Outline of the report Relevance of the topic Experimental limitations and our solutions Results on permeation measurements of Be coated Eurofer samples.

Outline of the report

• Relevance of the topic

• Experimental limitations and our solutions

• Results on permeation measurements of Be coated

Eurofer samples

• Conclusions and further plans

Activities of the Vacuum lab in relation to hydrogen /

deuterium gas interaction on ITER grade: stainelss

steel, Be and W

Study of hydrogen diffusivity, solubility and surface

parameters which deteremine the kinetics, by using:

1) infusion / outgassing techniques (until 2009)

or (and)

2) membrane techniques (from late 2009)

Any of them requires a careful selection of all experimental

details to get reliable results. Measurements consist of

time consumtion, failures, non-reproducible runs etc...

Processes involved in the hydrogen isotope retention and recycling in fusion reactor plasma-facing materials.

R. A. CAUSEY, J. Nucl. Mater., 300, p.91 (2002).

Relevance of permeation of H/D/T in fusion reactors

Experimental setup

permeation method using H2 or D2 (3 UHV chambers)

kovar glass and alumina thimbles:

infusion / outgassing techniques

(2 UHV chambers)

Instead of applying the more common

dynamic method, where the ion current of a characteristic

mass number applying mass spectrometer is recorded

we improved

the static method (gas accumulation), pressure recorded by

non-ionizing gauges in a small volume followed by

precise mass spectrometry

Classical permeation experiment using hydrogen or deuterium

2 CM heads:

1 mbar &

0.05 mbar FS pmin = 310-11 mbar

inverted

magnetron

All metal UHV system:- pressure rise measurement (vol. 0.5L)

- inert gauges

- post MS1 bar CM head

Vacuum system for permeation measurements using gas accumulation method

Membrane technique [permeation]• determine coefficients like

– permeability (P), solubility (S), diffusivity (D) etc.– what is the limiting step for hydrogen migration

• samples investigated in 2010:– Eurofer steel (40mm diam, 0.5mm thick) dr.R.Lindau, IMF-FZK

– Be coated Eurofer by TVA at NILPRP (dr.Lungu)– W coated Eurofer at NILPRP (dr.Ruset) High-Z

Conventional permeation cell to test a membrane in the

form of a disc

Vacuum side

High pressure side

Au gasket

40 mm O.D. membrane

Massive flanges guarantee uniform load of the seal and seems obvious

Thin-walled permeation cell low background H contribution

Vacuum side – pLO

High pressure side – pHI, H2

Au gasket

40 mm O.D. membrane

Thin-walled inset ~ 0.3 mm,

A ~ 66 cm2

Flanges are

not a part of

the UHV

All permeation measurements have been performed at 400°C.

Be / W coated surface is facing the high pressure (upstream) side.

External outlook of the permeation cell

Thin-walled inset – downstream side

Upstream side

Results on the cell testing A comparative test was done by two identical AISI316

membranes, 2 h to 400 °C and further heating at 400 °C for 72

hours. QMS analysis of accumulated gases done occasionally.

Conventional cell with massive flangesd = 12 mm thick area & A1 15 cm2;

d = 0.3 mm thin area & A2 20 cm2,

d = 0.5 mm membrane & A2 8.5 cm2

Thin-walled celld = 0.3 mm thin area & A2 66 cm2,

d = 0.5 mm membrane & A2 8.5 cm2

0.E+00

1.E-05

2.E-05

3.E-05

4.E-05

5.E-05

6.E-05

0 10000 20000 30000 40000 50000 60000 70000 80000 90000

t / s

mb

ar L

/s

thick wall

thin wall

Comparison of the conventional with the thin-walled cell

Outgassing rate at 400°C in the first 24 h

Comparison of the conventional with the thin-walled cell

Outgassing rate at 400°C in 5 - 24 h

30000 40000 50000 60000 70000 80000 900001E-8

1E-7

1E-6

1E-5

30000 40000 50000 60000 70000 80000 900001E-8

1E-7

1E-6

1E-5

thick wall cell

thin wall cell

q / m

ba

r L

/s

t / s

Comparison between conventional and improved cell

• Sensitivity of the setup increased, approx. 380 times

much lower permeation fluxes could be recorded.• Preparation time of the cell to achieve this value was

substantially shorter, 3 days instead of several weeks.

• Error due to the permeation of H2 to the air not observed

even at p = 0.1 mbar.• Minimum detectable flux j = 1.7×10-9 mbar L/(s cm2) or

4×1010 H2/(s cm2)

V.Nemanic, B.Zajec, M. Zumer, JVST A, 28(4) (2010) 578

Permeation measurements through Be coated

Eurofer membranes prepared by TVAIdea and motivation• Bulk Be sample tiles tested at JET successfully • Inconel coated Be tiles tested successfully • Similar tiles may be applied at ITER

What is the permeation rate of hydrogen H2/D2/T2 through such

Be films? Published bulk data scattered!

If porous, they may increase the tritium retention in Inconel

since both neutral gas and plasma enter its surface easier

JET ITER-like wall experiment

ITER

100m2 Tungsten • Low erosion • high melting T • Negligible T retention

Optimise lifetime & T- retentionBut high Z & melting

700m2 Beryllium first wall• low Z• Oxygen getter

Optimise plasma performanceBut large erosion & melting

50 m2 Graphite CFC • Lowish Z• No melting in transients• Superior heat shock behaviour

Optimise heat flux resistanceBut large erosion & T retention

W

CFC

JET

Published data on solubility and diffusivityfor hydrogen (H,D,T) in beryllium

• published data on diffusivity and solubility very

scattered and almost useless for prediction of results

(A.A. Pisarev, Fusion Techn., 28, (1995) 1262)

• no data about hydrogen amount in our samples available

• a few reports on the same Be quality found as a rough

guidance for scheduled measurements

Published data on solubility and diffusivity

A. Nikroo, H.W. et al, Fusion Science and Technology, 51, 4 (2007) 553-555

The only report we could

find on Be films

reveals that Be shells

are porous

Permeation of Be/Eurofer prepared by TVA (SEM)

As received Slightly polished by alumina

(1 m, 300 nm, 60 nm grains)

Be thickness 8 m, Eurofer 500 m

Permeation of Be/Eurofer prepared by TVA (AFM)

As received

RMS = 140 nm

Polished by alumina

RMS = 2 nm – 6 nm

Permeation of Be/Eurofer - fluxes

Sample flux, q PRF

H2/(s cm2)

1 3.21012 140

2 2.11011 2184 ?

3 2.81013 16

4 3.11013 15

5 9.01012 50

Results at 1 bar upstream pressure and 400 °C obtained

a short period after hydrogen was introduced

(5 of 8 samples displayed, 3 were too rough or failed)

Is there any reason for great scattering of the flux values?

Additional 2 samples with Ag interlayer below Be:

• Be coated membranes troublesome for permeation measurements:– Au gasket did not tightly seal the membrane fine polishing

– Sealing improved but not leak-tight (~10-5 mbarL/s) Be film is porous

for He between Be and Au. Is it porous for H2 too?

UHV [high pressure side]

UHV UHV

UHV [high pressure side]

Ag film

much better, but still leak

Permeation of 2 additional samples Ag film+Be

Sample flux, q PRF

H2/(s cm2)

6 9.01012 50

7 5.11012 83

Results at 1 bar upstream pressure and 400 °C

Is there any general reason for scattering of the flux values

for 8 µm Be prepared by TVA at the same bias?

The behaviour also different from bare membranes!?

Eurofer coated with Be:

at t = 0

pHI 0 1 bar

time / s

j /

10

-7 m

bar

H2 c

m-2 s

-1

Permeation reductionfactor (PRF) of steadyfluxes

coated

bare

j

jPRF

Permeation rate vs. time for 4 Be coated membranes (first exposure to 1 bar H2)

PRF=13

PRF=22

PRF=42

PRF=110

• Large scatter in kinetics (shape of curves) & PRF among identical samples.

• Subsequent H2 exposures more consistent but still large variations among samples. Steady flux for bare membrane achieved in 10s

Eurofer coated with Be:

10 20 30 400

1x10-7

2x10-7

3x10-7

4x10-7

92

188

437

724

1086

1294

pHI

0.5 / mbar0.5

j /

mol

H2 c

m-2 s

-1

permeation is limited by diffusionDiffusion limited regime (DLR)

HIpj mbar

mbar

Eurofer coated with Be:

porosity leak cannot be completely eliminated small air leak always present air could oxidize Be layer

Intentional oxidation by 1 bar of air (10min) increases the permeation rate for 2-5 times.

XPS profile of polished & oxidized Be sample

Eurofer coated with Be: could there be small holes (pinholes) in Be film ?

Low-angle (2°)polishing wedge

Imperfections doexist, but difficultto find.

Sample has ~cm2

TC29

Non-loaded

SEM by FZJ-IEF sent by dr.Lungu

Eurofer coated with Be: what can be deduced from permeation meas.?

PRF = ~10 - 100

Assume Be film is homogenous and dense (no holes) & DLR:

d1

P1

d2

P2

H

Pd

dPPRF

dddP

d

P

d

P

d

2

2

212

2

1

1 ,

P-1 R from electr.Similar to parallel resistorsand parallel insulation layers.However: H conc. is not continious

Permeation coef. P:

Time lag L:

0 500 1000 1500times0

10

20

30

40

50

60

pOL

Single layer:

Two layers:

Ash, Barrer, Palmer; Brit. J. Appl. Phys., 1965, VOL. 16, p.873

L

j must monotonically increase tothe steady state value

Eurofer coated with Be: what can be deduced from permeation meas.?

PRF = ~10 - 100

Assume Be film is homogenous and dense (no holes) & DLR:

PRF=10 PBe = 3.0x10-14 mol H2 / m2 s Pa0.5

PRF=100 PBe = 2.7x10-15 mol H2 / m2 s Pa0.5

P = D × SD = ?, S = ?

Measured time-lag not reliable to obtain D.

Take solubility S from published data:

400°C

SBe ~ 0.002 mol H2 / cm3 Pa0.5

, DBe = 1.5x10-7 cm2/s, L2= 26s

, DBe = 1.4x10-8 cm2/s, L2= 96s

400°C

DB

e /

cm

2s

-1

1000/T / K-1

Permeation results – conclusions and plans

• Permeation flux was measured on 10 samples

• Porosity on the microscopic scale was expressed as

intight seal (He) on Be/Au interface or in the bulk Be below Au

• The SEM evidence on porosity (pinholes) given (FZJ and JSI),

but confirmation that a low number of pinholes is indeed

responsible for the scattering in permeation flux could not be

given at this moment • W / Be films will be tested and compared to Be and W films • comparison to various models is already in progress

Eurofer coated with Be: what can be deduced from permeation meas.?

PRF = ~10 - 100

Assume Be film is homogenous and dense (no holes) & DLR:

• Measured PRF are in agreement with published range of Be D and S

• Observed time lag is considerably longer than predicted (26-96 s)

Eurofer coated with Be: what can be deduced from permeation meas.?

PRF = ~10 - 100

Assume Be film is impermeable, permeation occurs only through pinholes

H

d

,)18.11(

1

rdFPRF

r – pinhole radius

d – substrate thickness (500μm)F – fraction of free surface

W. Prins and J.J. Hermans, J. Phys. Chem., 63 (1959) p. 716

1

3.0/

F

rdpinhole

pinhole

pinhole

With known PRF we cannotdecouple r and pinhole area density n.

Pinholes with smaller radius coverlesser fraction of the membrane surface.

Equilibrium H concentration in the substrate at the bottom of the pinhole.

10-5 10-4 10-3 10-21

10

100

1000

small pinholes

PR

F

F

d/r = 100 d/r = 1000 d/r = 10000

large pinholes

Permeation through clean Eurofer

Could the air leak, Be porosity and permeation results be related?

Eurofer membrane was oxidized in water vapor and subsequently

also in the air. Tested at 1 bar (air) & 400 °C.

Duration, conditions PRF

Clean Eurofer well outgassed 1

Oxid. H2O, 400°C ~25 mbar, satur. at 25°C,18h 13.6

Oxid.air 400°C 1 bar, 18 h 0.75

Porous iron oxide even increases the permeation flux,

but Cr rich oxide seems to suppress it.

Permeation Be/Eurofer - searching for explanation

• A few Be coated Eurofer membranes were also oxidized in the air

at 1 bar and 400 °C.

• Results do not express so evident influence of the air, since all

samples have been previously (inevitably) exposed to air leaks

when a thin but dense oxides could be formed.

• The expected suppressing of hydrogen flux was not achieved as

expected since BeO should be even less permeable than Be

• It seems that Eurofer is exposed to the air in pinholes or voids in Be• Some facts should be verified by a new experiment.

01002003004005006000

0.5

1

1.5

2

2.5

Binding Energy (eV)

Nor

mal

ized

Inte

nsity

-C

1s

-A

r2p

-B

e1s

-O

1s

-O

2s

-B

e1s

Surface

After sputtering,subsurface region

-O

1s

XPS survey spectra on polished Be sample

104106108110112114116118120122124

0

0.5

1

1.5

2

2.5

x 104

Binding Energy (eV)

Inte

nsity

(c/s

)

dept

h

Be-oxide Be-metal

Be 1s

XPS depth profile of polished Be sample

0

10

20

30

40

50

60

70

80

90

100

0 20 40 60 80 100 120 140

Estimated depth (nm)

Co

nce

ntr

atio

n (

at.

%)

O

C

Be

Oxide film

non-zero O conc.:influence of roughness

XPS spectra Be 1s as a function of depth

X-ray photoelectron analysis -XPS• XPS: very surface sensitive technique

• XPS depth profiling (by Ar ion sputtering) => in-depth distribution of elements

Be covered by Be-oxide layer Be-oxide layer thickness ~ (35 ± 10) nm, native only 2-3 nm

Permeation of oxidized Be/Eurofer (SEM)

Polished oxidized Be heavily oxidized Eurofer