Current status of copper corrosion research at SKB

Transcription

1 Current status of copper corrosion research at SKB KYT Copper Corrosion Seminar Christina Lilja

2 Overview Status in the Swedish Programme Corrosion issues mentioned by SSM from safety assessment Ongoing corrosion research sulphide corrosion oxidising conditions stress corrosion cracking Recent results corrosion in pure water Summary

3 The Swedish system

4 Spent fuel management License application for spent fuel repository sent in March 2011 For 2 facilities spent fuel repository in Forsmark (Östhammar) encapsulation plant (extension of the existing interim storage, Clab) in Oskarshamn Will be tried according to 2 laws, in parallel Nuclear Activities Act handled by SSM (the Swedish Radiation Safety Authority) Environmental Code handled by the Environmental Court Östhammar Oskarshamn

5 Updated time schedule Environmental Court main hearing held September-October 2017 statement to government, according to the Environmental Act planned December 20 SSM Review comments preliminary evaluation issued June 2016 (sent to Environmental Court) final version will be sent to government around new year 2018 Also needed: decisions of acceptance from Östhammar and Oskarshamn municipalities (spring 2018) Construction license Government decision 2019(?) license to start constructing the repository (= start making tunnels etc) according to both Acts at the same time SKB to send in PSAR (Preliminary Safety Assessment Report) to SSM and get it approved before start of construction

6 SSM Review comments, June 2016 SSM Granskningsrapport, in Swedish, 683 pages a summary included in the decision document (17 pages), also in English Summary section focuses on unsaturated conditions [next slide] saturation time could be in the range years Main text touches on many corrosion issues probabilistic assessment of pitting mechanisms and time for consumtion of entrapped oxygen stress corrosion cracking in sulphide solution corrosion from high voltage direct current cables irradiation effects stress corrosion cracking at the inside of the copper

7 SSM list of most urgent issues Localised corrosion by sulphide under unsaturated buffer conditions Larger amounts of sulphide (than calculated for saturated conditions) could come from gaseous H2S (from microbial sulphate reduction) or from a biofilm formed at the canister surface. Could maybe cause a passive film, and when broken, lead to pitting. SCC under the same conditions Formation of a passive film is most probably needed. Creep in the copper shell The fundamental understanding of the effect of phosphorous on creep ductility is not enough to completely rule out a brittle failure mechanism. A possible effect of hydrogen on the deformation properties of the copper need to be considered. The presence of hydrogen in copper in coupled to processes like gamma radiolysis, sulphide corrosion and anoxic corrosion in pure oxygen free water. General corrosion at high chloride concentrations together with high sulphide concentrations, under unsaturated condition

8 Issues from SR-Site (I) Sulphide corrosion mass transport limitation? Measurements of sulphide depletion as a function of time Comparison with estimated flux through a diffusion layer gives a limit flux that is still rate determining (at even higher fluxes the kinetics of the electrochemical reactions can be rate determining) Conclusions for SA Also investigations of kinetics of reactions, and studies of film growth rates Comparison with fluxes of sulphide in the repository (modelling based on ground water fluxes and sulphide diffusion) => the repository fluxes are lower than these limiting fluxes, and thus sulphide corrosion is mass transport limited

9 Issues from SR-Site (II) SRB activity in buffer and backfill does it happen and what is the limiting factor? Part of Integrated Sulphide Project joint SKB-Posiva project WP2 experiments in bentonite => most bentonites show a threshold in density for microbial activity WP 3 modelling of the near-field 3 modelling teams (Unibern, Amphos+Claytech, Fraser King) Base case and Variant cases ongoing, reporting during spring 2018 => preliminary results indicate the precipitation of FeS(s) is very important and the limiting process on how much sulphide that reaches the canister

10 Ongoing sulphide (I) Sulphide (University of Western Ontario), mostly electrochemical studies (voltammetry, EIS, SEM etc) passivity is it possible and under what conditions? important to make scan in backward direction to distinguish passivity from onset of general corrosion all evidence points to passivity requires sulphide fluxes higher than in repository 5.0x10-5 M Na 2 S and 25 Hz 2.0x10-3 M Na 2 S and 25 Hz nature of cathodic reaction is it HS - (H 2 S) or H + (H 2 O) that is reduced? 2Cu + HS - + H 2 O Cu 2 S + H 2 + OH - will also be investigated with DFT (density functional theory) modelling at KTH

11 Ongoing sulphide (II) Stress corrosion cracking in sulphide (Swerea Kimab) attempt to repeat the Japanese study from 2008 (Tanaguchi and Kawasaki) only ductile fractures maybe some superficial cracks but the same as in creep experiments without corrosion? report under review internally SKB calculations of tensile stresses in the canister (never going through all the thickness), studies of possibilities for crack growth Localised corrosion (Swerea Kimab, Micans), exposure of copper to gaseous H 2 S, 10 and 10 4 ppm solution with microbs (10and 30 days, different amounts of lactate and sulphate) will look for pits with optical microscopy

12 Ongoing localised corrosion Localised corrosion under oxidising conditions SKB co-operation with NWMO measurements of active-passive regions for ph, Cl -, CO 3 2-, SO 4 2- (done at UWO) also work with multi-array electrodes, modelling etc

13 Recent research Äspö HRL Long-term experiments at Äspö Hard Rock Laboratory: Minican, LOT, Prototype etc not always designed to be corrosion experiments more difficult to interpret results concerning corrosion Some general experiences, example Minican first period, 1 month, of corrosion by oxygen (rate 1.5 µm/y) after that sulphide corrosion, 8.4 years (rate µm/y) misleading to use one average corrosion rate! O:S ~ 70:30 pits and surface defects => from manufacturing as the same in material stored at the shelf have similar appearance Prototyp Rör T58

14 Ongoing hydrogen Hydrogen from corrosion: irradiation, sulphide corrosion, (corrosion by pure water) intrusion in the copper? to be able to measure any difference at exposure - need to check accuracy in methods reproducibility in methods prepared by Kimab example from fusion analysis, SKB R Round-robin test large spread, also within some laboratories => very sensitive to surface treatment freshly prepared by lab

15 Corrosion in pure water - the issue Based mainly on experiments by Gunnar Hultquist, a group at KTH results published 1986, and claiming a reaction of copper with pure water driving force, a new, hypothetical, stable phase: Cu + yh 2 O H x CuO y + (y-x/2) H 2 with equilibrium pressure: p H2 =1 mbar (at around 45 C) Could be compared to established thermodynamic data the dominating reaction: Cu + H + Cu + + ½ H 2 equilibrium calculation including all known phases and species of Cu-O-H gives: p H2 = mbar, [Cu]= M (at 25 C) If this new driving force exists, the long-term durability of e.g. copper canisters in a KBS-3 repository would possibly be affected as the availability of water is more or less unlimited => SKB used several different approaches to investigate this issue

16 Uppsala University set-up Studies of hydrogen evolution in metal (stainless steel) contained system (similar to Gunnar Hultquist s set-up) Also studies of formation of copper corrosion products in separate set-ups of lower chambers sealed with Pd membrane in glove box

17 Uppsala University - results first phase of experiments ( ): hydrogen was detected but the same amount in exp. without copper! second phase ( ), with redesigned equipment: no hydrogen pressure above background separate chambers, exposed 0-29 months surfaces investigated with XPS: no oxide layer formed water analysed with ICP-MS: not increasing with time blue background with water and glass red with copper in water

18 Micans set-up Studies of hydrogen evolution in glass test tubes prepared in glove box, stored in N 2 atmosphere, at 70 C gas phase sampled with syringe through the rubber stopper analysed with GC different copper qualities and surfaces treatments used Series N8 after 806 days, no visual change of surface

19 Micans results Results with Cu-OFP 99,95% (SKB canister copper), samples mm 3 yields hydrogen gas, measured up to 800 days evacuated several times (new N 2 gas) new hydrogen is measured hydrogen also without water => out gassing from the material! this hypothesis corroborated by first heating copper and then exposing no hydrogen in the tubes H 2 (mbar) also polished Time (days) N8:1 N8:2 N8:3 N8:4 N8:5 N8:6 N8:7 N8:8 N8:9 N8:10 N8:11 N8:12 N8:13 N8:14 N8 K:1 N8 K:2 N8 K:3 N8 K:4 N8 K:5

20 Hydrogen partial pressure [mbar] Micans results Results with Cu-OF 99.95% (Goodfellow), foils mm 3 the same Cu as Gunnar Hultquist used (publications 2009, 2013, 2015) Conclusion: no hydrogen above background % Cu-OF; Goodfellow As received Electropolishing + reduction + heating to 400 C (Method I) Heating to 400 C only Water only; no copper Data from Hultquist et al Hultquist et al. [7] Time [days]

21 Other results looking for new phase KTH: Pavel Korzhavyi, Yunguo Li DFT calculations of possible configurations of CuOH structure of CuOH resembles structure of cuprite and cubic ice (ice-vii) calculation of G (and H ): less stable than Cu 2 O Cu + O 2- H + KTH: Inna Soroka synthesis and characterisation of CuOH recipe from 1955 reduction of Cu(II) with Fe(II) EDTA solution studied with XRD, FTIR, XPS, SEM results: a hydroxide exists: CuOH H 2 O, but it is less stable than Cu 2 O Cu + O 2- H + Conclusion: no new driving force for corrosion

22 Other results surface reactions KTH: Joakim Halldin Stenlid, Claudio Lousada DFT calculations of H 2 O on Cu and Cu 2 O surfaces ideal surfaces, and with defects Results ~½ monolayer OH can form ( G <0) from 1 mono-layer H 2 O, but not further H 2 dissociates from Cu; less clear for Cu 2 O corresponds to some ng H 2 per cm 2 Conclusion: surface reactivity can not explain the amount of H 2 in Hultquist s experiment (ca 100 ng H 2 per cm 2 )

23 Corrosion in pure water - conclusion SKB cannot find any support for the existence of a sustained corrosion of copper in pure, O 2 -free water above the very limited extent predicted by established thermodyamic data This extent described by thermodynamic data is completely negligible compared to other corrosion phenomena in a typical repository environment.

24 Summary current corrosion research Focus on unsaturated conditions SSM questions to be handled to PSAR Sulphide most important long-term corrodant prerequisites for formation of a passive film? stress corrosion cracking occurrence (stresses, material, environment)? mechanism? Corrosion i pure oxygen-free water SKB cannot find any support for a sustained corrosion process extending what established thermodynamics predict

25 Read more SKB reports, examples Minican: TR-16-12, TR Surface studies of copper: P-13-50, P Hydrogen Round robin: R Cu in pure water: TR-15-03, TR-16-01, TR Many papers recent examples from many of the areas may be found in the proceedings from the 6th Int. Workshop on Long-term Prediction of Corrosion Damage in Nuclear Waste Systems, published in Corrosion, Engineering, Science and Technology, 2017, Vol 52, No 51 (open access)

26 Thank you for your attention!