Erosion/redeposition analysis of Erosion/redeposition analysis of CMOD Molybdenum divertor and CMOD Molybdenum divertor and

NSTX Liquid Lithium DivertorNSTX Liquid Lithium Divertor

J.N. Brooks, J.P. Allain

Purdue University

PFC Meeting MIT, July 8-10, 2009

J.N Brooks, PFC/MIT 7/09

2

CMOD Mo tile divertor erosion/redeposition analysis CMOD Mo tile divertor erosion/redeposition analysis [ w/ D. Whyte, R. Ochoukov (MIT) ]

Puzzling results for Mo tile erosion--high net sputtering erosion in apparent contradiction of some models.

REDEP code package (rigorous) analysis of outer divertor conducted. {New probe diagnostic being implemented-MIT.} Sheath BPH-3D code applied to very near tangential (~0.6°) magnetic

field geometry. D, B, Mo on Mo, sputtering erosion/transport analyzed, for 600, 800,

1000, 1100 KA shots, and for OH and RF phases. Uses TRIM-SP sputter yield and velocity distribution simulations.

RF induced sheath effect studied.

W.R. Wampler et al. J. Nuc.Mat. 266-269(1999)217.

J.N Brooks, PFC/MIT 7/09

3

0

200

400

600

800

1000

1200

1400

1600

1800

2000

2200

0.55 0.54 0.53 0.52 0.51 0.5 0.49 0.48 0.47 0.46

Gro

ss M

o er

osio

n , n

m

Distance below midplane, m

WBC C-MOD outer divertor analysis. 1/28/098 plasma conditions (4 currents x (OH phase +RF phase), 1259 stotal, 1% B + D sputtering and self-sputtering, (no B near-surfacerecycle), TRIM-SP pure-Mo sputter simulation data. ~1e6 histories.

w/rf sheath

no rf sheath

-150

-100

-50

0

50

100

150

0.55 0.54 0.53 0.52 0.51 0.5 0.49 0.48 0.47 0.46N

et M

o er

osio

n , n

m

Distance below midplane, m

WBC C-MOD outer divertor analysis. 1/28/098 plasma conditions (4 currents x (OH phase +RF phase), 1259 stotal, 1% B + D sputtering and self-sputtering, (no B near-surfacerecycle), TRIM-SP pure-Mo sputter simulation data. ~1e6 histories.

w/rf sheath

no rf sheath

Ero

sion

, nm

Ero

sion

, nm

Predicted Mo erosion along CMOD divertor

Gross erosion Net erosion

J.N Brooks, PFC/MIT 7/09

4

DATA (post-exposuretile study-Wampler et al.)

Code/data comparison- CMOD divertor Net Erosion

• Reasonable code/data agreement

WBC

J.N Brooks, PFC/MIT 7/09

5

Code/data comparison-CMOD divertor Gross Erosion

• Poor code/data agreement (higher code values)

J.N Brooks, PFC/MIT 7/09

6

Goals-determine:

• Surface temperature limit set by lithium sputtering/runaway and/or evaporation

• Lithium density in plasma edge/sol; core plasma contamination potential

•Flux of sputtered lithium to carbon surfaces and D-pumping capability.

REDEP/WBC NSTX Liquid Lithium Divertor AnalysisREDEP/WBC NSTX Liquid Lithium Divertor Analysis[with D. Stotler, R. Maingi et al.][with D. Stotler, R. Maingi et al.]

J.N Brooks, PFC/MIT 7/09

7

REDEP/WBC LLD AnalysisREDEP/WBC LLD Analysis

REDEP/WBC code simulation of NSTX Liquid Lithium Divertor (LLD). Full kinetic, sub-gyro-orbit analysis. (100,000 sputtered histories per simulation).

Plasma parameters from UEDGE/DEGAS solution, “0.65 D+ reflection

coefficient” [D. Stotler, R. Maingi, et al. (2008)]. Includes LLD surface temperature profile, at t = 2.01020 seconds [L. Zakharov]. (Tmax = 281° C).

Incident particles: D+ ions, Li ions. (C on Li under analysis)

Energy-dependent and surface-temperature-dependent sputter yields for

D, Li, incidence, from TRIM-SP code runs (J.P. Allain), for D containing Li; 45° incidence.

Charged/neutral lithum sputtered fraction model (Allain) used (~2/3

charged, ~1/3 neutral). Reference WBC model used for recycle/re-sputter of sputtered Li+ (net atomic Li sputtered ~ ½ of total gross (charged + neutral).

J.N Brooks, PFC/MIT 7/09

8

REDEP/WBC LLD Analysis continuedREDEP/WBC LLD Analysis continued

Sputtered Li atom velocity distribution functions used in WBC per TRIM-SP results.

Li I density-dependent and Te-dependent electron impact ionization rate coefficients from ADAS, S.D. Loch et al., Atomic Data and Nuclear Data Tables 92(2006)813.

Sheath: BPHI-3D code run for NSTX conditions (Brooks/Ochuockov) Dual-structure (Debye sheath + magnetic sheath) found not to be present (due to weaker magnetic field, less oblique B-field incidence, viz. 5-10° NSTX vs. 1-2°. for ITER, CMOD, etc.). Debye-sheath-only model therefore used in WBC. Locally-varying sheath potential e = 3kTe.

J.N Brooks, PFC/MIT 7/09

9

BPHI-3D Code- NSTX Sheath Analysis at Liquid Lithium Divertor

• No magnetic sheath predicted; Debye sheath only

J.N Brooks, PFC/MIT 7/09

10

3-D, fully kinetic, Monte Carlo, treats multiple (~100) processes: Sputtering of plasma facing surface from D-T, He, self-sputtering, etc. Atom launched with given energy, azimuthal angle, elevation angle Elastic collisions between atom and near-surface plasma Electron impact ionization of atom→impurity ion Ionization of impurity ion to higher charge states Charge-exchange of ion with D0 etc. Recombination (usually low) q(E +VxB) Lorentz force motion of impurity ion Ion collisions with plasma Anomalous diffusion (e.g., Bohm) Convective force motion of ion Transport of atom/ion to core plasma, and/or to surfaces Upon hitting surface: redeposited ion can stick, reflect, or self-sputter Tritium co-deposition at surface, with redeposited material Chemical sputtering of carbon; atomic & hydrocarbon A&M processes

REDEP/WBC code package--computation of sputtered particle transportREDEP/WBC code package--computation of sputtered particle transport

J.N Brooks, PFC/MIT 7/09

11

Interesting erosion/redeposition physics for sputtered lithium Interesting erosion/redeposition physics for sputtered lithium in the NSTX low-recycle plasma regime studiedin the NSTX low-recycle plasma regime studied

Large sputtered-atom ionization mean free path, order of 10 cm

Large Li+1 gyroradius ( ~5 mm), due to relatively low B field

Low collisionality, due to high Te, low Ne

Kinetic, sub-gyro orbit analysis required (i.e. WBC code)

12

WBC Simulation of LLD sputtered lithium transport: 50 trajectories shown; 2-D view

• Long mean free paths seen for ionization; subsequent long, complex, ion transport

UEDGE/NSTX GRID

13

WBC Simulation of LLD sputtered lithium transport: 50 trajectories shown; 3-D view

x = distance along divertor (strike point at zero), y=distance along toroidal field

J.N Brooks, PFC/MIT 7/09

14

Gross and net lithium erosion rates along LLD

J.N Brooks, PFC/MIT 7/09

15

Liquid Lithium Divertor: Li Transport ResultsLiquid Lithium Divertor: Li Transport ResultsResults, run of 5/5/09 (PRELIMINARY)

Sputtering is “OK”. Peak gross rate = 10.3 nm/s, peak net rate = 5.7 nm/s. Average effective D+ sputter yield = 0.07

50.3% of sputtered (neutral) lithium is ionized within the computation

zone (LLD and associated near-surface grid). 49.7% of sputtered lithium “escapes”.

Of the zone-ionized material, 91% is redeposited on LLD, 9% escapes.

Fate of escaping lithium: ~13% (of total sputtered) goes towards core

plasma (not followed further), ~24% goes to outside (higher major radius) of LLD, ~17% goes to inside of LLD. Thus, roughly 50% of sputtered Li would likely impinge on the various carbon surfaces.

Total sputtered Li atom current = 1.38 x1020 s-1. About 9% is from self-

sputtering, rest from D+ sputtering.

J.N Brooks, PFC/MIT 7/09

16

Parameter (average across divertor) Value

Charge state 1.083

Energy 429 eV

Transit time 41 s

Angle (to surface normal) 21 deg

LLD redeposited lithium ion parameters

J.N Brooks, PFC/MIT 7/09

17

CMOD Molybdenum divertor Analysis• Analysis completed for complex, ~1200 sec campaign, with 8 plasma conditions.• Acceptable code/data comparison for net Mo erosion, not good for gross erosion-

puzzling result.• RF sheath significant but not a major effect.

Gross rate data via Mo photon emission being re-analyzed (Whyte et al.), WBC

analysis will continue as needed.

•NSTX Liquid Lithium Divertor Analysis•Results are encouraging-

--Moderate lithium sputtering; no runaway--About 50% of the lithium is transported to carbon surfaces--LLD could apparently handle much higher heat flux

• Several model enhancements being implemented• Higher temperature case analysis, at ~5.0 seconds • Carbon sputtering of lithium and C/Li material mixing; MD modeling of Li-D-C system (w/ P. Krstic ORNL)

• Higher power case-will need further UEDGE analysis

ConclusionsConclusions