Gamma radiation induces hydrogen absorption by copper in water

One of the most intricate issues of nuclear power is the long-term safety of repositories for radioactive waste. These repositories can have an impact on future generations for a period of time orders of magnitude longer than any known civilization. Several countries have considered copper as an outer corrosion barrier for canisters containing spent nuclear fuel. Among the many processes that must be considered in the safety assessments, radiation induced processes constitute a key-component. Here we show that copper metal immersed in water uptakes considerable amounts of hydrogen when exposed to γ-radiation. Additionally we show that the amount of hydrogen absorbed by copper depends on the total dose of radiation. At a dose of 69 kGy the uptake of hydrogen by metallic copper is 7 orders of magnitude higher than when the absorption is driven by H2(g) at a pressure of 1 atm in a non-irradiated dry system. Moreover, irradiation of copper in water causes corrosion of the metal and the formation of a variety of surface cavities, nanoparticle deposits, and islands of needle-shaped crystals. Hence, radiation enhanced uptake of hydrogen by spent nuclear fuel encapsulating materials should be taken into account in the safety assessments of nuclear waste repositories.

Scientific RepoRts | 6:24234 | DOI: 10.1038/srep24234 that are formed on different time scales 18 as depicted in Fig. 1. The radiation chemical yield (G-value) of a product of γ -radiolysis of water may be affected by the presence of chemical species or materials in the aqueous medium 19 . Several studies have shown that the yield of the aqueous radiolysis product H 2 can be significantly higher in water layers close to an oxide surface when compared to pure bulk water 20,21 . The magnitude of this effect depends on the nature of the oxide. However, no convincing mechanistic explanation that accounts for these observations has been given. In addition to the somewhat puzzling observations regarding the radiation chemical yield of H 2 , a number of studies on the interactions between other aqueous radiolysis products and oxide surfaces have been presented. It has been shown that H 2 O 2 is catalytically decomposed 22 to O 2 and H 2 O via the intermediate formation of hydroxyl radicals on most oxide surfaces, and that the hydroxyl radical has a high affinity for oxide surfaces 23 . Additionally, under certain conditions, the decomposition of H 2 O 2 at oxide surfaces can also lead to the formation of H 2 24 .
It is known that hydrogen has undesirable effects on the life-time and performance of many metallic materials 25 . Hydrogen can be absorbed by metals via the diffusion of H-atoms through preferred specific pathways that have low energy barriers for H-atom migration 26 . Upon radical recombination of the H-atoms, H 2 is formed and its effects on materials structure can be considerable. Such effects are the result of changes on vacancy-vacancy interactions 27 which ultimately leads to the formation of internal cavities and other defects that are accompanied by the deterioration of the physical-chemical properties of the materials 28 . Furthermore, γ -radiation can induce changes in nuclear spin states of H 2 and lead to the conversion of para-to ortho-hydrogen 29 . It is known that changes in the nuclear spin state of H 2 cause changes in its magnetic properties and this affects its interactions with metals leading to increased stability of hydrogen at some positions in the copper lattice 30 . Additionally, γ -radiation is known to cause low temperature annealing of metals 31 . The fact that γ -radiation is capable of inducing structural changes in the bulk material 32 and interface with water 21 -including corrosion 33-35 -lead us to develop this study, which is aimed at understanding whether γ -radiation is capable of triggering the uptake of hydrogen by high-purity metallic copper in aqueous medium. As shown above, one of the radiolysis products of water is hydrogen 36 . Hence, the undesirable effects that hydrogen has on copper could potentially hamper the barrier properties.

Radiation-induced hydrogen charging of copper metal in water.
The γ -radiation-induced hydrogen charging of copper was investigated as a function of total radiation dose (D). For each total radiation dose, blanks or background samples were prepared. The background samples consist of non-irradiated copper exposed to water for the same period of time as that of the irradiated samples. The background samples are necessary because copper has some amount of hydrogen from the manufacturing and even without the influence of ionizing radiation copper can induce the formation of minute amounts of H 2 (g) when exposed to water 37,38 . This is because water can adsorb dissociatively on some copper surfaces and this process leads to the formation of both HO • and H • , the latter is a precursor of H 2 38 . In Fig. 2 the amount of H 2 (g) measured in copper that was irradiated in aqueous media is plotted as a function of the total dose (D) corrected for the background values. It is known that desorption of dissociatively adsorbed water from metal and oxide surfaces can occur via a channel that leads to the formation of H 2 (g) 39 . This is due to the recombination of surface adsorbed H-atoms that occurs when the temperature is increased in the temperature-programmed desorption (TPD) experiment 40 . In order to exclude the hypothesis that the H 2 measured in our samples originates from the recombination of products of dissociatively adsorbed water at the copper surface, we also determined the amount of H 2 O that is adsorbed at the surface of the copper samples and investigated possible correlations with the amount of H 2 measured. In Fig. 2 it can be seen that the amount of H 2 O present in the copper samples does not correlate with the amount of H 2 detected. This fact excludes the possibility that the H 2 measured originates from the recombination of products of dissociatively adsorbed water at the copper surface. The H 2 (g) was measured post-irradiation which indicates that H • and/or H 2 are stable in copper. Previous investigations of desorption of H 2 from copper show that bulk H 2 desorbs in the same temperature range 41 as the H 2 that was detected in our irradiation experiments (Fig. 3). This indicates that the majority of the hydrogen measured originates from the copper bulk because surface-adsorbed hydrogen desorbs at lower temperatures in the TPD measurements. It can be seen in Fig. 3 that the irradiated sample has a considerably higher hydrogen content that desorbs at temperatures corresponding to bulk hydrogen, T > 500 K. The hydrogen content of the irradiated samples is also higher at the surface and at a few subsurface layers-this is for T < 500 K. A scanning electron microscopy investigation showed that the surfaces of both non-irradiated and irradiated samples exposed to water became covered with particles of approximately 100 nm in size (Fig. 4). There are no significant differences in morphology or size of the particles formed at the surface of non-irradiated and irradiated samples. However, besides the formation of nanoparticles, irradiated samples display extensive surface erosion and the presence of islands of needle-shaped crystals (Figs 4c,d, 5 and 6), which were not observed on non-irradiated samples. Erosion features are present all over the surface of the sample but are more pronounced in areas nearby islands of needle-shaped crystals (Fig. 5a-e). Here they reach considerable depth, affecting the material subsurface layers up to several hundred 33 nm. As a result of erosion, a diversity of porous structures, cavities and ravines are formed (Fig. 5b-e). Needle-shaped crystals occur either as large islands (of approximately 50-200 μm) or as very small crystals surrounding eroded areas (Fig. 5e). Interestingly, all observed islands of needle-shaped crystals have a core part (central nucleation point (Fig. 6d)), which is surrounded by a circle of nucleation points consisting of smaller crystal islands that are in turn encircled by very small crystals (indicated by white arrows on Fig. 6e). Energy-dispersive X-ray microanalysis and X-ray photoelectron spectroscopy performed at these areas confirmed the presence of oxygen, copper and a small amount of carbon in the crystals (Fig. 6e). According to previous studies on the radiation chemistry of copper/water interfaces, the corrosion  products formed are most likely oxygen-containing products-mostly cuprite, Cu 2 O-resulting from the oxidation of copper 33 . All islands of needle-shaped crystals have a light blue color (Fig. 6d). This suggests the presence of copper (II) hydroxide or copper (II) hydroxy carbonate because cuprite has a red-pink color 42 .

Discussion
According to the experimental data available, at an external pressure of 1 atm of H 2 (g), dry copper absorbs very little hydrogen at room temperature-in the order of 3 × 10 −6 ppm. Nevertheless, at a temperature of 1173 K, the uptake of hydrogen by copper increases to 40 ppm 43,44 . The driving force for this process has been attributed to the increase in the number of H-atoms at the Cu surface caused by the increase in temperature. This is because at higher temperatures a larger fraction of H 2 is dissociated into H-atoms. The uptaken hydrogen, which diffuses into the bulk of Cu as H-atoms, has been determined to be trapped at defects such as Cu atom vacancies and  : (a,b) non-irradiated sample, (c,d) sample which was exposed to a total dose of γ -radiation of 69 kGy. Figure 5. Scanning electron microscopy images of erosion at the surface of metallic copper kept in oxygenfree water and exposed to a radiation dose of 69 kGy. (a-e) An area next to an island of needle-shaped crystals; (f) particles at an area of the surface which is free of needle-shaped crystals. dislocations 26,45 . This implies that a material with a higher concentration of such defects will have a higher capacity to uptake H from its surroundings. Electrochemistry has also been used 46 for driving the uptake of hydrogen by Cu. In this process, hydrogen atoms are electrochemically produced on the copper surface resulting in a significant hydrogen uptake. A mechanism for this process has been proposed and the authors concluded that the surface properties are determinant for the diffusion of hydrogen into bulk copper.
Based on the observations presented above, it is plausible that the γ -radiation-induced hydrogen uptake presented here can be attributed to the radiolytic formation of hydrogen atoms. For γ -radiolysis of oxygen-free bulk water, the G-value (in μmol·J −1 ) for H • is 0.06 and for H 2 is 0.047, respectively 19 . However, as mentioned above, the interfacial G-value can be considerably higher. In the radiolytic system, the competing processes related with the uptake of H-atoms by copper are 19,47,48 : where Cu • H (s) denotes H-atoms dissolved in bulk copper, D f is the diffusion coefficient for the H-atom in bulk polycrystalline copper and k 2 is the second-order rate constant. R1 is fast, occurs in the diffusion limit and has no activation energy large enough that could allow its experimental determination. For H-atoms adsorbed at the Cu surface, R2 and R3 are competing processes. At a copper surface in contact with a gas-phase, R2 occurs without energy barrier at the (111) surface plane 49 . Yet, as highlighted by the authors of the cited study, the kinetics and energetics of such a process are dependent on the type of surface-the crystallographic plane-as well as the presence of surface defects. In aqueous media, R2 is expected to be slower-and have an energy barrier-due to the presence of surface defects and solvation effects as well as the hindering that these pose to the encounters between surface H-atoms when compared to the solid-gas interface 38 . The energy barrier for R3 corresponds to the transfer of an H-atom from solution into bulk Cu. This process is composed of several elementary steps that the H-atoms undergo and that occur in different time-scales: diffusion from solution to the interface; adsorption onto the Cu surface; and diffusion from the Cu surface into the bulk. The measured energy barrier of 43.5 kJ·mol −1 corresponds to the rate-determining step for the overall process described by reaction R3. For a surface adsorbed H-atom at a solid-liquid interface there is currently no kinetic data available in the literature for its diffusion onto the bulk Cu. Due to geometric factors, the growth of an irregular structure on the surface of Cu (Figs 5 and 6) can lead to a decrease in the yield of R2 and an increase in R3. This is because the oxygen-containing copper compounds formed on the metallic surface have a lower density and more defects than the metal surface. H-atoms adsorbed onto such irregular surfaces will have a lower probability for encounters when compared with the corresponding process occurring at the less defective metal surface. The rationale for this is that the H-atoms are more

Methods
Sample preparation and irradiation. All the solutions used in this study were prepared using water from a Millipore Milli-Q system. Samples of copper metal (Aldrich, 99.98%)-size: 5 × 5 × 0.25 mm-were subject to a surface treatment in order to remove traces of oxide, carbon dioxide and other possible contaminants. The samples were placed in an ultrasonic water bath for a period of 5 min and after removal were washed and transferred to a glove-box with an inert N 2 (g) atmosphere. Under the protective atmosphere the samples were washed with a solution of 5% amidosulfonic acid (H 3 NSO 3 ) (Aldrich, 99.9%) during 1 min. It has been shown previously that the amidosulfonic acid is an effective way for removing the products of copper oxidation from the copper surface 51 . After washing in the acidic solution, the copper samples were further washed with ethanol and water, and placed in vials containing 5 ml of water, sealed and irradiated. γ -Irradiation was performed using a MDS Nordion 1000 Elite 137 Cs γ -source with a dose rate of 0.135 Gy·s −1 , which was determined by Fricke dosimetry 52 .
Hydrogen measurements. The hydrogen content in the copper samples was determined using the thermal desorption spectroscopy (TDS) technique. The measurements were carried out with a thermal desorption apparatus designed and assembled at Aalto University, Finland. The apparatus consists of an ultra-high vacuum (UHV) chamber equipped with a vacuum furnace, coupled to a second air-lock vacuum chamber through which the samples are inserted. Additionally, there is a sample transportation mechanism with magnetic support for the sample basket. The pumping system has an effective pumping rate of 6.6 × 10 −2 m 3 s −1 . The mass spectrometer unit and the furnace heating system are controlled with a computer using the Lab View software. The vacuum in the UHV chamber is kept at 7 × 10 −9 bar. The heating system provides a direct control of the sample temperature in the temperature range from room temperature to 1123 K with a controlled heating rate that can be varied in the range 1-10 K·min −1 . Prior to the TDS measurements, the samples were cleaned with acetone in an ultrasonic bath for 1 min and dried in a He gas flow to remove the residual water from the samples surface. The size of the samples for TDS measurements was 0.9 × 4.0 × 10.0 mm. The hydrogen concentration in the copper samples was determined from integration of the area of the TDS peaks after background correction.
Scanning electron microscopy. For scanning electron microscopy experiments samples were taken from solution, dried with compressed air flow, placed onto stubs and immediately transferred to the microscope chamber, so that the total exposure of samples to air was less than 30 min. Images were collected using a Zeiss Gemini Ultra Plus scanning electron microscope and a FEI Quanta 3D Dual Beam system, equipped with an Oxford Instruments energy-dispersive detector.