1. EU-PWI TF meeting, Ljubljana, 13-15 November 2006 Progress with tritium removal and mitigation 2005-6 G Counsell 1, D Borodin 4, P Coad 1, J Ferreira 8, C Grisolia 2, C Hopf 3, W Jacob 3, A Kirschner 4, A Kreter 4, K Krieger 3, J. Likonen 5, A. Litnovsky 4, A.Markin 3, V Philipps 4, J Roth 3, M Rubel 6, E Salancon 3, A. Semerok 7, FL Tabares 8, C. Tomastik 9, A Widdowson 1 1 EURATOM/UKAEA Fusion Association, Culham Science Centre, Abingdon, OX14 3DB, UK 2 Association EURATOM-CEA, CEA/DSM/DRFC Cadarache, 13108 St.Paul lez Durance, France 3 Max-Planck-Institut für Plasmaphysik, EURATOM Association, D-85748 Garching, Germany 4 Institut für Plasmaphysik, Forschungszentrum Jülich, Association EURATOM-FZJ 5 Association EURATOM-TEKES, VTT Processes, PO Box 1608, 02044 VTT, Espoo, Finland 6 Alfvén Laboratory, Royal Institute of Technology (KTH), Association EURATOM-VR, 100 44 Stockholm, Sweden 7 CEA Saclay, DEN/DPC/SCP/LILM, Bat. 467,91191Gif sur Yvette, France 8 Association Euratom/Ciemat. Laboratorio Nacional de Fusión. 28040 Madrid, Spain 9 Institut für Allgemeine Physik, Vienna University of Technology, Wiedner Hauptstraße 8-10, A-1040 Wien, Austria

2. 2/28 Outline A reminder of the problem Advances in modelling tritium retention with mixed materials New challenges – retention in gaps, under layers and in the bulk Progress with characterising de-tritiation techniques and associated problems Can we avoid retention in the first place? Future thoughts

3. 3/28 Acceptable ITER operation ~2500 shots before maintenance period Long term T retention/shot must be < 0.14g/400s shot Strategies for T removal essential if CFC targets in DT phase Removal efficiency must be 80% - 98% Clear need for T removal schemes

4. 4/28 Preliminary modelling using local erosion & deposition model ERO (still many open questions) Model assumptions validated against TEXTOR C 13 injection experiments >80% of wall area in ITER is beryllium Eroded Be will transport to divertor (as ions) modify erosion and co-deposition Be concentration in plasma plays key role Balance of inc. Be coverage on target (dec. C erosion) and inc. C erosion due to Be flux For a range of Be conc., T/C and T/Be ratios 0.5g – 6.4gT/400s shot Be transport will impact C erosion

5. 5/28 Assumptions: Maximum possible stochiometric combination of Be 2 C formed No chemical erosion for Be 2 C Leads to non-linear evolution of relaxation time for reduction in chemical erosion - as observed Be 2 C formation potentially significant ERO modelling of erosion mitigation with Be seeding in PISCES

6. 6/28 All ITER plasma facing components will be castellated >2,000,000 Gaps in ITER (typ. 0.5-1mm x 10mm) Increases plasma exposed areas by factor 2 - 5 aC:H co-deposits form in tile gaps C x H y molecules and radicals form a:C:H co-deposits deep in gaps – how much and how deep is on-going research CFC tile segments from JET Mk1 divertor, 6mm gaps Retention in gaps twice that on plasma- facing surfaces (protected from re-erosion) CFC target (90,000 monoblocks):50 m 2 215 m 2 W baffle & dome (1.2M rods):100 m 2 460 m 2 Be main wall (300,000 tiles):680 m 2 1290 m 2 ITER mock-up

7. 7/28 Potential for significant T inventory TZM castellated monoblocks exposed to plasma for 200s in TEXTOR D retention fall-off dependent on D and T tile D gap ~ 0.4% - 4% of D, between low and high D at T tile ~200- 260 C Factor 10 decrease in D gap, 30 T tile 200 C Extrapolation to ITER based on D from B2-EIRENE modelling (Kukushkin, 2005): 0.5 – 5gT/400s shot Maybe other factors, however: strong function of gap width carbon source (local or remote) period of exposure TZM monoblocks from TEXTOR, 0.5mm gaps

8. 8/28 Planar and imbricated castellated tungsten blocks exposed in TEXTOR Co-deposition significantly reduced for imbricated blocks Also of note - Metal intermixing noted in deposits Some deposits buried under W layer Shaped castellations may help

9. 9/28 N11 ToreSupra: 75-85% D retention in short shots (<30s) Up to 100% D retention in long shots (>100s) Retention in short shots easily recovered by He glow Measurements of C erosion suggest co-deposition alone may not explain retention more than 1 mechanism? Retention in bulk CFC being considered for high fluence conditions Lab studies indicate D retention to several m in bulk D inventory fluence 0.5 Calcs. suggest this may initially exceed co-deposition in Tore Supra D/T retention in CFC bulk could affect choice of CFC for ITER Flux and time dependence needs more study

10. 10/28 Tritium trapped in aC:D/T co-deposits Oxidation an obvious candidate for detritiation through the reaction : aC:D/T + O CO x + DTO:D 2 O:T 2 O In-situ – no need for vessel entry Volatile products pumped from vessel Several schemes under investigation: Baking in O 2 ECR or ICR -wave plasma in O 2 or He/O 2 mix DC Glow discharge cleaning in He/O 2 mix Studies on-going in both laboratory and tokamak environments and both laboratory produced and tokamak co-deposited films T-removal through oxidation

11. 11/28 RF Antenna protection tile from TEXTOR 180 m thick aC:D co-deposit O 2 baking efficient, but at high T wall Molecular chemistry – O 2 penetrates all regions of deposition but …. Low D removal efficiency below ~300 C (cf ITER wall bakeout temp 240 C) High O 2 pressure needed for high removal efficiency Co-deposit not fully removed – becomes flaky and peels off Inhibited O penetration and release of volatiles due to carbide formation with impurities? WC and BeC may form in ITER … O 3 /O 2 mix effective at <200 C and low pressure but damage to bulk CFC seems to be too high 800 m

12. 12/28 O-Plasma – effective at room temp Erosion rate, E increases with T surf chemical reactions and with bias volts collisions 2 step process: surface damage by ion bombardment then chemical erosion ECR plasma in 100% O 2 Products: CO, CO 2, H 2, H 2 O He/O 2 mixture: E limited by He ion flux at high %O 2 E saturates above few %O 2

13. 13/28 He/O GDC in the tokamak environment Asdex Upgrade: 49h, 25g removed, 7x10 18 C-at/s E ~ 1.4x10 17 C-at m -2 s -1 TEXTOR: 3h, 5.2g removed, 2x10 19 C-at/s E ~ 5.7x10 17 C-at m -2 s -1 i.e. 0.075 - 0.3g T/h over 150m 2 CO and CO 2 dominant T 2 O 30 times higher than He GDC Production saturates at low O % No removal in boronised regions: B-coated sample coupons and boron coated co-deposited tiles unaffected - Impact of WC, BeC in ITER? or from shadowed areas: a:C:H coated samples behind first wall, deep in divertor untouched Tokamak and Lab studies less clear on removal from tile gaps 200nm and 350nm soft aC:H deposit on Si coupons in Asdex Upgrade Arcing and pitting occurred on boron dominated surface layers Cleaning not uniform on all surfaces

14. 14/28 Hydrocarbon coated elements at base of castellated structure exposed to He/O 2 discharge 2 orientations: Directly exposed Facing bottom of the chamber (shadowed from plasma) Cleaning rate nearly identical in both orientations Suggests atomic oxygen survives several collisions with the walls. Cleaning in shadowed areas O+/O may penetrate several mm into sufficiently wide gaps Castellation and tile gap design may be important for ITER

15. 15/28 Impurities in aC:H reduce efficiency Fully removed with 6.25 hours lab GDC i ~2.5x10 18 m -2 s -1, 8mbar, 20% O 2 in He Similar E to tokamak GDC E for tokamak co-deposits up to factor 10 less than for laboratory produced 80% co-deposit eroded during first 20% of plasma exposure Tokamak produced aC:H co-deposits on W substrate efficiently cleaned in He/O discharge effect of impurities in co-deposit building up at surface? W, Be will mix with aC:H in ITER

16. 16/28 Collateral damage and recovery OK Not all injected O 2 is pumped out of the vessel during GDC Some O retained in metal oxides O retention higher at larger V bias opportunity for optimisation H 2 discharge effective at removing oxides TEXTOR Recovery: 66h H 2 GDC, 0.5h He GDC & boronisation Asdex Upgrade Recovery: 72h baking at 150 C, 10h He GDC & boronisation) Will recovery extrapolate to BeO?

17. 17/28 BeO formation a challenge to oxidation Beryllium samples pre-oxidised to variety of oxide thicknesses (20-100 nm) Samples exposed to hydrogen plasma for 5 - 6 hours at 300 °C - 600 °C Residual oxide layer thickness always ~25 nm Thick layers are reduced but thin layers actually increase Development of an equilibrium state with the plasma Oxide thickness depends on the plasma conditions, especially oxygen impurities Complex (yet to be determined) consequences for formation and impact of oxygen based de-tritiation schemes

18. 18/28 N 2 injection into Asdex Upgrade sub-divertor Factor 5 reduction in aC:H net co-deposition rate No significant N retention Effect not seen with Ar (laboaratory studies) Scavenging proposed as one mechanism – moping-up of reactive radical pre-cursors But also alternative explanations - Synergistic interaction of H and N at surface – peaks at ~ 75%:25% Erosion rates high in H 2 /N 2 plasmas E up to 1 m/hour for lab deposits in ECR plasma less than O, but not optimised Alternative chemistry may have a role

19. 19/28 Without CH 4 : Chemical sputtering dominates HCN molecules are formed and released With CH 4 : N ions release nitrogenated compounds from surface – react with cracked CH 4 C 2 H 2 dominates, little HCN Role of atomic N as yet unknown Alternative chemistry may have a role Cryo-trapping assisted mass spectroscopy Quantitative analysis of compounds formed during discharge cleaning Pre–deposited hydrocarbon films exposed to N 2 /H 2 and CH 4 /N 2 /H 2 plasma glow

20. 20/28 T-removal through photonic cleaning aC:H co-deposits have poor thermal conductivity compared to substrates (CFC, Be, W) Surface heat flux leads to rapid temperature rise in co-deposit ablation or chemical bond-breaking Two photonic cleaning schemes under investigation: LASER Flash-lamp Requires vessel access, but can operate in high magnetic fields and in vacuuo, inert gas or atmospheric conditions Studies on-going in both laboratory and tokamak environments and both laboratory produced and tokamak co-deposited films

21. 21/28 100mm 2 in 2s with 20W Ytterbium fibre laser (1060nm, 120ns, 20kHz), 2J/cm 2 on 250 m spot @ 40cm 0.03 - 0.3g T/h over 150m 2 No difference between active and inert gas environment Laser cleaning of TEXTOR tile Energy density threshold for removal Threshold factor 5 lower for co-deposit compared to graphite selective removal

22. 22/28 Galvo-scanning fibre laser developed for use in JET Be handling facility Laser cleaning in JET BeHF Tritium release only accounts for ~10% of tritium in cleaned region micro-particulate? >200 mm of co-deposit easily removed by single scan at low scan rate (40s) and pitch At fastest rate (4s), not all deposit removed even with 4 passes

23. 23/28 JET 2004 trial showed engineering feasibility of flash-lamp technology Flash-lamp cleaning of tritiated aC:H Trials now conducted using flash-lamp in JET berylium handling facility Aim to clean thick, tritiated co-deposit from inner divertor CFC tile Photon flux from 500J, 140 s flash-lamp 3.6MW Rep. rate 5Hz Focused using semi-elliptical cavity – Footprint ~30cm 2 @ 30mm 375MWm -2, 6J/cm 2 Three positions treated (with varying co-deposit thickness and tritiation) Tritium release monitored and tile sent for SIMS/IBA/SEM analysis

24. 24/28 1 st demonstration of T-removal Total T release ~9 g. Decreasing efficiency with number of pulses 40% of T inventory & 70-90 m co-deposit, removed (off gas & SEM) 0.075g T/h over 150m 2 Untreated Treated 7 m de-tritiation at surface of treated zone Consistent with FE calcs of bulk heating above 700K Build-up of Ni at surface explanation for roll-over of tritium release/pulse? (similar results for Be on other treated tiles)

25. 25/28 Could prevention be the best cure? Hot liner added to PSI-2 linear plasma device CH 4 /H 2 added to H 2 plasma column Cold liner – CH 4 decomposes to form co-deposted layers on liner walls Hot liner – co-deposit re-eroded to form CH 4, C 2 H 2 and heavier species - ~100% C pumped through duct or returned to main chamber Relies on presence of sufficient atomic hydrogen (C 2 H 2 production saturates at high CH 4 flows)

26. 26/28 The year that was …. Significant advances during 2005-6 but also growing recognition of new challenges - Improvements in modelling of retention and co-deposition, including with beryllium fluxes Importance of retention in gaps, castellations and bulk CFC now better understood and characterised More machines prepared to run with oxidising plasmas but difficulties of shadowed regions, mixed materials and oxide formation now clear Further studies into alternative chemistry (apart from oxygen) Technological approaches continue to develop and first trials with tritiated samples A growing understanding that ITER may require a combination of techniques to operate within the inventory limit

27. 27/28 No single T-removal scheme likely to be sufficient – lets not close any doors Integration of different schemes on different timescales will probably be required – the good housekeeping approach Big toolbox needed for ITER Example of T-removal integrated into ITER operating schedule Extrapolated from predicted/measured T-removal rates allowing for future optimisation

28. 28/28 Characterise the impact of Beryllium Carbide formation in ITER relevant conditions CFC bulk retention – is it real? Need for confirmation of the effect and improved characterisation (impact of surface temperature, morphology etc.) Tile gaps – how does deposition in gaps scale the ITER? Resolve issue of shadowing on plasma removal schemes (e.g. how many wall collisions can the chemically active atoms survive) Demonstrate impact of repetitive oxidising plasmas and plasma recovery on Beryllium surfaces Explore further the impact of surface temperature on deposition in ITER relevant conditions Begin to develop a serious scheme for operating ITER with a CFC/Be/W materials mix in the DT phase Start to characterise retention in an all metal ITER – is it a problem or not? Key issues for the near future How can the EU-PWI and the SEWG help?

29. 29/28 Not a comprehensive list! - G Counsell et al33 rd EPS, 11th PFMC D Borodin et al11 th PFMC P Coad et al17 th PSI J Ferreira et al11 th PFMC C Grisolia, et al24 th SOFT, 17th PSI C Hopf, et al17 th PSI W Jacob, et al17 th PSI A Kirschner, et al17 th PSI A Kreter, et al17 th PSI K Krieger, et al17 th PSI J. Likonen, et al17 th PSI A. Litnovsky, et al11 th PFMC A Markin, et al11 th PFMC V Philipps, et al17 th PSI J Roth, et al17 th PSI M Rubel, et al17 th PSI, 21st IAEA E Salancon, et al17 th PSI A. Semerok, et al21 st IAEA FL Tabares et al17 th PSI, 21st IAEA C. Tomastik, et al11 th PFMC A Widdowson, et al17 th PSI Many presentations this year …..