Modeling hydrocarbon generation / transport In fusion experiments John Hogan

o

Modeling hydrocarbon generation / transportIn fusion experiments

John HoganFusion Energy Division

Oak Ridge National Laboratory

First Meeting Co-ordinated Research Program

” Atomic and Molecular Data for Plasma Modelling ”IAEA

Vienna International CentreSeptember 26-28, 2005

o

ITER tritium retention issues G. Federici, ITER GWS PSIF Workshop*

(to be published, Physica Scripta)

ITER predictions still uncertain due to – chemical erosion yields at high temperature and fluxes – effects of type I ELMs (ablation)– effects of gaps– effects of mixed materials– lack of code validation in detached plasma.

• T issues will be heavily scrutinised by licensing authorities.

• Scale-up of removal rate required is 104.• Potential options for T removal techniques for ITER.

1) Remove whole co-deposit by: • oxidation (maybe aided by RF)• ablation with pulsed energy (laser or flashlamp).

2) Release T by breaking C:T chemical bond:• Isotope exchange • Heating to high temperatures e.g. by laser, or ...

* “New directions for computer simulations and experiments in plasma–surface interactions for fusion”:

Report on the Workshop (Oak Ridge National Laboratory, 21–23 March 2005)

J.T. Hogan, P.S. Krstic and F.W. Meyer Nucl Fus 49 (2005) 1202

o

Fusion applications of hydrocarbon rate data

Erosion / re-deposition / tritium retention

(DIII-D, Tore Supra, JET examples)

- long discharges -> hydrocarbon films

- chemical erosion and T inventory

Use of presently available data

PISCES linear reflex arc (Erhardt-Langer)

throat of Tore Supra neutralizer

(compare E-L and Janev - Reiter)

DIII-D gaps

ELMs

o

a. 2/3D Kinetic codes

Code WBC ERO BBQ DIVIMP DORIS MCI EDDY

Geometry 3D 2/3D 3D 2D 3D 2D 3D

Model TR TR TR TR TR TR TR

Dynamics GO GO GC GO GC GO GC

b. 2D fluid codes

Code SOLPS EDGE2D UEDGE

Geometry 2D 2D 2D

Model SC SC SC

Dynamics FL FL FL

Codes in use for erosion/deposition/retention (D Coster, J Hogan PSIF Workshop Summary Nucl Fus 2005)

TR - trace impurity in fixed backgroundSC - self-consistent background and impurity solution

GO - gyro orbit following (classical diffusion)GC - guiding center fluid (anomalous transport)FL - fluid + kinetic corrections

o

Collision model like LIM, WBC, ERO codes but, includes 3-D, toroidal magnetic geometry

Background parameters assumed (ne, ,λn, Te, λT)

[ local D+ , flux amplification sheath ( ) , electrostatic es field E SOL

]

MZ dVZ/ = -dt Ffriction -Fes

+ // random and⊥ diffusion

Ffriction = -MZ (VSOL-VZ) / τ s ; Fes = Ze ESOL

F// = // , Random diffusion D//=(8EZ/3πMi) τ//F⊥ = Random⊥ anomalous ( diffusion D⊥ )

dEZ/ = (dt kTi -EZ)/τT +W friction+Wes

Molecular processes: - ,Erhardt Langer- - Allman Ruzic hydrocarbon break up rates

Atomic processes ( , ,ionization recombinationD0 ) , , , , , charge exchange for D C He Ne N Ar

- Birth gyro collisions with surface

BBQ: 3- , D velocity space

(3 )SOL Monte Carlo impurity transport D MIST ( . - )developed by R Hulse PPPL 1- ( ) , -D radial t multi species impurity transport code

:Atomic physics - ( )ADPAK original - ( . )STRAHL K Behringer - / / ( )JHU LLNL Frascati HULLAC ( . )collaboration M Mattioli

NCLASS ( . - )developed by W Houlberg ORNL - Full expressions for neo classical fluxes with- multi species impurities

(1- , )Core impurity transport D time

Modifications ( ) :'physics package' include - - ( self consistent heat flux from ne, Te, Φ) ( , )includes SEE thermionic - local impurity redeposition heat source

-2000 CASTEM ( - ).developed by CEA DEMT

(3- , )Wall thermal analysis D time

C O D E S

Wall equilibration (1-D, time)

WDIFFUSE (ORNL) 1-D (depth into a-C:H layer) , time multi-species wall deposition/retention/saturation code (combined with neutral CX calculation: Eirene or similar)




]

















C O D E S






]

















C O D E S






]

















C O D E S



Collision model

Background parameters assumed (ne, ln, Te, lT)

[ local D+ flux amplification, sheath electrostatic (es) field, ESOL

]

MZ dVZ/dt = -Ffriction -Fes

+ random // and ^ diffusion

Ffriction

= -MZ

(VSOL

-VZ

) / ts ; Fes = Ze ESOL

F// = Random // diffusion, D//=(8EZ/3pMi) t//F

^ = Random ^ anomalous diffusion ( D

^ )

dEZ/dt = (kTi -EZ)/tT +Wfriction +WesBirth gyro-collisions with surface

BBQScrape-off layer Monte Carlo impuritytransport (3-D, velocity space)

Hydrocarbon molecular processes

Erhardt-Langer CH4 database

Alman-Ruzic

Janev-Reiter

CASTEM / TOKAFLU (R Mitteau et al J Nucl Mater 1999)

Plasma-facing component thermal analysis (3-D FEM, time)

IMPFLUPhysics package: - self-consistent heat flux (from ne, Te, Φ) ( , )includes SEE thermionic - local impurity redeposition heat source - test sputtering models - effects of deposited films

C O D E S

B2-EireneAxisymmetric divertor simulation

Time / space (2D) -dependent

Divertor plasma fluid (B2), neutrals (Eirene)

Developed by B. Braams (Emory Univ.) R. Schneider, D. Coster (MP - IPP, Garching, Greifswald) D. Reiter (FZ-Juelich)

o

CD / CH Molecular Band is analyzed todetermine the H/D concentration in the divertor(G. Duxbury - Univ Strathclyde)

0

0.1

0.2

0.3

0.4

0.5

0.6

50 52 54 56 58 60

Divertor Spectroscopy Sub-Divertor - PenningCH-CD Spectroscopy

Time (s)

48280

Hydrogen Concentration

0.0

0.2

0.4

0.6

0.8

1.0

40 60 80 100 120 140 160 180

Divertor SpectroscopySub-Divertor-PenningCH-CD Spectroscopy

Hydrogen Concentration

Accumulated Time (s)

4829248284

4828348282

4828148280

275

H wall loading first 5 shotsOf D-->H changeover

H concentration as deduced from the CH-CD Molecular Band

Relation between chemical erosion processes and H/D/T retentionJET deuterium - hydrogen change-over experiment

D Hillis, J Hogan et al J Nucl Mater 2001

o

Interior view of Tore Supra

Tore Supra

Full toroidal limiter CIEL

poloidal direction

toroidal direction

machine axis

o

Coupled core/ SOL/PFC code system

CASTEM-TOKAFLU-IMPFLU BBQ

ITC-SANCO-MIST

Self- sputtering iteration

Deuterium sputtering stageBBQ

MIST/ITC

ne(ρ)Te(ρ)λΤe

ΓD+

Superstructureleading edgeneutralizer

physicalchemicalRES

ΓCin(Zi)

=> ΓCout(Zi) Core C efflux: D+ - sputtered

BBQ

MIST/ITC => ΓCout(Zi) Core C efflux: self - sputtered C

=> Core C influx Sum over processes and geometries

o

1.6

Tmax = 400KTmax = 825K

1.3 1016 pt/cm2/s6.0 1016 pt/cm2/sRoth et alNF supplement, (1991)

J. Roth et al11th EFPWHeraklion

Max. C flux

1.2 1017 pt/cm2/s2.7 1016 pt/cm2/sChemical sputtering:C erosion fluxes for two models

0.8

01 2 3 4 5 6 Incident Z

Self sputtering yieldSelf-sputtering sharp increase in yield with Zincident

CASTEM-2000 sputtering

C emissionflux distribution

Model

o

0.9 0.95 1.0 1.03 1.06 radius (ρ/a)

LCFS

0

1.2

0.6

V VI

VII

IV

III

Mid SOL heating< NC >

10 18 m-3

< N C > BBQ carbon density

averaged over,

o

8.8

0

4.4

0 0.5 1.0 ρ ( r/a)

2.3

0

1.15

0

1.5

0.75

1 9 iteration

< NC > (1016 m-3)NC (1016 m-3)

1.05

0

0.525

0 0.5 1.0 ρ ( r/a)

.NC (1018 m-3)

System evolution forlocalized heating ininner, mid- and far-SOLfor:- TRIM self-sputter (TOP)- enhanced self-sputter (BOTTOM)

o

Tore Supra heat, particle flux deposition is strongly influenced by magnetic field ripple (~7%)

R Mitteau et al J Nucl Mater 2001.

o

2.25

2.50

R(m)

0.0 10.0 20.0

Φ(°)

1.00000e+24 2.00000e+24initflux_D1_M5_34909_hdf 1.000e+24 2.000e+24 3.000e+24

initflux_D1_M5_34930_hdf

0.0 10.0 20.0

2.25

2.50

Φ(°)

R(m)

3493034909

Model for C emission from CIEL surfacenitial C flux: TOKAFLU/IMPFLU -> BBQ (physical sputtering)

0.0 10.0 20.0Φ(°)

2.25

2.50

R(m)

6.00e- 16 8.00e- 16 1.00e- 15c2rad_D1_M5_34909_hdf

2.25

2.50

R(m)

0.0 10.0 20.0Φ(°)

4.0e- 16 6.0e- 16 8.0e- 16 1.0e- 15c2rad_D1_M5_34930_hdf

Prad(R,φ) BBQ

34909 34930

o

34909 34930

Modelled C emission(TOKAFLU / BBQ)

Physical sputtering source

Te,b=35 eV, ne,b=0.67 1019 m-3

Pinj-Prad=0.7 MW

Measured CII radiation(E Dufour, C Lowry et al, EPS 2005)

Model validation requires attention to impurity generationfrom non-ideal features, e.g., intra-tile gaps

Tore Supra example

(°)0 20

Toroidal angle

R(m)

2.25

2.5

Majorradius

φ(°)0 20

Toroidal angle

o

Sources of BBQ CD4 Rate Data [W. Langer, A. Erhardt, PPPL Technical Report]

# Reaction Product Comment1 e- + CD4 CD4

+ + 2 e- M2 CD3

+ + D + 2 e- M3 CD3 + D + e- M4 e- + CD4

+ CD3 + D+ + e- NM: = 1/4 R1

5 CD3+ + D + e- NM: = 3/4 R1

6 CD3 + D Tot. R6+R7 M for E < 1 eV,

Ex (E > 1 eV) = 1/4 Rexp

7 CD2 + 2 D As for R6,R7 = 3/4 Rexp

8 e- + CD3 CD3+ + 2 e- CD3 M

Ex (E<15 eV)

9 CD2+ + D + 2 e- From CD3 (M)

10 CD2 + D + e- NM, = 3/4 R3

11 e- + CD3+ CD2 + D+ + e- NM, = 1/3 R10

M = Measured, NM = Not Measured, Ex = Extrapolated, A = Assumed

o

Sources of the BBQ CD4 Rate Data [W. Langer, A. Erhardt]

12 CD2+ + D + e- NM, = 2/3 R10

13 CD2 + D Total M for E < 1 eV, Ex (E>1 eV)neglect other channels

14 e- + CD2 CD2+ + 2 e-From CD2 M(E<200eV)

and CD4 for E> 200eV)15 CD+ + D + 2 e- M (CD2)16 CD + D + e- NM, = 1/2 R3

17 e- + CD2+ CD + D+ + e NM, = 1/2 R16

18 CD+ + D + e- NM = 1/2 R16

19 CD + D Total M (E < 1 eV)Ex (E>1eV) neglect other channels

20 e- + CD CD+ + 2 e- NM, adopted CD4 data21 C+ + D + 2 e- NM total R21+R22

A=1/2 R9; R21=1/2 total22 C + D+ + 2 e- NM, total R21+R22

A= 1/2 R9; R22=1/2 total23 C + D + e- NM = 1/4 R3

24 e- + CD+ C + D+ + e- NM = 1/2 R23

25 C+ + D + e- NM = 1/2 R23

M = Measured, NM = Not Measured, Ex = Extrapolated, A = Assumed

o

PISCES - A (UCLA) experiments: A. Pospieszczyk, FZ-Juelich

Reflex arc linear mirror plasma

Spectroscopic arrangementOptical multi channel analyzer

o

Code - experiment comparisonBBQ - Monte Carlo multi-species simulation ( Erhardt-Langer database)OMA ( A. Pospieszczyk, FZ-Juelich)

Nozzlelocation

Downstream distance(cm)

0 2 4 6

Rel. density

1.0

0.5

0

PISCES-A 13417

BBQ

CH2 axial density CH axial density

Rel. density

1.0

0.5

0

Nozzlelocation

Downstream distance(cm)

0 2 4 6

PISCES-A 13417

BBQ

Optical Multi-channel Analyzer(A. Pospieszczyk FZ-Juelich)

o

Pumping plenumpartial pressure(CnDm)

Langmuir probe D+flux measurement

Last closedmagnetic surface

plate separatingion / electron sides Langmuir probe (long)

electron side

Langmuir probe (short)electron side

Optical Multichannel AnalyzerCD, C2 band intensity measurement

Limiter NeutralizerplatesD+ ion flux

CnHm generationPlasma core

top view

scrape-off layer

Schematic diagram of experiment onTore Supra Midplane Limiter (Phase II)

E Gauthier, A Cambe J Hogan et alJ Nucl Mater 2003

o

Carbon production= f(D)Carbon from C3DYCarbon from C2DXCarbon from CD4

o

) Roth, Garcia-Rosales Mech et al Roth PSI 13, mono-energetic Roth PSI 13 Maxwell averaged Experiment

0

0.002

0.004

0.006

0.008

0.010

0 5 10 15

BBQ comparison

Average D+ flux density( 1022 pt / m2 / s )

OH discharges(Tsurf ~ 500K)

Model comparison using partial pressure- sensitivity to sputtering model

CD emission

CD4 partial pressure

510 670 830 Tsurf(K)

LHCD: High Tsurf

PCD4∝ PD20.70 (N.Hosogane)PCD4∝D0.73 JT-60UIncreased productionof CD4 with flux

o

Deuteron impact charge exchange data [Alman, Ruzic]D+ + CnDm --> D + CnD4+ Reaction Product Rate Known

(10-9 cm3/s) (total)CD4

D+ + CD4 D + CD4+ 1.880 4.150

D2 + CD3+ 1.880

D+ + CD3 D + CD3+ 1.800

D2 + CD2+ 1.800

D+ + CD2 D + CD2+ 1.700

D2 + CD+ 1.700D+ + CD D + CD+ 3.236C2D2

D+ + C2D2 D + C2D2+ 2.250 6.300

D2 + C2D+ 2.250 D+ + C2D D + C2D+ 4.358

o

C2D2 C2D2+ C2D+ C2+

Heavy hydrocarbon production: dominant species frombreak-up of C2D2.BBQ calculation using Alman-Ruzic database

o

Janev-Reiter database rates have been implemented in BBQ. A comparison, with the same background plasma conditions, for the Tore Supra pump limiter case, shows significant differences in comparison with the Erhardt-Langer rates

CH+ CH+ CH+

CH4 CH4 CH4

ne,LP= 2 1018 m-3

Te,LP= 20 eV Te,LP= 10 eV Te,LP= 5 eV

Janev-Reiter profiles

o

o

TeLP = 20 eV

ne,LP = 2 1018 m-3

E - L

J - R

CH+ density

Axial distance (m)0.04 0.08 0.12 0.16 0.20 0.24

TeLP = 10 eV

TeLP = 5 eV

CH+ density

Axial distance (m)0.04 0.08 0.12 0.16 0.20 0.24

E - L

J - R

Axial distance (m)0.04 0.08 0.12 0.16 0.20 0.24

CH+ density

E - L

J - R

o

CH+ CH+ CH+

CH4 CH4 CH4

ne,LP= 6 1018 m-3

Te,LP= 20 eV Te,LP= 10 eV Te,LP= 5 eV

Janev-Reiter profiles

o

TeLP = 20 eV

ne,LP = 6 1018 m-3

E - L

J - R

CH+ density

Axial distance (m)0.04 0.08 0.12 0.16 0.20 0.24

TeLP = 10 eV

TeLP = 5 eV

CH+ density

Axial distance (m)0.04 0.08 0.12 0.16 0.20 0.24

E - L

J - R

Axial distance (m)0.04 0.08 0.12 0.16 0.20 0.24

CH+ density

E - L

J - R

o

Given D+ flux (or Cn+for self-sputter) to surface and surfacetemperature, calculate impuritytype, rate and velocitydistribution .

Mechanisms: - physical sputtering, - chemical sputtering (thermal, athermal) - RES

Impurity generationmodels

0.09

0.045

0.0

250 875 1500

T surf (K)

T e (eV)

10

30

5070 90

90

70

50

30

T e (eV)

Chemical

RES

Yield

[C / ion ]

Surface-temperature dependence of chemical and RES yields forTe=10 -> 90 eV.

- 3-D, time-dependent finite-elements thermal analysis code

- developed by CEA-DEMT

Modifications (CSPUTTER) include:

- self-consistent heat flux (includes SEE, thermionic emission)

- local impurity redeposition generation due to T surf -dependent mechanisms as well as physical sputtering

- ablation cooling (Vieder model)

RDP-2000

RDP-1999

0.4 0.6 0.8 1.0 1.2 1.4 Time (s)

3

8

0

0

Carbon concentration (%)

Edge

Core

Filterscope(schematic)

0001

02

03

0411

13

12

DIII-D intrinsic impurities before/after new tile installation

Some evidence of T-dependent processes

o

0

200

400

600

800

1000

1200

1400

0 1 2 3 4 5 Time (s)

NBI

Maximum Tsurfon heated surface

2.5 mm

Ychem @ t ~ 4s

when Tmax = 1010K

Ychemmin = 1. 10-2

Ychemmax = 0.14

0 1 2 3 4 5 Time (s)

0.4

0.8

1.2

1.6

0 NBI

0.6

1.0

1.6

2.0

00 1 2 3 4 5 Time (s)

NBI

RES

Chemical

NBI

CASTEM-2000 simulation of time-dependent carbon generationfrom simulated DIII-D localized source

o

C2D2

CD+C2D+C2D2+

CIC2D

Axial (30 cm)

1:1 aspect ratio shown - actual 30:1

TC2

SOL

Axial distance (along B)(cm)

0 30

0 TC 1.0(eV)

2

Axial (along B) 15:1 aspect ratio shown; actual 30:1

TC2

SurfaceMolecular temperature in the region of max. molecular

density = 0.1 - 0.3 eV

Surface SOL15:1 aspect ratio shown; actual 30:1

SOL

1:1 aspect ratio shown - actual 30:1

Near-surfacemolecular density

nC D2

+2

BBQ calculations find near-surfacemolecular rotational temperaturesto be in this range, when respectivelocalization of density and temperatureis considered.

Issues: Spectroscopy is averaged over several ELMs To be quantitative: what are heavy hydrocarbon production rates?

BBQ comparison with spectroscopy is encouraging C2D2 calculation - Allman-Ruzic-Brooks rates - vicinity of the tile gap area

Band spectroscopic modeling givesmolecular energies ~ 0.1 - 0.3 eV (R. Isler)

Axial (along B)

Surface

o

edge / pedestal only,low-field side only

Semi-empirical model for ELM transport enhancement

Green: ELM-affected region in the model

Red: C neutrals and ionsYellow: D neutrals ion ions

1

2

34

ELM event

time

radius

1

2

3

separatrix

Transport time dependence(schematic)1. Pre-ELM (barrier)2. Strong enhancement (100 μ ) sec ELM3. , 2 - Loss of barrier x pre ELM value4. - reducing to pre ELM value as barrier -is re established

- Intra ELM transportradial dependence(schematic)1. barrier2. enhancement toward SOL3. SOL radial transport

o

C6+(ρ) vs time, solpsoutboard mid-planeSeparatrix

= 0.81Normalized Radius0.900.940.960.99 = 119434Shot

~ separatrix

0 0.005 0.01 ( )Time Relative to ELM s

0.00

0.02

0.04

0.06

0.08

0.10

0.12

Impurity Density

C6+ density CER

C6+ density solps

HFS separatrix

core

LFS separatrix

0 0.005 ( )Time Relative to ELM s

o

CIII 4650.1 evolutionsolps 'standard' model

Roth et al 'annealing' model

Inner leg pre-ELM detached during attached re-detaching after recovery of detachment

DIII-D experimental:fast (CID) camera

CIII evolution:M Groth et al,J Nucl Mater 2003

Modeling

solps 5.0 / Eirene99

IPP-Garching/Greifswald, FZ-Juelich

o

QuickTime™ and aYUV420 codec decompressor

are needed to see this picture.

Solps simulation of

CIII emission

seen by 240par

(lower divertor) camera

W Meyer,

M Fenstermacher,

M Groth

LLNL

o

0 1 2 time (msec)

Prad(MW)solps

Prad(MW)BOL4/TXPN

Dα( )rel

58139Pulse

61.47 61.52 61.57 ( )time sec

12.2

0

6.1

Prad( )MWELM heat flux mitigation byInjection of extrinsic impurities

o 10 100 1000

0.1

1

10

chemical sputtering yield (per ion)

energy (eV)

N+2

Ar+

Ne+

He+

H+2 = 2 H+

Flux ratio H/ion = 400

Chemical sputtering of carbon materials due to combined bombardment by ions and atomic hydrogen W Jacob, C. Hopf, M. Schlüter, T. Schwarz-SelingerMax-Planck-Institut für Plasmaphysik Garching

o

CONCLUSIONS

Intrinsic (carbon) impurity sources play a key role in ITER, both as regards erosion and for tritium retention

The ITER problem (addressed also by JET) requires a decision about the first wall material - carbon or an alternative Development of validated models for C generation, deposition and retention requires experimental comparison: this typically introduces multiple uncertainties; e.g., sputteringyield models vs hydrocarbon break-up rates

ITER relevant experimental scenarios involve fast timescale ELM events, for whichspectroscopy is an key tool

The importance of hydrocarbon generation processes has been seen in many experiments, and thus a quantitative, evaluated, integrated model for break-up processes in the plasma is needed.

Documents

Modeling hydrocarbon generation / transport In fusion experiments John Hogan