Hydrogen isotope retention and release in beryllium: The full picture from experiment and ab initio calculations

Transcription

1 Member of the Helmholtz Association Hydrogen isotope retention and release in beryllium: The full picture from experiment and ab initio calculations Christian Linsmeier With contributions from: A. Allouche, M. Oberkofler, M. Reinelt, R. Piechoczek, J. Brinkmann Forschungszentrum Jülich Institut für Energie- und Klimaforschung Plasmaphysik

2 Overview 1. Wall materials and particle loads 2. Retention mechanisms 3. Retention and thermal release: results 4. Lattice orientation: single vs. polycrystal 5. Outlook: Neutron-induced defects Multi-component materials 6. Conclusions

3 Materials in ITER and DEMO

4 ITER first wall Fusion power: 500 MW Plasma volume: 840 m 3 Plasma current: 15 MA Typical density: m -3 Typical temperature: 20 kev Beryllium 700 m 2 beryllium first wall 100 m 2 tungsten divertor Tungsten 50 m 2 CFC strike point Carbon

5 First wall loads: particle energy ASDEX Upgrade conditions

6 D retention in beryllium [Anderl 1999] Be: ~700 m 2 D,T Variation by 1-2 orders of magnitude!

7 Hydrogen in Be: saturation D retention [D/m 2 ] D ions -> Be 1000 ev, T irr = 300 K (Haasz and Davis, 1997) 1000 ev, T irr = 373 K (Alimov and Roth, 2006) 200 ev, T irr = 323 K (Alimov and Roth, 2006) 1000 ev, T irr = 300 K (Reinelt and Linsmeier, 2009) D retention demonstrates saturation. For 1 kev D ion irradiation, the saturation is observed at a fluence of D/m 2. For D is trapped in the implantation zone with negligible diffusion into the bulk Incident fluence [D/m 2 ] High fluence experiments D retention [D/m 2 ] When the irradiation temperature increases, the D retention decreases ev - TPE (Causey et al. 1997) 100 ev - PISCES-B (Doerner et al. 1997) Exposure temperature [K] From review of R. Anderl et al., J. Nucl. Mater. 273 (1999) 1

8 Strategy Possible reasons for quantitative uncertainties: undefined BeO coverage undefined crystal structure unclear retention mechanisms Need for well-defined experiments and theory! To be resolved: retention in pure Be influence of crystal structure/orientation influence of surface oxide retention and release mechanisms

9 Experimental concept QMS desorption rate Be poly,be(0001), Be(11-20) D implantation (mass and energy separated) TPD Temperature Programmed Desorption NRA 3 He(d,p) 4 He Sequential release of D Limited by combination of bulk + surface processes Energy barriers for... Diffusion Detrapping Recombination Retention mechanisms Hydrogen retention

10 Surface morphology

11 Annealing SEM (1120) T 1000 K, several hours: Formation of Be crystallites Substantial material transport Recrystallization to low indexed facets

12 Annealing sputtering cycles AFM 500 nm Cycles of Cleaning D Implantation Degassing 1000 K Recrystallization Erosion D-induced structural modifications

13 Erosion: orientation-dependent SE EBSD 100 μm Be PC + D shows crystallites with less damage... more damage Red White Black = Basal plane surface FAST diffusion to surface = Basal plane surface SLOW diffusion to surface = Not indexed

14 Retention and release mechanisms

15 Hydrogen isotope inventory Main mechanisms for retention Implantation and release Diffusion Desorption

16 Implantation Simulation of collision cascade with Monte Carlo codes (TRIM) Binary collision approximation Energy loss: stopping Creation of displaced atoms and vacancies Projectile is trapped Result: Depth profile Deuterium in polycrystalline beryllium Depth profile measured with 10 MeV 28 Si

17 Thermal release Elementary steps diffusion of H in the solid trapping and release of H from binding sites (defects) recombination and desorption of H 2 from the surface

18 Thermal release Described by Diffusion-Trapping Model Local elementary reactions (trapping / release of H) thermally activated processes with: n i : number density of each species k: reaction rate constant E a : activation energy Diffusion: driven by concentration gradient (1st Fick s law) local concentration change: Diffusion described by continuity equation (2nd Fick s law)

19 Particle flux at desorption Partial differential equation: considers both diffusion and elementary reactions Source Sink Example: n i : concentration of mobile (dissolved) H mobil R: rates for trapping and release at traps PDE describes H diffusion and the reaction: Consider arbitrary elementary reactions Description of the thermal release by a system of coupled differential equations

20 D retention

21 Saturation of retention Trapping of D at ion-induced defects Reemission above 25 at.% low fluence linear high fluence saturation (D/Be) max =const Be ion range Implantation: D at 1 kev into Be (1120)

22 Saturation of retention Trapping of D at ion-induced defects Reemission above 25 at.% Retention within ion range Low inventory in main wall of ITER (D/Be) max =const Be ion range Γ = D cm -2 Γ = D cm -2 Retained fraction Retained fraction Be(1120) Be poly Implantation energy [kev] Be(1120) - Be poly Implantation energy [kev]

23 Saturation of retention Trapping of D at ion-induced defects Reemission above 25 at.% Retention within ion range Low inventory in main wall of ITER Annihilation of traps at grain boundaries Anisotropic diffusion of interstitials (DFT) Γ = D cm -2 Γ = D cm -2 Retained fraction Retained fraction Be(1120) Be poly Implantation energy [kev] Be(1120) - Be poly Implantation energy [kev]

24 Thermal release

25 Trapping sites and desorption stages Temperature-programmed desorption spectra Desorption rate [a.u.] Be poly, E Impl =1 kev ITER main wall bake-out temperature Fluence [10 16 cm -2 ] Temperature [K] 2.5

26 Trapping sites and desorption stages Temperature-programmed desorption spectra Desorption rate [a.u.] Be poly, E Impl =1 kev Retention in the structural modifications [%] Fluence [10 16 cm -2 ] Temperature [K] Sample saturation Incident fluence [10 15 D cm -2 ]

27 Saturation: Simulation by SDTrim.SP Depth [nm] Retention in structural modifications [%] Supersaturation SDTrim.SP Calculation Incident fluence [10 15 D cm -2 ] 10 TPD Experiments Incident fluence [10 15 D cm -2 ] D Concentration (cut off) Bulk saturation concentration: 26 at% D Concentration build-up in a depth of 40 nm Supersaturation Structural modifications

28 Trapping sites and desorption stages Desorption rate [a.u.] Temperature-programmed desorption spectra Be poly, E Impl =1 kev Fluence [10 16 cm -2 ] Stable structures from density functional theory (DFT) calculations by A. Allouche Amorphous BeH 2-x Temperature [K] 2.5

29 Trapping sites and desorption stages Temperature-programmed desorption spectra Stable structures from density functional theory (DFT) calculations by A. Allouche Be poly, E Impl =1 kev Fluence [10 16 cm -2 ] Desorption rate [a.u.] Temperature [K] H Amorphous BeH 2-x Position of H inside a monovacancy

30 Modelling the release at low fluences Desorption flux [10 13 cm -2 s -1 ] Implantation energy 1 kev 3 kev - Be (1120) 2.7 * D cm Temperature [K] n t mob Reaction-diffusion equation n = D x x mob act E + R source/ sink Single processes not accessible in experiments Density functional theory Solve set of coupled differential equations numerically Reproduce both spectra (different initial conditions) with identical set of parameters

31 Modelling the release at low fluences Desorption flux [10 13 cm -2 s -1 ] Implantation energy 1 kev 3 kev - Be (1120) 2.7 * D cm Temperature [K] Model 1: static traps diffusion barrier from literature adapted barrier for detrapping Implemented processes: - diffusion of D E act = 0.3 ev - trapping E act = 0.3 ev D mob + trap D trap - detrapping E act = 1.9 ev D trap D mob + trap Shift with initial depth due to effective diffusion Assumed missing physics: Annealing of defects Simulated peaks broader than experiment! Implement trap mobility

32 Desorption flux [10 13 cm -2 s -1 ] Modelling the release at low fluences Implantation energy 1 kev 3 kev - Be (1120) 2.7 * D cm Temperature [K] Forward calculations yield Shift with initial depth profile Peak width Assymmetry Model 2: Assumption of trapping at single vacancies consistent with experimental desorption spectra trapping at mobile vacancies accumulation of D mob at D trap diffusion barriers from DFT adapted barrier for detrapping Implemented processes: - diffusion of vac E act = 0.70 ev - diffusion ( ) of D E act = 0.41 ev - trapping E act = 0.41 ev D mob + vac D trap - accumulation E act = 0.41 ev D mob + D trap 2 D trap - detrapping E act = 1.86 ev D trap D mob + vac (1.7 ev)

33 Lattice orientation

34 D diffusion in beryllium Be crystal lattice Be(0001) plane ǁ to surface Be(11-20) plane to surface Beryllium has not an ideal hcp structure. Be-Be distances are closer in <0001> than perpendicular to it (e.g. <11-20>) DFT (Density Functional Theory) calculations show anisotropy in transport processes with respect to Be basal planes

35 Modeling: 2D coupled rate equations Strategy Solve a coupled reaction diffusion system (CRDS) consisting of an arbitrary number of diffusing species i=a,b,c in 1, 2 or 3 dimensions Solve system of partial differential equations for the time dependent 2D depth profile ρ(x,z,t) [m -3 ] A(, xzt,) A(, xzt,) A(, xzt,) Dx( Tt ( )) Dz( Tt ( )) t x x z z ( ) ( ) ( xzt,, ) E 1 Γ = 1( A, ρb, t) k1ρ AρB exp kbt ( t) Form Annihilation Source i ABC,,... i ABC,,... i A + B C C A+ B ρ 2 Γ 2( ρc, t) = k2ρc exp B E ktt ()

36 TPD of D implanted in Be single crystals Surface desorption flux [10 17 D m -2 s -1 ] 2,5 2,0 1,5 Circles: Experiment Lines: 2D CRDS Model Be (11-20) 1 kev/d 2.7 x D cm -2 Be (11-20) 3 kev D 2.9 x D cm -2 Be (0001) 3 kev D 2.7 x D cm -2 1,0 0,5 0, Temperature [K] TPD spectra and 2D-CRDS simulations after implantation of D at 1 kev/d and 3 kev/d in Be(0001) and Be(11-20) Only one type of trap! R. Piechoczek et al., J. Nucl. Mater. 438 (2013) S Desorption temperature shift for different implantation energies 2. Desorption temperature shift for Be(0001) and Be(11-20) 3. Less D retention in Be(0001) than in Be(11-20) 1. Deeper implantation longer diffusion path to the surface higher TPD peak temperature 2. Anisotropy of Be lattice 2D anisotropic modelling required to implement results from DFT calculations 3. Dynamics during D implantation (MV+SI annihilation and trapping in MV) determines retention

37 Simulation phases 1000 Temperature [K] Implantation Relaxation Heating Time [s]

38 Simulation phases Surface flux [m -2 s -1 ] Surface flux [m -2 s -1 ] 1E20 1E18 1E16 2D-CRDS Simulation Be [11-20] Mono vacancies Be self interstitials Hydrogen E E20 2D-CRDS Simulation Time [s] Be [0001] E18 1E16 Mono vacancies Be self interstitials Hydrogen E Time [s] Temperature [K] Temperature [K] self interstitials during implantation mono vacancies stable after implantation, thermally instable self interstitials during implantation mono vacancies vanish quickly

39 TPD from Be polycrystal Identical parameters also applied for polycrystalline Be Be(11-20) Be(0001) 50 μm SEM (EBSD) Shift for Be poly reproduced with no retrapping: fast grain boundary diffusion heating ramp variation reproduced: no (self)trapping during heating Surface desorption flux [10 17 D m -2 s -1 ] Symbol: Experiment Poly crystal, 3keV/D, 2.0 K/s Lines: 2D CRDS Model [11-20] With re-trapping [0001] With re-trapping [11-20] No re-trapping [0001] No re-trapping Surface desorption flux [10 17 D m -2 s -1 ] Symbols: Experiment Lines: 2D CRDS Model 0.1 K/s 0.2 K/s 0.7 K/s 2.0 K/s Temperature [K] Temperature [K] R. Piechoczek et al., J. Nucl. Mater. 438 (2013) S1072

40 Parameters for D release and diffusion Experimentally and DFT determined parameters for full description Basal Plane Basal Plane Diffusivity Diffusivity ν ΔE act [ev] ΔE act [ev] CRDS CRDS DFT [Allouche] Mobile Hydrogen H mob 3.11E-06 <0.4* 0.2 Mono vacancy MV 3.11E Self interstitial SI 3.11E Trapped Hydrogen H trap - - H mob 7.68E-06 <0.4* 0.4 MV 7.68E SI 7.68E H trap - - Trapping H mob + MV H trap 1.00E Detrapping H trap H mob + MV 1.00E+11* Annihilation MV + SI E Self trapping H mob + 1 H trap 2 H trap + SI 1.00E Anisotropy in diffusion of self-interstitials crucial for modeling! R. Piechoczek et al., J. Nucl. Mater. 438 (2013) S1072

41 Surface oxide

42 Be with oxide layer Composition of first atomic layer: Ion scattering spectroscopy! (45, 500 ev He + ) Be + 3 ML BeO (closed surface layer) Be terminated surface + 3 ML BeO (covered) Segregation of Be at the surface Annealing (1000 K): Be-terminated surface (desorption via pure Be surface) No annealing: TPD via BeO covered surface

43 TPD: Influence of BeO-coverage desprtion rate [a.u.] 3.0 ML BeO 0.2 ML BeO BeO:D (surface) temperature [K] No change of E A of release from binding states No recombination limit Trapping in the bulk time [s] Formation of BeO:D at the surface

44 Implantation at elevated temperatures BeD 2 BeO:D (surface) Ion induced trap sites unaffected new trapping sites appear (deuteride, surface oxide) BeD 2 formation (decomposition >570 K) retention at 530 K: 90%, instead of expected 70% 320 K 530 K M. Reinelt et al., New J. Phys. 11 (2009)

45 Implantation into BeO Implantation of D within BeO layer changes D release: small amount of low-t trapping D release over wide temperature range only 75% of total D released until 850 K only small fraction of total D released until 513 K (ITER wall bake-out) Similar total retained D amounts M. Oberkofler et al., J. Nucl. Mater. 415 (2011) S724

46 Influence of neutron damage

47 Neutron-induced trap formation: Tungsten Damage simulation Example: Trapping in heavy-ion damage at rear surface Deuterium depth profiles measured on the front and rear side of W samples Damaged either by 20 MeV W ions The front side was exposed to D plasma to a fluence of D/m 2 at 550 K D retention decorates damage profiles, but also demonstrates long range damaging beyond ion range B. Tyburska et al., J. Nucl. Mater. 415 (2011) S680

48 Neutron-induced trap filling in ITER Trap density, at.% Modelling of ITER wall conditions Damage evolution: 10-7 dpa/s (from ITER Org.) saturation trap density 1% ev, I 0 =10 21 D/m 2 s, T=500 K plasma-induced defects n-induced defects 0.0 1x x x x x x10 2 Fluence (D/m 2 ) D concentration (at.%.) Diffusion and trap filling: Diffusive trap filling using established diffusion coefficient D/m D/m D/m D/m Filling of 1 mm W armour in about 3 years steady state operation Relevant to DEMO rather than ITER 60 ev, I 0 =10 21 D/m 2 s, T=500 K with n-generated traps without n-generated traps 0, Depth (µm) O. Ogorodnikova et al., J. Nucl. Mater. 415 (2011) S661

49 Multi-component materials

50 DEMO first wall: W alloy Power plant conceptual study Temperature profile in PPCS Model A, 10 days after accident with a total loss of all coolant. [Final Report of the European Fusion Power Plant Conceptual Study, 2004] Accidental loss of coolant: peak temperatures of first wall up to 1200 C due to nuclear afterheat Additional air ingress: formation of highly volatile WO 3 (Re, Os) Evaporation rate: order of kg/h at >1000 C in a reactor (1000 m 2 surface) large fraction of radioactive WO 3 may leave hot vessel Development of selfpassivating tungsten alloys

51 DEMO first wall: W alloy Rate [ m 2 / t [mg 2 /sec]] K W-Cr binary ~0.5% Ti, TGA 1273 K Chromium concentration [wt-%] W-Cr-Ti self-passivating alloy Optimum Cr concentration Ti active element: reduce oxidation rate at high T Ti: hydrogen retention?

52 W-Cr-Ti: D retention D depth profiles room temp. loading 100 V bias 3 He NRA depth profiles almost homogeneous D distribution D content vs. Ti fraction linear dependence 2 at-% threshold TiH 2? Replace Ti by non-hydride former!

53 Summary Hydrogen retention and release from beryllium Retention and release mechanisms identified and quantified ITER: Tritium retention in metallic Be only within ion range (small inventory) Release < 500 K: BeD 2-x supersaturated Be-D, D fraction ~0.25 Release > 500 K: vacancies, ion-induced traps Essential mechanisms: - effective diffusion - mobility of vacancies - accumulation of D in vacancies BeO Thin (several ML) oxide: not rate-limiting for release Different mechanisms than for Be Only 75% D released <850 K

54 Outlook DEMO first wall: neutrons and alloys Neutron wall loads: additional traps, deeper than for ion-induced defects DEMO PFM: most probably not a pure metal, but alloy New diffusion, retention and release mechanisms DFT data missing! Improve ability to deal with large unit cells