Download - Deuterium Retention in and Release from Beryllium ... · Deuterium Retention in and Release from Beryllium Resulting from High Flux Plasma Exposure aMax-Planck-Institut für Plasmaphysik,

D retention in and release from Be.ppt, © Thomas Schwarz-Selinger, PFMC13 I-19, 12. May 2011

U C S DUniversity of California San Diego

Deuterium Retention in and Release from Beryllium Resulting

from High Flux Plasma Exposure

aMax-Planck-Institut für Plasmaphysik, 85748 Garching, GermanybGeneral Atomics, San Diego, CA 92121, USA

cUniversity of California San Diego, CA 92093, USA

T. Schwarz-Selingera, M. Oberkoflera, H. Xub

R. Doernerc and the PISCES Teamc

Max-Planck-Institutfür Plasmaphysik

13th International Workshop on Plasma-Facing Materials and Components for Fusion Applications /

1st International Conference on Fusion Energy Materials Science



Motivation: T Removal in ITER

T Retention in pure Beryllium

implantation: saturation at low levels

codeposition: unlimited

main wall erosion dominated, but:

- gaps are potential deposition zones!

- resessed wall elements

assessing T removal by baking at

240°C / 350°C



K. Sugiyama et al., PSI-19(doi:10.1016/j.jnucmat.2010.09.043):

ITER baking temperatures 240ºC (main wall) 350ºC (divertor)

assessment of T removal in ITER by wall baking

Here: thick codeposits is removal efficiency increased for longer hold times?

(is release determined by diffusion or by trap/bond energies)

0 2000 40001E-3

0.01

0.1

1

100

200

300

400

500

600

D fl

ux (1

015 D

cm

-2s-1

)

time (sec)

0.3

K/se

c

ramp tem

perature (°C)

600 eV D3+ implanted in 300 nm thick

TVA deposited Be layers

350°C



outline

Deuterium Retention in and Release from Beryllium Resulting

from High Flux Plasma Exposure

D release from thick codeposits for long hold times at 240°C / 350°C

single / multilayer co-deposits

high flux codeposit

implanted Be

model system: magnetron sputtered D/Be

Comparing D release from



PISCES-B

W witness plate

B fieldBe target

optionalBe

seeding

target:bias: < -170V results in E < 150eV+ high flux:D ion flux: < 5·1018 cm-2s-1

Be seeding: nBe/nD < 3% - one sample at a time- implantation profile

witness plate:energy of sputtered Be: few eVenergy of reflected Be: < 10 eVenergy of reflected D: < 15-40 eV- low fluxes:Be flux: <3·1014 cm-2s-1

D flux: <3·1015 cm-2s-1

- radial inhomogeneity

DC reflex-arc discharge within airtight Be enclosure

high flux codeposit



H.W. Xu et al., Fusion Sci. Technol. 51, 547-552 (2007)

model system: magnetron sputtered Be

3 x 100Watt,6 mTorr,80% Ar, 20% D2

energy of sputtered Be: few eVenergy of reflected Ar: few eVenergy of ionized Ar: <120 eVenergy of D ions: <120eV

Be flux: 2.5·1015 cm-2s-1

+ up to 30 identical samples at a time

2 mm W spheres



D release from multilayer codeposits



D release from multilayer codeposits

Present predictions for T release in ITER are based on pure Be targets /

codeposits.

Actual ITER codeposits will be multilayers with BeO interfaces in between.

Be/D

Be/D

BeO

Question: Do BeO layers influence the D

release?

here: - collecting PISCES-B on W witness

- interrupted exposure for several hours / bring to air to allow an oxide layer

to grow

Diffusion barrier? Shift in release temperature?



experiment: low flux codeposit

- single layer: Be target + cap, bias = -100V ED = 22 eV = 1.2 · 1014 Be/cm-2s-1, t = 10000 sec

D release from single layer codeposit

W



100 200 300 400 500 600

1E-3

0.01

0.1

1

D fl

ux (1

016 D

/s)

sample temperature (°C)

PISCES-B, 300 K single layer

0.3 K/sec

2.41017D/cm2

tms2

2CW

B

experiment: low flux codeposit

- single layer: Be target + cap, bias = -100V ED = 22 eV = 1.2 · 1014 Be/cm-2s-1, t = 10000 sec

D release from single layer codeposit

TDS:

d = 100 nmD/Be = 0.2

W



100 200 300 400 500 600

1E-3

0.01

0.1

1

8.91017D/cm2

4.81017D/cm2

D fl

ux (1

016 D

/s)


PISCES-B, 300 K first layer second layer

0.3 K/sec

4.21017D/cm2

tms2

1CW

B / t

ms2

2CW

B

experiment

- 2 identical runs: Be target + cap, bias = -100V ED = 22 eV = 1.2 · 1014 Be/cm-2s-1, t = 10000 sec

- replacing one half piece of circular W witness plate after first run

D release from multilayer codeposit

TDS:

d = 2 x 100 nmD/Be = 0.2

BeOexpected 5-10 nm

W



100 200 300 400 500 600

1E-3

0.01

0.1

1

8.91017D/cm2

4.81017D/cm2

D fl

ux (1

016 D

/s)


PISCES-B, 300 K first layer second layer double layer

0.3 K/sec

4.21017D/cm2

tms2

1CW

A / t

ms2

1CW

B / t

ms2

2CW

B

experiment

- 2 identical runs: Be target + cap, bias = -100V ED = 22 eV = 1.2 · 1014 Be/cm-2s-1, t = 10000 sec

- replacing one half piece of circular W witness plate after first run


TDS:

d = 2 x 100 nmD/Be = 0.2

BeOexpected 5-10 nm

W



conclusion:

oxide interfaces (several nm thick) do not influence D release


100 200 300 400 500 600

1E-3

0.01

0.1

1

8.91017D/cm2

4.81017D/cm2

D fl

ux (1

016 D

/s)


PISCES-B, 300 K first layer second layer double layer sum first+second

0.3 K/sec

4.21017D/cm2

tms2

1CW

A / t

ms2

1CW

B / t

ms2

2CW

B

d = 2 x 100 nmD/Be = 0.2

BeOexpected 5-10 nm

W



comparing different codeposits



PISCES-B low flux (witness plate) codeposit

collecting Be and D (on W) sputtered/reflected from a Be target:

thermal desorption:

100 200 300 400 500 600

1E-4

1E-3

0.01

0.1

1

D fl

ux (1

016 D

cm

-2s-1

)


sum as D2

as HD

tms2

1CW

B

structurally modified, supersaturated(D2 bubbles / a-Be-D)

Be-D2 BeO defects bycoll. cascade



100 200 300 400 500 600

0.01

0.1

1

nD = 4.2 1017 D/cm2

low flux codeposit:D = 2.4 1015 D/(cm2s)

Be = 1.2 1014 Be/(cm2s)EBe< eV, ED= 21 eVT = 320 Kt = 10000 secD/Be = 0.34, d= 100 nm

rem

aini

ng D

(DT/D

0)

sample temperature T (°C)

tms-

22-C

-W-B240°C: 16%

350°C: 4%

PISCES-B low flux (witness plate) codeposit

collecting Be and D (on W) sputtered/reflected from a Be target:

thermal desorption:

100 200 300 400 500 600

1E-4

1E-3

0.01

0.1

1

D fl

ux (1

016 D

cm

-2s-1

)


sum as D2

as HD

tms2

1CW

B

ρ = 1.8 ·1017 D cm-2 / µmor: D/Be = 1.5%


Be-D2 BeO



thermal desorption:

high flux codeposit: collecting Be and D at floating target in Be seeded D discharge

PISCES-B high flux (target) codeposit

100 200 300 400 500 600

1E-4

1E-3

0.01

0.1

1 0.1 K/sec

D fl

ux (1

016 D

cm

-2s-1

)


sum as D2

as HD

BeL

SR

17gr

owth

100 200 300 400 500 600

0.01

0.1

1

nD = 1.54 1018 D/cm2

high flux codeposit:D = 3 1018 D/(cm2s)

Be = 1 1016 Be/(cm2s)ED=15 eVT = 320 Kt = 1000 secD/Be = 0.15, d = 825nm

rem

aini

ng D

(DT/D

0)


BeLS

R17

grow

th

240°C: 4%350°C: 2%

ρ = 5 ·1016 D cm-2 / µmor: D/Be = 0.5 %





100 200 300 400 500 600

0.01

0.1

1

nD = 1.3 1017 D/cm2 seeded target erosion:jD = 2.5 1018D/(cm2s)

jBe = 2.2 1016 Be/(cm2s)

= - 2.1 1016 Be/(cm2s)

ED=150 eVT = 320 Kt = 10400 secD/Be = ?, d < 8 nm

rem

aini

ng D

(DT/D

0)


BeLS

R12

240°C: 12%350°C: 6%

PISCES-B target

thermal desorption:

retention during net erosion:

ρ = 1.5 ·1017 D cm-2 / µmor: D/Be = 1%

100 200 300 400 500 6001E-5

1E-4

1E-3

0.01

0.1

1 0.3 K/sec

D fl

ux (1

016 D

cm

-2s-1

)


sum as D2

as HD

BeLS

R12





100 200 300 400 500 600

0.01

0.1

1

magnetron codeposit:D = ?

Be = 1.4 1015 Be/(cm2s)ED=120eVT = 320 Kt = 9000 sec=>3 at.% D, d= 1 m

rem

aini

ng D

(DT/D

0)


W-G

A70

240°C: 82%350°C: 36%

nD = 4.0 1017 D/cm2

magnetron sputtered films

thermal desorption:

collecting Be/D sputtered from a Be target in Ar/D2 :

ρ = 1.4 ·1017 D cm-2 / µmor: D/Be = 0.8%

100 200 300 400 500 600

0.01

0.1

1

D fl

ux (1

016 D

cm

-2s-1

)


sum as D2

as HD

W-G

A70





the total amount of D retained above 350°C is the similar for all codeposits investigated:

ρ ≈·1017 D cm-2 / µm

(D/Be ≈ 1%)



D release from thick codepositsduring long holds @ 240°C (510K) and 350°C (620K)



D release from magnetron sputtered D/Be @ 240°C

0 2000 4000

0.01

0.1

1

100

200

300

400

500

600

700D

flux

(1015

D c

m-2s-1

)

time (sec)

reference run no hold 1 h hold 5 h hold 24 h hold

0.3

K/se

c

ramp tem

perature (°C)

1st ramp + hold




0 2000 4000

0.01

0.1

1

100

200

300

400

500

600

700

D fl

ux (1

015 D

cm

-2s-1

)time (sec)


0.3

K/se

c

ramp tem

perature (°C)

2nd ramp

0 2000 4000

0.01

0.1

1

100

200

300

400

500

600

700

D fl

ux (1

015 D

cm

-2s-1

)

time (sec)


0.3

K/se

c

ramp tem

perature (°C)

1st ramp + hold

time constant of hours!




0 5 10 15 20 250

20

40

60

80

100

log(hold time)

rele

ased

D fr

actio

ns (%

)

hold time (h)

sum

1st ramp to 240°C/510 K

hold at 240°C / 510 K

2nd ramp to 650°C / 920 K

1.0 ·1017 D cm-2 remainsafter hold

or: D/Be = 0.8%

extrapolating to 690 m2 ITER main wall: 9 g D per µm remains after 24h @ 240°C




0 2000 4000

0.01

0.1

1

100

200

300

400

500

600

700

D fl

ux (1

015 D

cm

-2s-1

)time (sec)


0.3

K/se

c

ramp tem

perature (°C)

C:\Dokumente und Einstellungen\tms\Eigene Dateien\Projekte\PISCES\Daten\GA coater\ W-GA70s-holds-ramps.opj(350fluxtest2) 04.05.201

2nd ramp

0 2000 4000

0.01

0.1

1

100

200

300

400

500

600

700

D fl

ux (1

015 D

cm

-2s-1

)

time (sec)


0.3

K/se

c

ramp tem

perature (°C)

C:\Dokumente und Einstellungen\tms\Eigene Dateien\Projekte\PISCES\Daten\GA coater\ W-GA70s-holds-ramps.opj(350fluxtest) 04.05.2011 17:01:14

1st ramp + hold

time constant of hours!




0 5 10 15 20 250

102030405060708090

100110

2nd ramp to 650°C/920 K

sum

1st ramp to 350°C/620 K

hold at 350°C/620 K

rele

ased

D fr

actio

ns (%

)

hold time (h)

2.5 ·1016 D cm-2 remainsafter hold

or: D/Be = 0.2%

extrapolating to 690 m2 ITER main wall: 2.5 g D per µm remains after 24h @ 350°C



modeling D release with diffusion and trapping

implemented mechanisms:

- diffusion of D through BeD with

- trapping and de-trapping of D at 2 binding sites (one with, one without re-trapping)

CRDS-code (“coupled reaction diffusion system” by Matthias Reinelt*) = flexible code for simulation of such systems

Simple model: diffusion coupled with de-trapping (and re-trapping) from binding sites

D2

D

ED, nD

ET2, nT2

ET1, nT1

D0 e-ED/kBT

e-ET/kBT




CRDS-code (“coupled reaction diffusion system” by Matthias Reinelt*) = flexible code for simulation of such systems

Simple model: diffusion coupled with de-trapping (and re-trapping) from binding sites

0 1000 2000 3000

0.01

0.1

1

D fl

ux (1

015 D

cm

-2s-1

)

time (sec)

experiment model temperature

0.3 K

/sec

nD = 4 1017 D/cm2

0.01

0.1

1

100

200

300

400

500

600

700

tem

pera

ture

(°C

)

implemented mechanisms:

- diffusion of D through BeD with

- trapping and de-trapping of D at 2 binding sites (one with, one without re-trapping)



discussion of the fit:

- although ramp better described

- hold @ 240°C: model incorrectly predicts too much desorption during the hold


0 2000 4000 6000 8000

0.01

0.1

1

240°C

D fl

ux (1

015 D

cm

-2s-1

)

time (sec)

650°C


240°C hold

0.01

0.1

1

100

200

300

400

500

600

700

800

tem

pera

ture

(°C

)

analysis of independent measurements simultaneously




- ramp and hold desorption behaviorfairly well reproduced

- model shows sharper desorptionpeaks than observed in experiment

- outgassing above 450°C to fast insimulation

0 2000 4000 6000 8000

0.01

0.1

1

D fl

ux (1

015 D

cm

-2s-1

)

time (sec)

240°C hold

0.01

0.1

1


100

200

300

400

500

600

700

800

tem

pera

ture

(°C

)

present status:

Next step: implementing decomposition

tool to interpolate



summary

magnetron sputtered Be/D films are suited as a model system to study retention and release

release of D from Be at 240°C and 350°C has a very large time constant

even a several hours bake at 240°C or 350°C does not release all D(a D/Be of 0.8% or 0.2% remains, respectively)(deeper trap sites that cannot be drained)

codeposits grown in different ways (energies, growth rates) show similar release features

the total amount of D released above 350°C is the same for all codeposits investigated, equivalent to D/Be ≈ 1%

multilayer codeposits show the same release as single layer codeposits (BeO interface is no transport barrier)



Thanks to:

PISCES team:

Russ DoernerMatt BaldwinDaisuke NishijimaTimo DittmarJonathan YuTyler LynchKarl UmstadterRay SeraydarianLeo ChousalRolando Hernandez