Communication—A Method to Measure Extremely Low Corrosion Rates of Copper Metal in Anoxic Aqueous Media

Corrosion rate calculations

It is assumed that any corrosion of copper in anoxic aqueous media involves the production of hydrogen gas, according to the acidic and alkaline cathodic reactions, respectively:

By accurately measuring the rate of H₂ release over a series of monthly measurements, a bounding maximum is determined for the corresponding oxidation of copper, according to the anodic reaction:

This assumes that: i) the only source of H₂ is copper corrosion and ii) copper oxidizes by a one-electron process. These are very conservative assumptions that may lead to overestimation of the true corrosion rate. Many forms of copper contain trapped hydrogen, which could significantly affect the measurements; thus the assumption that all hydrogen arises from corrosion is particularly conservative.¹¹ It must be stressed that the method in this work provides an upper limit to the corrosion rate, and that the true rate will approach zero, as per the thermodynamic arguments. Nonetheless, a firm maximum value holds practical value for UFC safety predictions.

From the measured moles of H₂, n_H2, the known surface area of copper, A_Cu, and the measurement time interval, Δt, the corresponding maximum corrosion rate is calculated using the molar volume, V_m,Cu, of copper:

Cell design

Two cell designs were used, shown in Figures 1a and 1b. Glass was used where possible, rather than stainless steel, to avoid the possibility of the outgassing of H₂ and/or its evolution from corrosion.²² Top components were fitted with two ports for the introduction and exchange of solutions and headspace gas. Cells were charged with 5 cm segments of copper wire (≥99.9%, temper annealed, 250 μm diameter, Goodfellow, UK). Each cell contained 2,600 segments, corresponding to 1,024 cm² total surface area. Notably, no effort was made to control impurities in the copper or to quantify the level of hydrogen present in the copper wires prior to immersion, which could contribute to the measured H₂ and artificially inflate calculated corrosion rates. The wires were rinsed with acetone, high purity water (18.2 MΩ cm), 1 M HNO_3(aq), high purity water again, and then cleaned by ultrasonication in absolute ethyl alcohol. All solutions were prepared with high purity water and were deaerated with certified ultra-high purity N₂. Salt solutions were prepared with reagent grade NaCl which may contain trace impurities and therefore may have contributed to hydrogen production. Temperature control was performed using a water bath, where values were selected based on thermal modelling of a repository, which sees a brief thermal elevation above 75°C that lasts < 200 a, before dropping below 50°C then 30°C after ∼7,500 and ∼20,000 a, respectively.²³

Zoom In Zoom Out Reset image size

Molecular hydrogen quantification

The H₂ production rate from copper in anoxic pure water was measured in duplicate cells at 75°C (Figure 2). For both of the cells, the H₂ production rates were low, resulting in the equivalent of approximately 1 nm of cumulative corrosion after many months. In addition, for the cell where the initial rate was higher (Blue circles, Figure 2), hydrogen production appeared to be leveling off. Both of these demonstrate that bounding corrosion rates are significantly lower than some estimates of the copper corrosion rate in anoxic water, which range as high as ∼1 nm d⁻¹.⁸

Zoom In Zoom Out Reset image size

In contrast, these results suggest that if copper corrosion occurs in pure water, it is at very slow initial rates (≤ 2 nm a⁻¹) that may decrease over time. Yet for nuclear waste management, corrosion rates in pure water are more academic than they are of actual practical use as containers would be exposed to groundwaters that contain dissolved minerals.² Of particular concern are chloride anions, from dissolved halite, that could form solution-soluble complexes with copper ions. Therefore, the measurements were repeated using aqueous sodium chloride solutions.

Copper exposed to 0.25 mol kg⁻¹ NaCl_(aq) at 30°C or 50°C produced negligible amounts of H₂ that was below the detection limit for all except one transient measurement (Figure 3, red diamonds). At 75°C, H₂ was produced for the duration of the six month experiment to produce an equivalent damage of <5 nm. Similarly, copper exposed to 2.5 mol kg⁻¹ NaCl_(aq) at 50°C produced no detectable H₂ (Figure 3, blue circles), while the 6 month exposure at 75°C, saw H₂ released at a rate corresponding to 5 nm of cumulative copper corrosion. This experiment also demonstrated that when the temperature was returned to 50°C, the H₂ production immediately returned below the detection limit.

Zoom In Zoom Out Reset image size

Finally, copper was exposed to 5 mol kg⁻¹ NaCl_(aq) solution at 75°C (Figure 4). There was nearly an order of magnitude variation in the H₂ production rate between duplicate cells at the start of the experiment. Although these cells contained copper from different production batches, the manufacturer's specifications were identical for the two spools, but this does likely reveal the importance of the starting condition of copper, and the hydrogen trapped therein. The results indicate early spikes of H₂, particularly for one of the samples, but both cells saw leveling off of their H₂ production rates. While the cumulative damage exceeded 15 nm within 6 months for the higher hydrogen producing cell, it did not reach 18 nm after an additional 10 months. For the other cell, reduction of the temperature to 50°C saw a drop in hydrogen release to below the detection limit. Future work is encouraged to perform additional replicates complemented with detailed materials analysis, such as comparing the initial and final hydrogen content of the copper metal itself.

Zoom In Zoom Out Reset image size

A new method is presented using a H₂ probe to establish bounding corrosion rates for copper metal in anoxic aqueous media. This approach allows the thermodynamics of the system to be disregarded for the purpose of establishing practical, data-based values to support the safety arguments for copper in nuclear waste containers. Such values have previously remained elusive. Preliminary results are shown for copper in pure water and buffered NaCl_(aq) solutions over the range 0.25–5.0 mol kg⁻¹. The results show excellent precision of the measurement itself and reveal an ability to capture short-lived spikes of hydrogen evolution throughout the multi-month experiment. Negligible hydrogen was detected below 50°C, the temperature below which repositories will persist after 7500 a. At 75°C, hydrogen evolution rates tend to decrease over time, even if occasional high transient values occur.

This work was funded by the Nuclear Waste Management Organization (NWMO).

W. J. Binns 0000-0003-2472-1251

P. G. Keech 0000-0001-8435-628X

D. S. Hall 0000-0001-9632-0399