Comparison of tungsten fuzz growth in Alcator C-Mod and linear plasma devices

Comparison of tungsten fuzz growth in Alcator C-Mod and linear plasma devices G.M. Wright 1, D. Brunner 1, M.J. Baldwin 2, K. Bystrov 3, R. Doerner 2, B. LaBombard 1, B. Lipschultz 1, G. de Temmerman 3, J.L. Terry 1, D.G. Whyte 1, and K.B. Woller 1 1 Plasma Science & Fusion Center, MIT, Cambridge USA 2 Center for Energy Research, UCSD, San Diego, USA 3 DIFFER, Nieuwegein, NL UCSD University of California San Diego 1

Growth conditions for W fuzz are well-defined from experiments on linear plasma devices 1. Bare W or Mo surface 2. 1000 K < T surface < 2000 K 3. Flux of He-ions with E He 20 ev 4. Layer thickness depends on t 1/2 Bubble Formation Fuzz Formation [1] S. Kajita et al. Nucl. Fusion 49 (2009) 095005 [1] S. Kajita et al. Nucl. Fusion 49 (2009) 095005 W fuzz grown in NAGDIS-II 2

Will surface morphology of tungsten modify into fuzz under helium bombardment in ITER and reactors? 1. Bare W surfaces? Bare W surface in net erosion regions of ITER divertor Entire first wall is potentially bare W in a fusion reactor 2. Surface temperatures of 1000-2000 K? For ITER divertor, 5 MW/m 2 830 K, 15 MW/m 2 1500 K Reactor W will operate above ~800 K for efficiency and below W recrystallization temperature (~1600 K) Bubble Formation W re-crystallization Fuzz Formation ITER divertor with 15 MW/m 2 ITER divertor with 10 MW/m 2 ITER divertor with 5 MW/m 2 Minimum T wall for efficient reactor 3

Will surface morphology of tungsten modify into fuzz under helium bombardment in ITER and reactors? 3. Flux of He ions with E 20 ev? He ions born from D-T reactions Growth rate saturates at fluxes >10 22 He/m 2 s [2] For total ion fluxes of ~10 24 m -2 s -1, only 1 % He content needed for maximum fuzz growth rate 4. Growth time? ITER has 300 s pulse length Fusion reactors operate in steadystate All basic growth conditions will be met in the ITER divertor and a fusion reactor. [2] M.J. Baldwin et al. J. Nucl. Mater. 390-391 (2009) 886 4

Will inherent differences between tokamak plasmas and linear device plasmas prevent fuzz growth in a tokamak? B-field C-Mod Tokamak 5.4 T, Grazing incidence Parallel Heat flux ~200 MW/m 2 ~1 MW/m 2 Exposure stability Transient, Disruptions Ionization MFP, Re-deposition Short, prompt redeposition Recycling flux High Low Generic Linear Device ~0.1 T, Typically normal incidence Steady-state Typically > plasma column radius, little or no redeposition Ion incidence Grazing Closer to normal Exploit ITER/reactor-like densities in the C-Mod divertor to find the answer! 5

Alcator C-Mod helium plasmas produced necessary plasma conditions for fuzz growth at the outer strike point He plasma discharges Strike point run above vertical divertor face to reduce flux expansion allowing for higher local surface temperatures. W Langmuir probe 14 repeated L-mode discharges T e,divertor 20-25 ev, q > 0.2 GW/m 2 ~13 s of total exposure at appropriate growth conditions Heat flux instrumented tiles 6

Tungsten Langmuir probe reached and exceeded surface temperatures required for W fuzz growth W Langmuir probe ramped ~11 o into parallel heat flux and is electrically/thermally isolated. Bubble regime Nano-tendril growth regime W Langmuir probe intercepts significant parallel heat flux and rapidly reaches high surface temperatures. W Langmuir probe surface heat flux is obtained directly from Langmuir probe measurements, T surf is determined from 1-D finite element heat flux modeling. 7

Nano-tendrils are fully formed on surface of the tungsten probe exposed to heat fluxes of 30-40 MW/m 2 After exposure W Langmuir Probe Mo ramped tiles Thickness of individual tendril is ~100 nm. 8

Is the empirical growth rate determined on PISCES applicable to the fuzz grown in Alcator C-Mod? 900 ka disruptions Baldwin formula Sputtering of bulk W Growth is estimated through t 1/2 -dependence: layer depth = δ G(T surf ) t 1/2 where G exp(-e act /kt surf ), E act = 0.71 ev [3] M.J. Baldwin, R.P. Doerner, Nucl. Fusion 48 (2008) 035001 Calculated cumulative layer depth of ~520 nm for W probe Sputtering only a small contribution in W case (~35 nm of calculated bulk W gross sputtering) QUESTION: Baldwin formula is based on data set at temperatures of 1120 K and 1320 K. Experimental results from Kajita [4] and de Temmerman [5] show this formula to over-estimate layer thicknesses for T > 1400 K. Why is the agreement with C- Mod results (2000+ K) so good? Short pulses? Transient surface temperatures? [4] S. Kajita et al. J. Nucl. Mater. 418 (2011) 152 [5] G. de Temmerman et al. ICFRM-15, Charleston, USA, October 2011 10

Pilot-PSI is capable of re-creating many of the plasma and surface conditions in the Alcator C-Mod divertor Similarities to Alcator C-Mod: High plasma fluxes and densities Plasma heated surfaces/high heat flux Able to operate in short pulses Central flux density ~3x10 24 m -2 s -1, off-center (10 mm) still >10 23 m -2 s -1 Exposure consists of 14 x 1.5 s shots = 22.5 s of total plasma exposure time. 11

Pilot-PSI exposure allows for the examination of the surface morphology at several different temperatures 1 µm T surf, max = 1800 K 12

Pilot-PSI exposure allows for the examination of the surface morphology at several different temperatures 1 µm T surf, max = 1700 K 13

Pilot-PSI exposure allows for the examination of the surface morphology at several different temperatures 1 µm T surf, max = 1550 K 14

Pilot-PSI exposure allows for the examination of the surface morphology at several different temperatures 1 µm T surf, max = 1350 K 15

Pilot-PSI exposure allows for the examination of the surface morphology at several different temperatures 1 µm T surf,max = 1120 K 16

Pilot-PSI exposure allows for the examination of the surface morphology at several different temperatures 1 µm T surf, max = 1000 K 17

Fuzz tendril thickness increases with increasing temperature and void size Similar dependence on surface temperature for both tendrils and voids. 18

Morphology very similar to C-Mod fuzz is found on the Pilot-PSI target, but at significantly lower surface temperatures Alcator C-Mod Pilot-PSI 1 µm Tendril thickness = 100 ± 10 nm T surf, max = 2200 ± 300 K No tendrils formed T surf, max = 2350 ± 100 K 19

Morphology very similar to C-Mod fuzz is found on the Pilot-PSI target, but at significantly lower surface temperatures Alcator C-Mod Pilot-PSI Tendril thickness = 100 ± 10 nm T surf, max = 2200 ± 300 K Tendril thickness = 95 ± 20 nm T surf, max = 1550 ± 100 K 20

Fuzz layer depth is also very similar between C-Mod and Pilot-PSI fuzz despite the different temperatures Alcator C-Mod Pilot-PSI T surf, max = 2200 ± 300 K Typical bubble size ~ 35 nm Prediction from PISCES empirical formula 520 nm T surf, max = 1500 ± 100 K Typical bubble size ~ 30 nm Prediction from PISCES empirical formula 230 nm Despite different temperatures, nearly identical surface morphology grown in Alcator C-Mod and Pilot-PSI. 21

What is causing the different temperature dependence of the fuzz morphology between C-Mod and Pilot-PSI exposures? Alcator C-Mod Target Biasing -100 to +50 V, 100 Hz Pilot-PSI -40 V, constant Grains Re-crystallized Unannealled Magnetic field 5.4 T, grazing Target cooling Inertial 0.8 T, normal Active water cooling Incident He ion energy is known to impact fuzz surface morphology [6]. Re-crystallized W has shown to grow much deeper layers of W fuzz than other forms of W [7]. [6] M.J. Baldwin et al. J. Nucl. Mater. 415 (2011) S104 [7] M.J. Baldwin et al. J. Nucl. Mater 404 (2010) 165. 22

He voids/bubbles play a critical role in the formation and morphology of these fuzz layers. How do they form? Pressure-driven model: He solubility in W is extremely low Lattice imperfections (vacancies, grain boundaries, impurities, etc) can act as seed sites for He precipitation. For more work on He bubbles and their effects see: P2-082, F. Sefta/Brian Wirth, Today! P3-021, Niklas Juslin, Thursday 23

He voids/bubbles play a critical role in the formation and morphology of these fuzz layers. How do they form? If W is in an oversaturated state, He will continue to precipitate and collect at the imperfection. Pressure builds and stresses are induced in the surrounding lattice. 24

He voids/bubbles play a critical role in the formation and morphology of these fuzz layers. How do they form? When critical pressure is reached (typically ~few GPa), loop-punching is induced and a void develops. He continues to precipitate into the larger void until loop-punching pressure is reached again and the void continues to grow. This is the basic premise behind the pressuredriven loop-punching model for void growth. 25

What do we expect the helium concentration to be in these fuzz layers? Assume all He is trapped in voids Void pressure determined by equilibrium state: P = 2γ/r void P is void pressure γ is surface tension r void is void radius at T W = 1350 K, r void ~ 5 nm, P He ~ 1GPa n He = P He /kt W = 5.4 x 10 28 m -3 ~ 0.85n W From other work, TEM imaging [8] and ellipsometry [9] shows typical void fractions of 20-50% in W fuzz. TEM of W fuzz [8] Expected He concentration in W fuzz layer is 17-43 at.% [8] S. Kajita et al. J. Nucl. Mater. 421 (2012) 22. [9] M. Miyamoto et al. J. Nucl. Mater. 415 (2011) S657. 26

Heavy ion elastic recoil detection allows us to measure the He content in the surface layer/fuzz E o =4.0 MeV 16 O 3+ θ inc = 75 θ exit = 75 Aperture 3.2 μm Al Range Foil θ recoil = 30 He Solid State Detector More details P2-83 by Kevin Woller (today!) 27

He content in tungsten targets exposed in Pilot-PSI and PISCES indicates a lack of high He pressures in voids He solubility in W ~0.00001 at.% * C-Mod W probe surface was contaminated before true He concentration could be measured. Concentration is well above natural solubility, but below what might be expected for ~GPa He pressures in all voids. No clear trend vs. surface temperature, despite varying void sizes. 28

Where did the He go? If pressure-driven loop punching is the growth mechanism, the post-mortem He pressures in these voids have been greatly reduced from their equilibrium pressures. The void fractions of these targets was not measured directly, but FIB/SEM cross-sections indicate void fraction is significant. TEM observation is needed to properly account for all voids. He outgassing from W fuzz occurs at 1000-1600 K [10], meaning there could be outgassing during target cool down Dynamic measurement of He concentration during fuzz growth is needed to confirm presence of pressure-driven loop-punching. Work planned for the DIONISOS experiment. [10] M.J. Baldwin et al. Nucl. Fusion 51 (2011) 103021 29

Summary: W fuzz can be grown in a tokamak environment Tokamak plasmas have been shown to be capable of growing W fuzz if the proper growth conditions are met. W fuzz layer depth is in good agreement with empirical formula from Baldwin et al. even at high surface temperatures. Comparison of W fuzz from Pilot-PSI linear plasma device shows nearly identical surface morphology as the C-Mod W fuzz, but with a different temperature dependence on the morphology characteristics. Measured He concentrations of 1-4 at.% in modified layers are well below what might be expected if sub-surface voids are filled with ~GPa of He gas. 30

Open questions and future work Is the different temperature dependence on surface morphology reproducible? Planned hot divertor in Alcator C-Mod would be greatly beneficial to W fuzz studies in tokamaks. What is the He concentration during He implantation? Confirm pressure-driven void model? Will irradiation damage (neutrons) have any effect on fuzz formation or layer growth? Due to high temperatures, this can only be studied dynamically What is the influence of ELMs and impurity seeding on fuzz growth in tokamaks? Not covered in C-Mod exposures, but work has been done in Linear devices. 31