Table of Contents
A research investigation of “ant nest” corrosion (ANC) on copper tube was conducted in terms of the variables of the corrosion potential and pH value in 103 ppm copper formate solution over 20 days. The paper presents the surface and cross-sectional observations and examines Cu2O and H2O as the stable chemical species produced. A Cannizzaro reaction as a disproportionation reaction from formic acid and a comproportionation reaction from the metallic copper tube and copper formate solution critically influenced the ANC mechanism. The paper also categorizes the ANC attack as a rapid reaction system from the electrochemical point of view by using a polarization resistance curve.
Global warming contributes dramatically to an increasing number of applications of air conditioners and refrigerators . Due to the many advantages of copper tube, especially its thermal conductivity, copper is commonly used in an air conditioner and refrigerator in a household or public building. Therefore, many researchers in the company and public research realms have been studying the problems of copper tubes such as corrosion leakage. Ant nest corrosion (ANC) on the copper tube is one of the most severe problems occuring in air conditioners these days. However, the ANC issue is important to solve to maintain the efficiency of the air conditioner. This issue also can contribute to the energy use that can influence global warming.
The ANC or formicary corrosion has corroded pitting hole traces of several micron size morphologies that are similar to the nest of ants. This typical corrosion was firstly reported by Edwards et al. in the 1970s. Characteristics of the ANC are: (1) the ANC can not be observed by bare eyes’ inspection because its size is about several microns; (2) forms of its corrosion are complicated; it looks like a tunnel and has a number of wormholes randomly in the thickness direction of the copper tube; (3) the corrosion rate might be fast; for example, the corrosion rate of the phosphorus deoxidized copper tube was shown as approximately 10 μm/day (about 0.3 mA cm−2) .
Carboxylic acids such as formic and acetic acid are the agent behind the mechanism of ANC. However, the form of corrosion, branching pits, and wormholes attack are depended on the compound agent in the environment. The formic acid atmosphere provokes ANC on the copper tube to be a more perplexing branch hole rather than the acetic acid atmosphere. In vapor test conditions, acetic acid has a higher corrosion rate than formic acid in copper tube . However, in some conditions, ANC phenomena have a closeness to stress corrosion cracking and stress cracking.
There are two kinds of copper tube commonly used in daily life; a phosphorus deoxidized copper tube (Cu-PDC) and an oxygen-free copper tube (Cu-OFC) without phosphorus. Sakai et al. conducted research on the ANC of both copper tubes in acetic acid solution, a formic acid solution. This found that phosphorus in copper is not a crucial factor for the generation of ANC. In addition, there are many research works on the prevention of ANC in copper tubes such as the development of surface treatments on copper tube. However, ANC still occurs in copper that has had surface treatment. The surface treatment only postpones the ANC attacking the copper tube.
The corrosion mechanism of ANC in copper has been reported by many researchers but the phenomena of ANC which is complicated makes necessary an explanation about the corrosion mechanism in more complex solutions, except acetic acid or formic acid. O. Seri et al. explain that in a neutral or alkali environment the reaction between copper and formic acid can produce copper formate. The proportionate reactions may be involved between metallic copper with copper formate ions. Moreover, this can be used as a considering factor to explain the mechanism.
In order to understand the underlying behavior of the ANC, this research is focused on the corrosive environment of copper formate solution. The reasons why copper formate solution is employed as a test solution are: (1) copper reacts with formic acid to form copper formate as a corrosion product; therefore, the pitting cavities will be occupied with the copper formate solution; (2) the replacement of the pit solution to the bulk solution is complicated due to the narrow pitting mouth of the ANC; therefore, the copper formate solution probably resides in the pitting cavities; and (3) the comproportionation reaction between a copper formate and metallic copper is indispensable for the explanation of the ANC that will be discussed later.
2. Materials and Methods
2.1. Test Material
The sample specimen was commercial phosphorus deoxidized copper tube (JIS1220, UACJ Cooperation, Tokyo, Japan, ∅15 mm × 0.7 mm, Cu ≥ 99.90%, P: 0.015~0.040%). All samples were rinsed with acetone as a pretreatment and then washed with deionized water. All the specimens were air-dried before the test.
2.2. Test Solution
Test solutions were copper formate (98 wt % Cu(COOH)2, Wako Pure Chemical Ltd., Tokyo, Japan) as a solvent of the ion-exchanged water. The test solution was adjusted to 103 ppm Cu(COOH)2 concentration. The solution temperature was room temperature (about 298 K). In the initial condition, the solution pH was 5.5, electric conductivity κ was 86 mS/m, and the dissolved oxygen (DO) was around 5 ppm. For altering the pH 5.5 of solution to be pH 3, reagent grade 98 wt % formic acid (HCOOH, Wako Pure Chemical Ltd., Tokyo, Japan) was added.
2.3. Observation Method
Copper tube specimen was immersed in 103 ppm Cu(COOH)2 solutions, respectively, of pH 5.5 and 3 (added HCOOH). The sample immersed in the test solutions was periodically observed at 0 day, 5 days, 10 days, 15 days, 20 days and 30 days after cleaning by ion-exchanged water and air-dried. The surface observation was carried out using a Microscope Lasertec Optelics Hybrid (Lasertec L3SMZ, Lasertec cooperation, Tokyo, Japan). In a cross-section observation, a part of the sample was cut, embedded in the epoxy resin, then polished with the fine polishing machine (Polis IMT-P2, IMT Cooperation, Tokyo, Japan) and finally using microscope BX51M-33MB (Olympus cooperation, Tokyo, Japan). The cross-section observation was carried out whether the ANC occurs or not. The areas of copper tube attacked by ANC were investigated then with an electron probe micro-analyzer (EPMA) (JEOL JXA8900R, JEOL Ltd., Tokyo, Japan).
2.4. Measurement Method
An electrochemical measurement system (HZ7000 Hokuto Denko Ltd., Tokyo, Japan) was employed for polarization measurements. The scan rate was 0.1 mV/s. The surface of the copper specimen was masked with insulating tape and silicon resin, except for the exposed surface area of 5.6 cm2. As a reference electrode, an Ag/AgCl electrode (DKK-TOA Co., Tokyo, Japan) in the saturated potassium chloride solution was used. In this paper, the electrode potential related to the Ag/AgCl reference electrode (V vs. SSE) was simplified as V unless otherwise noted.
3.1. Specimen Observation
Figure 1 shows the results of surface and cross-section observation of copper tube in 103 ppm Cu(COOH)2 solution with periodic time observation. The surface colors of the copper tube in the visual observation were reddish brown, reddish purple, or black as presented in the optical microscope of the left table of Figure 1. It took 10 days until all of the surfaces were covered by corrosion. The small stain black holes were found on the copper tube surface after being immersed for 10 days as shown in the white circle in Figure 2. It was assumed as the initial stage of the pitting attack of ANC which was confirmed by the cross-section observation. The embryo of ANC started after 10 days’ immersion, and the ANC was obviously observed on the copper tube after being immersed for 20 days. Therefore, the measurement of natural corrosion potential and the polarization curve were done after 20 days. Increasing immersion time increased the number of ANC on the copper tube.
In order to study the influence of pH on the copper tube in 103 ppm Cu(COOH)2 solution, the copper tube was immersed in 103 ppm Cu(COOH)2 solution that was adjusted to pH = 3 using HCOOH. Figure 3 shows the surface and cross-section observation of the copper tube in 103 ppm Cu(COOH)2 solution adjusted to pH = 3 with periodical observations. Compared to pH = 5.5; the corrosion occurred more prominently on the surface of the copper tube in this condition. The surface of the copper was covered by the corrosion after 20 days immersion and spread on the surface rather than a pitting attack into the copper tube specimen, as confirmed by the cross-section observation. The embryo of the ANC started after 20 days’ immersion, and the ANC was obviously observed on the copper tube after being immersed for 30 days.
Figure 4 shows the result of a scanning electron microscope (SEM) picture and EPMA analysis of the cross-section of the copper tube that was immersed in 103 ppm Cu(COOH)2 solution with pH 5.5 after 15 days immersion time. This EPMA analysis was conducted to detect the presence of copper, oxygen, and phosphorus elements. Results show that the presence of phosphorus was at a low level in all areas. However, the presence of copper was almost in all areas and was higher in the border of the pitting hole, whereas the presence of oxygen was only detected in the pitting hole of the copper tube and the surface. This EPMA analysis result shows that copper and oxygen had an important role in the pitting attack. In addition, the EPMA analysis was also conducted on the copper tube after 30 days immersion time to detect the level differences of copper and oxygen elements on the copper tube as shown in Figure 5. As a result, It was found that oxygen was detected at a low level in the pitting hole after 30 days immersion time compared to 15 days. Therefore, after long immersion periods of copper tube it was concluded that the oxygen found it difficult to exist deep in the pitting hole.
3.2. Ecorr, pH-Time
The monitoring results of Ecorr and pH of copper in 103 ppm Cu(COOH)2 solution with pH 5.5 and 3 are shown in Figure 6 and Figure 7. Ecorr for pH 5.5 was in the range 0.01 to 0.08 V vs. SSE for 20 days immersion time. Ecorr started from 0.01 V vs. SSE and then increased rapidly to 0.06 V vs. SSE in 3 days and after that was almost constant in the range 0.06 to 0.08 V vs. SSE. The solution pH measurement shows that at first immersion time the solution pH was 5.58 and it decreased to 4.59 in 3 days, then after that it increased slowly until the end of the measurement time to 5.29. Ecorr for pH solution 3 was in the range 0.04 to 0.11 V for 20 days immersion time. Ecorr started from 0.04, then increased to 0.1 in 3 days immersion time, and decreased slowly to reach 0.07 V until the end of the measurement. The solution pH measurement shows that at the first immersion time the solution pH was 3 and continuously increased to 3.85 until the end of the measurement time.