Hydrogen isotopes permeation through the tungsten-coated F82H first wall

Transcription

1 Hydrogen isotopes permeation through the tungsten-coated F82H first wall Presented at the IAEA-CRP meeting Oct. 16 th ~Oct.18 th, 2017 International Center, Vienna, Austria Yoshi Hirooka 1,2 and Yue Xu 2 1 National Institute for Fusion Science, Japan 2 Graduate University for Advanced Studies (currently Hefei Univ. Tech.)

2 1. Motivation Technical issues: PDP/GDP through the 1 st wall 2. Experimental setup 3. Highlight data 3-1. VPS tungsten coatings effects 3-2. SP tungsten coatings effects 3-3. VPS+SP tungsten coatings effects 4. Summary Table of Contents 2

3 1. Motivation 3

4 The First Wall of a fusion power reactor Definition of the first wall: All the fusion experimental devices: a vacuum chamber wall. Fusion reactors: the plasma-facing surface of breeding blanket units. First wall JET LHD The first walls of existing fusion devices. FFHR (Force-Free Helical Reactor) A. Sagara et al., research report, NIFS-MEMO-64 (2013). 4

5 Breeding blanket structure Blanket concepts include: Water-cooled solid breeder (e.g. Li 2 TiO 3 ), He-cooled solid breeder (e.g. Li 4 SiO 4 ), Self-cooled liquid breeder (e.g. FLiBe), Water-cooled liquid breeder (e.g. Li-Pb) First wall (F82H ) Vacuum chamber wall Characteristics of the first wall of a fusion power reactor are: A large surface area: ~1000 m 2, Made of reduced activation materials (e.g. RAFS (F82H ), V-alloy, and SiC f /SiC), Operated at high temperatures (e.g. ~500 for ferritic steel alloys), Thin wall design to reduce thermomechanical stress (~5 mm for RAFS). First wall Flibe blanket for FFHR After A. Sagara [1] Vacuum chamber wall [1] A. Sagara et al., Fusion Technol. 39 (2001)

6 Optimum thickness of the first wall Optimum first wall thickness: 5 mm or even less, although these concepts employ various first wall materials. [1] Thickness optimization [2] [1] Y. Ueda et al., J. Nucl. Mater (2003) [2] A. Sagara et al., Fusion Technol. 39 (2001)

7 T 2 pressure in breeder For blanket employing FLiBe, the tritium thermodynamic equilibrium pressure is ~10 4 Pa at 800 K at a tritium concentration (at%) of ~0.1 ppm (~12 g tritium for FFHR). ~1 torr liter/s/m 2 ~1000 torr liter/s 800 K S. Fukada, et al., Fusion Eng. Des. 83, 747 (2008). 7

8 T + T + For the self-cooled breeder blankets, hydrogen isotopes will penetrate through Do thewe first wall want by Plasma-Driven bi-directional Permeation permeation? (PDP) in one direction and Gas-Driven Permeation (GDP) No! in the opposite direction, referred to as bi-directional permeation. Edge plasma D + D + D + T + D + D Implantation + T + T + D + PDP T + T + T + T + D + re-emission Bi-directional permeation First wall ~5 mm, ~500 T-GDP Diffusion D, T-PDP Diffusion Breeding blanket n + Li T + He T 2 T 2 T 2 Dissociation T & Solution 2 T 2 T 2 T2 T 2 T 2 T 2 T 2 GDP: Bi-directional permeation 1. Important W as a plasma-facing parameters Potential issues: material [1,2] : GDP: PDP D lowers the High Solubility thermal conductivity Diffusivity recovery efficiency Low sputtering yield Low External ofactivation T pressure from the 2. breeder. PDP: Applications: Divertor: JET, EAST, ITER Surface First GDP-T wall: ASDEX-U permeation recombination 3. W leads coatings to methods: coefficient Diffusivity uncontrolled (VPS-W) Implantation fueling. flux Reflection coefficient High melting point (3410 ) Vacuum Plasma-Sprayed W SPutter-deposited W (SP-W) Chemical Vapor Deposition W (CVD-W) [1] A. Sagara et al., Fusion Technol. 39 (2001) [2] R. A. Pitts et al., J. Nucl. Mater. 438 (2013) S48 S56. 8

9 2. Experimental facility and setup 9

10 VEHICLE-1 Experimental facility Plasma parameters: n e : ~10 10 cm -3 T e : 3~10 ev Ion flux : ~10 16 cm -2 s -1 (nn HH22 : ~10 13 cm -3 ) QMS II TMP VEHICLE-1 test chamber Sample Gate valve Bias Gauge II V Variable leak valve Pneumatic shutter Gas side ECR plasma source Plasma side Insulator TMP Resistive heater Magnet Langmuir probe I Orifice Gauge I Optical spectrometer Langmuir probe II TMP QMS I Y. Hirooka et al., J. Nucl. Mater (2005) QMS: Quadrupole Mass Spectrometer TMP: Turbo Molecular Pump 10

11 Experimental setup Permeation membrane sample 70 mm Hα spectrometer Plasma Membrane Thermocouple φ35 mm Heater Gas Bias Samples: F82H (Fe-8Cr-2W) VPS-W coated F82H SP-W coated F82H F82H thickness: mm VPS-W thickness: µm SP-W coatings thickness: µm VPS-W: Vacuum Plasma-Sprayed W SP-W: SPutter-deposited W PDP and GDP fluxes are measured by two QMS (Quadrupole Mass Spectrometer), respectively. Ion bombardment energy is provided by a DC bias (-100 V). Temperature: ~ Argon plasma bombardment pre-treatment to reduce surface contamination. 11

12 Ion flux to the first wall D + ion flux to the first wall has been estimated to be of the order of D cm -2 s -1. calculated ion fluxes (cm -2 s -1 ) A B C D E D + B A C ITER E D R. Behrisch, et al., J. Nucl. Mater (2003)

13 Gas-driven permeation (GDP) model (single layer) Solubility of gases in metals (Sieverts' law): C= ST ( ) p H 2 S(T): Sieverts constant as a function of temperature; p H2 : hydrogen pressure. Diffusion limited gas-driven permeation flux: JJ = DD TT dddd dddd = DD TT SS TT LL Permeability: HHHH pp H2 L: membrane thickness; D(T): diffusion coefficient; pp HHHH HH22 : H 2 pressure at the high pressure side. Φ(T) D(T)S(T) Low pressure LLLL pp H2 ( 0, usually) HH 22 HH 22 HH 22 Permeation HH 22 HH 22 downstream Membrane HH HH HH HH HH Diffusion X HH HH L C HH 22 HH 22 HH 22 HH 22 High pressure HH 22 HHHH pph2 HH 22 HH 22 HH 22 HH 22 HH 22 Dissociation & Solution upstream HH 22 HH 22 F. Waelbroeck et al., J. Nucl. Mater (1979)

14 F82H: pressure and thickness effects on GDP GDP (0.5 mm thick F82H) GDP at ~500 o C JJ pp H2 JJ 1 LL The GDP flux is proportional to the square-root of upstream pressure, and inversely proportional to the membrane thickness. GDP is diffusion-limited: ( ) ( ) J = DT ( ) dc = D T S T p dx L GDP H 2 14

15 Plasma-driven permeation (PDP) model Upstream Edge Plasma J (DD) J 0 Implantation Re-emission Diffusion c 0 c 0 (RD) Recombination Membrane d (~nm) Diffusion c p c p x c 1 c 1 d c p x c 0 c 1 Downstream J + Permeation Diffusion Diffusion Particle balance: JJ 00 = JJ + JJ + implantation re-emission permeation c 0 0, c 1 0 JJ + = dd JJ JJ LL JJ c p c 0, c 1 0 JJ + = DD JJ = KK rr CC 0 LL JJ 0 KK rr (RR) Recombination Kr d L x Recombination KL c 0 and c 1 : H concentrations of the plasma side and back side; d: implantation range; c p : H concentrations at x=d; K r and K L : H recombination coefficients. B. L. Doyle, J. Nucl. Mater. 111 & 112 (1982) c p c 0 c 1 JJ = KK rr JJ + = JJ JJ KK KK rr + KK 0 rr + KK 0 LL LL KK LL 15

16 F82H: temperature and thickness effects on PDP PDP flux (D/cm 2 /s) PDP flux increases as temp. increases ( o C) ~1x10 16 D/cm 2 /s, 100 ev 1 mm F82H 0.5 mm F82H /T (K -1 ) PDP flux (x D/cm 2 /s) PDP at 445 o C and 500 o C ~1 x D/cm 2 /s, 100 ev 2 JJ 1 LL Thickness (mm) PDP in RD-regime [1] : J + = 445 o C 510 o C D L J K 0 r The steady-state PDP flux is inversely proportional to the membrane thickness, meaning that hydrogen PDP is in the RD-regime. [1] B. L. Doyle, J. Nucl. Mater. 111 & 112 (1982)

17 Permeation studies by modelling: DIFFUSE code Hydrogen transport in solids can be described as: Mobile hydrogen: Trapped hydrogen: Cxt C (,) xt = ( ) + Gxt (, ) 2 (,) Cxt (,) t DT 2 t x t Mobile e Ct(,) xt CxtC (,) t (,) xt = D( T ) C (, ) 2 t x t ν 0 exp( U t / kt ) t λ C xt C x C xt e 0 t (,) = t () t(,) Trapped Implantation profile C(x,t) and C t (x,t): concentrations of mobile and trapped; D: diffusion coefficient; G(x,t): hydrogen implantation profile; C t0 (x) and C te (x): concentrations of intrinsic and empty trapping sites; λ: mean distance between trapping sites; v 0 : jumping frequency; k: Boltzman s constant; U t : the de-trapping energy. Isotopic effects: I. Isotope dependence of diffusivity, trapping energy, etc. II. Coupling of isotopes through the process of surface recombination. III. Competition of the various isotopes for traps. 17

18 One-dimensional modelling by DIFFUSE (2) Bi-directional PDP-D/T and GDP-T 2 flows. PDP-T PDP-D First wall GDP-T 2 GDP-T 2 pressure: ~1 Pa Plasma: D cm -2 s T cm-2 s-1 Membrane: 5 mm thick α-fe Temperature: 527 o C Boundary conditions: Gas side: Sieverts law Plasma side: recombination Intrinsic trap density: 1% Trapping energy: 0.62 ev The tritium concentration profiles interact with each other in the two counter flows. Deuterium flow appears to be independent of these tritium flows, driven by its own concentration gradient, i.e. random walk diffusion. 18

19 Demonstration of simultaneous bi-directional H/D permeation for the first time H 2 partial pressure D-plasma Hα Spectrometer Membrane D PDP: T e : ~10 ev n e : ~10 10 cm -3 H 2 gas D 2 partial pressure Flux: ~10 16 D cm -2 s -1 H - GDP: ~10 4 Pa (Fukada [1] ) Membrane: H-GDP D-PDP 0.5 mm F82H Temperature: ~470 Hα intensity (a.u.) H 2 filled in for GDP Hα intensity H 2 partial pressure GDP H 2 pumped Time (s) Temperature 0.5 mm F82H Time (hour) 19 [1] S. Fukada, et al., Fusion Eng. Des. 83, 747 (2008). Y. Xu et al., Fusion Eng. Des. DOI: /j.fusengdes (2017). Temperature ( o C) Pressure (x10-6 Pa) Pressure (x10-6 Pa) HD HD D 2 D 2 Evidence of H-GDP Evidence of D-PDP pressure build-up pressure build-up H 2 partial pressure (x10-7 Pa)

20 3. Tungsten coatings effects on hydrogen permeation 3.1. VPS-tungsten coated F82H 20

21 Microstructure of VPS-W coatings (a) F82H surface, polished 1 µm (c) W/F82H cross-section, polished (b) VPS-W surface, as-received unmelted W particles melted W pores 50 µm (d) VPS-W cross-section, fractured Deposition temperature: 600 W powder particles: 25 μm Density: ~90% of bulk W lamellar structure columnar grains cracks pores TEM W interface 5 µm F82H 10 µm 50 µm 2 µm 21

22 Experimental proof of connected pores of VPS-W VPS-W: 170 µm H 2 flows through VPS-W coatings at room temperature Φ5 mm Open pores H 2 gas VPS-W 50 µm Permeation flux (x10 15 H/cm 2 /s) H 2 RT Pressure (Torr) Φ5 mm hole F82H Machining down to the coating surface by removing F82H substrate. The permeation flux was detected immediately after gas loading. The permeation fluxes increase proportional to the pressure difference, indicating that gas molecules penetrate through the connected pores in the coating but not through the W bulk. 22

23 VPS-W coatings suppress PDP PDP flux (H/cm 2 /s) Hydrogen PDP: at ~500 Temperatures VPS-W coated F82H F82H 90 µm VPS-W, 0.5 mm F82H 0.5 mm F82H Time (hour) Temperature ( o C) PDP flux (H/cm 2 /s) Hydrogen PDP through F82H with/without VPS-W coatings F82H ( o C) VPS-W coated F82H 90 µm VPS-W, 0.5 mm F82H /T (K -1 ) VPS-W coatings effects on PDP: VPS-W suppresses DT-PDP from the plasma side. 23

24 VPS-W coatings do not suppress GDP Hydrogen GDP through F82H with/without VPS-W coatings ( o C) Open pores VPS-W Pa GDP flux (H/cm 2 /s) F82H VPS-W coated F82H (F82H upstream) H 2 gas (F82H upstream) F82H µm VPS-W, 0.5mm F82H /T (K -1 ) VPS-W coatings effects on GDP: VPS-W does not suppress T-GDP from the blanket side. Y. Xu, et al., Plasma Fusion Res. 11 (2016)

25 3.2. SP-tungsten coated F82H 25

26 Microstructure of SP-W coatings SP-W surface, as-deposited (300, 0.19 Pa) SP-W/F82H cross-section, polished 100 nm columnar grains W F82H 1 µm 100 nm Intensity (a. u.) XRD interface 1 µm Density: 19.2 g/cm3, ~99.5% of bulk W (a) SP-W Fine grain. (110) (b) Bulk-W (200) (211) (220) 70 2 Theta (degree) Dense structure. Preferred orientation. 26

27 SP-W coatings suppress GDP GDP flux (D/cm 2 /s) Deuterium GDP through F82H with/without SP-W coatings ( o C) 1.5 µm W, 1 mm F82H SP-W coated F82H (F82H upstream) /T (K -1 ) F82H D 2 gas (F82H upstream) SP-W F82H Effective permeability Φ eff [1] : LL ΦΦ eeeeee = LL 11 ΦΦ 11 + LL 22 ΦΦ 22 SP-W coatings effects on GDP: SP-W suppresses T-GDP from the blanket side. [1] J. Crank, The Mathematics of Diffusion, 2nd Edition,

28 SP-W coatings enhance PDP in VEHICLE-1 Permeation rate (D/m 2 /s) Literature data [1] : Deuterium IDP at ~ 500 SS: 0.5 mm, SP-W: 1 µm, VPS-W: 21 µm After R. Anderl [1] Time (min) SP-W coated SS SS VPS-W coated SS D/cm 2 /s 3 kev D 3 + IDP: Ion-Driven Permeation SS: Stainless Steel SP-W: SPutter-deposited W VPS-W: Vacuum Plasma-Sprayed W PDP flux (H/cm 2 /s) Present data [2] : Hydrogen PDP: at ~ F82H: 0.5 mm, SP-W: 0.5 µm, VPS-W: 90 µm Temperatures SP-W coated F82H H/cm 2 /s ev H-plasma Time (hour) F82H VPS-W coated F82H SP-W coatings effects on PDP: Temperature ( o C) SP-W enhances DT-PDP from the plasma side. [1] R. Anderl et al., J. Nucl. Mater (1994) [2] Y. Xu et al., Plasma Fusion Res. 12 (2017)

29 Surface recombination coefficient K r is a key parameter relating to PDP flux PDP flux (D/cm 2 /s) Steady-sate D-PDP fluxes ( o C) SP-W coated F82H F82H SP-W: 1.5 µm, F82H: 1 mm /T (K -1 ) PDP in RD-regime [2] : JJ + = DD LL Materials with small K r tend to suppress re-emission, increasing the hydrogen concentration. As a result permeation flux increases. Particle balance: JJ 00 = JJ + JJ + JJ 00 KK rr implantation re-emission permeation Recombination coefficient (cm 4 /s) Recombination coefficient ( o C) Nagasaki [1] Andrel [3] D-Fe D-bulk W Doyle [2] H-Fe This work D-SP-W This work D-F82H /T (K -1 ) KK rr DD WW = eeee eeeeee kkkk KK rr (DD FFFFFFFF) = eeee eeeeee kktt cm 4 s -1 cm 4 s -1 [1] T. Nagasaki et al., J. Nucl. Mater. 202 (1993) [2] B. L. Doyle et al., J. Nucl. Mater. 123 (1984) [3] R. Anderl et al., Fusion Sci. Technol. 21 (1992)

30 Increased D retention in SP-W coated F82H has been observed: trapping effects [1]? PDP (implantation) TDS (desorption) SP-W coated F82H SP-W 10 coatings effects: 14 PDP flux (D/cm 2 /s) Steady-state D-PDP fluxes ( o C) increase 10 fuel retention. 13 F82H SP-W:1.5 µm, F82H: 1 mm /T (K -1 ) Deuterium retention (D/cm 2 ) [1] B. Zajec, et al., J. Nucl. Mater. 412 (2011) D retention by TDS with coating SP-W coated F82H (1.5 µm/1 mm) F82H (1 mm) ( o C) without coating /T (K -1 ) 30

31 SIMS deuterium conc. depth profile SIMS depth profile D retention data by TDS D-PDP: D/cm 2 /s, ~250, 2 h D retention: D/cm 2 SIMS: O 2 +, 5 kev, 250 na (Secondary Ion Mass Spectrometry) SIMS depth profile shows a good agreement with the D retention data and exhibits a sign of uphill diffusion of D in the W/F82H bi-layer structure. CC 1,ii = SS 1 CC S 2,ii SS 2 eff 31

32 Evidence Possible of trapping sites effects in in SP-W SP-W GDP flux (x10 13 D/cm 2 /s) Vacancy cluster H Time (s) 1st run 2nd run 3rd run 4th run Hatano et al., data presented on PSI-22, Rome (2016). Grain boundary Dislocation W Possible trapping sites: 1) Grain boundaries; 2) Dislocations; 3) Vacancy and vacancy clusters Y. Xu et al., Fusion Eng. Des. DOI: /j.fusengdes (2017). 32

33 3.3. VPS+SP double-layer tungsten coated F82H Edge plasma D + T + D + D + T + T + D + T + VPS-W SP-W F82H First wall Breeding blanket T-GDP D, T-PDP T 2 T 2 T 2 T 2 T 2 T 2 T 2 33

34 Double-layer W coatings suppress PDP PDP flux (atoms/cm 2 /s) Hydrogen PDP at ~ o C SP-W coated F82H F82H VPS-W coated F82H VPS/SP-W coated F82H Time (hour) PDP flux (D/cm 2 /s) Steady-state PDP fluxes ( o C) F82H VPS/SP-W coated F82H /T (K -1 ) Bi-layer W coatings effects on PDP: VPS/SP-W suppresses DT-PDP from the plasma side. 34

35 Double-layer W coatings suppress GDP F82H: 1 mm, SP-W: 1.5 µm, VPS-W: 90 µm ( o C) Steady-state GDP fluxes GDP flux (D/cm 2 /s) VPS/SP-W coated F82H F82H /T (K -1 ) Bi-layer W coatings effects on GDP: VPS/SP-W suppresses T-GDP from the blanket side. 35

36 Summary I. VPS-W coatings have been found to have a low density structure with connected pores, while SP-W coatings have a dense and pore-free structure. II. Effects of W coatings on GDP and PDP have been investigated: 1) VPS-W coatings suppress PDP, but does not suppress GDP. 2) SP-W coatings suppress GDP, but tend to enhance PDP. III. Based on these observations, the use of overlaid VPS+SP coatings has been proposed for the first wall design. IV. It has been demonstrated that VPS+SP double layer W-coatings can suppress both GDP and PDP of hydrogen isotopes through the first wall. 36