plasma interactions with be surfaces

U C S DU ni v e rs it y of C ali fo rn i a S a n D i eg o

R. P. Doerner, 2nd PMIF Meeting, Juelich, Sept. 19-21, 2011

Plasma interactions with Be surfaces

R. P. Doerner, D. Nishijima, T. Schwarz-Selinger and members of the PISCES Team

Center for Energy Research, University of California – San Diego, USA

Work performed as part of: Plasma-Surface Interaction Science Center (MIT and UTenn)

US-EU Collaboration on Mixed-Material PMI Effects for ITER



The PISCES-B divertor plasma simulator is used to investigate ITER mixed materials PSI.

PISCES ITER (edge)

Ion flux (cm2s–1) 1017–1019 ~1019 - 1020 Ion energy (eV) 20–300 (bias) 10–300 (thermal) Te (eV) 4–40 1–100 ne (cm–3) 1012–1013 ~1013 Be Imp. fraction (%) Up to a few % 1–10 (ITER) Pulse length (s) Steady state 1000 PSI materials C, W, Be C, W, Be .. Plasma species H, D, He H, D, T, He

• PISCES-B is contained within an isolated safety enclosure to prevent the release of Be dust.



PISCES-B has been modified to allow exposure of samples to Be seeded plasma

Radial transport guard

102 mm

153 mm

76 mm

195 mm

45 o

Cooled target holder

Heatable deposition probe assembly

Thermocouple

Thermocouple

Water cooled Mo heat dump

Resistive heating coils

High temperature MBE effusion cell used to seed plasma with evaporated Be

12 °

PISCES-B PlasmaTarget

Depositionprobesample

Axial spectroscopic field of view

Berylliumimpurityseeding

Radial transport guard

102 mm

153 mm

76 mm

195 mm

45 o

Cooled target holder

Heatable deposition probe assembly

Thermocouple

Thermocouple

Water cooled Mo heat dump

Resistive heating coils

High temperature MBE effusion cell used to seed plasma with evaporated Be

12 °

PISCES-B PlasmaTarget

Depositionprobesample

Axial spectroscopic field of view

Berylliumimpurityseeding

P-B experiments simulateBe erosion from ITER wall,subsequent sol transport and interaction with W bafflesor C dump plates, as well asinvestigation of codepositedmaterials using witness plates

PISCES



Outline of Presentation

• Erosion in the plasma environment– Comparison to TRIM and ion beam data– Surface characterization– Role of morphology

• Redeposition/sticking efficiency

• Summary

PISCES



Be erosion yield measurements in PISCES

Two techniques are used to measure physical sputtering yield

• Weight loss measurements, use low density plasma to reduce redeposition (i.e. long ionization mean free path).

• Line emission spectroscopy, uses high density plasma to minimize geometrical loss terms (i.e. short ionization mean free path).

These presentation will focus on the weight loss technique and changes to spectroscopic measurements that are normalized to weight loss measurements. [He plasma on Be, weight loss are 5-10 times lower than TRIM calculations, He plasma on C, weight loss agree with TRIM calculations].

PISCES



Significant variations in the Be sputtering yield are measured

Incident ion energy ~100 eV

J. Roth et al., FED 37(1997)465.

PISCESdiscrepancy between - JET - PISCES-B - ion beam – TRIM - sputter yields

(< 45%) (< 0.7%) (< 8%) (< 3.5%)



Uncertainties in sputtering yield measurements

Uncertainties in weight loss measurements:• Surface contamination• Incident ion energy• Ion flux measurement• Ion species mix (molecular ions)• Redeposition fraction• Surface morphology changes• Atomic D adsorption on the surface (secondary ion yield)

Uncertainties in absolute spectroscopic measurements:• Atomic physics database• Electron energy distribution (non-Maxwellian)• Angular distribution of sputtered particles• Geometrical loss fraction

PISCES



Native beryllium oxide surface is removed early during the plasma exposure

• Distinctive oxygen lines near 777 nm can monitor erosion of O from surface• Background helium plasma does not change (second order He I line)• Larger ion energy will remove oxide layer quicker• BeO (0,0) molecular band emission (@ 470.86 nm) is not detectable

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0 200 400 600 800 1000

OI@7771/HeI@3888

Time[s]

He plasma (30 eV ions)

Helium/oxygenplasma

(30 eV ions)

0

500

1000

1500

2000

776.5 777 777.5 778 778.5

t = 265 st = 641 s

Wavelength (nm)

PISCES



AES reveals a relatively ‘clean’ Be surface after sputtering yield measurements

Time from plasma shut-off (s)

0 1000 2000 3000 4000 5000

Sur

face

com

posi

tion

(%)

020

4060

8010

0

Be CO

(7.0)

(0.9)

(92.1)

PISCES

500 1000 1500

before

dN(E

)

1.25·1022 Dx+/cm2 , -100V bias

FeB O

Be

after

kinetic energy (eV)

C Cu



Incident ion energy is corrected for plasma potential. Ion flux is uniform over sample surface.

From B. LaBombard et al, JNM 162-164 (1989) 314.

PISCES-A space potential measurements show Vpl ~ 1-2 Te

Eion = |Vbias | - 1.5Te

Target

From D. Whyte et al, NF 41 (2001) 47.

Ion flux calculated from upstream Langmuir probe agrees with total current collected on the sample manipulator when biased into Isat (with ~10%)

PISCES



Molecular ion fractions are calculated from zero-d rate balanced model, based on measurements

• Model uses the atomic and molecular processes to the right

• Rate equations predict molecular ion fractions based on ne, Te, B, NH2

• Turbulent radial transport is assumed (1/B scaling)

• Model is verified with molecular ion species measurements

From E. Hollmann et al., Phys. Plasmas 9 (2002) 4330.

PISCES



Molecular ion effects change both the shape and magnitude of sputtering yield curve

PISCES

10-5

10-4

10-3

10-2

10-1

0 20 40 60 80 100 120 140

Measured Yield in D plasma (*5)Y_D(Ei)2*Y_D(Ei/2)3*Y_D(Ei/3)total_Y(0.25,0.47,0.28)

Ei [eV]

10-5

10-4

10-3

10-2

10-1

0 20 40 60 80 100 120 140

Measured Yield in D plasmaY_D(Ei)2*Y_D(Ei/2)3*Y_D(Ei/3)total_Y(0.25,0.47,0.28)

Ei [eV]

Shape of measured yield agrees with molecular ion model predictions.Magnitude is a factor of ~5 too low.



ERO calculates 10-20% ionization of sputtered Be in the PISCES-B plasma

• Shape of Be I (457 nm) axial profile agrees well with experimental profiles

• Magnitude of profiles are normalized

• Molecular ion fractions provided by previous model

• No Be deposition observed on W or C targets exposed to D plasma (no wall source)

PISCES

From D. Borodin et al, Phys. Scr. T128 (2007) 127.



Be surfacebefore plasma exposure

after

Surface morphology evolution with time / fluence

spectroscopy:

mass loss:

0 1 2 3 4 50

2

4

6

8

10

12

sput

ter y

ield

(10-3 B

e pe

r ion

)

ion fluence (1022/(cm2s))

D2, <330K, -100V

morphology change can account for a factor of 2 in reduction of the yield

PISCES



Similar yield evolution with time/fluence is documented in the

literature

morphology change can account for a factor of 2 reduction of the yield

1keV, H2+

7.3E21 ions/cm2

Mattox and Sharp, J. Nucl. Mater. 1979:PISCES



1) “maximum” – static TRIM + MD2) “minimum” – SDTrimSP with 50% of D (reasonable limit)

Plasma atoms remaining in the near surface also can reduce the sputtering yield by a factor of 2-3

From C. Björkas

PISCES



Gross erosion is expected to remain constant with increasing Be redeposition/influx

• N = G * (1-R), where N = net erosion, G = gross erosion, R = redeposited fraction

• In PISCES-B, Gross = net erosion with no Be seeding (λion is large compared to rplasma, so redeposition is small)

• Ion influx from Be oven is identical to sputtered atoms that are ionized in the plasma and redeposited on the target, this allow a controlled and independent variation of the Be influx to the target

PISCES

0

0.2

0.4

0.6

0.8

1

1.2

0 0.2 0.4 0.6 0.8 1

Current theoretical picture

Gross (theory)Net (theory)

Redeposited fraction or influx (normalized units)



Erosion/deposition balance in Be seeded high flux D discharges

• Use Be oven seeding to balance surface erosion to test input parameters of material migration models

• Mass loss measures net erosion• Spectroscopy measures gross

erosion (Be I line)• Y Be→Be ≈ Y D→Be, and low

concentration of Be• When incident/seeded Be ion

flux = sputtered flux of Be, net erosion should = 0.

40 60 80 100 120 140 160 180

0.01

0.1

1

yiel

d (%

Be

per i

on)

-bias (V)

pure D2

1/6 TRIM

Mass loss

ion fluence: 1022/cm2

target temperature < 320K

PISCES



No change in mass loss is measured when Be seeding flux equals sputtering of Be by D

• ADAS database is used • Be flux from Be II (313.1 nm) and

background plasma flow velocity (E. Hollman JNM, PSI-19)

• Be ion flux is verified during no bias discharges, when weight gain is measured (net deposition)

• Net erosion stays constant, implying gross erosion must increase

• Erosion yield of 0.15% can only be compensated by seeding 2.8% Be

40 60 80 100 120 140 160 180

0.01

0.1

1

0.1<0.03<0.06<0.01

<0.01

0.26

0.10.9 0.9

eff.

yiel

d (%

Be

per i

on)

-bias (V)

Be seeded pure D2

Be concentration seeded / non-seeded (%)

0.9

Mass loss

ion fluence: 1022/cm2

D/Be plasmatarget temperature < 320K

2.8%

PISCES



Beryllium seeded He discharges

target:bias: < -40V results in E ≈ 30eVHe ion flux: 5·1018 cm-2s-1

Be seeding: nBe/nD = 0 – 4 %

sputter yield He on Be @ 30eV: Y = 0.15 % (measured)

0 20 40 60 80

0.01

0.1

1

10

BeI

(457

.3nm

) int

ensi

ty (a

.u.)

z [mm]

4%

1%0.3%

0.06%0.02%

no seeding

target oven

PISCES

: During ‘no Be seeding’ discharge,Gross ≈ Net erosion

Grosserosion



Different behavior is observed experimentally

• Net erosion stays constant until influx >> sputtering rate

• Gross erosion increases with increasing Be influx to target

• Reduced sticking of depositing Be, or increased re-erosion could explain observations

• Similar reduced sticking needed to model CD4 and WF6 injection experiments in TEXTOR (from A. Kirschner)

• Possibly similar to C-MOD modeling difficulties reported by J. Brooks (PSI19)

PISCES

0.1

1

10

0 2 4 6 8 10

Artisitc impression of measured results

Gross (theory)Net (theory)Net (exp)Gross (exp)

Influx (normalized units)



What PMI issues are still unresolved

• Sputtering Yield – we believe we can explain the observed, low sputtering yields in PISCES-B, due to surface morphology and fuel atoms within the target surface (no Be seeding)– How many gas atoms are in the surface during exposure?

• Gross/Net Erosion – drastic differences in behavior of gross erosion during Be seeding indicates the simple theory regarding the benefits of prompt redeposition may need to be revisited– 10 times more influx is needed to balance erosion and force net

erosion to zero– Low sticking probability or high re-erosion possible explanations

PISCES



0 50 100 150 200

1E-4

1E-3

0.01

0.1

parti

cle

refe

lcet

ion

coef

ficie

nt

energy (eV)

D on Be Be on Be

TRIM(Eckstein 2002)

10 100 1000

10-4

10-3

10-2

10-1

D on Be Be on Be

SPU

TTER

ING

YIE

LD (a

t/ion

)ENERGY (eV)

TRIM(Eckstein 2002)

Be seeding fraction (i.e. influx) is at most in the percent range, so Be self-sputtering and Be reflection can be neglected

plasma interactions with be surfaces

Documents

pmif meeting

yield measurements pisces

plasma environmentcomparison

weight loss technique

erosion yield measurements

pisces teamcenter

jet pisces

detectable pisces