The effects of chloride and sulphate on the growth of sulphide films on copper in anoxic sulphide solutions

The corrosion of copper in chloride‐ and sulphate‐containing solutions containing sulphide (ranging from 5 × 10−5 M to 10−3 M) was studied by electrochemical impedance spectroscopy and the structure and physical properties of the copper sulphide deposits (Cu2S) were analysed by scanning electron microscopy. At low [SH−], a porous and non‐protective deposit was formed, and the low corrosion rates were limited by sulphide transport, partially within the deposit and partially in the bulk of the solution. At higher [SH−], the corrosion rates were much higher and initially more rapid in chloride‐containing than in sulphate‐containing solutions, suggesting a direct role for chloride in the interfacial corrosion process. When the copper sulphide deposit thickened, the rate became limited by CuI transport within the more compact and more crystalline deposit formed. Changes in the morphology of the deposit suggest that chloride adsorption on the surfaces of the deposit inhibited the incorporation of Cu3S3 clusters, transported from the corroding copper surface, into the growing crystals.


| INTRODUCTION
[6] Within this sequence of barriers, comprising the geologic formation, compacted bentonite, container and spent fuel, the sealed waste container plays a vital role by providing the primary barrier against radionuclide release to the groundwater. [7]lthough the initial exposure conditions within the DGR will be aerated and oxidising, the corrosion damage to Cu will be minor (<0.09 mm [4,5,7] ), since most of the entrained O 2 trapped in the DGR upon sealing will be consumed rapidly by mineral and microbial reactions occurring in the bentonite clay compacted around the container or by the corrosion of structural steel within the DGR. [5]12] Previous studies [12][13][14][15][16][17][18][19][20][21][22][23][24][25][26][27] showed that Cu would corrode rapidly in the presence of SH − , leading to the deposition of chalcocite (Cu 2 S [13] ) on the Cu surface.The structure and properties of the Cu 2 S film were shown to depend on the [SH − ], the transport flux of SH − to the corroding Cu surface, and the influence of other groundwater species (i.e., Cl − , SO 4  2− and HCO 3 − /CO 3 2− ), which interfered with sulphide film formation by competition with SH − for surface sites on the Cu surface. [26]This competition limited the formation of the chemisorbed Cu(SH) ads , which is the essential precursor for film formation.It has been shown that, under anodically polarised conditions, Cu 2 S film growth in SH − solutions with low ionic strength is controlled by ionic migration within pores in the film, while at high ionic strength, film growth is dependent on the [SH − ].In solutions with high ionic strength but low [SH − ], Cu 2 S film growth was limited by the rate of the interfacial anodic reaction to produce Cu (SH) ads at the base of pores in the film, and in solutions with high ionic strength and high [SH − ], Cu 2 S film growth was limited by the transport of Cu species in a more compact film.The presence of SO 4 2− was found to suppress the growth of porous Cu 2 S films to a greater extent than Cl − does under these anodic conditions.It was suggested that this could be attributed to the strong co-adsorption of SO 4 2− and H 2 O on the Cu surface. [26]ng-term studies [13][14][15]17,18,[20][21][22][23] have demonstrated that, under corrosion conditions, the structure and properties of the Cu 2 S film vary greatly with the exposure conditions. When the bul [SH -] (i.e., 5 × 10 -5 M), the SH -flux and the [Cl − ] (i.e., 0.1 M) were low, leading to even lower [SH − ] at the Cu surface, the Cu 2 S film was very porous and nonprotective, and its growth was controlled by the flux of SH − to the Cu 2 S film/electrolyte interface.However, when the bulk [SH -] was high (i.e., ≥5 × 10 -4 M), creating a higher SH -flux at the Cu surface, and the [Cl − ] was low (i.e., 0.1 M), the Cu 2 S film formed was compact and at least partially protective.Under these conditions, Cu corrosion was controlled by Cu + diffusion in the Cu 2 S film.With both high [SH − ] (i.e., 1 × 10 -3 M) and high [Cl − ] (i.e., ≥ 1.0 M), the structure and properties of the Cu 2 S film changed from compact to porous, with Cl − interfering with film growth in a number of ways: (1) displacing adsorbed SH − from the Cu surface to inhibit the first step in the overall corrosion process; (2) inducing and maintaining the porosity in the Cu 2 S film; and (3) facilitating the transport of Cu into the solution when the [Cl − ] was extremely high (i.e., 5.0 M). [20] However, when SO 4 2− was present rather than Cl − in SH − solutions, it was uncertain whether SO 4 2− exhibited an influence under corrosion conditions similar to that observed in electrochemical experiments.[26] In addition, there is a need to validate whether the rate-determining step for Cu canister corrosion under DGR conditions remains the same as that observed in a solution containing Cl − .In this study, we compared the corrosion potential (E CORR ) and the electrochemical impedance spectroscopic (EIS) responses of Cu in anoxic SH − solutions containing either SO 4 2− or Cl − at similar concentrations and characterised the surface and cross-sectional morphologies of the Cu 2 S formed.The primary aim of this work was to determine how the structure and properties of the Cu 2 S films changed with the type of anion present and to shed light on the impact of groundwater anions on the corrosion of nuclear waste containers under Swedish/ Finnish/Canadian DGR conditions.

| Sample preparation
The material studied was P-doped (30-100 wt.ppm), O-free Cu (<5 wt.ppm O, Cu-OFP), provided by the Swedish Nuclear Fuel and Waste Management Company (SKB).Cu disk working electrodes (1 cm in diameter) were cut from the copper block and connected to a stainless-steel shaft.The connection was painted with a nonconductive epoxy to prevent the solution from contacting the Cu/steel junction during the experiments.
The electrode was then heated to 60°C for 12 h to promote adhesion of the epoxy.The exposed flat Cu surface was ground successively with 240, 600, 800, 1000, and 1200 grit SiC paper and then polished to a mirror finish using 1 μm, 0.3 μm, and, finally, 0.05 μm Al 2 O 3 suspensions.Before each experiment, electrodes were washed with Type I water (resistivity = 18.2 MΩ•cm, obtained from a Thermo Scientific Barnstead Nanopure 7143 ultrapure water system), ultrasonically cleaned using reagent-grade methanol, and finally washed with Type I water, and dried using a stream of Ar.

| Long-term corrosion experiments
To ensure anoxic conditions, the long-term corrosion experiments were performed in an Ar-purged anaerobic chamber (Canadian Vacuum Systems Ltd.), maintained at a positive pressure (2-4 mbar) using an MBraun glove box control system.31][32] The solutions used were anoxic 0. Electrodes were immersed in the anoxic SH − solutions for various exposure periods at room temperature, 25 ± 2°C.A conventional three-electrode cell was employed, with the Cu electrode as the working electrode, a Pt plate as the counter electrode, and a saturated calomel reference electrode (SCE).Before each experiment, electrodes were cathodically cleaned at −1.5 V/SCE for 1 min and then at −1.15 V/SCE for 1 min to reduce the air-formed oxides during sample preparation.The corrosion potential (E CORR ) was monitored, and electrochemical impedance spectroscopy (EIS) measurements were performed intermittently over the full exposure periods.The EIS measurements were carried out over the frequency range from 10 5 to 10 −3 Hz using a Solarton 1287 electrochemical interface and a Solarton 1255B frequency response analyser at an ac potential amplitude of 10 mV.The validity of the impedance spectra was checked using the Kramers-Krönig transform.
Since the exposure periods were long, the [SH − ] was monitored weekly by measuring the pH.If SH − depletion had occurred, as indicated by a decrease in pH (since the formation of copper sulphide clusters would release protons [33] ), SH − was replenished to the initial concentration as described previously. [13]3 | Analyses of the Cu 2 S films Once removed from the solution, Cu electrodes were rinsed with Type I water for 1 min and dried with Ar gas.Analyses were then performed after a minimum period of interim storage in the anaerobic chamber (<30 min).The surface and cross-sectional morphologies of these corroded Cu electrodes were observed using a Leo 1540 scanning electron microscope (SEM) equipped with a focussed ion beam (FIB) (Zeiss Nano Technology Systems Division).The composition of the films formed on the surface of these Cu electrodes was qualitatively analysed by energy dispersive X-ray spectroscopy (EDS) using a Leo 1540 FIB/SEM microscope (with an O detection limit of 1 at.%).

| E CORR measurements
The evolution of the E CORR in anoxic 0.1 M Cl − -containing solutions containing different [SH − ] ranging from 5 × 10 −5 M to 1 × 10 −3 M as a function of immersion time is shown in Figure 1a.On switching to an open circuit, the potential rose rapidly from the applied cathodic cleaning value (i.e., ≤12 min).Generally, the value of E CORR then established was more negative at a higher [SH − ].All E CORR values exhibited positive shifts with the consumption of SH − , due to the deposition of Cu 2 S on the Cu surface as corrosion proceeded.SH − depletion was particularly rapid at the lowest [SH − ] (i.e., 5 × 10 −5 M), as indicated by the significant decrease in pH, Figure 1a.When the SH − was replenished, E CORR decreased to the original value of ~−0.82 V/SCE, indicating a competition between SH − and Cl − for surface adsorption sites. [20]At high [SH − ] (i.e., ≥ 5 × 10 −4 M), E CORR slowly increased to a slightly more positive value over an extended exposure time, indicating that minimal SH − depletion occurred, suggesting that the structure and properties of the Cu 2 S films may be different from those formed at the low [SH − ], since their structure and properties are highly dependent on the sulphide concentration, sulphide flux and the chloride to sulphide concentration ratio. [17] CORR -time plots recorded in SO 4

2−
-containing solutions over the same [SH − ] range are shown in Figure 1b.At low [SH − ], the solution pH did not decrease, nor was there a positive shift of E CORR , meaning that there was no depletion of the bulk [SH − ].This suggested that there is a lower SH − consumption rate, and hence a lower Cu 2 S film growth rate, than in Cl − -containing solution, indicating differences in the influences of SO 4 2− and Cl − on the Cu surface. [20]The strong co-adsorption of SO 4 2− and H 2 O indicated in electrochemical experiments [26] may inhibit the initial SH − adsorption step required to initiate Cu 2 S formation.At high [SH − ], the E CORR responses were similar to those observed in Cl − -containing solution, suggesting a similar film growth process.

| EIS measurements
EIS measurements were conducted over the exposure periods shown in Figure 1 to characterise the corrosion (film growth) process occurring.Figure 2 shows EIS spectra (in the form of Bode plots) recorded as a function of immersion time in an anoxic 0.1 M SO 4 2− + 5 × 10 −5 M SH − solution.As previously observed in anoxic 0.1 M Cl − + 5 × 10 −5 M SH − solutions, [14] the spectra exhibited two poorly-defined capacitive loops, identified by the red and blue arrows in Figure 2b.In the study in Cl − -containing solutions, the response at high frequencies (marked by the red arrow) was attributed to the interfacial charge transfer process involving the creation of the Cu I species at the Cu surface and the associated double layer capacitance distributed across the whole surface, either exposed metal or Cu 2 S since Cu 2 S is electrically conducting. [13,14]The Cu I species was stabilised by complexation with adsorbed SH − , and this complex is the precursor to the formation of the Cu 2 S film.The lower-frequency response was attributed to the formation of an adsorbed surface layer.The linear increase in log|Z| (the logarithm of the impedance modulus) at frequencies ≤10 −1 Hz (Figure 2a) and the increase in phase angle (theta) to a value in the range 45°-60°(Figure 2b) indicate a significant contribution to the control of the film growth process by SH − diffusion, both within the growing Cu 2 S film and in the bulk of the solution, as was also observed in the Cl −containing solution. [14]The impedance response in the  [13,14,20]  high-frequency range (Figure 2b, red arrow) showed only a small change in position on the frequency scale with exposure time, indicating only a minor modification of the interfacial impedance response as the Cu 2 S film accumulated on the surface, as previously observed in the Cl − -containing solution. [14]he electrical equivalent circuit (Figure 3a) proposed previously [14] was used to fit these spectra.In this circuit, R S represents the solution resistance, CPE 1 is a constant phase element (CPE) representing the double-layer capacitance across the Cu 2 S/electrolyte interface, R 1 is the charge transfer resistance for the formation of Cu (SH) ads species at the Cu/electrolyte interface, CPE 2 and R 2 are the capacitance and resistance of the adsorbed surface layer, respectively, and W inf is an infinite Warburg element attributed to the SH − diffusion in solution.These fits yielded values of the overall interfacial polarisation resistance (R P ), with R P −1 proportional to the film growth rate (which can be considered as directly proportional to the corrosion rate).The change of the polarisation resistance is plotted as a function of exposure period for both SO previous study demonstrated that the film growth was proceeding under diffusion control in Cl − -containing solutions, [14] and that film deposition occurred at the Cu 2 S/electrolyte interface, [23] the dependence of film growth rate on SH − supply is consistent with a porous film which does not act as a significant corrosion barrier as it thickens.
When [SH − ] was increased to 5 × 10 −4 M, two time constants similar to, but better defined than those observed at the lower [SH − ], were observed, as shown in Figure 5a,b.This behaviour was similar to that observed previously in Cl − -containing solutions with the same [SH − ], [13] and shows that the overall film growth process proceeded by a mechanism similar to that observed at the lower [SH − ], but with a number of observable differences.The location of the two time constants on the frequency axis shifted to slightly lower frequencies, consistent with both a decrease in the rate of the interfacial (Cu/Cu 2 S) process (as indicated also by an increase in |Z| as the frequency decreased, Figure 5a) and a higher resistance (increased protectiveness) of the Cu 2 S layer.The phase angle at the low-frequency limit was much lower, as shown in Figure 5b than that observed at [SH − ] = 5 × 10 −5 M, as shown in Figure 2b, indicating a significantly lower contribution of SH − diffusion to the impedance at this higher [SH − ].Comparison to previously recorded spectra over a similar exposure period showed that at short exposure times (i.e., ≤100 h), the phase angle recorded in Cl − -containing solution approached zero at 10 −3 Hz, indicating a minimal contribution of SH − diffusion to the early stages of film formation.As the exposure time increased (up to 1691 h) 2− -(Figure 2a) and in Cl − -containing [14] solutions.[Color figure can be viewed at wileyonlinelibrary.com] in Cl − -containing solution, the phase angle increased to values approaching 15°-20° [ 13] indicating an increasing contribution to the impedance from SH − diffusion as the Cu 2 S film accumulated.This difference in phase angle response indicated a more significant effect of SH − diffusion on the film growth process in SO 4
When the [SH − ] was ≥5 × 10 −4 M, the change in film structure from porous to compact required a change in the electrical equivalent circuit (Figure 3b) required to fit the EIS spectra in Figure 5 as previously used. [13]In this circuit, R S represents the solution resistance, C 1 is the double-layer capacitance across the Cu 2 S/electrolyte interface since the sulphide film is electrically conducting, R 1 is the charge-transfer resistance for the formation of Cu(I) species at the Cu/Cu 2 S interface, CPE 2 and R 2 are the capacitance and resistance of the adsorbed surface layer, respectively, and W S is a short Warburg element attributed to a diffusion process of Cu + through the sulphide film.The sum of R 1 and R 2 would yield the value of the polarisation resistance, R P .Figure 6 shows the R P values obtained in SO 4 2−and Cl − -containing solutions obtained from EIS fitting (Figure 5a and Chen et al. [13] ) plotted as a function -containing solutions.This difference provides a measure of the major influence of diffusive transport in determining the rate of film growth at the lower [SH − ].The R P values at short exposure times (≤200 h) show film growth was more rapid (i.e., R P was lower) in Cl − -containing than SO 4 2− containing solutions, but with exposure time in Cl −containing solutions, decreased to a value just slightly less than that in SO 4 2− -containing solution.Based on the differences in the values of the phase angle at the lowfrequency limit, this appears attributable mainly to a much more significant contribution of SH − diffusion in the films present on Cu in Cl − -containing solutions than in SO 4 2− -containing solutions, suggesting that the film present in Cl − -containing solution was initially significantly less of a barrier to SH − transport.This is consistent with electrochemical results recorded under transportcontrolled conditions in which lower anodic limiting currents were recorded in SO 4 2− -containing solutions than in Cl − -containing solutions. [26]At longer exposure times (>500 h), the film growth rate remained slightly lower in the SO 4 2− -containing solution (R P was slightly higher), suggesting a slightly, but not significantly, lower barrier for SH − diffusion to the corroding Cu surface.-(Figure 4a) and Cl − -containing solutions [13]  -containing solution, the impedance spectra were very similar to those recorded at the lower [SH − ] of 5 × 10 −4 M and did not change significantly as the exposure time increased.
Figure 9 shows that the values of R P as a function of immersion time were very different in Cl −and SO 4 2− containing solutions at short exposure periods but effectively the same beyond an exposure period of ~1000 h.At short exposure periods, the film growth rate in Cl − -containing solutions was considerably higher than in SO 4 2− -containing solutions, as observed at the lower − before its conversion to Cu 2 S. [14] However, once Cu 2 S film formation commenced, the corrosion rate decreased as the film The coincidence in the film growth rates at longer times demonstrated that, for sufficient film coverage, the corrosion process became dominated by the properties of the film, with any influence of the kinetics at the Cu/ Cu 2 S interface obscured.This contrasted with the observations at the lower [SH − ] of 5 × 10 −5 M, when the difference in film growth rate at long times, Figure 4, remained sensitive to the nature of the anion.For the two higher [SH − ] the differences in film growth rate at longer exposure times were marginal, suggesting that there was little difference in the ability of the Cu 2 S films present to control the interfacial corrosion rate by transport of species through the sulphide film.This could indicate that the transport of Cu + species through the sulphide film, to be incorporated into the film at the Cu 2 S/ electrolyte interface, limits the overall rate at these high [SH − ] due to the structure and properties of the deposited sulphide film.

| Surface and cross-sectional morphologies of films grown at a low ionic strength
Figure 10a,b shows SEM micrographs of the surface and a FIB-cut cross section of a specimen corroded for 3015 h in a 0.1 M SO 4 2− -containing solution containing 5 × 10 − 5 M SH − .The deposit comprised a mixture of small crystals and agglomerated small particulates apparently less than ~100 nm in dimension, with the cross section demonstrating that the film was porous and non-protective.While the Cu/Cu 2 S interface was slightly roughened, there was no evidence of any locally penetrated sites, confirming that corrosion had progressed generally uniformly across the surface and that a significant area of the Cu substrate remained exposed to the solution.A similar crystalline/nanoparticulate deposit was observed after corrosion in a 0.1 M Cl −containing solution containing the same [SH − ], [14] although the Cu 2 S film was less uniformly distributed and visibly more porous.These observations are consistent with the impedance analyses (Figures 2 and 4) and our previous results obtained in 0.1 M Cl − -containing solution containing 5 × 10 −5 M SH − which demonstrated film growth controlled by SH − diffusion, [14] with the process occurring slightly faster at longer exposure times in the Cl −containing solution.Diffusion-controlled film growth was demonstrated at this [SH − ] in Cl − -containing solution by the observation of a linear film growth law (based on measured film thicknesses [14] ).The mixture of crystals and nanoscale particles within the film was to be expected, since it has been demonstrated that Cu I transport from the corroding surface occurs via a combination of ionic complexes (Cu(SH) 2 − ) and Cu 3 S 3 clusters. [18,19]The agglomeration of clusters was to be expected in the solutions used in this study, which had significant ionic strengths (see section 3.4 below).When the [SH − ] was increased to 5 × 10 −4 M, the Cu 2 S deposit was much thicker, consistent with the increased flux of SH − , and more crystalline than observed at the lower [SH − ], Figure 11a,b.Inspection of the Cu/Cu 2 S interface showed no evidence of localised penetrations.A similar more crystalline and dense film was also observed after growth for a similar exposure period in Cl − -containing solution, [13] although the surface morphology of the films was different, indicating an influence of the anions on the Cu 2 S deposition process.Additionally, in Cl − -containing solution, the microgalvanic coupling could have led to nonuniform film growth and localised penetrations when [SH − ] was increased to this concentration. [18]lose inspection of the surfaces of the deposited crystals, Figure 11a, shows they were decorated with a fine particulate layer.Such a layer was much more obvious on the surface of Cu 2 S crystals grown in Cl −containing solutions. [18]Since these large crystals grew by the incorporation of Cu(SH) 2 − /Cu 3 S 3 species transported from the corroding surface, this difference suggests that the incorporation of Cu 3 S 3 species may be hindered more by the adsorption of Cl − than SO 4 2− ions at the Cu 2 S surfaces.This could partially account for the apparently more extensive film growth observed in SO 4 2− -containing solutions, based on film thickness measurements, shown in Figure 12, despite the evidence from the impedance measurements that the corrosion rate was slightly lower in the presence of SO 4 2− as shown in Figure 6.In the presence of either of the two anions, the film growth rate (based on film thickness) exhibited a logarithmic law, indicating the formation of a more compact and protective film than observed at 5 × 10 −5 M SH − .This possible explanation is supported by the analytical detection of Cu in Cl − -containing solutions, which would suggest the transport of some clusters into the solution.If this was the case, the film thickness measurement would have underestimated the corrosion rate.At the highest [SH − ] employed in this study (10 −3 M), the EIS data, shown in Figure 9, showed that, despite significant differences in the film growth rate at short exposure times, the rates beyond ~100 h were effectively the same, suggesting that film growth was dominated by the transport of Cu I through Cu 2 S deposits with similar properties, leading to a uniform, predominantly crystalline deposit.Again, an inspection of the Cu/Cu 2 S interface showed no evidence of localised penetrations as shown in Figure 13.
By contrast, after a similar exposure period the film deposited in a Cl − -containing solution containing the same [SH − ] showed areas only sparsely covered by a deposit adjacent to areas covered by very large crystals, indicating a nonuniform deposition process.Such a situation has been shown to lead to localised penetrations that can be attributed to micro-galvanic coupling of sparsely covered, net anodic areas to densely covered, and net cathodic areas at which Cu(SH) 2 − /Cu 3 S 3 deposition occurs, a process which can lead to localised penetrations of the Cu substrate. [20]A more thorough study would be required to determine whether such a process was avoided in SO 4 2− -containing solutions.The blue plot is cited from Chen et al. [13] to compare the growth of the Cu 2 S film in SO 4 2− -containing solution.
[Color figure can be viewed at wileyonlinelibrary.com] | 1673 The corrosion rate data suggest micro-galvanic coupling may be confined to Cl − -containing solutions in which the initial corrosion rates, when only minor amounts of Cu 2 S have been deposited, are much higher than in SO 4 2−containing solution.This higher corrosion rate in Cl −containing solutions would facilitate the formation of more copper sulphide clusters and complexes, and thus would accelerate the growth of the deposited Cu 2 S crystals, leading to their non-uniform distribution on the Cu surface [20] (Figure 14).

| The influence of high ionic strength on film morphology
The above results, and those in Cl − -containing solutions published previously, [13][14][15][16][17][18][19][20] show that the nature of the anion had an influence on both the morphology and porosity of the deposited Cu  1]), which may reflect a slight increase in film growth rate at the higher concentration, although it is not possible to verify this without impedance measurements.After 3015 h of exposure, the deposited film was composed almost exclusively of fine particulates, with only a smattering of small crystals as shown in Figure 15a.
This film morphology was radically different from that observed when the deposit was grown in a solution containing the lower concentration of 0.1 M, Figure 11a.The cross section in Figure 15b shows that this finely particulate outer layer obscured a sublayer of irregular crystal growth, with some areas of the Cu surface covered by thick crystals and some only by a thin layer whose morphology was difficult to discern but appeared crystalline.It was clear that, at the high ionic strength prevailing in the 0.5 M solution, the agglomeration of Cu 3 S 3 clusters limited their incorporation into growing crystals, as observed at [SO 4 2− ] = 0.1 M. While the transformation of Cu 3 S 3 clusters into large Cu 2 S crystals is not well understood, these results show that it is strongly inhibited at high ionic strength, consistent with the results obtained in concentrated Cl − -containing solutions (i.e., 5.0 M). [20] In a Swedish/Finnish DGR, the groundwater anions of dominant interest will be chloride and sulphate, and their expected concentrations 10 000 years after DGR closure are in the range from 3 × 10 −4 M to 0.154 M (chloride) and from 2 × 10 −4 M to 7.2 × 10 −3 M (sulphate), respectively. [5]In Canada, the maximum [Cl − ] in the reference crystalline rock (CR-10) and reference sedimentary rock groundwaters (SR-270) is 0.17 M and 4.74 M, respectively, with the respective maximum [SO 4 2− ] being 1.04 × 10 −2 M and 1.86 × 10 −2 M. [33] Since [SH − ] in Swedish/Finnish/Canadian DGRs will be low (e.g., [SH − ] in a Swedish DGR is expected to be in the range from 10 −7 M to 10 −4 M, [5] and that in a Canadian DGR is not expected to exceed 10 −5 M [34] ), the sulphide film formed on the Cu container would be porous, with growth controlled by SH − transport to the container surface from remote locations.-containing solution as the flux of SH − to the corroding surface increased.At short exposure times, before the significant accumulation of a surface deposit, the corrosion rate was considerably higher in Cl − -containing solution, suggesting that the formation of Cu I chloride complexes may have accelerated corrosion before Cu 2 S deposition eventually suppressed it.At longer exposure times, as the deposit accumulated, the corrosion rates in both solutions converged, indicating that the Cu I transport through the more compact crystalline deposit eventually became the rate-controlling step./Cl − concentrations significantly higher in Swedish/Finnish/ Canadian groundwaters, the copper sulphide deposits formed on waste containers in a DGR would be porous and non-protective, leading to corrosion under SH − transport control.

F
I G U R E 1 The corrosion potentials of P-doped O-free Cu as a function of immersion time in: (a) anoxic 0.1 M Cl − -containing solution containing different [SH − ] (from references ); and (b) anoxic 0.1 M SO 4 2− -containing solution containing the same [SH − ]. [Color figure can be viewed at wileyonlinelibrary.com]F I G U R E 2 Electrochemical impedance spectroscopic (EIS) spectra (in the form of Bode plots, [a] and [b]) recorded on P-doped O-free Cu after immersion in an anoxic 5 × 10 −5 M SH − + 0.1 M SO 4 2− solution for different exposure times.[Color figure can be viewed at wileyonlinelibrary.com]

F I G U R E 3
The electrical equivalent circuits used to fit the experimentally acquired electrochemical impedance spectroscopic (EIS) spectra: (a) in anoxic 5 × 10 −5 M SH − solution; and (b) in anoxic ≥5 × 10 −4 M SH − solutions.F I G U R E 4 The polarisation resistance obtained from the electrochemical impedance spectroscopic (EIS) spectra recorded as a function of immersion time in anoxic SO 4 of exposure time.The values of |Z| were between 1 and 2 orders of magnitude lower than those recorded at the lower [SH − ] (5 × 10 −5 M) in both Cl −and SO 4 2−

F I G U R E 5
Electrochemical impedance spectroscopic (EIS) spectra (in the form of Bode plots, [a] and [b]) recorded on P-doped O-free Cu after immersion in anoxic 5 × 10 −4 M SH − + 0.1 M SO 4 2− solution for different exposure times.[Color figure can be viewed at wileyonlinelibrary.com]F I G U R E 6 The polarisation resistance obtained from electrochemical impedance spectroscopic (EIS) spectra recorded in anoxic SO 4 2− as a function of immersion time.[Color figure can be viewed at wileyonlinelibrary.com]EIS spectra were obtained in both Cl −and SO 4 2− -containing solutions when [SH − ] was further increased to 10 −3 M, Figures 7 and 8.At lower [SH − ], the same two time constants were observed.In the Cl −containing solution, only the high-frequency time constant was initially observed.This time constant remained visible in the spectra as the low-frequency time constant developed, confirming that the interfacial impedance response remained detectable despite the accumulation of the Cu 2 S deposit.At very short exposure times (up to 21 h), the phase angle at the low-frequency limit approached zero, indicating minimal control of the film growth process by SH − diffusion.At intermediate exposure times (up to 83 h), the phase angle at 10 −3 Hz was markedly higher, indicating an increased control of film growth by SH − transport in the accumulating deposit.At the longest exposure periods (>167 h), the phase angle approached ~60°(Figure 7b), indicating that the film was becoming a substantial barrier to corrosion.

[
SH − ] of 5 × 10 −4 M. It confirms that, in the absence of an accumulation of Cu 2 S, the film growth rate is high in Cl − -containing solution, as observed electrochemically in anodic polarisation experiments.These observations suggest Cl − may have initially accelerated corrosion, possibly by facilitating dissolution as CuCl 2

F I G U R E 7
Electrochemical impedance spectroscopic (EIS) spectra (in the form of Bode plots, [a] and [b]) recorded on P-doped oxygen-free Cu after immersion in anoxic 1 × 10 −3 M SH − + 0.1 M Cl − solution for different exposure times.[Color figure can be viewed at wileyonlinelibrary.com]F I G U R E 8 Electrochemical impedance spectroscopic (EIS) spectra (in the form of Bode plots, [a] and [b]) recorded on P-doped oxygen-free Cu after immersion in anoxic 1 × 10 −3 M SH − + 0.1 M SO 4 2− solution for different exposure times.[Color figure can be viewed at wileyonlinelibrary.com] accumulated.By contrast, SO 4 2− initially inhibited Cu 2 S formation, as observed electrochemically, with the corrosion rate increasing slightly with exposure time.

F
I G U R E 10 Scanning electron microscope surface (a) and cross-sectional (b) images of P-doped O-free Cu after 3015 h immersion in anoxic 5 × 10 −5 M SH − + 0.1 M SO 4 2− solution.The red dashed line shows the interface of Cu 2 S film and Cu matrix.[Color figure can be viewed at wileyonlinelibrary.com]F I G U R E 9 The polarisation resistance from electrochemical impedance spectroscopic (EIS) spectra recorded in anoxic 1 × 10 −3 M SH − + 0.1 M SO 4 2− and 1 × 10 −3 M SH − + 0.1 M Cl − solutions as a function of immersion time.[Color figure can be viewed at wileyonlinelibrary.com]

F
I G U R E 11 Scanning electron microscope surface (a) and cross-sectional (b) images of P-doped O-free Cu after 1691 h immersion in an anoxic 5 × 10 −4 M SH − + 0.1 M SO 4 2 − solution.F I G U R E 12 Growth laws and average thicknesses of the Cu 2 S films formed on P-doped O-free Cu after exposure to anoxic SH − solutions containing different [SO 4 2− ] or [Cl − ] as a function of immersion time.
2 S layer.To investigate this influence further, a long-term corrosion experiment was performed in a solution containing [SH − ] = 5 × 10 −4 M with a [SO 4 2− ] increased to 0.5 M. The evolution of E CORR with exposure time in this solution exhibited a larger shift to more positive values (than that in [

F
I G U R E 14 The corrosion potentials recorded on P-doped O-free Cu as a function of immersion time in anoxic 5 × 10 −4 M SH − solutions containing different [SO 4 2− ]. [Color figure can be viewed at wileyonlinelibrary.com]F I G U R E 15 Scanning electron microscope surface (a) and cross-sectional (b) images of P-doped O-free Cu after 3015 h immersion in an anoxic 5 × 10 −4 M SH − + 0.5 M SO 4 2− solution.F I G U R E 13 Scanning electron microscope surface (a) and cross-sectional (b) morphologies of P-doped oxygen-free Cu after 1691 h immersion in anoxic 1 × 10 −3 M SH − + 0.1 M SO 4 2− solution.

4 | 2 −.
SUMMARY AND CONCLUSIONS• The corrosion rate of Cu varied with both [SH − ] and the presence of either Cl − or SO 4 At low [SH − ], the Cu 2 S deposit was finely particulate in the presence of either anion.The film growth (corrosion) rates were determined by SH − transport to the Cu surface, partially within the deposit and partially in the bulk of the solution, with long-term rates higher in Cl − -containing solutions, suggesting that this anion maintains a more open porosity in the Cu 2 S deposit than does SO 4 2− .• As the [SH − ] was increased, the corrosion rate increased markedly in both Cl −and SO 4 2−

•
The morphology of the Cu 2 S deposit formed at [SH − ] ≥ 5 × 10 −4 M was influenced by the nature of the anion, with a less uniform deposit formed in Cl − -containing solution.The more extensive decoration of the surfaces of the deposit by nanoparticles in Cl − -containing solutions than in SO 4 2− containing solutions suggests that Cl − adsorption on the Cu 2 S surfaces may have inhibited the incorporation of Cu 3 S 3 clusters (transported from the corroding Cu surface) into the surface Cu 2 S crystals and led to their release to the bulk of the solution.At high ionic strength (i.e., in 0.5 M SO 4 2− -containing or 5.0 M Cl −containing solution) agglomeration of such clusters significantly limited their incorporation into the growing crystals.• The much higher initial corrosion rates in Cl −containing solutions than in SO 4 2− -containing solutions ([SH − ] ≥ 5 × 10 −4 M) may account for the more uneven distribution of deposited crystals observed in Cl − -containing solution.Since such uneven distributions of deposited crystals have been associated with the occurrence of localised penetration into the Cu surface due to micro-galvanic coupling, this suggests that such a coupling process was less likely in SO 4 2− containing solutions.• Since the [SH − ] is expected to be ≤10 −4 M and SO 4 2− 3. The anaerobic chamber was maintained at a total [O 2 ] ≤4.2 mg/m 3 , which included the contribution from both O 2 and water vapour residues.Using Henry's law, the [O 2 ] dissolved in the solution was calculated to be ≤1.3 × 10 −9 M. The actual [O 2 ] in any solution in the chamber would be less than this value.Despite the possibility of minor traces of O 2 being present, it is unlikely that copper oxides played any role, since Cu 2 S is more stable in SH − solution than Cu 2 O, based on thermodynamic data for the conversion from Cu 2 O to Cu 2 S in SH − solutions: Cu 2 O (s) + SH − (aq) → Cu 2 S (s) + OH − (aq), ΔG°= −101.46 kJ/mol at 298 K