Long Term Reduction of Divertor Carbon Sources in DIII-D

Long Term Reduction of Divertor Carbon Sources in DIII-D D.G Whyte, UCSD R. Doerner, W.P. West, R.L. Lee, N.H. Brooks, R.D. Isler, M.R. Wade, G.D. Porter APS-DPP, Seattle, Nov. 1999 NATIONAL FUSION FACILITY #1

Abstract Submitted for the DPP99 Meeting of The American Physical Society Sorting Category: 5.1.1.2(Experimental) Long-term Reduction of Divertor Carbon Sources in DIII-D 1 D.G. WHYTE, R. DOERNER, University of California, San Diego, W.P. WEST, R.L. LEE, N.H. BROOKS, General Atomics, R.C. ISLER, M.R. WADE, Oak Ridge National Laboratory, G.D. PORTER, Lawrence Livermore National Laboratory The long-term evolution of carbon sputtering from the tokamaks graphite-covered lower divertor is studied using in-situ spectroscopy and ex-situ laboratory erosion yield measurements. The total effective yield and, hence, the impurity source amplitude has gradually decreased a factor > 4 since installation of the tiles in 1992. This reduction is likely caused by the 2-fold reduction in the chemical erosion yield of the graphite, the effect of > 3 applied boronizations. Despite this great success in wall conditioning, there has been no concomitant decrease in the core plasma carbon contamination during the same period. Two explanations are being explored: (1) the nature of parallel impurity transport decouples the wall source and content, and (2) the first wall is the principal source of carbon in the core. 1 Supported by U.S. DOE Grant DE-FG3-95ER-54294 and Contracts DE-AC3-99ER54463, DE-AC5-96OR22464, and W-745-ENG-48. X Prefer Oral Session Prefer Poster Session D.G. Whyte whyte@gav.gat.com University of California, San Diego Special instructions: DIII-D Poster Session 2, immediately following RJ Colchin Date printed: July 16, 1999 Electronic form version 1.4

Outline Background & Motivation History of chemical and total erosion yields in DIII-D lower divertor (1992- present). Effect of lower divertor carbon erosion yield on core plasma carbon contamination. Laboratory studies of DIII-D graphite tiles Summary & discussion. Reduction of divertor carbon sources, APS, Nov. 1999, D. Whyte #2

Background & Motivation Previous results suggest chemical erosion of carbon plays an important role in DIII-D Helium plasmas: reduced carbon radiation in core and divertor. UEDGE modeling: a distributed chemical carbon source (using University of Toronto Y chem ) provided radiation to induce detachment and match C spectroscopy. Previous results suggest that chemical erosion of carbon plays a minor role in DIII-D CD-band emission not observed in divertor (Y chem < 1-3 ). Divertor detachment, which eliminates physical sputtering, showed only carbon deposition in divertor (DiMES, PSI 98)..total yield reduced by x1. Current carbon-based tokamaks show divertor erosion and tritium retention will seriously limit operations of next-step DT devices. Reduction of divertor carbon sources, APS, Nov. 1999, D. Whyte #3

5 1 2 1 23 25 T ( C) Divertor detachment eliminates net carbon erosion in the entire lower divertor!...yet carbon density in core plasma remains constant? T e (ev) n e (m -3 ) i (m -2 s -1 ) q (MW m -2 ) 2 1 Attached vs. Detached Carbon Erosion Rate (cm / burn-year) 3 3-4 Inner leg Outer leg erosion deposition private flux exposure 4 8 R strikepoint (cm) -1 1 2 R OSP (cm) Reduction of divertor carbon sources, APS, Nov. 1999, D. Whyte #4 D 2 Torr L /s I p MA P aux MW W th MW n e /n C 1 19 m -3 f C % 1.4 7.2 1.2 4 /.16 4 ~1 1.35 6.9.9 1 /.15 1.5

Brightness ( x 1 17 ph/s/m 2 /sr/å) Brightness ( x 1 17 ph/s/m 2 /sr/å) 4 2 2 1 Atomic and molecular spectroscopy suggest chemical erosion plays a minor role in DIII-D divertor. C I 8335 Å detached plasma C II detached: OSP ( ) attached: outer leg attached: top of baffle CD band chord viewing DiMES 833 Wavelength (Å) 834 BD band chord viewing inner corner 426 43 434 Wavelength (Å) D Ar II Atomic Carbon SXB ~ 3 Molecular Carbon (C x H y ) DXB ~ 1 Assumption Y chem Ions only 1-3 Ions=neutrals 1-4 Particle Flux (m-2 s-1) Detached DiMES exp t 1 22 1 21 1 2 1 19 Inner Strikepoint Atomic carbon 1. 1.4 1.8 R chord at divertor floor (m) 1 2 3 4 5 Outer Strikepoint 6 7 OSP exposure DiMES sample Carbon in molecules PF exposure Reduction of divertor carbon sources, APS, Nov. 1999, D. Whyte #5

Based on carbon-based tokamaks standard operation, erosion will severely limit operation of D-T burning device. Pervasive in/out asymmetry. Outer divertor net erosion rate: > 1 cm / burn-year! T trapping in erosion caused layers: ~ 1 kg / m 2 / burn-year! Net change in tile thickness ( m) DAC 1-1 8 4 ISP DIII-D net redeposition net erosion OSP 1. 1.4 1.8 Major Radius (m) <T surf > (K) DAC 4 35 3 4 2 (f) (g) ASDEX Upgrade Inner Wall Heat shield Divertor plates ISP ICRH- Antenna OSP.2.4.6.8 Distance along divertor plates (m) DAC 1 2 JET (MkI) Tile # 19 1 17 3 5 7 9 11 15 13 R=2.3 m 334 ISP Be target OSP 5 1 15 Tile number R=3.1 m Plasmafacing Only Entire top surface Carbon target Poloidal erosion / redeposition patterns from 3 tokamaks: > 1 year of operations. DAC: trapped D concentration Reduction of divertor carbon sources, APS, Nov. 1999, D. Whyte #6 D.G. Whyte, et al. Nucl Fusion (1999).

History of carbon chemical erosion yields in the DIII-D lower divertor (1992- present). #7

Relative spectroscopy can be used to measure carbon erosion yields in situ Y Technique relies on S/XB (loss events per photon) and relative intensities: Fairly insensitive to T e and n e. Insensitive to photometric calibration. Chemical erosion yield Y chem D D B = B Total erosion yield total ΓC x = Γ ( Γ = C x D y Γ y +Γ D C CD D/ XB S / XB CD D γ D γ ) B B D CII S S / / XB XB CII D α, γ α, γ Intensity (a.u.) 1 2 3 4 5 6 7 OSP exposure DiMES sample C CD band (Å) PF exposure D 426 43 434 Reduction of divertor carbon sources, APS, Nov. 1999, D. Whyte #8

S/XB and D/XB values are fairly constant for attached plasma (T e > 1 ev) conditions. 1 1 n e =3x1 13 n e =5x1 13 n e =1x1 14 ADAS Ionizations per photon for D Experiment (Attached plasmas) 1 1 1 1 T e (ev) Derived values for D γ S/XB Atomic rate calculations. Experimental comparison to probes & DTS. Agreement / scatter to within a factor of 2. Assumes recycling ~ influx. D/XB is derived from controlled CD 4 injections: JET: D/XB ~4-5 for attached plasmas. PISCES: D/XB ~5-1 for T e > 1 ev. Reduction of divertor carbon sources, APS, Nov. 1999, D. Whyte #9

DIII-D tiles have accumulated ~2.5 µm of boronization layers & ~1 5 seconds (25, shots) of plasma exposure without significant alterations to lower divertor tiles. Accumulated boron deposition ( µm ) 3 2 1 6% 1% graphite coverage Existing tiles removed, cleaned & re-installed Exposure time (s) 1.2E+5 8.E+4 4.E+4.E+ 24 hours Jan-91 Jan-93 Jan-95 Jan-97 Jan-99 7 8 9 1 Boronization shot number Single boronization = ~1 nm uniform B layer deposited Reduction of divertor carbon sources, APS, Nov. 1999, D. Whyte #1

Graphite tiles are visibly altered by tokamak exposure. replacement tiles Plasma-exposed tile removed from DIII-D for PISCES exposure DiMES DIII-D Lower Divertor near DiMES (August 1998) Scratches from plasma arc Reduction of divertor carbon sources, APS, Nov. 1999, D. Whyte #11

In 1992-1993, a single boronization would cause an immediate, but temporary, reduction in Y chem and Y total. Relative brightness to D γ 1.2 1.8.6.4.2 BD CII CD #7677 #7667 Direct comparison before and after boronization. Y chem reduced ~x4 Y total reduced ~x2 BD band reduced, but not eliminated by plasma exposure. 1 1 1 1 Shots since boronization Averaged over lower divertor chords Substantial removal of 1 nm boron layer consistent with measured OSP net erosion rates ~ 2 nm/s Reduction of divertor carbon sources, APS, Nov. 1999, D. Whyte #12

Comparison of nearly identical plasmas show that Y chem has decreased significantly in DIII-D since 1993. CD / Dγ 1 1.1.1.1 ISP 1 12 14 16 18 2 R (cm) OSP Reduction of divertor carbon sources, APS, Nov. 1999, D. Whyte #13 Mar-93 Jan-99 Spectra normalized to D γ intensity & for different detector responses (intensified vs. non-intensified) #76952 #9842 I p (MA) 1.5 1.4 n e (cm -3 ) 7.6x1 13 6.8x1 13 P inj (MW) 6.9 6.6 W dia (MJ) 1.2 1.1 Intensity (a.u.) 1 #76952 #9842 CD band BD band 43 432 434 436 (Å) D

A database of lower divertor CD band spectroscopy (1992-1999) clearly shows a large reduction in Y chem. All chords, all plasma types 3 2.5 2 1.5 1.5 CD / D γ Includes all discharges where spectrometer was tuned to 43 nm: LSN & DN Ohmic, L, ELMy H Attached & detached. 7 8 9 1 shot number Reduction of divertor carbon sources, APS, Nov. 1999, D. Whyte #14

Average lower divertor Y chem for attached ELMy H-mode plasmas has gradually decreased a factor of 2 from ~2% to ~.1% since 1992..1.1 >1 shots post-boronization Yield measurement fairly reliable since for attached divertor T e > 1 ev: D/XB CD ~ 5 S/XB ~ 2 Dγ.1 Y chem Y chem ΓC x = Γ D D y B = B CD D/ XB S / XB D γ D CD γ.1 75 8 85 9 95 1 Shot Scatter in yield probably indicative of uncertainty in T plate, Γ & S/XB. Reduction of divertor carbon sources, APS, Nov. 1999, D. Whyte #15

Observed yield reductions are not due to any systematic drift of plasma parameters used in spectroscopy database. Attached ELMy H-mode >1 shots post-boronization 1 14 n e (cm -3 ) P inj (W) 1 7 1 13 I p (A) 1 6 Average values 1 5 75 8 85 9 95 1 Reduction of divertor carbon sources, APS, Nov. 1999, D. Whyte #16 Shot

Y chem reduction is even larger at ISP, consistent with plasma redeposition of both boron and carbon at the inner divertor leg reducing the yield. Attached ELMy H-mode >1 shots post-boronization 1 1.1.1 CD / D γ Outer Strikepoint Inner strikepoint Absolute yield comparison not available because of poor plasma diagnosis at ISP..1 75 8 85 9 95 1 Shot Reduction of divertor carbon sources, APS, Nov. 1999, D. Whyte #17

Smaller subset of detached plasmas show a lower erosion yield than attached plasmas, consistent with lower incident energy at tiles..1.1 ELMy H-mode >1 shots post-boronization Attached Detached Yield comparison depends on changes due to recombining divertor plasma (T e < 2 ev): D/XB CD ~ 2 (PISCES) S/XB Dγ ~ 35 (Experiment).1.1 Y chem 75 8 85 9 95 1 Shot DiMES Y chem ΓC x = Γ D D y B = B CD D/ XB S / XB CD D γ D γ Reduction of divertor carbon sources, APS, Nov. 1999, D. Whyte #18

The average BD band in the lower divertor has not changed significantly with accumulative boronizations. 1 BD / D γ.1.1.1 Attached ELMy H-mode >1 shots post-boronization Tiles cleaned 75 8 85 9 95 1 Shot BD band is a measurement of boron chemical erosion yield B II measurements would better indicate relative boron content but rarely tuned to lines with spectrometer. Results could indicate substantial boron-carbide bond formation, which would be resistant to chemical erosion. Reduction of divertor carbon sources, APS, Nov. 1999, D. Whyte #19

Reduction of total lower divertor carbon sources & core plasma carbon contamination #2

Simultaneous with the reduction in Y chem of ~x1, the lower divertor Y total has decreased at least x4. Attached ELMy H-mode >1 shots post-boronization Maximum value in lower divertor Y total ( Γ = C x D y Γ +Γ D C ) B B D CII S / XB S / XB CII D α, γ α, γ.5.4.3 CII / D α CII / D γ 2.5 2 1.5.2 1.1.5 75 8 85 9 95 1 Shot Reduction of divertor carbon sources, APS, Nov. 1999, D. Whyte #21

Comparison of nearly identical ELMy H-mode plasmas show that the total divertor carbon source has decreased significantly in DIII-D since 1993, without any subsequent change in core carbon concentration! Intensity (a.u.) 1 D ELMy H-mode Outer Strikepoint #781 (6-93) #945 (1-97) 656 657 658 (Å) C II doublet Spectra normalized to D α intensity & for different detector responses (intensified vs. non-intensified) LSN #781 #945 plasmas I p (MA) 1. 1.3 n e (cm -3 ) 4.6x1 13 5x1 13 P inj (MW) 5 4.8 W dia (MJ).85 1.2 R OSP (cm) 158 146 CII / D α (OSP).23.2 f core, C (%) 2.3 2.3 Reduction of divertor carbon sources, APS, Nov. 1999, D. Whyte #22

1 The outer mid-plane carbon source has remained large and constant, in contrast to the significant decrease in lower divertor yield / sources. Attached ELMy H-mode >1 shots post-boronization OVI / D Lyβ CIII / D Lyβ Relative midplane XUV spectroscopy of impurity resonance lines: CIII (977 Å) OVI (13 Å) D Lyβ (125 Å) 1 Average carbon yield using S/XB technique and absolute calibration: Y midplane ~ 1-2% 1 75 8 85 9 95 1 Shot Absolute CIII & CII intensities and S/XB technique indicate significant carbon influx from first wall (~2 A). Reduction of divertor carbon sources, APS, Nov. 1999, D. Whyte #23

Despite the significant decrease in lower divertor yield / source, the core carbon fraction has not decreased. Attached ELMy H-mode >1 shots post-boronization Core carbon fraction (%) 8 7 Detached ELMy H-mode 6 Attached ELMy H-mode 5 Average 4 3 2 1 75 8 85 9 95 1 shot number Based on r/a =.72 carbon concentration from visible CER. Consistent with larger f carbon database compiled by M. Wade & day-to-day experience. Some shots unavailable. Reduction of divertor carbon sources, APS, Nov. 1999, D. Whyte #24

Fractional radiated power, which is typically dominated by carbon, has not significantly decreased either. Attached ELMy H-mode >1 shots post-boronization 1.8.6.4.2 Lower divertor Core (LSN only) Agrees qualitatively with CER core carbon spectroscopy results. 75 8 85 9 95 1 Shot Reduction of divertor carbon sources, APS, Nov. 1999, D. Whyte #25

Summary: DIII-D Y chem Study Old DIII-D results for Y chem are in agreement with other tokamaks... JET: Y chem = 2 % T surf =45K ASDEX-U: Y chem = 1-2 % T surf =35K DIII-D (1992-93) Y chem ~ 2 % T surf ~4 K But present chemical and total yields are greatly reduced DIII-D (1999): Y chem ~.1% T surf ~4K DIII-D tile: Y chem =.16-.4 % T surf =6K (PISCES) Boronizations are the most likely cause for the Y chem reduction 3+ boronizations have applied ~2.5 microns of boron. No major tile changes or clean-ups in lower divertor in over six years both JET and ASDEX have refit lower divertor several times. Reduction of divertor carbon sources, APS, Nov. 1999, D. Whyte #26

Summary on Y total & Carbon sources Reduced Y chem (and other effects?) have significantly reduced the measured total carbon outflux from the lower divertor. Chemical erosion is now a minor player in lower divertor erosion with Y chem ~.1% << Y phys ~1-5%. No simultaneous decrease in core carbon fraction is found. Two explanations are possible: 1. Carbon ion SOL transport always adjusts to equilibrate thermal gradient vs. diffusive forces, making the source amplitude unimportant (preliminary UEDGE results show this effect). 2. The divertor is the not the principal source of carbon in the core plasma (e.g. All Be divertor in JET had carbon as main core impurity). Reduction of divertor carbon sources, APS, Nov. 1999, D. Whyte #27

Laboratory studies of DIII-D graphite tiles effect of boronization on chemical erosion yield. #28

Plan: Use PISCES at UCSD to directly assess Y chem of DIII-D divertor tiles Two DIII-D lower divertor tiles removed (Summer 98) SEM and profilimetry revealed a >micron-depth low-z deposition layer at the inner divertor. ~8 discharges since last boronization. Visual inspection shows a metallic-like surface substantially different than virgin ATJ graphite. Section of tile exposed in PISCES-B linear plasma device Steady-state exposure with similar Γ and T e to DIII-D divertor. 3 ev incident energy fixed by target bias minimal Y phys. In-situ surface analysis station. Reduction of divertor carbon sources, APS, Nov. 1999, D. Whyte #29

PISCES-B Experimental Configuration Surface Analysis System Sample Manipulator Target C, W, Be Plasma Source PISCES-B Linear Plasma Device Reduction of divertor carbon sources, APS, Nov. 1999, D. Whyte #3

Y chem Measurements on PISCES Weight loss of sample net yield -> gross yield correction from modeling. CD band spectroscopy Calibrated visible spectrometer. Benchmarked with CD 4 puffing through target. Reduction of divertor carbon sources, APS, Nov. 1999, D. Whyte #31

Calibrated CD 4 injection through PISCES target & High resolution spectroscopy provide detailed information about hydrocarbon transport & Y chem Sample Face Spectrometer Slit Image Axis for Spatially Resolved Emission, I(z) CD radical band (43 nm) axial profile 8 1 14 6 1 14 4 1 14 2 1 14 Background Level CD 4 injection 1 2 3 4 5 Axial Position (cm) Reduction of divertor carbon sources, APS, Nov. 1999, D. Whyte #32

Summary of PISCES & UTIAS erosion tests on DIII-D tiles DIII-D CD-band molecular excitation rates measured Apparent Y chem < 1-3 in DIII-D. Energy dependence: Y chem Ε (.5 1) Surface composition 5% boron initially on surface. Y chem reduced x4 compared to virgin porous graphite. < 1 minutes of plasma exposure removed most of the carbon from the surface! Flux dependence?: apparently not needed to explain low Y chem. Preliminary result from UTIAS indicates portion of DIII-D redep film is resistant to O 2 chemical removal Reduction of divertor carbon sources, APS, Nov. 1999, D. Whyte #33

D/XB CD 43 nm 1 1 1 PISCES-B D/XB experiment PISCES-A This experiment 2e12 cm -3 3e12 cm -3 1 1 1 Te (ev) D/XB Measurements (Molecules / Photon ratio) D/XB~1 is a maximum. Reproduces previous result. Implication: DIII-D CD-band spectroscopy used D/XB=1...but T e < 5 ev in PDD. Revised upper limit from CD-band spectroscopy: Y chem < 3-4x1-4 Reduction of divertor carbon sources, APS, Nov. 1999, D. Whyte #34

WBC Erosion Code Has Been Thoroughly Benchmarked Against PISCES Data. 1 e-folding distance for CD (cm) 1 CD penetration vs. density, with T e ~ 4 ev 4.5 ev 4 ev 3.5 ev 1.E+11 1.E+12 1.E+13 electron density near target (cm -3 ) PISCES-B WBC prediction 3.5 ev 3.5 ev D/XB 1 1 1 1 D/XB vs. density, with T e ~ 4 ev WBC /w Bethe- Born PISCES-B 4.5 ev 4 ev 3.5 ev 3.5 ev 3.5 ev 1.E+11 1.E+12 1.E+13 electron density (cm-3) WBC sub-gyro Monte-Carlo model (J. Brooks, ANL) includes full dissociative reactions for CD 4. Model matches both CD penetration and photon efficiency. Strongly non-linear penetration vs. n e Very large n e dependence on D/XB for approx. constant T e. Questions constant D/XB used in tokamak experiments during density scan Reduction of divertor carbon sources, APS, Nov. 1999, D. Whyte #35

7.E-3 Chemical erosion yield 6.E-3 5.E-3 4.E-3 3.E-3 2.E-3 1.E-3 T < 15 C Power law fit y = 1.E-3xE.4 Energy dependence on Y chem Y chem E -.4 Γ= 1 23 s -1 m -2 T e ~ 4 ev.e+ 2 4 6 8 Bias Voltage (V) Different energy dependence on chemical erosion yield from one experiment to another. PISCES data is more reliable in the sense that bias voltage is directly controlled...and all other exposure conditions are constant. Reduction of divertor carbon sources, APS, Nov. 1999, D. Whyte #36 Chemical Erosion Yield (CH/ion) 1-1 1-2 1-3 3 K YsurfD A Up PISCES JET 1 1 1 1 Average Energy (ev)

1% 9% 8% 7% 6% 5% 4% Oxygen Nitrogen Carbon Boron Auger surface analysis of DIII-D tile surface pre and post exposure on PISCES-B 3% 2% 1% % Preexposure Postexposure 2.5 Virgin graphite Exposure conditions: E = 3 ev Γ= 1 23 s -1 m -2 T e ~ 5 ev n e ~ 1.5x1 13 cm -3 CD band intensity (a.u.) 2 1.5 1.5 Y chem =2.8x1-3 DIII-D tile with boron Y chem =1.1x1-2 1 2 3 4 5 6 Axial distance from target (cm) Reduction of divertor carbon sources, APS, Nov. 1999, D. Whyte #37

Discussion Chemical erosion in the lower divertor is NOT the source of carbon in detached plasmas. Yet substantial atomic carbon radiation / flux is measured despite T e < 5 ev, i.e. no physical sputtering? Net erosion is eliminated everywhere in lower divertor! Strong indications that boronizations have substantially reduced the chemical and total effective yields of carbon in the DIII-D tokamak. In-situ and laboratory experiments show Y chem reduced ~4-1x compared to virgin graphite. No need to invoke flux dependence as method of reduced Y chem Future work Test more tile samples (PISCES and UofT) Remove more complete poloidal set of tiles during next vent? Incorporate Y chem into DiMES / WBC modeling & mixed material modelling Reduction of divertor carbon sources, APS, Nov. 1999, D. Whyte #38

Discussion (continued). These results are very encouraging for the use of carbon in the learning phase of any next-step, high duty-cycle DT device. Graphite is very forgiving of disruptions, ELMs, etc. Detached plasmas necessary for power handling Y phys = But large Y chem for graphite produces significant erosion / redeposition and Tritium retention. Based on the DIII-D results, ~24 hours of plasma exposure, including frequent boronizations, should condition graphite divertor tiles to obtain order of magnitude reduction in Y chem. Reduction of divertor carbon sources, APS, Nov. 1999, D. Whyte #39