R. A. Pitts

1
Centre de Recherches en Physique des Plasmas
Material erosion and migration in tokamaks
R. A. Pitts CRPP, Association-EURATOM
Confédération Suisse, EPFL Lausanne, Switzerland
with many thanks for contributions from N.
Asakura1, S. Brezinsek2, C. Brosset3, J. P.
Coad4, D. Coster5, E. Dufour3, G. Federici6, R.
Felton4, M. E. Fenstermacher7, R. S. Granetz8, A.
Herrmann5, J. Horacek, A. Kirschner2, K.
Krieger5, A. Loarte9, J.Likonen10, B.
Lipschultz8, A. Kukushkin6, G. F. Matthews4, M.
Mayer5, R. Neu5, J. Pamela11, B. Pégourié3, V.
Philipps2, J. Roth5, M. Rubel12, L. L. Snead13,
P. C. Stangeby14, J.D. Strachan15, E. Tsitrone3,
W. Wampler16, D. Whyte17 1JAERI, 2FZJ-Jülich,
3CEA Cadarache, 4UKAEA, 5IPP Garching, 6ITER,
7LLNL, 8PSFC-MIT, 9EFDA CSU Garching,
10VTT-TEKES, 11EFDA CSU Culham, 12Alfvén Lab.
RIT, 13ORNL, 14UTIAS, 15PPPL, 16SNL,17Univ.
Wisconsin,
2
Outline of the talk

• Introduction
• The components of migration
• Global migration accounting
• Material choices for the next step
• Conclusions

3
What is migration?
Migration

Transport
Erosion
Deposition
Re-erosion
Not an operational issue in todays tokamaks, but
certainly will be in ITER and beyond
4
Migration will be important

• Co-deposition
• High erosion rates and long term migration of
  carbon yield high levels of Tritium retention

• ITER 50 g T per pulse
• 0.01-0.2 g per pulse now
• ITER operation suspended once 350 g T accumulated

Could be fewer than 100 pulses No proven T
clean-up technology

• Material mixing, properties
• Formation of compounds and alloys through the
  interaction of pure materials
• Change of material properties

• Be on W forms BeW alloys already at 800C

Surface melting point could be 2000C lower than
for pure W
5
Where do erosion and migration occur?
JET 62218 plasma visible light emission
At specific structures to protect the vacuum
vessel walls or isolate the plasma-surface
interaction
6
Some terminology
Poloidal cross-section

• Scrape-off layer (SOL)
• Cool plasma on open field lines
• SOL width 1 cm (? B)
• Length usually 10s m ( B)

Core plasma

• Divertor
• Plasma guided along field lines to targets remote
  from core plasma low T and high n

Separatrix
Private flux region
Inner
Outer
ITER will be a divertor tokamak
Divertor targets
7
Materials in todays tokamaks
Low Z (Carbon) Low Z (Carbon) High Z
Divertor TCV, MAST, NSTX, DIII-D, JT-60U, JET AUG (CW) AUG (CW) C-Mod (Mo)
Limiter TEXTOR, Tore Supra FTU (Mo)

• The majority of todays medium to large size
  tokamaks favour Carbon ? extensive operational
  experience
• No melting / low core radiation / high edge
  radiation

But T-retention problem and high erosion rates of
low Z mean that high Z may be the only long term
solution
8
Upscale to ITER is a big step
Parameter JET MkIIGB(1999-2001) ITER
Integral time in diverted phase 14 hours 0.1 hours
Number of pulses 5748 1
Energy Input 220 GJ 60 GJ
Average power 4.5 MW 150 MW
Divertor ion fluence 1.8x1027 6x1027
Matthews et al., EPS 2003
1 ITER pulse 0.5 JET years energy input
1 ITER pulse 6 JET years divertor fluence
code calculation
9
Migration

Transport
Erosion
Deposition
Re-erosion
10
Principal erosion mechanisms

• Sputtering
• Ions and neutrals
• Physical and chemical (for carbon)
• Macroscopic - transients
• Melt layer losses
• Evaporation, sublimation
• Not generally observed in present experiments
  currently the main reason for Carbon being used
  in the ITER divertor
• Arcing, Dust

11
Physical and chemical sputtering
Chemical (carbon)
Physical
No threshold Dependent on bombarding energy, flux
and surface temperature ? More optimistic
prediction for ITER
Energy threshold ? higher for higher Z
substrate Much higher yields for high Z
projectiles
12
ELMs an example of transient erosion
H-mode ? Edge MHD instabilities ? Periodic bursts
of particles and energy into the SOL. Type I
ELMing H-mode is baseline ITER scenario
Time (s)
13
ELMs can ablate Carbon on JET
Range of energies expected per Type I ELM in ITER
0.6 ? 3.5 MJm-2
Loarte et al, Phys. Plasmas 11 (2004) 2668
14
ELM ablation limits ITER divertor lifetime

• Acceptable lifetime before target change
  required
• 3000 full power shots ? 1 x 106 ELMs

• Both low and high Z target materials marginal on
  present scalings
• Significant effort in the community towards ELM
  mitigation

15
ELMs might also erode the main walls

• Main chamber thermography on AUG

A. Herrmann, AUG

• Type I ELMs 25 of stored energy drop deposited
  on non-divertor components

• ELM ion energies measured at JET walls agree with
  recent theory
• SuggestsEion gt 1 keV on ITER ? erosion
  problem, even for high Z wall

16
Migration

Transport
Erosion
Deposition
Re-erosion
17
Transport creates moves impurities
Ions
Cross-field transport high ion fluxes can
extend into far SOL? recycled neutrals? direct
impurity releaseELMs ..
Eroded Impurity ions leak out of the divertor
(?T forces)
SOL and divertor ion fluid flows can entrain
impurities
18
Experimentally, strong SOL flows
D-flows parallel Mach Number, M v/cs.
POSITIVE towards inner target
M
M
JET
(Tore-Supra)
C-Mod
JT-60U
(TCV)
M
N. Asakura, NF 44 (2004) 503 B. LaBombard, NF 44
(2004) 1047 S. K. Erents, PPCF 46 (2004) 1757
JT-60U
M
M
JT-60U
Distance to separatrix (mm)
Distance to separatrix (mm)
19
Can SOL ion flows transport material?
Yes, but picture is complex theory and
experiment not yet reconciled
Poloidal
Bj
Parallel
Simplified shown in the poloidal plane only
20
Using tracers to study the transport
13CH4 markers are being increasingly used to get
a handle on migration
? gas puff just before vent and tile retrieval
pioneered on TEXTOR
0.2g 13C, L-mode
2.8g 13C, ohmic
0.2g 13C, H-mode
JET
DIII-D
AUG
0.0025g 13C H-mode
9.3g 13C H-mode
21
Top injection C13 ? inner target
Wampler et al, JNM 337-339 (2005) 134
Likonen et al, Fus. Eng. Design 66-68 (2003) 219
JET
Start
End
DIII-D

• Simple conditions ohmic, L-mode, no ELMs
• DIII-D toroidally symmetric injection, JET
  toroidally localised
• Data and modelling demonstrate fast flow to inner
  divertor
• Situation more complex in H-mode and other
  injection points

22
Migration

Transport
Erosion
Deposition
Re-erosion
23
Deposition sensitive to local conditions
DIII-D

• Outer divertor usually hotter ? favours C erosion
  (phys. chem.)
• Inner divertor usually colder ? favours C
  deposition (chem. only)
• C transport by SOL flows
• Similar picture on most other carbon machines

Observations consistent with a contribution to
the carbon source from outside the divertor
Detached
Whyte et al., NF 41 (2001) 1243, NF 39 (1999) 1025
24
Re-erosion important for C-migration
Esser et al., JNM 337-339 (2005) 84
JET
L-mode
C-deposition (nm/s)
ERO code

• Reproduced by transport modelling
• Large increase on baseplate requires enhanced C
  re-erosion

Chemical erosion
Migration to remote areas due to magnetic and
divertor geometry
Kirschner et al, JNM 337-339 (2005) 17
25
Global migration accounting

Transport
Erosion
Deposition
Re-erosion
26
A non-trivial task!
Spectroscopic methods in plasma, post-mortem
surface analysis and just plain old scraping and
sweeping up ? extremely rigorous balance achieved
first on TEXTOR (Wienhold et al., JNM 313-316
(2003) 311)
Tore Supra balance see Dufour et al, P5.002
Friday
27
JET migration accounting (I)
Use spectroscopic methods modelling to compute
C sources
EDGE2D/NIMBUS DIVIMP/OSMSimulation of CIII
emission ? intrinsic sources
1 ton/year
Divertor C-source 5-10 x Wall source
Strachan et al, NF 43 (2003) 922
Carbon recycles
28
JET migration accounting (II)
Make balance for period 1999-2001 with MarkII
GasBox divertor 14 hours plasmain diverted
phase (50400 s, 5748 shots)
Spectroscopy Modelling

• Post mortem surface analysis
• Deposition all at inner target

215 kg/year ? strong T co-deposition
(1 year 3.2 x 107 secs)
Very similar result for AUG, but overall
C-balance more complex
Mayer et al, JNM 337-339 (2005) 119
Likonen et al, JNM 337-339 (2005) 60, Matthews et
al., EPS 2003
29
Tungsten migration in AUG
2002-2003 Campaign 1.4 hours in diverted phase
(4680 s, 1205 shots)

• Post mortem surface analysis
• Only 12 of inboard W source deposited in
  divertor
• few to upper divertor and other main chamber
  surfaces

W-coated (40 of total area)
1.3x1018s-1
W erosion not balanced by non-local deposition
most is promptly redeposited ? simpler than C
picture
Larger Larmor radius helps at higher mass
1.5 kg/year
0.5x1017s-1
1.1x1017s-1
Krieger et al, JNM 337-339 (2005) 10
30
Material choices for the next step
An ITER-like first wall at JET
31
Current materials choice for ITER

• Be for the first wall
• Low T-retention
• Low Z
• Good oxygen getter

W

• C for the targets
• Low Z
• Does not melt

350 MJ stored energy

• W for the baffles
• High threshold for CX neutral sputtering

CFC

• Fallback option
• Be wall, all-W divertor

Castellations for stress relief ? co-deposition
in gaps?
Driven by the need for operational flexibility
32
An ITER-like wall in JET

• Option 1 or 2 to be chosen in 2006 Objectives
• Demonstrate low T-retention
• Study melt layer loss (walls and divertor) ? ELMs
  and disruptions
• Study effect of Be on W erosion
• Be and W migration
• Demonstrate operation without C radiation
• Refine control/mitigation techniques? ELMs and
  disruptions

Demonstrate routine / safe operation of fully
integrated ITER compatible scenarios at 3-5MA?
Power upgrade to 40-45 MW ? Experiments from 2009
onwards
33
Conclusions

• Erosion and migration Complex materials and
  physics
• Not an operational issue now
• But will be in ITER and beyond
• Optimisation of core plasma performance and wall
  lifetime cannot be decoupled
• Refine predictive capability

Still significant uncertainties .
? Full wall materials tests in current machines
34
Reserve slides
35
Carbon balance TEXTOR, Tore Supra
Carbon Sources (g/h)
von Seggern et al, Mayer et al., Phys. Scripta
T111 (2004) Wienhold et al, von. Seggern et al.,
JNM 313-316 (2003)Brosset et al., JNM 337-339
(2005) 311, E. Tsitrone et al., IAEA 2004
TEXTOR
TS
Toroidal limiters 22 7
Carbon Sinks (g/h)
Toroidal limiters 10 1
Obstacles 6 0.5
Low sticking also AUG
Bumper 1 ?
Neutralisers 1 1-2
Very good balance considering the scope for error
TEXTOR deposition extrapolates to220 kg/year
of plasma Tore-Supra balance still preliminary
Pump ducts 0.02 ?
Pumped out 1-2 0.2-2
Total 19-20 2.7-5.5
36
Similar observations at JET
Net inner divertor deposition and little net
erosion in outer divertor implies net wall source
Macroscopic flakes in regions not generally
visible to plasma ? migration to remote areas ?
high levels of T-retention
Flakes
Coad et al., JNM 313-316 (2003) 419
37
JET migration accounting (II)
Make balance for period 1999-2001 with MarkII
GasBox divertor 16 hours plasma
Spectroscopy Modelling

• Post mortem surface analysis
• Deposition all at inner divertor
• Surface layers are Be rich ? C chemically eroded
  and migrates, Be stays put

215 kg/year ? strong T co-deposition
Very similar result for AUG, but overall
C-balance more complex
Mayer et al, JNM 337-339 (2005) 119
Likonen et al, JNM 337-339 (2005) 60, Matthews et
al., EPS 2003