m.g. bell nstx supported by nstx physics operator’s course february 2010 1.conditioning the vacuum...

M.G. Bell

NSTXNSTX Supported by

NSTX Physics Operator’s CourseFebruary 2010

1. Conditioning the Vacuum System and PFCs for Plasma Operation

2. Physics Operator’s Roles and Responsibilities

M.G. Bell / PhysOps Course / 2010 2

NSTX Vacuum System and PFCs

• Vacuum vessel– Mostly stainless steel; inconel inner tube

– ~600 primary vacuum seals!– Volume 29 m3

• Pumped (at Bay L) by 2 1.5 m3s-1 TMPs 1.6 m3s-1 for D2 in vessel

• NB cryo-pumps add 35 m3s-1 for D2 when NBI valve open (no He pumping)– Characteristic pump-out time ~ 1 s

• Plasma-Facing Components (PFCs) of graphite (ATJ grade) and some carbon-fiber composite (CFC) on CS– ~3200 tiles 1/2 – 2 in. thick attached by T-slots on back

– plasma facing area ~40 m2

• HHFW antenna screen is inconel coated with TiC layer


The “Subway” Tour of the Vessel Interior

• Taken with camera mounted on the model train used for transporting a neutron source to calibrate the neutron detectors in December 2009

QuickTime™ and aMotion JPEG OpenDML decompressor

are needed to see this picture.


Graphite Tiles Absorb Atmospheric Gases, Particularly Water, While Vessel Is Open

• Graphite is relatively porous– Outgassing from graphite is a slow process – heating increases diffusion

• Stainless-steel surfaces also adsorb water on surface

W. Blanchard

RGA Spectra Prior to BakeoutJanuary 2009 February 2010

Note lowerCO2 Peak

Dominatedby water

Ion gauge3.7E-6 Torr


NoticeableCO2 Peak

• Residual CO2 (44 AMU) peak in 2009 thought to be due to Li2CO3 on tiles– Tiles were cleaned of lithium residue during most recent outage

Note: RGA pressures

are “indicator only”


Perform High-Temperature Bakeout After a Major Opening of the Vacuum Vessel

• Use combination of techniques to heat PFCs and internal structures

– 350°C helium through network of pipes for outer PFCs and (now) lower divertor

– Resistive heating of central inconel tube for inner PFCs and upper divertor

• Current flows through jumpers across upper CHI gap and returns via outer vacuum vessel

– 150°C pressurized water to heat outer vacuum vessel

• Bakeouts are reproducible and additive within a year

1.E-08

1.E-07

1.E-06

1.E-05

1.E-04

-2 0 2 4 6 8 10 12Days since bake began

Base Pressure (Torr)

2005

2006

2007

• Initial transient ~1 x 10-3 Torr

– Residual gases during bakeout: H2, H2O, CO, CO2 (“water-gas”)

• Initial rapid fall from hottest carbon

• Final slope due to colder regions with much lower outgassing rate

• Performing 3-week bake this year to improve condition for starting run

D. Gates


Bakeout is Effective at Removing Water

• Even after bakeout, it takes time to produce satisfactory plasma conditions

28 AMU Peakfrom CO, N2


February 2009 After Bakeout

Appearanceof D2, HDO

Note: RGA pressures

are “indicator only”


Glow Discharge Cleaning and Boronization

• In past, performed extensive glow-discharge cleaning to remove contaminants and residual plasma constituents from PFCs– 2 stainless-steel anodes at Bays G, K on outer VV near midplane

– ~3 A total current at ~550 V with gas pressure 2 - 5 mTorr

– Deuterium GDC used to remove oxygen-bearing contaminants at end of bake

– Helium GDC to remove hydrogenic species

• After DGDC (~2hr); at start of each run day (30min); between shots (7–15 min)

• Recently, eliminated DGDC and cut back on HeGDC

• Also used “boronization” at end of bake and periodically during each run

– Run GDC in mixture of 5% deuterated trimethyl boron (TMB - (CD3)3B), 95% He

– Generally used half or full bottle containing 10 g TMB over 2 - 3 hours• TMB is toxic, pyrophoric, expensive (bottle costs ~$3K)

– Apply ~1hr pure HeGDC afterwards to deplete D from deposited B/C/D layer

– B has high affinity for oxygen and sequesters it as borates

– B layer does not sequester impinging hydrogenic species significantly

• Use of boronization also reduced - will be eliminated during this run with LLD

M.G. Bell / PhysOps Course / 2010

NSTX has a Continuing Research Program into Effects of Lithium-Coated PFCs in

Divertor Plasmas

8

2005: Injected lithium pellets into He discharges prior to D NBI shot

2006: LIThium EvaporatoR (LITER) deposited lithium on room-temperature center column and lower divertor

2007: Larger evaporator re-aimed to increase deposition rate on lower divertor

2008: Dual LITERs to eliminate shadowed regions on lower divertor

– First use of “lithium powder dropper” to introduce lithium through SOL

2009: Routine use of dual LITERs

• At end of run, 80% of discharges had lithium applied beforehand

– Also conducted experiment using dual lithium powder droppers

2010: Liquid lithium coating on section of lower outer divertor plate


Dual LITERs Replenish Lithium Layer on Lower Divertor Between Tokamak Discharges

9

LITER Canisters

ROTATABLESHUTTER

• Electrically-heated stainless-steel canisters with re-entrant exit ducts

• Mounted 150° apart on probes behind gaps between upper divertor plates

• Each evaporates 1 – 40 mg/min with lithium reservoir at 520 – 630°C

• Rotatable shutters interrupt lithium deposition during discharges & HeGDC

• Withdrawn behind airlocks for reloading and initial melting of lithium charge

• Reloaded LITERs 6 times during 2009 run (Mar - Aug): ~250g on PFCs

H. Kugel


Dual LITERs Deposit Lithium on Lower PFCs Including Divertor Plates

10

• Measured deposition pattern in laboratory tests with scannable quartz-crystal micro-balance (QMB)

– Plumes of lithium vapor are roughly Gaussian in angular distribution– Good agreement with model based on molecular flow through exit duct

• Lithium applied between discharges typically 20 – 600 mg– More than needed to react all injected D2, typically 5 – 15 mg

• In-situ QMB data implies deposited lithium thickness is 5 – 160 nm on inner divertor plate near strike point of standard NSTX plasmas

Modeled deposition pattern

LITERs

H. Kugel, L. Zakharov


Lithium Coating Reduces Deuterium Recycling, Suppresses ELMs, Improves Confinement

No lithium (129239); 260mg lithium (129245)

Without ELMs, impurity accumulationincreases radiated power and Zeff

H. Kugel, B. LeBlanc, R.E. Bell

Without ELMs, impurity accumulationincreases radiated power and Zeff


Solid Lithium Does Pump Deuterium but Normally We Increase Fueling to Avoid Early Locked Modes

• Tangentially viewing camera for edge D emission shows greatly reduced neutral D density across outboard midplane with lithium from LITER

• Lower density is achievable early in discharges but likelihood of deleterious locked modes increases: we need to learn to avoid locked-modes to exploit lithium

Measured at 0.22s

B. LeBlanc, P. Ross

CS Limiter D discharges (0.9MA, 0.45T, 4MW NBI)

with same gas fueling afterHelium conditioning and

lithium pellets (~30mg) injected into preceding 10 OH He discharges

0.14 – 0.18 s0.51.01.5ne(1019m-3)

Ist shotafter LiBefore Li(after He)Radius (m)0.00.51.02.01.5

2008-9 LSN discharges100 – 600 mg fresh Li from LITER

(25 – 160 nm thick at inner divertor)

ND ~ 0.8 1021


Analysis of Carbon Tile Surfaces Confirms Migration of Lithium Under Plasma Fluxes

• Analysis performed on surface of carbon tiles as removed from vessel • Used ion-beam nuclear-reaction analysis for lithium and deuterium areal

density in surface layer

W. Wampler (SNL), C. Skinner, H. Kugel

0

1

2

3

4

5

6

0 20 40 60 80 100 120 Position along scan (cm)

Lit

hiu

m d

ens

ity

(1

019 c

m-2)

Bay L Bay F

Center columnInner

divertor Outer divertorScan across lower divertor

• Peak lithium density remaining on inner divertor ~0.6 mg·cm-2

• Total deposition there estimated at ~8 mg·cm-2

• Less than 1% of deposited lithium remains in high heat flux region


0.15 mmPorous Mo

0.25 mm SS Li Barrier Bonded

to 16 mm Thick Cu Backing Plate

Micrograph of porous Mo layer

Note the small poorly connected surface features

Back side of plate with heaters and thermocouples installed

We Must Learn to Operate With and Make Best Use of the Liquid Lithium Divertor Now Installed

January 11, 2010

H. Kugel, R. Nygren (SNL), S. O’Dell (PPI), E. Starkman

M.G. Bell / PhysOps Course / 2010 15L. Zakharov, H. Kugel

• Rely on liquid wetting the porous Mo surface to spread the lithium• Only 7% of lithium evaporated by LITERs reaches LLD-1 plates• Estimate ~40g lithium required to fill porous volume in Mo coating

– ~600g evaporation to fill 22 days at maximum rate & ~7 loadings

• Wettable area in porous Mo estimated at ~8 times plate area– 1.1g lithium on LLD would coat wettable area to 250nm penetration depth of

incident D+ 15g evaporated ~ 1 day at normal evaporation rate

LLD loading at 60mg/min total rateLITERs

LLD-1

Plan to Fill LLD-1 with Lithium from Dual LITERs, Possibly Supplemented by Lithium Droppers

0.5

0.4

0.3

0.2

0.1

0.0

Lit

hiu

m d

epo

siti

on

rat

e (m

on

laye

rs/s

)


To Give the LLD a Fair Trial, We Must Protect It From the Plasma and vice versa, At Least Initially

• Direct plasma impingement could melt the porous surface very rapidly and affect its porosity and ability to hold lithium

– During plasma startup when control is not tight and energetic ions are poorly confined

– Uncontrolled transients (e.g. disruptions and collapses)

– We have seen strong divertor interactions during HHFW, particularly without lithium to reduce the SOL density

• NSTX is susceptible to metal radiation in its operating regime

– More a problem in ELM-free H-modes produced by lithium

– Sputtering, particularly by impurities, could lead to runaway Mo influx

• We will not employ boronization which could affect lithium wetting

• To protect its engineered surface and reduce the exposure of the plasma to Mo, we will partially fill the LLD with lithium before running any plasma shots


Can We Distinguish Effects of the LLD from Lithium Applied Elsewhere While Filling the LLD?

• Solid lithium ceases to absorb D when the surface to the depth of penetration (~250 nm) becomes saturated at some D:Li ratio < 1

– 250 nm Li film contains 1.2 1022 Li/m2, so absorbing 1 1021 D in a typical NSTX shot would fill the active area of the lower divertor, ~1m2, with ~8% D

• A thick liquid lithium film is expected to retain its D-sequestration capability because adsorbed D remains mobile in the liquid

– Measurements suggest that precipitation of LiD begins for D/Li ~ 0.1

• If we prefill the LLD, we can “deactivate” the lithium surface on it and surrounding PFCs by running a series of 3 – 10 standard shots

• We can then heat the LLD above the lithium melting point (180°C) and hope to restore its activity, assuming that

– A saturated layer of D in solid Li will be absorbed into the Li below as it melts

– Migration of the Li and Li-D from the contact areas on the divertor to the LLD during the discharges will not overwhelm or “poison” the LLD layer


Physics Operator’s Role and Responsibilities

18

• In consultation with the Session Leader, execute the plans described in NSTX Experimental Proposals (XPs)

– Produce plasmas with the required properties as efficiently as possible

– Set up the Plasma Control System and monitor its operation

– Interact as needed with the• Chief Operating Engineer (Bill Blanchard, Ray Camp, Steve Langish)

• Neutral Beam Operator (by dedicated phone at PO station)

• HHFW Physicist (Joel Hosea, Gary Taylor, Phil Ryan, Randy Wilson)

• Vacuum Operator (Joe Winston, Gerry Gething, Scott Gifford)

• LITER Operator (Hans Schneider, Doug Westover, ...)

• LLD Operator (Hans Schneider, Twyman Plumb, ...)

• Diagnostic physicists

• Monitor the basic diagnostic data to ensure that the machine systems are operating satisfactorily and safely

• Document progress and problems shot-by-shot in the online log


Each XP Contains a Physics Operations Request


Documenting What Was Done Operationally is an Important Part of the Physics Operator’s Job

• The Logbook is the first record that we use to find out what was done on a particular shot or day

– http://nstx.pppl.gov/nstx/Software/logbook.html

• Make entries for the daily test shots and any calibration shots

• Note what shot was used in the PCS as the basis

• Note what changes were made and to what end– Some changes are apparent in diagnostic data, but it’s a lot quicker to find if you

know where to look!

• Note when systems like the LITERs, or the LLD are introduced

• Record what happened, particularly if the result was unexpected

• If the shot failed to meet expectations because of mis-programming, say so!– We want to learn from our mistakes and identify common causes of error

• The test: another operator should be able to reproduce an experiment based on the PO log– Log in under your own name - don’t use a session left by your predecessor

m.g. bell nstx supported by nstx physics operator’s course february 2010 1.conditioning the vacuum...

Documents