag, Cu-Ni, Cu-Ti, and Cu-Fe were used for the electrochemically accelerated corrosion test applying 1.0 V to the electrochemical cell as shown in Fig. 3, and the current changes for a week is shown in Fig. 5. In Fig. 5(a), Cu-Ag stabilized at a constant current at first and then fluctuated around 0.108 mA. The Cu-Ni also fluctuated around 0.104 mA, and the Cu-Ti fluctuated around 0.110 mA as shown in Figs. 5(b) and 5(c). This fluctuation was interpreted as a repetitive action of the copper oxide film formation and falling off. Copper oxide film may form a passive barrier, but fall off at some growth by volume expansion, and then new oxide film forms on it again. In the electrochemically accelerated corrosion test of Cu-Ag, Cu-Ni, and Cu-Ti, the overall current values showed no linear increase for test period. However, in the Cu-Fe specimen of Fig. 5(d), the corrosion current typically increased linearly for the test period. This corrosion behavior was attributed to an increase in specific surface area as the surface becomes rough. Consequently, it was thought that iron oxide layer did not play a protection barrier.

Fig. 5

Current variation curves of galvanic corrosion specimens from one-week static voltammetry test.

3.3 Morphological Observations

Four galvanic specimens after a week of electrochemically accelerated corrosion tests are shown in Fig. 6. There were not any signs of corrosion on the silver and nickel. Some green patina was seen on the nickel surface, but that was due to the corrosion of copper residue which was not removed clearly during specimen preparation. But in case of titanium, some white corrosion product (presumably TiO₂) was observed around the interface. In these galvanic specimens, it was found that a lot of green patina and pits was formed on copper. In addition, black oxide (presumably CuO) was observed under the green patina. So it was verified that the galvanic corrosion occurred mainly on copper side in Cu-Ag, Cu-Ni, Cu-Ti specimens. Meanwhile, light red rust presumed to be Cu(I) was formed on the copper of Cu-Fe specimen, and no green patina and pits were seen. In case of carbon steel, a black rust was formed thickly on the surface. So it was verified that galvanic corrosion occurred mainly on iron side in Cu-Fe specimen.

Fig. 6

Galvanic corrosion specimens after one-week static voltammetry test.

The rust of the galvanic specimen was ultrasonically removed in the 2.5% aqueous HCl solution. The rustremoved specimens are shown in Fig. 7. The white oxide (presumably TiO₂), on titanium was clearly removed, and so it seemed that the white rust was poorly adhered on titanium surface. On all Cu surface, there were traces of pitting corrosion except Fig. 7(d) Cu-Fe. But the copper surface showed no sign of any corrosion feature in Fig. 7(d) Cu-Fe. Instead, the crown-bottle cap shaded corrosion boundary was found at the edge side of iron. So it was considered that some copper was damaged at the boundary line between copper and iron.

Fig. 7

Rust removed galvanic corrosion specimens of Fig. 6.

Fig. 8 shows the cross sections of galvanic specimens by a wire-cut electric discharge machining. It was confirmed that the boundary between dissimilar metals made by cold spray coating was well bonded without any cracks or pores. Fitting marks can also be seen along the copper cutting line and copper surface of Cu-Ag, Cu-Ni, and Cu-Ti, but there were no clear corrosion signs on the silver, nickel, and titanium cutting lines. However, in the case of Cu-Fe, it was confirmed that carbon steel was heavily corroded, and making a deep eave at the boundary to copper. On the other hand, in the case of Cu-Ti, even though white rust observed on the surface, but there was no sign of titanium corrosion around the boundary. From this cross section observations, it was believed that silver, nickel, and titanium were inactive to corrosion in their galvanic specimens with copper.

Fig. 8

Cross sections of two metal boundary of the corroded galvanic specimens in Fig. 7; P: pitting mark, E: eaves.

Fig. 9 shows the magnified SEM image of the green patina on copper in Fig. 6(b), in which numerous needle shaped crystals of hundreds nm long were seen. The needle shaped crystals were commonly observed in all galvanic specimens of Cu-Ag, Cu-Ni, and Cu-Ti. It is reported that the needle shaped crystal is typical form of Cu(OH)₂ or Cu(OH)₂-CuO, and grow on the copper surface by electrical anodization [14,15]. As an elemental analysis of EDAX (Energy Dispersive X-ray microanalysis) for the needle shaped crystals, it was verified that the crystals were composed of oxygen 42.5at%, copper 55.4at%, and the others. After removal of the needle-like crystals using aqueous HCl solution, numerous pits of several hundred μm sized were found on copper as shown in Fig. 9(b).

Fig. 9

Magnified SEM image of (a) green patina and (b) pits after removal of green patina on copper.

The copper in Cu-Fe galvanic specimen was hardly corroded in the electrochemically accelerated corrosion test, but the carbon steel was heavily corroded. It was confirmed that a thick dark brown corrosion product covered the carbon steel surface as shown in the EDAX image of Fig. 10(a). As results of elemental analysis, the corrosion product was composed of oxygen 53.1at%, iron 30.2at%, copper 0.7at% and the others. Therefore, the main corrosion product was presumed to be Fe₃O₄. On the copper surface, copper oxide film was not found, but some iron oxide residue was found as shown in Fig. 10(b). So it was verified that galvanic corrosion occurred on carbon steel in Cu-Fe specimen.

Fig. 10

EDAX images of (a) cast iron and (b) copper in a Cu-Fe galvanic specimen.

Meanwhile, a Widmanstäthen pattern showing lamella grid crossing in three directions was disclosed on Ag surface in the corroded Cu-Ag galvanic specimen when removing the green patina on it, as shown in Fig. 11(a). Widmanstäthen pattern is known as a geometric pattern that occurs when a new metal phase is formed on an existing metal pattern [16-18]. In the case of Silver, the Widmanstäten pattern has been reported in aluminum-silver [17] and copper-silver alloys [18]. It is known that the formation of Widmanstäten pattern requires distinguished conditions that the crystal pattern could form separately. As a result of elemental analysis, the Widmanstäthen pattern was composed of copper 84.9at%, oxygen 13.3at%, and about silver 0.9at%, which means the main component was copper. It was considered that this copper pattern was formed by the reduction of copper cation on silver surface, which act as a cathode.

Fig. 11

Views of (a) Widmannstäthen pattern formed on Ag in Cu-Ag Galvani specimen, and (b) copper film on the Cu-Fe specimen formed after completion of corrosion test.

Another copper reduction phenomenon was found on iron surface of the corroded Cu-Fe galvanic specimen which was left untreated for a month after the completion of corrosion test. By removing the rust on that, a thin copper film was found covering all iron surface as shown in Fig. 11(b). This phenomenon was not observed right after the corrosion test as shown in Fig. 7(d). It was considered that the iron oxidation makes the reduction of copper cation to form the copper plate. However, this reduction behavior was not found on the Cu-Ti and Cu-Ni specimens. This reduction phenomenon seemed to be worth to give an attention. If the internal cast iron is exposed by the copper corrosion, and copper could be plated again on the cast iron, it will greatly help the corrosion resistance of the copper-cast iron canister. For this to happen, it seemed necessary that the copper corrosion product must remain on the canister.

4. Conclusions

To increase the corrosion resistance of the copper-cast iron double structured disposal container, inserting a galvanic inactive metal such as silver, nickel, and titanium between the copper and cast iron layer was proposed in this paper. The new metal insertions were made using cold spray coating technique, and some electrochemical corrosion tests were performed to evaluate the corrosion potentials and corrosion behavior of them.

As a result of the corrosion potential measurement in KURT groundwater, the corrosion potentials of the galvanic specimens were higher than that of copper only. Therefore, it was considered that copper is likely to be corroded when copper is present with silver, nickel, and titanium metals. Examining the corrosion surfaces of the galvanic specimen, which electrochemically corroded for one week at 1.0 V, all kind of galvanic specimens except Cu-Fe showed galvanic copper corrosion, and the insert metals like silver, nickel, and titanium were hardly corroded. On the contrary, copper was hardly corroded in the Cu-Fe galvanic specimen, and only iron was corroded. Examining the interface corrosion between copper and insert metal through cross-sectioning of the corroded galvanic specimens, a local corrosion at interface was not observed in Cu-Ag, Cu-Ni, and Cu-Ti, but Cu- Fe specimen exhibited severe corrosion of iron at the interface, and the corrosion front penetrated into the interface.

On the other hand, a Widmanstäthen pattern of copper was found on the silver surface of the Cu-Ag Galvani specimen after rust removal. The copper pattern seemed to be the result of cathodic reduction of copper cation on silver. In addition, such copper cation reduction was also observed for the iron surface of corroded Cu-Fe specimen. The copper cation reduction phenomenon was seemed to be beneficial to the corrosion resistance of a disposal canister.

From the electrochemically accelerated corrosion tests, it was proved that the insertion of an inactive metal layer can promote the corrosion of residual copper even if copper barrier gets penetrated, and may diminish the severe concern about the local corrosion of a copper canister. This galvanic insert may not need to be treated on the entire surface of cast iron canister, and could be applied effectively to the areas where partial corrosion is concerned like welded parts. As a follow-up study, a long-term field corrosion experiment has launched recently at KURT site to prove the galvanic corrosion resistance of some galvanic specimens, in which a pin-hole was made on a copper layer to simulate pitting corrosion.