Pr

1
Fuel retention in tokamaks
Th Loarer with special thanks to N Bekris, S
Brezinsek, C Brosset, J Bucalossi, P Coad, G
Esser, W Fundamenski, E Gauthier, A Kreter, J
Likonen, B Lipschultz, M Mayer, P Monier-Garbet,
Ph Morgan, R Neu, B Pégourié, V Philipps, R
Pitts, V Rohde, J Roth, M Rubel, C Skinner, J
Strachan, E Tsitrone. And also EU TF on PWI and
JET EFDA contributors
2
OUTLINE - Introduction - Evaluation of fuel
retention in tokamaks - Gas Balance - Post
mortem analysis - Extrapolation to ITER - Summary

3
Introduction

• Evaluation of hydrogenic retention in present
  tokamaks is of high priority to establish a
  database and a reference for ITER (400 susually
  10-20 s today).
• T retention constitutes an outstanding problem
  for ITER operation particularly for the choice of
  the materials (carbon ?)
• A retention rate of 10 of the T injected in
  ITER would lead to the in-vessel T-limit
  (350/700g) in 35/70 pulses.
• Retention rates of this order or higher (20)
  are regularly found using gas balance.
• - Retention rate often lower (3-4) are obtained
  using post mortem analysis

4
Mechanisms for fuel retention
Two basic mechanisms for Long term fuel retention
C, Be
C, Be, D ,T
Codeposition
Short term retention (Adsorption dynamic
retention) Recovered by outgassing
Deep Implantation, Diffusion/Migration, Trapping
In tokamaks today ? Two complementary methods Gas
balance ? How much ? During the pulse /
integrated pulse/days. Post mortem ? Where and
how? Integrated over experimental campaign
5
Gas balance
Calibrated Particle Source (Gas, NBI, pellets)
Plasma
Retention Short Long Term
Exhaust by NBI boxes, Diag
Divertor cryo-pumps
Gas Balance ? Retention Injection - Pumped
Plasma
1 - Function of time - Measurement of the
Injected and pumped fluxes (Pressure gauges and
pumping speed) 2 - Integral of Pumped flux
intershot outgassing ( Short Term Ret )
(Collected in a calibrated volume). ?
Separate Short and long term retention
6
Retention Short and long term
Actively cooled device? Steady state
operation--gtLong term retention
Short term retention - Depends on plasma
scenario, wall conditioning and Material (Be,
C) -Limited to fast reservoir and recovered
in between pulses (outgassing)
Dynamic retention 5 x1021D JET 2.5 1022 D
wall area ratio
Long term retention - Co-deposition ?
Correlated to C production - Implantation ? Edge
plasma, material structure
7
Plasma scenario C prodution
Series of repetitive and consecutive discharges
(no history effect)
Pulse type Divertor phase (s) Injection(Ds-1) Long term retention (Ds-1)
L-mode 2 MW 126 1.8?1022 1.34?1021
Type III 6 MW - lt5-10 kJ 350 0.6?1022 0.8?1021
Type I 13 MW 100 kJ 50 1.7?1022 2.08?1021
T Loarer et al., EPS 2007

• Long term retention
• Drop by 60 when moving from L-mode to Type III
  H-mode.
• Increases by 60 when moving from L-mode to
  Type I H-mode.
• ? Retention correlated to increase of C erosion

8
Carbon production and scenario
J Strachan et al. Nuc. Fus. 2003
- Increase by a factor of 2 of carbon source
from L to type I ELMy H-Mode
- Increase of carbon source depends on scenario
(ELMs, recycling flux) ? enhanced retention
by co-deposition
9
Implantation and wall saturation
Nakano et al. IAEA 2004, Nuc Fus 2006
- As Tsurf increases ? Strong outgassing of D,
CxDy from target plates (eventually lost of ne
control) - Outgassing ?signature that
implantation is saturated for overheated PFCs
- Also observed on long discharges in
TRIAM-1M, Tore Supra. (Sakamoto et al. IAEA 2006,
JNM 2007 Grisolia, JNM 1999, ).
But, with gas balance analysis this strong
outgassing hides co-deposition. And since
carbon source increases this enhances retention
by co-deposition.
10
Inventory proportional to duration

• Recent exp. campaign of 10 days on TS repeating
  the same long pulse (2min 2MW)
• - Total of 5 h of plasma w/o conditioning

11
Retention in carbon devices

• Retention by co-deposition dominates

Limit/cancel co-deposition ? ? High Z
12
High Z Alcator C-Mod

• 16 repeated pulses (w/o disrp) in cleaned Mo
  walls
• Retained D fluence remains linear with incident
  D ion ?3.5x1020Ds-1

D Whyte et al. IAEA 2006, B Lipschultz et al.,
PSI 2008
13
High Z - W in ASDEX Upgrade
AUG from all C to all W (Carbon free)

• - With 70 of W the retention was still around
  10-20 (C dominated)
• 100 W significant drop of retention below 1

14
Where is the fuel retained? ? Post mortem
analysis
15
Post Mortem Analysis
Wide range of methods for different objectives
? Surface analysis, Structure, Depth profile,
Composition
Method Principle Quantity measured
SIMS Secondary Ion Mass Spectroscopy Incident ion flux (Cs, Ar, O) with an angle 20-30 with respect to the surface Analysis of sputtered ions CH3, CH4 and/or ionised fragments Depth mm
RBS Rutherford Back Scattering He incident beam (protons) _at_1-3MeV) Measurement of the energy of the reflected beam Analysis of the elemental surface. From nm to mm
TDS Thermodesorption Temperature ramp up (room up to 1600K) of a sample. Under vacuum or inert atmosphere (Mostly Ar or He) w or w/o H2 Total quantity of H, D, O, CxDy trapped in the material, coupled to Mass Spectrometer Activation Energy
NRA Nuclear Reaction Analysis Nuclear reaction triggered above an energy threshold of the incident 3He beam1-3MeV 2D(3He,p) 4He Target 2D, particle analysed p, product 4He 10Be(3He,p) 12B -- 12C(3He,p) 14N -- 13C(3He,p) 15N Ratio D/C, Be/C Depth profile 1Mev?1.5mm 2.5MeV? 7.5mm
PIGE Proton Induced Gamma Emission 13C(p,g) N14 Use the narrow resonance for protons at 1.748MeV giving a 9.17MeV gamma (g) Concentration of 13C vs depth Smaller depth resolution, but detection limit 10 times better than NRA)
NMR Nuclear Magnetic Resonance Resonance frequency proportional to the distance CH of the element connected to H(D) Used for liquids, in organic chemistry
16
Where and how ? Implantation in CFC
D retained in the samples (by TDS)
Test limiter with material stripes exposed in
TEXTOR
Comparison with PISCES-A data
J Roth et al. PSI 2006.
TEXTOR
A.Kreter et al. 2007
No saturation observed for these fluences and a F
0.5
Test limiter for dedicated experiments since
removed at the end of the experiment also the
case of marker exp (13C)
17
Where and how ? co-deposition with C
Different Carbon erosion-transport and eventually
co-deposition with plasma scenario L and H mode
Also 13C experiments in TEXTOR P WIenhold et al.
JNM 2001, JET J Likonen et al. FED 2003, AUG
M Mayer et al. JNM 2005, JET P Coad et al. Nuc.
Fus. 2006.
? Dedicated 13C experiment to localize deposition
and associated retention
18
Carbon deposition on PFCs
C-erosion/deposition JET 2001-2004

• From deposit thickness
• (r 1.0 gcm-3 - 1.8 for the substrate)
• Total C deposition
• Inner 625 g Outer 507 g

J Likonen P1-70
19
D/C ratio and retention
D/C ratios JET 2001-2004
0.02
0.14

• Injected 1800g (5.381x1026D)
• In the divertor area
• Total D 66 g 3.7 of Ginj
• (2.2x1020Ds-1)
• Retention 70 Inner 30 Outer

0.42
0.11
0.15
0.91
0.25
0.12
0.08
0.17
0.79
J Likonen P1-70
20
Retention at First wall
(JET MkII-SRP, 2001-2004)
- Total D retention 0.3g (0.02) - Outer
poloidal limiters have a minor contribution to D
retention
J Likonen P1-70
21
Temperature effect - JT-60U

• JT-60U Normalised to NBI time (8h20)
• - Carbon inner dome 550 g ? 1.0x1021/sec (2.7
  times higher than in JET 3.7x1020s-1 )
• - Carbon outer erosion 340 g
• 210 g comes from the main chamber

PFC ? D/C 0.01-0.15
Remote? D/C 0.75 Same as JET 50C

• Higher Twall ? lower D/C ratio on PFCs
• No drop of co-deposition in remote area

Masaki IAEA 2006, Hayashi PSI 2006, Sugiyama PSI
2006, Hirohata PSI 2006
22
AUG From all C to all W ? D inventory
2002 2003 4940 s
C-dominated campaign 2002/2003 Normalised to 3000
s - D on divertor tiles 0.9 1.3 g - D below
roof baffle 0.4 g ? 1.3-1.7g
C
W
C
6A
C
4
23
AUG all W Analysed tile D inventory
M Mayer I-13
2007 2620 s
Full W camp 2007 No Boronisation Norm. to in
3000 s - D on divertor tiles 0.15 0.23 g - D
below roof baffle 0.03 g ? 0.18-0.26g (drop
7-10)
W
W
W
? D retention in inner divertor still dominated
by C-codeposition drop 10-15

• D retention in outer divertor dominated by
  trapping in W Drop 5-10

6A
W
From C-dominated to all W
D inventory reduced 5-10 ? Retentionlt1
4
24
Gas balance Post mortem
When analysing the same plasma? Gas Balance
and Post mortem lead to similar evaluation
? Extrapolation to ITER
25
Fuel inventory estimates for ITER
Evaluations based on ion CX fluxes to the wall
and resulting - Implantation - PFC erosion and
associated co-deposition
J Roth R-1
- More RD required for evaluation of
n-effects - Need to improve modeling in
retention by co-deposition/trapping (fluence)
All C reaches tritium limit (700g) in less than
30-40 discharges All-W reduces tritium problem,
but n-effects need to be considered
26
Summary

• Gas balance
• Long term retention for C machine depends on
  Plasma scenario
• As far as C source exists ? co-deposition
  dominates, increases with C production recycling,
  ELMs (AUG, JET, JT-60U, TFTR, TS) and a to pulse
  duration.
• Retention by implantation saturates for
  overheated PFCs (JT-60U)
• Long term recovery (outgassing and disruption)
  is weak (TFTR, JET and TS)
• Mo exhibits retention a to plasma duration, but
  D recovery from disruption.
• Full W shows a significant drop of the retention
  1.

• Post Mortem analysis
• Confirms long term retention in PFCs is low but
  high in remote areas
• In carbon? AUG-JET (3-4), JT-60U8, TEXTOR
  TS 10-15
• Significant drop of the retention below 1 in
  AUG with full W configuration

Complementary and reliable methods ? retention in
full metal wall (ILW)
Extrapolation to ITER - A full C machine would
reach the limit in few discharges of 400s - A
high Z device would limit the co-deposition and
strongly reduce retention
27

• END

28
Short term plasma scenario
Ip 2.0 MA, Bj 2.4 T
69260
PTOT (MW) NBIICRH13MW
DWELM 100 kJ 60 Hz
Da (in)
Da (out)

Time (s)
T. Loarer et al., EPS 2007
- Short term retention limited to fast
reservoir and recovered in between pulses
(outgasing) - Long term retention Co-deposition
and implantation Slow process compared to short
term over 5-10 sec.
29
Retention in gaps
TFTR
T Tanabe et al. Fus Sci and Tech 2005

• 2 decay lengths in TFTR
• l12mm ? hydrogen rich deposition ((DT)/C0.2
• prompt redeposition of C2Dx (1,3,5) ... with high
  sticking coefficient
• ? l26-12mm influenced by the gap width (up to
  30mm with large gaps) ? migration of neutral
  hydrocarbon with low sticking coefficient

30
Retention in gaps
Distribution of D on side surface in the gaps
between the Mk-I CFC tiles.
Be limiter tiles of Mk-I divertor
M Rubel et al, JNM 2007
31
Gas Balance Accuracy
DT experiments in JETover the first week the
most difficult part to evaluate is the
outgassing. In actual tokamak discharges.

• 65 of retention during the pulses (10sec)
• 40 of retention over the campaign
• 17 after intensive cleaning (6 months)

Loarer JNM 2005
T particle balances from Gas balance analysis ?
Good agreement between gas balance and cryopump
reg. (green points)
32
JET DTE Campaign
JET DTE Campaign 1997-1998 T Vessel inventory
End of DTE1 campaign Injected-Exhausted (35g 23.5g) 11.5 g
Clean up phase (D, H, He, Disrp, GDC) ?Remove 5.3g 6.2 g
Venting ? Remove 2.5g 3.7 g
Flakes ?Remove 0.5g 3.2 g
Inner and Outer wall ? Remove 0.1g 3.1 g
480 Tiles of Divertor ? Remove 0.1g 3.0 g still remaining (flakes)
After the campaign and intense cleaning campaign
(pulses, conditioning, venting, the T inventory
in the PFCs is negligible (0.2 g). Trapped
tritium in flakes.
Bekris et al. JNM 2005
33
JETand TFTR DTE Campaign
Drop by a factor of 2 in 11/2 year
Skinner et al. EPS 2001
34
High Z material Alcator C-Mod

• OSP (net erosion) low D retention ? consistent
  with results from Mo erosion and B deposition.
• ISP Boron coverage is very small whilst D content
  is also very small
• - Boron surface layers from boronisation was
  found at all locations except near the OSP

Dominant impurity Boron and some amount of Mo
Wampler, IEEE 1999, PSI 1998
35
Gas balance AUG - All W
36
DTE TFTR and JET
Effect of oxygen on samples removed from device
C Skinner et al. 2002
37
DTE TFTR
C Skinner et al. 2002
38
Location of TFTR tritium inventory
C Skinner et al. 2002
39
Co-deposition and Carbon production
Non-linear carbon deposition on ELM energy ?
Thermal decomposition of surface layers

• Increase of carbon source (ELMs, recycling
  flux)
• ? enhanced retention by co-deposition

40
High Z exp in Alcator C-Mod
B Liptschultz I-14

• Although cleaning process allows to remove the
  boron layers, there is still a non negligible
  amount of boron at the surface
• - If experiments are carried out in AUG with full
  W and boronisation, this could clarify the
  possible effect of boron in the retention
  behaviour observed in Alcator C-Mod

41
Alcator C-Mod and ASDEX Upgrade
Roth R-1
Alcator C-mod, Mo ASDEX Upgrade, W
Gas balance Variation between net outgassing and up to 50 retention, linear with number of discharges 1 retention,
Post-mortem 1 lt 0.2 retention (to be confirmed)
Modelling 1-2 retention, linear with number of discharges


One possible explanation are plasma impurities,
largely B (1 ), that are present in the C-Mod
plasma and in the first 500 nm Mo tiles.
Overall, in both high-Z devices the campaign
integrated retention on D is small, much smaller
than for carbon or boron co-deposition.