Sie können Operatoren mit Ihrer Suchanfrage kombinieren, um diese noch präziser einzugrenzen. Klicken Sie auf den Suchoperator, um eine Erklärung seiner Funktionsweise anzuzeigen.
Findet Dokumente, in denen beide Begriffe in beliebiger Reihenfolge innerhalb von maximal n Worten zueinander stehen. Empfehlung: Wählen Sie zwischen 15 und 30 als maximale Wortanzahl (z.B. NEAR(hybrid, antrieb, 20)).
Findet Dokumente, in denen der Begriff in Wortvarianten vorkommt, wobei diese VOR, HINTER oder VOR und HINTER dem Suchbegriff anschließen können (z.B., leichtbau*, *leichtbau, *leichtbau*).
Diese Studie untersucht die Leistungsfähigkeit von sauerstofffreiem phosphordotiertem (OFE + P) und hochleitfähigem phosphordotiertem (HCP) Kupfer in ammoniakhaltigem Grundwasser, wobei der Schwerpunkt auf der Korrosionsbeständigkeit und der Anfälligkeit für Spannungsrisskorrosion (SCC) liegt. Die Forschung vergleicht das elektrochemische Verhalten, die Korrosionsraten und die SCC-Anfälligkeit beider Kupferqualitäten unter simulierten Endlagerbedingungen. Es untersucht auch die Wasserstoffaufnahme und ihre möglichen Auswirkungen auf die materielle Leistung. Die Studie kommt zu dem Schluss, dass HCP-Kupfer eine ähnliche Korrosionsbeständigkeit und SCC-Anfälligkeit gegenüber OFE + P aufweist, was auf sein Potenzial als alternatives Behältermaterial hindeutet. Darüber hinaus bieten molekulardynamische Simulationen Einblicke in die Wasserstoffdiffusion in Kupfer und zeigen, dass Diffusionskoeffizienten in polykristallinen Strukturen denen in perfekten Kristallen nahe kommen. Diese umfassende Analyse bietet wertvolle Erkenntnisse über die langfristige Leistung von Kupferlegierungen in der Endlagerung von Atommüll.
KI-Generiert
Diese Zusammenfassung des Fachinhalts wurde mit Hilfe von KI generiert.
Abstract
High-purity copper is used for high-level waste (HLW) canister material due to its mechanical strength and corrosion resistance. While oxygen-free phosphorous-doped copper (OFE + P) is the current standard material in Finland and Sweden, high-conductivity phosphorous-doped (HCP) copper has emerged as a potential alternative. This study aimed to evaluate the corrosion performance, stress corrosion cracking (SCC) susceptibility, and hydrogen uptake of HCP relative to OFE + P under ammonia-disturbed groundwater conditions. Both copper grades were subjected to a 3-month exposure in an autoclave containing simulated groundwater and 100 mg/L of ammonia at room temperature. Constant potentials were applied to shift surface potentials into the Cu₂O/CuO stability region, mimicking thermodynamic conditions associated with SCC. Hydrogen content was assessed by hot-melt mass spectroscopy and thermal desorption spectroscopy supported by molecular dynamics simulations of hydrogen diffusion. U-bend specimens, constantly polarized at − 50 mV versus saturated calomel electrode, were analysed using plasma-focused ion beam (PFIB) and electron backscatter diffraction (EBSD) to characterize microstructural degradation. Both materials presented very-low corrosion rates (0.2–0.4 µm/year), and no hydrogen uptake was detected. However, PFIB and EBSD results revealed fully oxidized, intergranular cracks penetrating up to 10 µm, along random grain boundaries. These shallow features were attributed to ammonia-induced degradation facilitated by localized strain, even under anoxic conditions and room temperature. HCP copper exhibited comparable corrosion behaviour and susceptibility to ammonia-induced intergranular attack to that of OFE + P copper. These results support the potential use of HCP copper as a viable alternative for HLW canisters in ammonia-containing environments.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Introduction
Finland and Sweden have chosen highly pure oxygen-free phosphorus-doped copper as the high-level waste (HLW) final disposal canister material (KBS-3 concept), and in Canada, electrodeposited or cold-sprayed copper coatings are being developed as corrosion barriers in their final disposal solution [1‐5]. The main advantages of copper as a nuclear waste canister material include sufficient mechanical properties and, most importantly, predictable general corrosion behaviour [1]. Copper impurities and microstructure significantly influence its properties, performance, and inspectability. For this reason, HLW containers follow the standard EN 1976:2012 (UNS C10100), which strictly limits the container material to high-purity copper grades (≥ 99.3% (w/w)), with additional stipulations for critical elements and impurities, e.g. O < 5 wt-ppm, P = 30–100 wt-ppm, H < 0.6 wt-ppm, and S < 8 wt-ppm [6]. Oxygen-free phosphorous-doped copper (OFE + P) fulfils the above criteria and is the current choice for the HLW canisters. Nonetheless, copper grades with less strict requirements, such as the high-conductivity phosphorous-doped (HCP) copper (UNS C10300) studied in this work, could enable a robust alternative for the OFE + P but require exhaustive testing to demonstrate equal performance.
Most research on copper for HLW disposal has focused on OFP-copper, but cost and availability considerations suggest that alternative grades, e.g. HCP, should also be evaluated. OFE + P production is more costly than HCP due to additional refining steps to eliminate oxygen content. Its ultra-high-purity requirements often limit the use of scrap feedstock, increasing reliance on primary sources and raising the environmental footprint per ton. Alternative copper grades offer several advantages, including lower cost, broader supplier availability, and better sustainability through recycled content. However, the long-term performance of HCP under repository-specific conditions has not been fully explored. The transition to alternative copper grades also requires deep understanding on their microstructural characteristics during annealing, age hardening, and stress relieving [7]. These treatments, as well as the presence of different minor impurities, may significantly affect the mechanical properties such as creep resistance, which are relevant in repository concepts using dual canister systems [1].
Anzeige
Stress corrosion cracking (SCC) is an unlikely but possible degradation mode for copper canister materials, requiring several prerequisite conditions to occur [8]. For SCC to take place, three conditions must be met: the container surface must be wet, the corrosion agents must be present in the near field at sufficient concentrations, and the material needs to be under sufficient tensile stress. Additionally, a uniform passive oxide layer must be formed within a certain potential-pH range, as SCC develops from the passive layer breakdown [9, 10]. In other words, SCC occurs when the corrosion potential and interfacial pH exceed certain threshold values. For pure copper, these thresholds have been defined above the Cu2O/CuO equilibrium line on a Pourbaix diagram (Fig. S1) [5, 11, 12]. The outcome of SCC is usually an intergranular or transgranular crack that propagates perpendicular to the load. Other types of cracking development can result from high-work-hardening and/or specific environments (e.g. presence of chlorides, concentrated ammoniacal solutions), which favour SCC due to the high density of dislocations at grain boundaries and breakdown of the passive layer, respectively [13, 14].
Ammonia (NH3) is a known SCC agent and can be introduced into the repository during the construction phase by residues from explosives, by human or microbial activities, and by gas-phase radiolysis of atmospheric N2 [15]. The two most proposed mechanisms of copper SCC in ammonia-containing environments are tarnish rupture and slip–anodic dissolution. Both mechanisms are related to the oxide layer formed on a strained surface of copper exposed to elevated pH (> 8–9) in oxic redox conditions with low temperatures and low Cl− concentration, but they differ in crack propagation mechanisms [11]. The thickness of the oxide determines the extent of crack progress in the tarnish rupture process. When the passive layer breaks, revealing the underlying metal, crack propagation only occurs if subsequent oxide layers form and fracture again [11, 16]. In the anodic dissolution mechanism, crack progress is restrained by film dissolution, stopping briefly upon repassivation at the crack tip before the film dissolves again [11, 17‐19].
To understand the full degradation risk, it is also essential to consider processes that may act in parallel with SCC. For instance, other processes can trigger or boost SCC, such as localized corrosion (pitting) at the crack tip in electrochemically active zones and hydrogen embrittlement, which results from the entry of atomic hydrogen into the metal [8, 13]. One way to quantify the temperature-dependent diffusion constant of hydrogen in copper and assess structural changes in crystallinity is through the analysis of atomic trajectories and energy distributions. Molecular dynamics (MD) simulations can provide relevant insights into a scenario with possible hydrogen embrittlement and to understand material performance in various thermal conditions, such as the repository environment. MD simulations enable the calculation of temperature‐dependent hydrogen diffusion coefficients in copper by analysing atomic trajectories and energy distributions [20]. Previous works on MD for H in Cu have shown to match experimental diffusivities across a wide temperature range [20]. This approach also captures hydrogen-induced lattice changes (such as the nucleation of H‐filled cavities) that can affect mechanical performance [21]. Here, we use MD to predict hydrogen mobility and microstructural evolution in copper at elevated temperatures.
Both SCC and hydrogen-related degradation depend heavily on the evolving redox conditions at the canister surface. The probability of SCC will rely on the evolving near-field environment during the final disposal. Laboratory studies and simulations suggest that trapped atmospheric oxygen in a bentonite-backfilled repository is gradually consumed within months, or a few years, at most [4]. This results in a corrosion potential (Ecorr) shift towards more negative values in the absence of O2, leading to lower corrosion rates compared to the oxic conditions [8]. Conceptually, the canister in contact with groundwater (at ambient temperature) will experience a potential window between − 390 and − 500 mV versus saturated calomel electrode (SCE) and pH between 7 and 9 (blue-dashed rectangle in Fig. S1) [11, 15]. Although ammonia is primarily recognized as an SCC agent under oxic conditions [5, 11], the objective of this work was to evaluate whether SCC can still be triggered in its presence under otherwise anoxic conditions by shifting the electrochemical potential into SCC-relevant regions. Our approach was based on narrowing down the window where SCC occurs in the presence of ammonia. This means that, even though the tests were conducted under anoxic conditions, a constant anodic current was applied to artificially raise the surface potential to values comparable to those in oxic environments. This allowed us to correlate the anoxic environment with the thermodynamic driving force of corrosion based on E-pH diagrams (Fig. S1). In doing so, we explored how different copper grades behave near the Cu(OH)₂/Cu₂O/CuO stability transition, where SCC of copper is known to initiate under controlled and simplified conditions.
Anzeige
Extensive research has demonstrated that SCC is unlikely to occur in Cu-OFE + P under several oxic and anoxic conditions. However, the behaviour of other copper grades, such as Cu-HCP, remains largely unexplored in these environments. The objective of this study was threefold: to examine the impact of ammonia on the degradation of copper, to compare the performance of OFE + P and HCP, and to clarify whether exposure to NH3-containing groundwater induces H-uptake. These results are expected to be highly beneficial for reducing the uncertainties related to final disposal via a better understanding of the degradation mechanisms and determining further research directions and requirements for copper materials. The work was mainly experimental in nature but also included a brief modelling part aiming to support the understanding of H-content in copper.
Experimental
Test materials and conditions
All experiments were performed for OFE + P and HCP with roughly similar P content but, most notably, HCP having significantly higher O content than OFE + P. The materials were provided by Posiva Oy, Finland. The plates were delivered in either hot-rolled (flat, F) or hot-rolled and 90° bent (bent, B) state. The main properties of the copper alloys examined in this study are compiled in Table 1.
Table 1
Properties of copper alloys investigated in this work [22]
The received material was subjected to metallographic sectioning for triplicate mass-loss tests (20 mm × 20 mm × 3 mm) and rectangular specimens (150 mm × 15 mm × 3 mm) for triplicate U-bends samples in which 180° deformation was induced with a 20 mm-radius loading pin, resulting in 15% total strain. Six specimens of each grade were employed in the H-content measurements (4 mm × 4 mm × 8 mm), and a single specimen of each grade was used in electrochemical assays (10 mm × 10 mm × 3 mm). Prior to immersion, the surfaces were metallographically polished using gradually finer silicon carbide papers up #500, followed by cleaning with acetone and ethanol. Micro-indentation tests (Vickers hardness) were performed on the as-received specimens. The samples’ hardnesses are indicated in Table 1 with HV1 mean values falling within a narrow range of 57.5–60.8 HV1, with overlapping standard deviations (Table 1). All values agree with the range specified by the manufacturer of 40–65 HV1 [22].
The grain size of the samples was investigated along with a detailed examination of their integrity (Fig. 1a, b), e.g. confirming the absence of grain boundary corrosion [23]. The mean grain sizes were calculated according to the ASTM E112-2010 using the intercept method. The analysis shows that all samples exhibit relatively coarse microstructures (Fig. 1b), with mean grain sizes ranging from approximately 94 to 152 µm (Fig. 1a, Table 1). HCP_B displayed the largest average grain size, while HCP_F had the smallest. The two OFE + P specimens showed similar mean grain sizes, with only a slight increase observed in the bent condition. Although these differences are measurable, they are not pronounced, and given that the metallographic evaluation was conducted on a limited area, the results may not be representative of the bulk material. The optical micrographs support the quantitative data, showing generally equiaxed and coarse grains across all conditions, with some local variation.
Figure 1
a Mean grain size and standard deviation for the HCP and OFE + P copper samples and b representative optical micrographs showing grain structures.
The corrosive environment of the copper canister evolves over time under the final disposal. The evolution is usually divided into five stages: initial state, early evolution, remaining temperature period, next permafrost and glaciation period, and long-term evolution (up to 106 years). The experimental conditions selected in this work mimic the stages following the early evolution phase (> 10,000 years after the closure). The environment is anoxic, and there is no microbial activity in the near field. The canister is under high isostatic load, causing plastic deformation and creep [5, 8]. If present and in contact with the surface, species such as halides, ammonia, carbonate, and phosphates can form complexes with Cu(I) and Cu(II) and thermodynamically activate copper corrosion. In the thermal power industry, ammonia significantly corrodes copper alloys in steam surface condensers due to the formation of soluble complexes such as (\(Cu\left( {NH_{3} } \right)_{2}^{2 + }\) and \(Cu\left( {NH_{3} } \right)^{2 + }\)) [24]. Moreover, atomic hydrogen can diffuse into the metal, accumulate at grain boundaries or defects, weakening the structure, during, for example, welding, annealing, or exposure to hydrogen-rich environments. Hydrogen gas can be produced at the surface of copper by a sequence of reactions that occur when it is exposed to anoxic conditions and excess of ammonia. At the anode, copper metal oxidizes to release copper ions into the solution:
$$Cu \to Cu^{2 + } + 2e^{ - }$$
At the cathode, ammonia is reduced, and hydrogen gas is produced:
$$6NH_{3} + 6e^{ - } \to 3NH_{2}^{ + } + 3H_{2}$$
Overall, the global cell reaction can be represented as follows:
Previous research studied hydrogen uptake in the presence of sulphides, and the results show that a fraction of hydrogen atoms can ingress the copper structure through grain boundaries and dislocations [24, 25]. However, the solubility of hydrogen into copper is very low, and electrolytic charging has been shown to produce only thin (~ 50 µm) hydrogen-rich layers on the copper surface [26]. While these thin hydrogen-rich layers would not be detrimental to the mechanical properties of the copper canister, they may influence the SCC crack initiation and crack propagation. Therefore, despite the concentration of hydrogen gas in the final repository being low (theoretically ~ 10–6 M) [24] as is the concentration of ammonia, it is relevant to investigate hydrogen uptake under the studied conditions.
In this work, autoclave tests were carried out during three months at room temperature employing simulated groundwater deaerated by nitrogen purging (5N N2) to which 100 mg/l of ammonia hydroxide (NH4OH) was added. The simulated groundwater used in this study (Table 2) was prepared based on earlier studies [27]. It represents the equilibrium chemistry of groundwater at the Olkiluoto disposal site after ion exchange with bentonite. The pH and ammonia concentration of the water were weekly monitored using a UV–vis spectrophotometer (Hach Lange) and an ammonia test kit Hach LCK302. Additions of NH4OH were made to keep the pH value between 10 and 10.4 and the ammonia concentration between 90 and 130 mg/l. Three potentials (− 200, − 125, and − 50 mV vs. SCE) were constantly applied to the U-bend samples, sorted into three isolated racks. These potentials were selected to correspond to Cu2O, CuO, and Cu(OH)2 oxidation states (yellow dots in Fig. S1 [11]), respectively, where the SCC threshold is expected to be reached at the maximum anodic potential (− 50 mV).
Table 2
Chemical composition of the simulated groundwater [27]
K
Ca
Cl
Na
SO4
Br
HCO3
Mg
Sr
Si
B
F
Mn
PO4
lactate
mg/L
54.7
280
5274
3180.2
595
42.3
13.7
100
8.8
3.1
1.1
0.8
0.2
0.1
1
It is noteworthy that SCC does not occur below a certain critical potential because the adsorption of aggressive species is strongly influenced by the redox potential [13]. Adsorption of an anion occurs at potentials higher than the point of zero charge on a surface in a particular environment, which influences the selectivity of species for adsorption [13]. It is also noted that the selected conditions applied in this study do not correspond to realistic conditions under the final disposal but were intentionally exaggerated in terms of ammonia concentration, mechanical stress, and corrosion potential to serve as a stress test. The goal was to evaluate whether degradation mechanisms (if any) take place under accelerated conditions at the laboratory scale. For context, detailed site investigation studies have reported notably smaller ammonia concentrations of 3 and 1.1 mg/l at Hästholmen (Finland) and Olkiluoto (Finland), respectively [28]. The limitations of extrapolating these findings to realistic repository scenarios are discussed in the conclusion.
Autoclave exposures and analyses
Figure S2 depicts the 3D design and mounting of the racks containing the U-bends as well as the mass-loss samples, right after their placement at the bottom of the autoclave. A saturated Ag/AgCl was used as the reference electrode for the electrochemical tests, and a stainless-steel mesh was used as the counter electrode (CE), ensuring an area ratio to the working electrodes higher than 10. Three constant different potentials (− 200, − 125, and − 50 mV vs. SCE) were applied to the immersed U-bends connected by copper wires using Wenking potentiostat model LB 81. The pressure and temperature of the solution in the autoclave were continuously monitored at > 0.3 bar and 21 °C, respectively.
Electrochemical impedance spectroscopy (EIS) data were obtained using a PalmSens4 Potentiostat. A potential of 10 Vrms was applied after measuring the open-circuit potential (OCP) stability for 2 min, varying the frequency between 10 kHz and 1 mHz. Polarization resistance values over time were calculated by fitting the EIS curves using an electrical equivalent circuit (EEC) with the PSTrace 5.9 software provided by PalmSens®. One of the most used models to fit EIS data is the simplified Randles cell. It consists of a charge transfer resistance (Rct) in parallel with a double-layer capacitor (CPEdl), added to the solution resistance (Rs) [29]. It is worth noting that charge transfer resistance (Rct) and polarization resistance (Rp) are different properties. The impedance that charge transfer encounters at electrode–electrolyte interfaces is known as polarization resistance (Rp). It arises when an external voltage pushes the electrode potential out of equilibrium, prompting electrochemical processes to transfer negative charges, leading to current flow [29]. However, when studying simple redox reactions, Rct and Rp can be considered the same because they can be modelled with a simple resistance. Information on the polarization resistance of the samples was extracted by fitting the impedance results using an electrical equivalent circuit (EEC) or simplified Randles. In this study, the EEC was composed of solution resistance (Rs) in series with one time constant (Rct + CPEdl).
The rate of general corrosion was evaluated from non-polarized mass-loss specimens. Dimensions and masses of the samples were determined using callipers and an analytical scale before and after the autoclave testing. During the exposure, the samples were placed at the bottom of the autoclaves (Fig. S2). Right after opening the autoclaves, the specimens were rinsed with ethanol and placed in a desiccator. A deoxidizing procedure (also known as pickling) was carried out by immersing the specimens and the reference (non-immersed sample) in a solution containing 500 ml H2O + 500 ml HCl (37%) + 3.5 g hexamethylenetetramine for 5 min. They were subsequently rinsed with ethanol, dried, and weighed on an analytical scale, which offers a readability of up to 0.00001 g. This procedure was repeated until the weight loss was stable, allowing the calculation of the corrosion rate (CR) in µm/year using Eq. 1 (ASTM G 31-72) [30]:
where K is a constant that defines the units for the corrosion rate (8.76 × 107 for µm/year), Wloss is the equivalent weight in grams, t is the time of exposure in hours (2088), d is the density of Cu (8.94 g/cm3), and A is the surface area of the sample in cm2 [30].
Six non-polarized specimens of each alloy (4 × 4 × 8 mm) were included into the 3-month autoclave exposure to examine the H uptake into the copper. After the immersion, the samples were immediately polished slightly using silicon carbide #220 to remove the oxide layer, rinsed with Milli-Q water, ethanol, and dried with hot air blow. They were then immersed in liquid nitrogen and analysed using the hot-melt extraction (Bruker G8 Galileo). The reference specimens, which were not immersed, were prepared in the same manner and investigated under identical conditions.
After autoclave testing, all the samples were examined using a light optical microscope (LOM) (Axio Zoom V16 ZenCore 3.8). The microstructure and grain size of the six alloys were determined by etching the cross sections and the images acquired in a reverse light stereomicroscope (Leica MEF4). U-bend surfaces were investigated using a Zeiss Merlin field emission scanning electron microscope (FESEM) and a Tescan Amber X Xenon plasma-focused ion beam (PFIB). The electron beam backscattering diffraction (EBSD) maps were recorded using a Zeiss Crossbeam 540 FESEM. Cross sections for the EBSD analyses were prepared by cutting the 180° locations from the U-bend specimen and embedding them in a conductive resin. The embedded specimens were then gradually polished down to 0.25 µm diamond paste and followed by polishing with colloidal silica. Raman analysis of the U-bends was carried out using a Cora 5001 Fiber from Anton Paar® scanning the top of the samples from − 400 to 3500 cm−1 to elucidate differences in chemical composition.
Additional testing on H-uptake and molecular dynamics modelling
To further explore the mechanism of H-uptake in copper, electrolytic H-charging was conducted in a three-electrode cell with a calomel reference electrode, a platinum wire as a counter electrode, and the specimen of interest (9 mm × 4 mm × 0.8 mm) as the working electrode. The 1 N H2SO4 (28 mL/L) electrolyte, with 10 mg/l of thiourea as hydrogen recombination poison, was deaerated using N2. The samples were charged with a Gamry potentiostat under a constant cathodic potential of − 1.15 V (vs. SCE) for 4 h at 50 °C, as reported by Sahiluoma et al. [31] to ensure H-concentration growth in deoxygenated copper materials. Before the analysis, the specimens were mechanically polished with emery paper #2000. After H-charging, the specimens were rinsed with distilled water and dried under He flow to remove surface moisture. The partial pressure of hydrogen was measured in an ultra-high vacuum (UHV) chamber (1 × 10–9 mbar) with a mass spectrometer. To maintain the required pressure in the UHV chamber, the specimen was initially placed in an airlock compartment and pumped to an intermediate pressure of 1 × 10–6 mbar. The specimen was then transferred to the UHV chamber, and the measurement was initiated. The total time from the end of H-charging to the measurement initiation was under 7 min. All thermal desorption spectroscopy (TDS) measurements were performed at a 10 K/min heating rate. To ensure that the results are reproducible and statistically viable, the H-charging and TDS were repeated four times for each condition.
Molecular dynamics modelling was applied to study atomic hydrogen diffusion in copper with an ambitious goal of comparing the results with experiments. The modelling approach included setting up models for a perfect Cu crystal, Cu crystal with point defects, and polycrystalline Cu with a random set of grain sizes and orientations. In these systems, diffusion of atomic hydrogen was studied to check that the transport properties of hydrogen can be estimated from simulations, and the effect of point defects and grain boundaries on the transport properties can be evaluated. The work did not attempt to assess the chemistry leading to the formation of hydrogen on the Cu surface, nor the penetration of hydrogen into the bulk sample. Thus, all simulations were performed for bulk conditions. Different Cu grades were not explicitly considered, and all simulated systems consisted of pure Cu with hydrogen. Although surface chemistry of hydrogen on Cu was not attempted in the current work, it was seen as important to ensure it could be done in the future. Furthermore, it was considered important to include the potential association/dissociation reactions of hydrogen in bulk Cu. For these reasons, a force field was required, which has the capability to describe the dissociation of the hydrogen molecule, as well as the recombination of two hydrogen atoms into a hydrogen molecule, when the hydrogen is inside a Cu lattice. Such force fields include the ReaxFF force field [32], the COMB force field [33], and the Cu-H bond-order potential [34]. After an evaluation of these options, we concluded that the Cu-H bond-order potential (hereinafter BOP) was the best-suited for our purposes. This force field was available in the MD software LAMMPS [35].
The bulk face-centred-cubic (FCC) copper lattice was established using LAMMPS. For bulk FCC copper, we verified the elastic constants C11 = 176, C12 = 125 and C44 = 82 GPa given in [34] using a system of 4000 Cu atoms. The creation of point defects in the lattice was done by randomly removing 1% of Cu atoms. The operation yielded elastic constants of C11 = 174, C12 = 122, and C44 = 81 GPa, while removing 5% of Cu atoms yielded elastic constants of C11 = 161, C12 = 108, and C44 = 75 GPa.
The introduction of grain boundaries to the structure is more challenging due to the vast number of different possibilities. Creating a series of single, well-defined grain boundaries is relatively easy [36], but achieving a single system with a grain boundary structure that would resemble the grain boundary structure of a real polycrystalline material is practically impossible, as this implies a system with a very large (computationally prohibitive) number of atoms to accommodate all boundaries in appropriate proportions, as dictated by their energetics. In practice, the creation of moderate-size polycrystalline systems makes use of randomness. In the popular Voronoi tessellation method [37], a space containing a set of random points is divided into cells surrounding these points such that a segment of a cell boundary between two points is always equidistant from these two points. The cell volumes can then be filled with structures of a (randomly) chosen orientation. However, in the current work, we chose an approach based on melt solidification. First, a relatively large (500,000 atoms) perfect crystal was created. It was subsequently heated to T = 2000 K, which is well above the melting point Tm = 1390 K for the BOP potential. After the system had completely melted, the temperature was brought again below the melting point, which resulted in the formation of a polycrystalline system.
Results and discussion
Electrochemical analysis
Four specimens were weekly monitored by EIS to assess their electrochemical behaviour over 3 months of autoclave testing in ammonia-containing simulated groundwater. Figure S3 displays the Nyquist and Bode plots after 3, 45, and 79 days. The comparative analysis (Fig. 2a, b) shows that all alloys present a similar behaviour with no significant variation of impedance values within the period, indicating that the formed passive oxide on the surface of the samples is stable under the studied conditions. Figure 2a depicts the Rp values calculated for HCP, and Fig. 2b shows an identical analysis of OFE + P. The time evolution of Rp of HCP (Fig. 2a) presents values between 100 and 200 kΩ cm2 with upspikes in the line trend. These spikes might have been triggered by different reasons, e.g. noisy EIS data, the presence of oxidizing substances (nitrite, nitrate, or other water components) that provide oxygen for corrosion, or even some small variation of oxygen content in the solution that resulted in redox reactions on the Cu surfaces; however, no relevant variations in terms of corrosion behaviour are observed. Furthermore, no correlation was found between grain size (ranging from 94 to 152 µm) and electrochemical response. According to Li et al. [38], only nanosized grains and a high density of special grain boundaries (particularly Σ3 twins) significantly influence corrosion resistance. Since none of the specimens fall within this regime, it is reasonable to conclude that the grain size had a negligible effect on their electrochemical performance under the tested conditions. The OFE + P alloy (Fig. 2b) shows, in particular for OFE + P_B, a higher Rp in the first month of immersion than the HCP, with values between 200 and 350 kΩ cm2. After this period, the Rp decreases and remains similar to OFE + P_F until the end of the immersion. This suggests that the OFE + P_B sample has higher corrosion resistance in ammonia environments, but only at the initial stage of immersion. However, further work is required to clarify the effect of deformation on the corrosion behaviour.
Figure 2
Time evolution of polarization resistance (Rp) values after fitting EIS data for a Cu-HCP and b Cu-OFE + P; c Mean corrosion rate values and standard deviations for both copper alloys calculated with the weight losses after the pickling procedure.
Mean corrosion rate (CR) values and standard deviations obtained for duplicates are shown in Fig. 2c. The small range of CR values between 0.19 and 0.33 µm/year and large standard deviations suggest that all alloys corrode similarly and are stable against disturbance introduced by ammonia under repository conditions. According to King [11], dissolution rates above, e.g. 0.6 mm/year, would be necessary to sustain crack initiation, but these rates are unlikely in anaerobic environments.
3.2. Chemical composition
Raman spectra were recorded for the autoclave tested and non-exposed OFE + P_F and HCP_F specimens. One specimen per applied potential (− 200, − 125, and − 50 mV) was studied. A clear attenuation in signal intensity after testing can be observed in Fig. S4, which might be related to the more amorphous nature of the oxide layer formed on the surfaces. No shifts in the Raman peak positions were observed for the different potentials. The broad peak around 762 cm−1 can be attributed to Cu2O, while the one at 890 cm−1 is characteristic of Cu(OH)2 [39]. Further surface analysis by EDS, spectra shown in Fig. 3, was performed to identify specific changes in the surface oxide composition for the copper U-bends before and after the autoclave testing. The semi-quantitative results from the EDS spectra, as shown in Table 3, reveal that the reference specimen (non-immersed) consists of 99.6 wt.% Cu and 0.4 wt.% O, as expected. In general, the amount of oxygen increases with increasing potential, in line with the increasing oxidation state. However, some discrepancies are detected, which are likely related to elements such as Si, Cl, and Mg being incorporated in the oxide film (Table 3).
Figure 3
EDS spectra of reference and autoclave-tested of a HCP_F and b OFE + P_F copper samples. The semi-quantitative results of the EDS analyses are listed in Table 3.
Semi-quantitative EDS elemental composition of OFE + P_F and CU-HCP_F before (reference) and after autoclave testing in ammonia-containing groundwater for 3 months. Values are represented in wt.%
Sample/potential
O
Si
Cl
Cu
Mg
Cu/o
Reference
0.4
–
–
99.6
–
–
HCP_F/ − 200 mv
13.7
–
1.8
84.5
–
6.1
HCP_F/ − 125 mv
16.1
–
1.8
81.8
0.2
5.0
HCP_F/ − 50 mv
36.3
1.6
0.8
56.1
5.1
1.5
OFE + P_F/ − 200 mv
20.5
–
1.3
76.8
1.4
3.7
OFE + P_F/ − 125 mv
15.9
1.2
1.4
81.5
–
5.1
OFE + P_F/ − 50 mv
26.7
1.1
2.2
60.3
9.6
2.2
3.3. Visual inspection and cross-sectional analyses
After removing the samples from the autoclaves, the specimens were visually inspected. A reference, non-immersed U-bend sample is included in Fig. S5 (left) for comparison. While the application of − 200 mV yielded a brownish layer, − 125 mV resulted in a reddish colour and − 50 mV produced a blueish coloration, characteristics of the increasing oxidation state of copper alloys. This agrees with the EDS results presented in Fig. 3 and with the semi-quantitative values shown in Table 3.
Vickers hardness values were measured at ten different positions of the U-bend cross sections, as illustrated in Fig. 4. As expected, all HV1 values are higher for the U-bends than the as-received materials, reflecting the effect of the 180° plastic deformation (15% strain). The results reveal a clear hardness variation across the cross section: the inner surface, subjected to compressive strain, consistently shows higher hardness than the intermediate and outer regions. This trend is observed across all samples. Additionally, a horizontal gradient is evident, with hardness values decreasing from the top towards the sides of the bend, which aligns with the location of maximum tensile strain.
Figure 4
Stereomicrographs of Cu U-bends cross sections after 3 months of immersion with the Vickers hardness measurements HV1 in different spots.
Microscopic inspection using a stereomicroscope and plan-view SEM demonstrated the absence of surface defects when − 200 and − 125 mV versus SCE were applied to the U-bends. However, the samples under − 50 mV showed the presence of a small density of surface defects in SEM. In order to study the features in more detail, cross-sectional samples were prepared using metallographic methods and imaged in SEM. Figure 5 displays the metal–oxide interface and the presence of surface defects for all the studied samples. The images confirm that although defects approximately 1 µm thick were present, they mainly correspond to the oxide layer, with only a few reaching the underlying metal. To further analyse the small features observed in the plan-view SEM (Figs. S6, S7), micro-cross sections were prepared directly at the defects using focused ion beam (FIB).
Figure 5
Cross-sectional images of Cu U-bends after 3 months of immersion in ammonia-containing groundwater with application of − 50 mV.
Figure 6 presents plan-view SEM images of two distinct areas from the OFE + P_F sample, both displaying crack-like surface features. Upon focused ion-beam (FIB) milling (centre and right images), it is evident that these features are primarily associated with the surface oxide layer and/or corrosion along grain boundaries, with a maximum penetration depth of approximately 1 µm into the bulk material. Similar features were observed on the HCP_F sample, as shown in Fig. 7. However, the defects on HCP_F appear slightly deeper, extending up to 4 µm, suggesting a somewhat more advanced stage of localized corrosion or surface degradation. FIB cross-sectional analysis confirms that all observed surface defects are fully oxidized, indicating that these are not cracks propagating into the metallic matrix but rather oxidized surface intrusions. The short lateral propagation length of these features, ranging from 1 to 5 µm, suggests that the oxide layer plays a significant role in limiting further defect growth. This behaviour supports the notion that the oxide scale acts as an effective barrier, slowing the ingress of corrosive species and mitigating deeper material degradation under the conditions studied [40].
Two surface defects observed in the OFE + P_F sample anodically polarized at − 50 mV versus SCE were further examined using SEM-EBSD. The collected maps are shown in Fig. 8. The analysis clearly indicates that the defects beneath the oxide film develop preferentially along random high-angle grain boundaries rather than special boundaries such as twin or low-angle boundaries. These intrusions, with penetration depths of less than 10 µm, are morphologically similar to those reported by Forsström et al. [41] after a slow strain rate testing, although they are notably shorter than the cracks described in [42] (both of which were observed in sulphide-rich, chloride-containing anoxic water at 90 °C). While those earlier studies attributed crack propagation to local corrosion of oxide inclusions along grain boundaries in samples welded in air [41, 42], such conditions are not applicable in the present investigation. No welding was involved, and the testing environment differed significantly, lacking elevated temperature or sulphide species. Nevertheless, these findings reinforce the understanding that the formation of brittle surface CuO under certain electrochemical and mechanical conditions can contribute to the onset of SCC [14]. It is proposed that, under these conditions, the surface oxide can act as a cathodic layer, while the crack tip or defect front becomes anodic. This localized galvanic coupling may promote copper dissolution, particularly in the presence of complexing agents like ammonia [11, 14]. Although the observed features are shallow and do not appear to propagate extensively, their alignment with high-energy grain boundaries and their full oxidation suggest that even in relatively benign conditions, localized oxidation along microstructural weak points can initiate early-stage degradation.
Figure 8
SEM and EBSD maps of surface defects found in a Cu-OFE + P sample; inverse pole figures (IPF-Y direction).
The cracking mechanisms of copper have long been a subject of investigation and debate within the materials science community. Various studies have tried to elucidate the conditions under which pure copper may develop cracks, particularly under anoxic environments corresponding to the fourth phase in the canister lifecycle [43]. Despite several researchers have reported evidence of SCC in pure copper, the microstructural features described in these studies remain under controversy. It is unclear whether the observed cracks indeed represent classical SCC, or if they are more accurately attributed to intergranular corrosion phenomena, which may subsequently lead to grain boundary decohesion facilitated by plastic deformation [15, 25, 44‐46]. The distinction is crucial, as SCC involves a synergistic interaction between mechanical stress, corrosive environment, and electrochemical processes, whereas intergranular corrosion followed by plastic strain typically signifies a distinct degradation pathway. In the present study, the observed surface defects in both copper alloys appear to be shallow, fully oxidized, and preferentially located at grain boundaries, consistent with a corrosion-driven process. These characteristics, combined with the mechanical loading imposed by the 180° U-bending, suggest that the most plausible mechanism behind defect formation is anodic dissolution facilitated by localized strain. This interpretation aligns with the view that mechanical deformation plays a key role in increasing susceptibility to surface breakdown and defect initiation, even in nominally reducing environments.
3.4 Hydrogen uptake
Hydrogen uptake analysis by hot-melt mass spectroscopy (HMMS) and thermal desorption spectroscopy (TDS)
Hot-melt mass spectroscopy (HMMS) was employed to determine the average values of hydrogen concentration in different samples before and after the autoclave testing. The HMMS results are shown in Fig. 9. After 3 months of immersion, all samples showed lower H-content values than before immersion, varying between 0.5 and 0.8 wt.ppm, indicating that no hydrogen uptake occurred under the applied conditions. The absence of hydrogen uptake during these tests might be associated with the low concentration (or non-formation) of H2 under anoxic conditions or the temperature of the tests, as H2 solubility increases only above about 600 °C [47]. The atmospheric pressure of the tests also limits the solubility of H2 compared to disposal “in situ” conditions. Hydrostatic pressure is about 4 MPa at the repository depth, but it can increase up to 40 MPa during periods of glaciation [5]. It is worth noting that the reduced hydrogen content post-autoclave is consistent with MD simulations and HMMS observations, which indicates that hydrogen diffuses outward and exits the copper more slowly as grain size increases.
Figure 9
Average hydrogen content values before (orange) and after (green) immersion in ammonia-containing simulated groundwater.
To further understand the hydrogen uptake characteristics of copper, an OFE + P_F specimen was electrochemically charged, and the hydrogen content was measured by thermal desorption spectroscopy (TDS). The average hydrogen content measured by TDS for as-supplied and H-charged specimens is found to be 1.42 and 3.55 wt.ppm, respectively (Fig. 10). The difference in measured hydrogen content in the as-supplied specimens, by TDS (≈ 1.42 wt.ppm) and by HMMS (≈ 2.10 wt.ppm), arises from the differing measurement techniques; nonetheless, they are comparable when considering the temperature regimes involved, TDS measured desorbing hydrogen up to 600 °C, whereas HMMS measured desorbing hydrogen up to the specimen’s melting point at approximately 1085 °C. This is consistent with the work by N. Senior et al. [48], who reported continued hydrogen release from copper up to 971 °C.
Figure 10
a Comparison between as-supplied and H-charged specimens; b corresponding TDS curves (all TDS curves are shown in Fig. S8).
Electrochemical H-charging consequently led to a significant (~ 150%) growth of hydrogen content in the OFE + P_F sample unlike in the autoclave experiment. We believe that the observed behaviour could be due to one or a combination of two of the following scenarios:
(i)
The electrochemical conditions of the immersion environment and specimen were not adequate to trigger a hydrogen evolution reaction (HER) with sufficient fugacity to cause a significant hydrogen concentration gradient between the surface and the bulk of the samples leading to H-diffusion into the specimen. It is worth noting, however, that even if this was the case, at least the hydrogen content of the immersed specimen should remain the same compared to the as-supplied specimens, not less.
(ii)
The second scenario considers that HER from corrosion was sufficient to cause hydrogen ingress into the specimen. But upon entry, hydrogen combines with existing metallurgical hydrogen increasing the pressure at specific trapping sites forming blisters/voids that grow and burst, therefore releasing hydrogen (Fig. S9). This can safely be deduced by comparing the obtained TDS curves for as-supplied and H-charged specimens shown in Fig. 10a, b. Despite the complexity of the spectra, it is notable that the curves of H-charged specimen consist of spikes (Fig. 10b) that are absent from the TDS curve of as-supplied specimens. These spikes are representative of hydrogen blisters bursting from the surface and sub-surface vicinities during the measurement. Similar results have been reported in the literature emphasizing that the size and density of these spikes correspond to the size and density of void openings [47, 49, 50]. It is worth noting that, despite that these spikes are more pronounced between 400 and 550 K on the spectra, they are also present from ≈ RT to 320 K. These openings releasing hydrogen may have caused a concentration gradient, facilitating diffusion of metallurgical hydrogen towards the surface. This progressive phenomenon may be responsible for the lower concentration of hydrogen measured after immersion under no cathodic potential for the studied materials.
(iii)
The third scenario considers that the temperature of the immersion environment is not sufficient for hydrogen from corrosion products to diffuse into the specimens. The solubility and diffusivity of hydrogen are relatively low in copper at RT.
The hydrogen content for the specimen charged at RT (red data point in Fig. 10a) was about half of the average of specimens charged at 50 °C. It is widely reported in the literature that the possibility of hydrogen transport in copper is limited at low temperatures [46]. A deeper observation of the desorption spectra (violet colour in supplementary Fig. S8) shows that there was a first big peak at about 400 K, signifying hydrogen desorption from shallow traps and sub-surface vicinities. This peak temperature is lower compared to the first peak (450 K) of all specimens charged at 50 °C. Additionally, the specimens charged at 50 °C have subsequent peaks between 625 and 800 K. On the contrary, in these temperature ranges, the specimen charged at RT does not manifest any comparable peaks. This may be indicative that at RT, hydrogen does not readily diffuse into the bulk of the specimen to occupy deep traps that require higher temperatures to desorb. Therefore, facilitating a high sub-surface concentration paves the way for scenarios (iii) and (ii).
Molecular dynamics (MD) simulations
Figure 11a presents the evolution of crystallinity as a function of time at two temperatures (600 and 800 K). Specifically, the quantity plotted is the fraction of atoms belonging to a FCC crystalline phase, as given by the common neighbour analysis (CNA) [51]. Following a rapid period (< 1 ns) of initial nucleation, the effect of temperature on the crystallinity and crystallization rate is clearly observed, with a higher temperature corresponding to a slower crystallization rate and higher crystallinity.
Figure 11
a Time evolution of the FCC crystalline phase during melt solidification; b Experimental and simulated diffusion constant for hydrogen atoms. Polycrystalline Cu structures from melt solidification at c 600 K, and d 800 K.
Figure 11c, d shows visualizations of the polycrystalline structures. It is observed that the FCC crystallites (green) are somewhat bigger in the 800 K case. The red portions are designated as the hexagonal close-packed phase by the common neighbour analysis algorithm, and they may correspond to glide planes in the FCC structure. The HCP phase represents about 15% of all atoms. The rest (grey) is classified as “other” by the CNA, and it contains the grain boundaries. Obviously, using a higher temperature would have resulted in yet larger crystallites and likely more well-defined grain boundary regions, but the process would have been too slow to observe computationally, as a larger system would have been required to accommodate the larger crystallites. The structures shown in Fig. 11c, d may be thought to represent “worst-case” scenarios when considering the effect of grain boundaries on hydrogen diffusion.
The diffusion of hydrogen was studied in a single crystal and two polycrystalline systems by randomly inserting hydrogen atoms into the structure at a concentration of 1%. This concentration is much higher than the true hydrogen content of copper, but it is used for computational reasons. True hydrogen concentrations of the order of ppm would mean a single hydrogen atom in a structure of 500,000 copper atoms. Observing the diffusion of a single atom for diffusion, while possible, would lead to extremely poor statistics, as well as spending most of the computational resources on the movement of copper atoms. It is expected that a concentration of 1% is high enough to provide meaningful statistics on diffusion, but low enough so that interactions (and possibly reactions) between hydrogens do not affect the results.
The diffusion of hydrogen atoms was measured by recording the mean square displacement (MSD) in the temperature range of 300–1000 K as a function of time, and computing the diffusion constant from the slope of the MSD(t) curve according to
$$\overline{r}^{2} \left( t \right) = \frac{1}{N}\mathop \sum \limits_{N} [\overline{r}\left( t \right) - \overline{r}\left( 0 \right)]^{2} = 6Dt$$
where N is the number of hydrogen atoms, and D is the diffusion constant. At each temperature, the simulation was run until a clear linear dependence of the MSD on time was observed and the slope of the curve could be accurately determined.
The simulated diffusion constants are plotted in Fig. 11b with a line fitted to a compilation of experimental results [47]. Simulations were performed for a perfectly crystalline FCC Cu, as well as for the two polycrystalline Cu structures shown in Fig. 11c, d. All simulated diffusion constants are slightly below experimental ones. As expected, diffusion is slowest in the perfect FCC crystal. The atomic disorder due to grain boundaries increases the diffusion constant, and it is noted in particular that the structure produced by melt solidification at 600 K, containing smaller crystallites and wider grain boundaries, yields the fastest diffusion. However, the effect of grain boundaries is perhaps surprisingly small, considering that diffusion along grain boundaries is often thought to be dominating, as the defects in grain boundaries generally provide more free volume for diffusion [52]. On the other hand, the tendency of hydrogen to be trapped in Cu lattice defects is well known [53], but traps not involving impurities (such as oxygen) are shallow and should not be effective at temperatures above 500 K. The current simulations were not designed to analyse hydrogen trapping in detail.
Conclusions
A 3-month autoclave experiment with ammonia-containing simulated groundwater under anoxic conditions (ammonia and pH disturbed corresponding to the fourth phase in the canister evolution) and applied potentials revealed a similar electrochemical behaviour for OFE + P and HCP copper alloys. While OFE + P copper initially exhibited higher polarization resistance (Rp) values, which is interpreted to be a signal of higher corrosion resistance, the average polarization resistance and corrosion rates for the test showed a similar result for HCP. The reason behind the higher initial Rp values was not assessed in detail, but they might be related to the different manufacturing parameters. Additionally, both alloys present very-low corrosion rates, ranging from 0.2 to 0.4 µm/year.
The study successfully met its objective of evaluating SCC susceptibility of OFE + P and HCP copper under controlled anoxic conditions by applying a thermodynamically guided polarization strategy to mimic environments where SCC can be triggered. The application of different potentials to the U-bends resulted in various oxidation states, with the highest anodic potential (− 50 mV vs. SCE) leading to the highest oxidation of copper (Cu(OH)2), showing a distinctive blueish surface layer. Small surface defects were observed in both OFE + P and HCP specimens after this exposure. These defects were blunt in nature and filled with oxide and occurred on random grain boundaries. It is suggested that the formation mechanism of these defects was stress-assisted local anodic dissolution. However, the SCC susceptibility of HCP and OFE + P was found to be small under the studied conditions.
Hydrogen concentration measurements showed no hydrogen uptake for all the materials after a 3-month autoclave experiment. However, electrochemical H-charging of OFE + P_F under cathodic potential at 50 °C showed significant hydrogen concentration growth up to 3.55 wt.ppm. This indicated that the lack of hydrogen uptake in the autoclave experiment is likely attributable to an insufficient thermo-electrochemical immersion environment or non-formation of H2 in anoxic circumstances. MD simulations reproduced the experimentally observed temperature dependence of hydrogen diffusion in copper and revealed that diffusion coefficients in polycrystalline structures are close to those in perfect crystals, indicating limited trapping and minimal hindrance from grain boundaries.
The research hypothesis was that HCP copper would exhibit corrosion resistance and mechanical performance against SCC in ammonia-containing groundwater comparable to OFE + P, due to their similar composition and microstructure. All results indicate that HCP copper performed equally well as compared to OFE + P in NH3-containing groundwater. While the results are encouraging, it is important to note that the ammonia concentration used in this study (100 mg/L) is significantly higher than what is expected in actual repository environments (1–3 mg/L). This makes the test a “stress-test” scenario designed to evaluate worst-case material behaviour. As such, the results should not be directly extrapolated to long-term repository conditions without further validation. Furthermore, future work should focus on several other aspects, such as creep properties or alternative environments, e.g. nitrites, acetate, sulphides, before HCP can be considered as an alternative canister material.
Declaration
Conflict of interest
We declare that we do not have any commercial or associative interest that represents a conflict of interest in connection with the work submitted.
Acknowledgements
The authors acknowledge the financial support from the European Joint Programme on Radioactive Waste Management (EURAD). EURAD has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 847593. Additionally, the funding obtained from VTT Technical Research Centre of Finland Ltd for the project is acknowledged. We thank Posiva Oy for providing materials for this study and for the insightful discussions. We kindly thank colleagues and staff for their advice and support in the experiments: Dr. Timo Saario, Dr. Supriya Nandy, Dr. Aloshious Lambai, Mr. Jukka Maunumäki, Mrs. Hanna Iitti, Mr. Tuomo Kinnunen, Mrs. Taru Lehtikuusi, and Mr. Pasi Väisänen.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Hall DS, Behazin M, Jeffrey Binns W, Keech PG (2021) An evaluation of corrosion processes affecting copper-coated nuclear waste containers in a deep geological repository. Prog Mater Sci 118:100766. https://doi.org/10.1016/j.pmatsci.2020.100766CrossRef
2.
Keech PG, Vo P, Ramamurthy S et al (2014) Design and development of copper coatings for long term storage of used nuclear fuel. Corros Eng Sci Technol 49:425–430CrossRef
3.
Keech PG, Behazin M, Binns WJ, Briggs S (2021) An update on the copper corrosion program for the long-term management of used nuclear fuel in Canada. Mater Corros 72:25–31. https://doi.org/10.1002/maco.202011763CrossRef
4.
King F, Kolàř M, Briggs S et al (2024) Review of the modelling of corrosion processes and lifetime prediction for HLW/SF containers—part 1: process models. Corros Mater Degrad 5:124–199
5.
Posiva Oy (2021) Canister evolution. Working Report 2021–06. Posiva Oy, Olkiluoto
6.
(2012) EN 1976:2012 Copper and copper alloys: Cast unwrought copper products. European Committee for Standardization (CEN), Brussels
Bojinov M, Goel S, Ikäläinen T, Saario T (2024) Effect of sulfide addition on the corrosion mechanism of copper in saline groundwater solution. J Electrochem Soc 171:041505. https://doi.org/10.1149/1945-7111/ad3fedCrossRef
11.
King F (2021) Assessment of the stress corrosion cracking of copper canisters. WR 2021–11. Posiva Oy, Olkiluoto
12.
King F, Lilja C (2013) Localised corrosion of copper canisters in bentonite pore water TR-13–27. Swedish Nuclear Fuel and Waste Management Company (SKB), Stockholm
Miyamoto H, Saburi D, Fujiwara H (2012) A microstructural aspect of intergranular stress corrosion cracking of semi-hard U-bend tubes of commercially pure copper in cooling systems. Eng Fail Anal 26:108–119. https://doi.org/10.1016/j.engfailanal.2012.07.006CrossRef
15.
King F, Lilja C, Pedersen K et al (2010) An update of the state-of-the-art report on the corrosion of copper under expected conditions in a deep geologic repository. TR-10–67. Swedish Nuclear Fuel and Waste Management Company (SKB), Stockholm
16.
Fujimoto S, Tsuchiya H, Ogawa S et al (2021) Stress corrosion cracking of copper in swollen bentonite simulating nuclear waste disposal environment. Mater Corros 72:333–338. https://doi.org/10.1002/maco.202011878CrossRef
17.
Lazzari L (2019) Mechanistic model for stress corrosion cracking-anodic dissolution mechanism. Metall Ital 111:21–25
18.
Tromans D, Sun R-H (1996) Intergranular/transgranular fatigue of copper: influence of environment on crack path and propagation rates. Mater Sci Eng A 219:56–65. https://doi.org/10.1016/S0921-5093(96)10422-6CrossRef
Sami SN, Sanati M, Joshi RP (2021) Simulations of hydrogen outgassing and sticking coefficients at a copper electrode surface: dependencies on temperature, incident angle and energy. Phys Rev Res. https://doi.org/10.1103/PhysRevResearch.3.013203CrossRef
21.
Williams DC, Riahi A, Carcea A et al (2024) Hydrogen embrittlement and strain rate sensitivity of electrodeposited copper: part I – the effect of hydrogen content. Npj Mater Degrad. https://doi.org/10.1038/s41529-024-00498-yCrossRef
Gubner R, Andersson U, Linder M et al (2006) Grain boundary corrosion of copper canister weld material. TR-06–01. Swedish Nuclear Fuel and Waste Management Company (SKB), Stockholm
24.
Macdonald DD, Sharifi-Asl S, Engelhardt GR, Urquidi-Macdonald M (2012) Issues in the corrosion of copper in a Swedish high level nuclear waste repository. 2012:11. Swedish Radiation Safety Authority (SSM), Stockholm
25.
Becker R, Forsström A, Yagodzinskyy Y et al (2020) Sulphide-induced stress corrosion cracking and hydrogen absorption in copper exposed to sulphide and chloride containing deoxygenated water at 90 °C. Mater Corros. https://doi.org/10.1002/maco.202011695CrossRef
26.
Martinsson A, Sandström R (2012) Hydrogen depth profile in phosphorus-doped, oxygen-free copper after cathodic charging. J Mater Sci 47:6768–6776. https://doi.org/10.1007/s10853-012-6592-yCrossRef
27.
Ratia-Hanby V, Isotahdon E, Yue X et al (2023) Characterization of surface films that develop on pre-oxidized copper in anoxic simulated groundwater with sulphide. Colloids Surf A Physicochem Eng Asp. https://doi.org/10.1016/j.colsurfa.2023.132214CrossRef
28.
Anttila P, Ahokas H, Front K et al (1999) Final disposal of spent nuclear fuel in Finnish bedrock: Olkiluoto site report 99–09. Posiva Oy, Helsinki
29.
Loveday D, Peterson P, Rodgers B (2004) Evaluation of organic coatings with electrochemical impedance spectroscopy. JCT Coat Tech 8:46–52
30.
(2004) Standard practice for laboratory immersion corrosion testing of metals G 31–72. ASTM International, West Conshohocken
31.
Sahiluoma P, Yagodzinskyy Y, Bossyut S, Hänninen H (2023) Hydrogen-induced micro-void formation in copper used for spent nuclear fuel disposal canisters. J Nucl Mater 574:154177. https://doi.org/10.1016/j.jnucmat.2022.154177CrossRef
32.
van Duin ACT, Dasgupta S, Lorant F, Goddard WA (2001) Reaxff: a reactive force field for hydrocarbons. J Phys Chem A 105:9396–9409. https://doi.org/10.1021/jp004368uCrossRef
33.
Liang T, Shan T-R, Cheng Y-T et al (2013) Classical atomistic simulations of surfaces and heterogeneous interfaces with the charge-optimized many body (COMB) potentials. Mater Sci Eng R Rep 74:255–279. https://doi.org/10.1016/j.mser.2013.07.001CrossRef
34.
Zhou XW, Ward DK, Foster M, Zimmerman JA (2015) An analytical bond-order potential for the copper–hydrogen binary system. J Mater Sci 50:2859–2875. https://doi.org/10.1007/s10853-015-8848-9CrossRef
35.
Thompson AP, Aktulga HM, Berger R et al (2022) LAMMPS - a flexible simulation tool for particle-based materials modeling at the atomic, meso, and continuum scales. Comput Phys Commun 271:108171. https://doi.org/10.1016/j.cpc.2021.108171CrossRef
36.
Tschopp MA, Coleman SP, McDowell DL (2015) Symmetric and asymmetric tilt grain boundary structure and energy in Cu and Al (and transferability to other fcc metals). Integr Mater Manuf Innov 4:176–189. https://doi.org/10.1186/s40192-015-0040-1CrossRef
37.
Ito Y (2015) Voronoi Tessellation Encyclopedia of Applied and Computational Mathematics. Springer, Heidelberg, pp 1546–1547CrossRef
38.
Li W, Yu B, Tam J et al (2020) Microstructural characterization of copper coatings in development for application to used nuclear fuel containers. J Nucl Mater. https://doi.org/10.1016/j.jnucmat.2020.152039CrossRef
Dinu A, Chicinas I, Abrudeanu M et al (2013) Stress corrosion cracking initiation of oxidized Incoloy 800 in caustic environment. INIS-RO-0004. In: Nuclear 2013. Institute for Nuclear Research, Pitesti, pp 117–124.
41.
Forsström A, Becker R, Hänninen H et al (2021) Sulphide-induced stress corrosion cracking and hydrogen absorption of copper in deoxygenated water at 90 °C. Mater Corros 72:317–332. https://doi.org/10.1002/maco.202011695CrossRef
42.
Becker R, Öijerholm J (2017) Slow strain rate testing of copper in sulfide rich chloride containing deoxygenated water at 90 °C. 2017:02. Swedish Radiation Safety Authority (SSM), Stockholm
43.
Ikeda BM, Litke CD, Kwong G (2011) Stress corrosion cracking of pure copper under possible nuclear fuel waste management conditions. ECS Trans 33:25–34. https://doi.org/10.1149/1.3557749CrossRef
44.
Taxén C, Flyg J, Bergqvist H (2018) Stress corrosion testing of copper in near neutral sulfide solutions. TR 19–13. Swedish Nuclear Fuel and Waste Management Company (SKB), Stockholm
45.
Taxén C, Flyg J, Bergqvist H (2018) Stress corrosion testing of copper in sulfide solutions. Swedish Nuclear Fuel and Waste Management Company (SKB), TR-17–16, Stockholm
46.
Suzuki Y, Hisamatsu Y (1981) Stress corrosion cracking of pure copper in dilute ammoniacal solutions. Corros Sci 21:353–368CrossRef
Yagodzinskyy Y, Malitckii E, Tuomisto F, Hänninen H (2018) Hydrogen-induced strain localisation in oxygen-free copper in the initial stage of plastic deformation. Philos Mag 98:727–740. https://doi.org/10.1080/14786435.2017.1417647CrossRef
Ganchenkova MG, Yagodzinskyy YN, Borodin VA, Hänninen H (2014) Effects of hydrogen and impurities on void nucleation in copper: simulation point of view. Philos Mag 94:3522–3548. https://doi.org/10.1080/14786435.2014.962642CrossRef
