SANDIA REPORTSAN D99-8222Unlimited ReleasePrinted February 1999

Proceedings of the 4thInternationalWorkshop on Tritium Effects in PlasmaFacing Components

~~@l~F~@zgm

R. A. Causey Qsr/

SF2900Q(8-81 )

●

“

Issued by Sandia National Laboratories, operated for the United States Depart-ment of Energy by Sandia Corporation.

NOTICE: This report was prepared as an account of work sponsored by anagency of the United States Government. Neither the United States Govern-ment, nor any agency thereof, nor any of their employees, nor any of theircontractors, subcontractors, or their employees, make any warranty, express orimplied, or assume any legal liability or responsibility for the accuracy,completeness, or usefulness of any information, apparatus, product, or processdisclosed, or represent that its use would not infringe privately owned rights.Reference herein to any specific commercial product, process, or service bytrade name, trademark, manufacturer, or otherwise, does not necessarilyconstitute or imply its endorsement, recommendation, or favoring by the UnitedStates Government, any agency thereof, or any of their contractors orsubcontractors. The views and opinions expressed herein do not necessarilystate or reflect those of the United States Government, any agency thereof, orany of their contractors.

●

Printed in the United States of America. This report has been reproduceddirectly from the best available copy.

Available to DOE and DOE contractors bornOffice of Scientific and Technical InformationP.O. BOX 62Oak Ridge, TN 37831

Prices available from (615) 576-8401, FTS 626-8401

Available to the public fromNational Technical Information ServiceU.S. Department of Commerce5285 Port Royal RdSpringfield, VA 22161

NTIS price codesPrinted copy: A04Microfiche copy: AO1

.

.

DISCLAIMER

Portions of this document may be illegiblein electronic image products. Images areproduced from the best available originaldocument.

*

.

●

SAND99-8222UnlimitedRelease

Printed February 1999

Proceedings of

The 4th International Workshop on TritiumComponents

Santa Fe, New MexicoMay 14-15,1998

Rion A. CauseySurface Chemistry Department

SandiaNational LaboratoriesP.O. BOX 969

Livermore, CA 94551-0969

Abstract

Effects in Piasma Facing

The 4th International Workshop on Tritium Effects in Plasma Facing Components washeld in Santa Fe, New Mexico on May 14-15, 1998. This workshop occurs every twoyears, and has previously been held in Livermore/California, Nagoya/Japan, and the JRC-Ispra Site in Italy. The purpose of the workshop is to gather researchers involved in thetopic of tritium migration, retention, and recycling in materials used to line magneticfbsion reactor walls and provide a forum for presentations and discussions in this area.This document provides an overall summary of the workshop, the workshop agenda, asummary of the presentations, and a list of attendees.

3/4

.

.

.

.

.Contents

Workshop Summary

Workshop Agenda

Presentations

W. Jacob and B. Landkammer, “Removal of Codeposited Layersby ECRDischarge Cleaning”

H. Grote, W. Bohmeyer, P. Kornejew, H.-D. Reiner,G. Fussmann, R. Schlogl, G. Weinberg, C.H. Wu, “FluxDependence of Chemical Sputtering Yields”

E. Vietzke, “D, O Removal from a-C:D and a-C:C)Films byAtomic Hydrogen”

W. Jacob, “Basic Mechanisms of the Deposition of CodepositedLayers”

R. Anderl and R.J. Pawelko, “BET Surface Area and Tritiumhalysis for TFTR Carbon Dust Samples”

.& . R.P. Doemer and PISCES Experimental Team, “Mixed-MaterialEffects on Plasma-Facing Components”

R. Rolli, H. Werle, C.H. Wu, “Influence of Neutron Darnageon the Tritium Retention of ITER Representative Carbon FibreComposites (CFC’S)”

M. Rubel, V. Philipps, T. Tanabe, “Deuterium Retention andDistribution on Tungsten Test Limiters from TEXTOR”

M. Rubel, A Vevecka-Priflaj, J. von Seggern, V. Philipps“Deuterium Analysis in Thick Codepposited Layers on PlasmaFacing Components”

11

13

16

18

20

23

26

31

33

I

D. Hildebrandt, M. Akbi, B. Juttner, W. Schneider, “Influenceof Surface Roughness on the Deuterium Inventory ofASDEX-Upgrade Devertor Tiles”

R. Wampler, “Analysis of Invessel Components in TFTR Priorto DT Operations and in DIII-D”

C.H. Skinner, “Tritium Experience in Large Tokamaks;Applications to ITER”

Y. Hirohata, T. Nakamura, Y. Aihara, T. Hino, “Dependenceof Stiace Oxidation on Hydrogen Absorption and theResorption Properties of Ti-6Al-4V Alloy”

K. I&u, and T. Tanabe, “Co- and Counter-permeation ofHydrogen and Deuterium in Metals”

K. Yamaguchi, K. Ohkoshi, O. Onoue, M. Yamawaki, “Releaseof Ion-Driven Hydrogen Isotopes from Molybdenum”

T.A. Burtseva, I.V. Mazul, A.E. Gorodetslq, A.P. Zakharov,N.N. Shipkov, “Study and Development of Carbon-BasedMaterials with Reduced Hydrogen Isotope Retention”

T. Tanabe, “Reconsideration of Hydrogen Recycling Processat the First Wall”

V. Kh. Alimov, V.N. Chernikov, A.E. Gorodetsky, A.P.Zakharov, “Depth Distribution of Deuterium Atoms andMolecules in Beryllium Oxide Implanted with D Ions”

T. Horikawa and K. Morita, “Isotope Difference in ThermalRe-emission of Hydrogen Implanted into WC Layers Depositedon Graphite”

List of Attendees

Distribution

35

37

38

39

41

42

44

47

50

53

56

61

.

6

.

a

.

Summary

4th International Workshop on Tritium Effects in Plasma Facing Components

The 4th International Workshop on Tritium Effects in Plasma Facing Componentswas held on May 14/15, 1998 at Rancho Encantado located near Santa Fe, New Mexico.This workshop provides a forum for presentations and discussions on tritiurn migration,retention, recycling, and other effects on materials used to line magnetic fusion reactors.The workshop is held every two years, immediately before or after the InternationalConference on Plasma Surface Interactions in Controlled Fusion Devices. Theinternational committee for the workshop is composed of Dr. R. C!awey, Sandia NationalLaboratories/US, Dr. C. Wu, Max Pkmck Institute for Plasma Physics/Germany, andProfessor M. Yamawaki, University of Tokyo/ Japan. Previously, the workshop hasbeen hosted by Dr. R. Causey in Livermore, California, by Professor S. Morita at NagoyaUniversity, and by Dr. A. Perujo at JRC-Ispra Site.

Twenty six participants from the US, Germany, Japan, Russia, and Sweden attendedthis year’s workshop. While the topics were varied, most of the presentations dealt withtritium migration and retention in metals, carbon, mixed materials, and the codepositedlayer. There were also open discussions on tritium inventory problems associated withcarbon use in tokamaks and on the fiture of tritium research for fusion energy.

The next International Conference on Plasma Surface Interactions in Controlled FusionDevices will be held in Europe. It is presently anticipated that the 5th InternationalWorkshop on Tritium Effects in Plasma Facing Components will be hosted either inGermany or Sweden.

II

Agenda4thInternationa1 Workshop on TritiumEffectsin PIasmaFacing

Components

Day 1

7:00 to 8:45 Breakfwt

8:30 to 9:00 Continue Registration

9:00 to 9:20 Opening remarks and announcements

Session 1: Carbon and codeposited layers Chairman-M. Rubel

9:20 to 9:40

9:40 to 10:00

10:00 to 10:20

10:20 to 10:40

10:40 to 11:00

Il:ooto 11:20

ll:20to 11:40

11:40 to 12:00

W. Jacob, Max-Planck-Institute for Plasma PhysicsRemoval of Codeposited Layers by ECR Discharge Cleaning

H. Grote, Max-Planck-Institute for Plasma Physics, BerlinChemical Sputtering of CFC’s at High Ion Flux Densities

E. Vietzke, KFA Institut fuer PlasmaphysikRemoval of Hydrogen and Oxygen Containing Carbon Films byAtomic Hydrogen

Break

W. Jacob, Max-Planck-Institute for Plasma PhysicsBasic Mechanisms of the Deposition of Codeposited Layers

R. Anderl, Idaho National Engineering LaboratoryBET Surface Area and Tritium Analysis for TFTR Carbon DustSamples

R. Doemer, University of California, San DiegoBe/C/W Mixed-material Investigations

Chung Wu, NET Team, Max Planck Institute of Plasma PhysicsInfluence of Neutron Damage on the Tritium Retention of XTERRepresentative Carbon Fibre Composites (CFC’S)

.

.

8

1:40 to 2:00

2:00to 2:20

2:20 to 2:40

2:40 to 3:00

12:00 to 1:20 Lunch, Rancho Encantado

Chairman-E. V]etzke

1:20 to 1:40 M. Rubel, Royal Institute of Technology, StockholmDeuterium Retention in the Codeposits on the Tungsten Limitersused in TEXTOR

D. Hildebrandt, Max-Planck-.Tnstitute for Plasma PhysicsInfluence of Surface Roughness on the Deuterium Inventory ofASDEX-UPGRADE Divertor Tiles

Bill Warnpler, SandiaNational LaboratoriesRetention of Deuterium and Tritium in Tokamaks

Break

C. Skinner, Princeton Plasma Physics LaboratorySummary of the IEA Workshop on Tritium Experience in LargeTokamaks

3:00 to 4:00 All Chairman-G. LonghurstDiscussion of tritium inventory problems associated with carbonuse in tokamaks

6:00 to 10:00 AllDrinks and dinner at the “barn”

Day 2

7:00to 9:00 Breakfmt 7

Session 2: Metals Chairman-T. Tanabe

9:00 to 9:20 Y. Hirohata, Hokkaiclo UniversityDependence of Surface Oxidation on Hydrogen Resorption andAbsorption Properties of Ti-6Al-4V “Alloy

9

9:20 to 9:40 K. K&u, Nagoya UniversityCo- and Counter-permeation of Hydrogen and Deuterium inMetals

9:40 to 10:00 K. Yamaguchi, Universi@ of TokyoRelease of Ion-driven Hydrogen Isotopes from Molybdenum

10:00 to 10:20 Break

Session 3: Other Chairman-C. Wu

10:20 to 10:40

10:40 to 11:00

ll:ooto 11:20

ll:20to 11:40

11:40 to 12:00

12:00 to 1:30

1:30 to 3:00

3:00

T. Burtseva, D.V. Efiemov InstituteStudy and Development of Carbon-Based Materials with ReducedHydrogen Retention

T. Tanabe, Nagoya UniversityReconsideration of hydrogen recycling at the frostwall

S. Ohtsu, Universi& of TokyoIsotope Effects of Particle Refection on Linear Plasma Facility

V. Alimov, Institute of Physical ChemistryDeuterium Retention in Beryllium Oxide Irradiated with D-ions

K. Morita and T. Horikawa, University of NagoyaIsotope Difference in Thermal Re-emission of Hydrogen Implantedin WC Layers Deposited on Graphite

Lunch Rancho Encantado

Open forum on tritium in tokamaks and fusion reactor materialsresearch Chairmen C. Skinner and R Causey

Adjourn

.

.

.

10

Removal of Codeposited Layers by ECR Discharge Cleaning

W. Jacob and B. Landkammer

Max-Planck-institute for Piasma Physics, D-85748 Garching, Germany, EURATOM Association,

The deposition and erosion of amorphous hydrogenated carbon (a-C:H) films wasinvestigated in laboratory experiments. Plasma-deposited a-C:H films with a widerange of physical properties, such as stoichiometry, hardness, and opticalproperties, were deposited from a methane electron cyclotron resonance (ECR)low-pressure discharge. These a-C:H films represent a model system for thecodeposited layers found in fusion experiments. Depending on depositionconditions, these films can be soft, polymer-like or dense and hard [1-4]. Hard a-C:H films are deposited if energetic ions with energies above about 50 eVpaflicipate in the deposition process and at low-energy ion bombardment (E <20-30 eV) polymer-like, soft films are produced. Hard and soft films are easilydistinguished by their mechanical properties. Hard films are very scratch resistantwhile soft films can easi}y be scratched with a metallic pin. The films showsignificant differences in almost ail of their physical properties. For example, harda-C:H has typically an H/C ratio of 0.4, a density around about 1.5-2 g cma, and arefractive index, n, of 2 and higher. Soft films, on the other hand, possess an H/Cratio of 1 and higher, densities down to 1 g cm+, and n around 1.5 [1-4]. However,the transition between soft and hard films is gradual and prevents a cleardistinction.

Real-time in-situ ellipsometry was applied to measure the optical response of thelayers during deposition and erosion. This allows a precise determination of theoptical constants and the film thickness. From these data the deposition anderosion rates, respectively, are calculated [5].

To erode carbon films and deplete the T concentration in the layers anexperimental investigation was performed using 02, D2, H2, H20, and 02/H2 asworking gases [5,6].The dependence of the erosion rates on substrate temperature, applied ion energy,gas composition and discharge pressure for two different types of a-C:H films wasmeasured. The erosion rates increase with increasing ion energy (at T = 300 K)and with increasing substrate temperature (at Us~ = O V). However, at UsB = 100 Vno increase of the rate with T was found. This indicates that the simultaneousincrease of T and USE does not lead to a simple multiplication of the enhancementfactors observed when only one parameter is changed. On the other hand, the ratemeasured at UsB = 100 V and T = 300 K is already much higher than the rate at UsB=,0 V and T = 650 K showing that an increase of the bias voltage is much moreefficient in increasing the erosion rates than an increase of the substratetemperature.

The following gases and gas mixtures were investigated using identical externalplasma parameters: 02, D2, t+2, HZO, and 02/H2 (1:2). In all investigated casesoxygen shows always the highest rate. At T = 300 K the relative rates behave as Oz :Dz : H2 = 10:2: 1.The highest rates achieved in oxygen discharges at 300 K with

11

about 1) and 1.7 yin/h for hard a-C:H films (H/C about 0.5). The erosion ratesincrease with increasing substrate temperature (roughly by a factor of 3-4, 300 to650 K) and with increasing ion energy (by about a factor of 6-8, 0 to 200 eV).

.

Best conditions for a fast and effective removal of tritium from codeposited Iayersare an oxygen plasma at low pressure to achieve high ion fluxes and at increased

.

wall temperature. At components with a high codeposition rate, e.g. in shadedregions of the divertor, the application of a DC self-bias to increase the ion energy,if technically feasible, could generate locally higher erosion rates. At a bias voltageof 200 V, corresponding to ion energies slightly above 200 V, we measured in our

setup erosion rates of 9.9 yin/h and 21.6 Urn/h for hard and soft a-C:H films,respectively [6]. We anticipate that a hydrogen plasma applied after an oxygencleaning discharge will efficiently decrease the oxygen content in the firstmonolayer. But this is a process that has to be investigated in more detail.

increased wall temperatures (e.g. 600 K) offer the possibility to removecodeposited layers by thermally-induced chemical erosion with atomic hydrogen[3,6,7]. This process relies only on the efficient production of atomic hydrogen in theplasma and does not require direct line-of-sight with the cleaning plasma. It is onlyabout a factor of 3 slower than oxygen at these high temperatures and low ion

.

energies. It offers the advantage that no detrimental effects on other-than-carbonmaterials have to be anticipated. However, much more work has to be done toallow a finally assessment of the possibilities and shortcomings of each option.

.

References

[1] P. Koidl, C.Wild, R. Lecher, and R.H. Sah, in ‘Diamond and Diamond-Like Films

and Coatings’ R. E.Clausing, L.L. Horton, J.C. Angus, and P. Koidl (Eds.), NATO-

ASI Series B, 266, Plenum, New York (1991) 243.

[2] A. von Keudell, W. Jacob, and W. Fukarek, Appl. Phys. Lett. 66 (1995) 1322.

[3] A. von Keudell and W. Jacob, J. Appl. Phys.79(1996) 1092.

[4] A. von Keudell and W. Jacob, J. Appl. Phys. 81 (1997) 1531.

[5] B. Landkammer and W. Jacob, J. Nucl. Mater, submitted.

[6] W. Jacob, B. Landkammer, and C.H. Wu, (Proceedings PSI 13, San Diego 1998)

J. NUCL Mater, submitted.

[7] W. Jacob, Thin Solid Films, in print (1998).

12

Flux dependence of chemical sputtering yields

.

.

.

.

.

H. Grote, W. Bohmeyer, P. Kornejew, H.-D. Reiner, G. Fussmann,R. Schlogl#, G. Weinberg#, C. H. WU*

Max-Planck-Instituter Plasmaphysi~ Bereich Plasmadiagnostik, EURATi?MAssociution, Mohrenstr. 41,D-10117 Berlin, Germany

#Fritz-Haber-Institut der Max-Planck Gesellschq?, Faradayweg 4, D-14195 Berlin, Germany*NET Team, Max-Planck-Institut @r Plasmuphysik, EURATOM Association, Boltzmannstr. 2,

D-85748 Garching, Germany

The plasma generator PSI- 1 was used for experiments in steady state plasmas to investigate the chemicalerosion of graphite and carbon fibre composites (CFC) [1,2].

CFC-samples of CONCEPT II (DUNLOP) and silicon-doped NS 31 (SEP) as well as fine grain graphiteEK 98 (Ringsdorff - Werke GmbH) - for comparison - were exposed to hydrogen and deuterium plasmas afterappropriate cleaning and outgassing procedures.

The electron density Q was varied between (0.05 - 2.5)*1018 m-3, the electron temperature T. was kept atabout 5-10 eV, in order to realise a particle impact energy onto the samples of less than 30 eV, i.e. below ornear the threshold for physical sputtering [3]. Higher density values are usually correlated with the higher

temperatures.Sotheionfluxdensitywasvariedin therange (0.1 - 5)*1022 m-2s-l. During exposure the surfacetemperature of the samples passed the maximum for chemical sputtering, reached 850-900 K about 15 min afterthe beginning of the exposure and was fixed at this level by active cooling. During temperature rise and the whole

exposure time the axial dependence of the CH- (CD-) band intensity at431 nm and the H7- (Dy-) line weremeasured with up to 8 channels in front of the sampIe (5 -30 mm). The CH- (CD-) intensity appeared to beindependent of position and can thus be taken as a measure of the CH- (CD-) density. The analysis then used the

ratio ~(CH)dA / \(HY)dA or ~(CD)dk / J(DJdl to compensate for possible transmission changes. Furthermore, the

CXHY-(CXDY-)formation was monitored by using a differentially pumped gas analyser. Both methods werecalibrated for each ion flux density with the known fluxes of Cm (CD4), the main product of chemical erosion[5]. These gases were blown into the discharge either through a hole in a molybdenum target or through a nozzlenear the target under almost identical conditions as for the graphite sample exposure. This is a direct calibrationbecause no assumptions concerning photon eftlciencies and fractions of hydrocarbons reaching the massspectrometer have to be taken into account. To reduce the background, the amount of graphite exposed to theplasma was minimised. In addition to the in-situ methods weight loss measurements and scans with an opticalprotilometer over the etched crater area were carried out to determine the eroded mass. Finally, surface analysismethods (SIMS, AES, SEM and EDX) were used to check for possible impurities deposited on the samplesurface and to detect morphological changes.

Within the error bars of the three different methods of analysis there is no difference between theerosion yields in hydrogen and deuterium. Therefore, in what follows we will not distinguish between exposuresin hydrogen and deuterium. Within 4070 of the average value more than 90 ‘ZO of all single measurements are

found. The decrease of the erosion yield with incoming ion flux density is fitted by a power law of Y cc r- 0“6.

The differences in the erosion yields for the investigated specimen are shown in Fig. 1. The decreasewith increasing ion flux density is the same for all materials indicating the same underlying physical effect. Theerosion values for the silicon-doped CFC NS 31 are generally lower than for CONCEPT II and reach valuesbelow 1 % at the highest flux densities investigated. The values for EK 98 are almost identical with NS 31 andtherefore difficult to distinguish in the figure. The data at 3*102] m-2S-l show an unuswd large deviation from theoverall trend for both CFC-materials, These were the i%st in a series of experiments and are possibly erroneousdue to plasma impurities or poor sample conditioning. Therefore three additional measurements were carried outto proof these data, resulting in the data points at 1.4, 3.6 and 6.2*1021 m-2s-], that fit the overall trend very well.

EDX-spectra and SEM micrographs show a preferential erosion of the amorphous C-matrix and of thetransition region between matrix and fibre. Due to the silicon between the fibres the morphology of the Si-dopedmaterial is different before and especially after exposure to the plasma. There is also a preferential erosion of theamorphous C-matrix, within the matrix / fibre transition zone and on areas of lattice defects of the graphite.However, since Si and SiC are sputtered physically only, we observe an enrichment of silicon on the surface afterexposure.

13

The steep decrease of the erosion yield with increasing flux density is in good agreement withexperiments at TEXTOR [4] and ASDEX-Upgrade [5],Thedifferences in the absolute values can be explainedby the lower surface temperature at ASDEX-Upgrade. The measurements at TEXTOR give only the flux ratio

@(CD~)/@(D) rather than erosion yields. The analytic description by Roth and Garcia-Resales [6,7] predicts a

remarkable decrease of the erosion yield just for the range of ion flux densities investigated here. Neverthelessthe theoretical slope is still too small compared to our experimental results (Fig. 2). According to the model Ytimand Y~ti give almost equal contributions to the total erosion yield. A significant contribution of Y~ti, however,would result in a dependence of erosion on the mass of the incident particles. The lacking isotope effect in ourmeasurements is therefore also in disagreement with the model.

The differences in the erosion yields maybe explained as follows: Preferential erosion is observed atcrystal imperfections or amorphous areas in the CFC’s. This generally leads to higher erosion rates of thecomposites compared to crystalline graphite. The Si-doping partly compensates this effect, so that the erosionrate of the silicon-doped CFC NS 31 is similar to that of EK 98. To reduce the erosion yield of doped CFC’sbelow the values of graphite it seems necessary to improve the process of graphitisation of the CFC-matrixduring manufacturing.

Possible sources of errors in the determination of the erosion yields are the impact contributions ofneutral atoms and molecular ions. Furthermore impurity sputtering as well as redeposition of eroded materialmust be taken into consideration. The latter effect is particularly critical in the case of high u and T. when theeroded material is ionised near to the sample. The difference between total and net erosion would cause theerosion rates determined by the loss of weight method to be systematically smaller than those determined by theother methods. This, however, is not found in our data. Estimations show that for our conditions (plasmaparameters, geometry) about 4 % of the eroded material will return to the target [8]. Similarly, physical sputteringwould cause the loss of weight measurements to show significantly higher erosion yields than the in-situ methods,since the latter do not detect the sputtered carbon. The influence of plasma impurities, especially oxygen, on theerosion rates is discarded because of the very low concentration (nti% < 10-3) [2]. The impact of molecular ionsand neutral atoms additionally to ~- (D+-) ions would lead to an increase of the impinging particle flux andhence decrease the effective erosion yield. As far as the ion - neutral density ratio stays constant the general

parameter tendencies would be unaffected. The amount of neutral atoms was estimated to = 20 % of the electron

density [2], but further investigations will have to be performed to clear this issue.

References

[1] W. Bohmeyer et al., in: Proc. 22nd EPS Conf. Contr. Fusion and Plasma Phys., Bournemouth 1995, ECA 19CEur. Phys. Sot., Geneva, 1995, Pt. II, p. 297-300[2] H. Grote et al., J. Nucl. Mater. 241-243 (1997) 1152[3] W. Eckstein et al., Sputtering Data, Report IPP 9/82[4] A. Pospieszczyk et al., in: Proc. 22nd EPS Conf. Contr, Fusion and Plasma Phys., Bournemouth 1995, ECA19C Eur. Phys. Sot., Geneva, 1995, Pt. II, p. 309-312[5] A. Kallenbach et al., submitted to Nuclear Fusion[6] J. Roth and C. Garcia-Resales, Nuclear Fusion, 36 (1996) 1647[7] J. Roth and C. Garcia-Resales, Nuclear Fusion, 37 (1997) 897[8] D. Naujoks, private communication

14

.

.

.

.

.

15

0

+ A EK 98 H NS 31 + CONCEPTII

— IEK98 -- NS31 —CONCEPT II

1E+21 1E+22 -. 1E+23Ion flux density r [m-z s-’]

Fig. 1. Chemical erosion yield vs. incident ion flux density for three different materials.

D=>C — Ytota\(Eo=Eth)[Roth]T=T~a(r)

■ NS 31, average values

+I

1E+20 1E+21 1E+22 1E+23 1E+24Ion flux density 17[m-2s-1]

Fig. 2. Total erosion yield VS.ion flux density. Comparison of experimental data with the analytical expression

given by Roth and Garcfa-Resales [6,7]

15

D, 0 removal from a-C:D and a-C:O films by atomic hydrogen

E. Vietzke

Institut* Plasmaphys~ForschungszentrumJiilichGmbH,EuratomAssociation,

TrilateralEuregioCluster,D-52425JiilichGermany

There are not many ways discussed for tritium (T) removal from walls in fision devices,

Erosion or combustion of the surface layers and thermal resorption seems to be a possible

way since more than 99°/0of retained T or deuterium (D) was found in the first pm layers.

This was shown in flakes from protection limiters of the RF antennas in TEXTOR [1] and

also from flakes from JET limiters [2].

The experiments were perilormed in an atomic beam apparatus and the products were detected

by a line-of-sight quadruple mass spectrometer [3]. In the reaction of a thermal If’ beam

(2800 K) with a hard amorphous deuterium-containing carbon film (a-C:D) mainly the radical

CHSand a wide spectrum of hydrocarbons are formed [4]. The Cl, C2 and C3 hydrocarbons

dominate ( [CHI] : [CZHX]: [CSHX]= 1: 0.8:0.5 at 500 K), the higher hydrocarbons appear

only in traces. Since for the carbon erosion the number of carbon atoms in each molecule has

to be taken into account, the chemical erosion of a-C:H(D) films by HOis dominated by C2HX

and C3HXformation. The total chemical erosion yield reach its maximum at 750 K and a value

of around 0.1 C/HO(DO).The hydrocarbon spectrum and the chemical reaction yield is very

sensitive to the structure of the carbon film. In reaetions of II’ with hard a-C:D films the D

bonded in the film is nearly exclusively released in form of HD and only in traces in

hydrocarbons [4]. This behavior is understood from the model of Kuppers and his group [5]

and also observed in the vice-versa reaction Do with a-C:H. When polymerlike films are

eroded significantly more moIecules of mixed hydrogen isotope are formed and the erosion

yield is extremely increased (see table) [4].

Table 1: Some data of D“ reaction with RF plasma deposited a-C:H films (13.6 MHz, p = 3 Pa,

T =50 –150°C) with difTerent bias voltages at 520 K [4].

Fihn I ErosionType/Bias 1 Density 1 WC I Yield

very hard/1000V 1.80 0.30 0.014 C/D”

hard, 400 V 1.65 0,45 0.017 cm”

polymerlike, 100V 1.3 0.7 0.140 cm”

16

.

.

Spectrum

HD, C,DY.

HQ GDY

I-ID(?),CXHYDZ

,-

.

,..

H“ reaction on a-C:O film: The films were prepared by 5 keV 1s02+implantation in graphite

at room temperature up to saturation. The amount of implanted oxygen was determined by

thermal resorption. In the reaction of W’with this film hydrocarbons are formed in a similar

spectrum as described above. In additio~ also CO, H20 and C02 are liberated proportional to

the hydrocarbon formation. A&r a total erosion of the film all implanted oxygen was emitted

in form of CO, H20 and C02. The total erosion T during this procedure was determined by

summing up all components. The product ratios [CO]/T, ~20]/T and [C02]/T depend on the

temperature as shown in Fig. 1. This result clearly shows that kinetics is involved in forming

H20 and C02 with the maximum at 450 K. The CO formation is nearly temperature

independent up to 1000 K where the resorption of the implanted oxygen starts.

Summarv: By a H’ impact on a-C:D and a-C:O films the liberation of bonded D or O in these

films occurs only by erosion of the film. k-the case of a-C:H the bonded D is mainIy emitted

as HD, whereas in polymerlike films D is mainly emitted in hydrocarbons (as CXHYDZ).The

bonded O is temperature dependent emitted as CO, H20 and COZproportional to the formed

hydrocarbons.

Referenem

[1]V. Philippa,E. Vietzke, H.-G. Esser et al.,

Proe. 13* PSI san Diego 1998.

[2] R.-D. Penzho~ private conununication 1998,

[3] E. Vietzke, K. Flaskamp, V.Philipp~

J. Nucl. Mater. 111-112 (1982) 763.

[4] E. Vietzke, K. Flaskamp, V.Philipp~

J. I&d. Mater 145-147 (1987) 443.

[5] A. HOW A. Sche& J. Biener et al.,

App]. Phys. L&, 61 (1992).

s

6

4

3

2

1

0

A

260 600 760 1000

T[~~

Fig. 1 The releasedoxygencompoundsnormalizedtothetotalerosionasfunctionof the target temperature in tbe

reactionof1? withana-C:Ofilm.Thefilmwaspreparedby5 keV ‘s02+implantationingraphiteat roomtemperatureupto saturation.

17

}

Basic Mechanisms of the Deposition of

W. Jacob

Codeposited Layers

Max-Planck-h?stitute for P/asma Physics, D-85748 Garctvhg, Germany, EURATOM Association,.

Plasma-deposited hydrocarbon films represent a model system for thecodeposited layers found in fusion experiments. Depending on depositionconditions, these films can be soft, polymer-like or dense and hard [1-4]. Hard a-C:H films are deposited if energetic ions with energies above about 50 eVparticipate in the deposition process and at low-energy ion bombardment (E <20-30 ev) polymer-like, soft films are produced.

The chemical reactions and physical processes which occur at the surface ofhydrocarbon films during deposition from low-temperature hydrocarbon plasmaswere reviewed. Special emphasis was placed on the chemical reactions of atomichydrogen, the interaction of energetic particles with the solid, and the synergisticeffects between them.

The interaction of energetic particles with the surface of the growing film has beensimulated in the binary collision approximation by means of the TRIM.SP computercode. The sputtering and displacement yields were calculated for hydrogen andcarbon ions in the energy range from 25 to 500 eV and the depth distributions of the .

damage and implantation profiles were shown. The dominant ion-induced effect inthis energy range is displacement of bonded hydrogen atoms [5].

The microscopic processes of atomic hydrogen interacting with a carbonaceoussurface, such as adsorption, abstraction, addition, and etching, were brieflyreviewed and summarized.

The film growth of hydrocarbon films from hydrocarbon plasmas was investigatedexperimentally by real-time, in-situ ellipsometry and a variety of ex-situ analyses [2-7]. The real-time possibilities and the submonolayer sensitivity of ellipsometryallow a detailed investigation of the growth process in the plasma environment [6].The temperature dependence of the growth and the interaction of atomic hydrogenand energetic particles with the film surface were thoroughly investigated. Theexperimentally observable net deposition rate represents a competition between atemperature-independent deposition process and the temperature-dependenterosion by atomic hydrogen. [3]

At very low ion energies the synergistic interaction between atomic hydrogen andthe ions leads to ‘ion-assisted chemica/ erosion’ [7]. {n contrast to the true,thermally-induced chemical erosion step this process depends on the momentumtransfer of low-energy ions to surface groups. It should therefore exhjbit a clearisotope effect which was found experimentally (see discussion in Ref. [5]). On the *

other side, the energy threshold for this ‘ion-assisted chemical erosion’ step issignificantly lower than that for physical sputtering. At these very low ion energies,the measured erosion rate depends cri~cally on the internal structure of the layer.Soft, polymer-like films show markedly higher erosion rates than hard, dense. .- -.

18

.

The interaction of ions with the surface generates a modified layer on top of the filmsurface. This modified layer is an intrinsic property of the deposition as we!l as theerosion process at ion energies above about 30 eV. Its properties are determinedby the impinging particle fluxes from the plasma. A consequence of this modifiedlayer is that the erosion rates in a plasma depend to a larger extend on theproperties of this modified layer than on the properties of the underlying bulkmaterial [3,5]. The thickness of this modified layer is in good agreement with thepenetration depths of hydrogen ions as calculated by TRIM.SP [2,5].

The topics discussed in this presentation are thoroughly reviewed in a recentreview [5].

References

[1] P. Koidl, C.Wild, R. Lecher, and R.H. Sah, in ‘Diamond and Diamond-Like Films

and Coatings’ R. E.Clausing, L.L. Horton, J.C. Angus, and P. Koidl (Eds.), NATO-

ASI Series B, 266, Plenum, New York (1997) 243.

[2] A. von Keudell, W. Jacob, and W. Fukarek, Appl. Phys. Lett.. 66 (1995) 1322.

[3] A. von Keudell and W. Jacob, J. Appl. Phys. 79 (1996) 1092.

[4] A. von Keudell and W. Jacob, J. AppL Phys. 81 (1997) 1531.

[5] W. Jacob, Thin Solid Films, in print (1998).

[6] W. Fukarek and A. von Keudell, Rev. Sci. Intrum. Meth. 66 (1995) 3545.

[7] A. Annen and W. Jacob, Appl. Phys. Lett. 71 (1997) 1326.

[8] W. Jacob, B. Landkammer, and C.H. Wu, (Proceedings PSI 13, San Diego 1998)

J. Nucl. Mater, submitted.

19

BET Surface Area and Tritium Analyses for TFTR Carbon Dust Samples

R. A. Anderl and R. J. PawelkoFusion Safety Program

Lockheed Martin Idaho Technologies CompanyIdaho National Engineering and Environmental Laboratory

Idaho Falls, ID 84315-7113

The purpose of this work is twofold: (1) to measure the specific surface areas for carbon dustsamples retrieved from various locations in the Tokamak Fusion Test Reactor (TFTR) and (2) tomeasure the tritium and deuterium contents of the dust samples. Such characterization oftokamak dust is important to improving an understanding of dust-generation mechanisms and toquanti~ing the potential dust source term that is relevant to accident and de-commissioningsafety analyses. For accident safety analyses, particulate size and quantity information isimportant because size affects chemical reactivity rates and tritium mobilization rates, and thetotal quantity of dust affects hazardous material generation and transport under accidentconditions. Dust quantities, particulate size and tritium content are important to assessment ofcleanup methods and disposal approach for de-commissioning activities,

Two types of TFTR dust samples were analyzed: (1) dust samples from the MIRI windows and(2) dust samples obtained by vacuuming in three regions .inside the vacuum vessel (VV). TheMIRI window samples corresponded to sample collection from windows associated with the 10diagnostic pipes that enter the floor at Bay S, with pipes 1-3 penetrating the limiter and pipes 5-10 located in the nozzle trench. These pipes are arrayed across the minor diameter of the plasma.Samples we tested were obtained born MIRI windows 1,2,6, 8 and 10. The dust samples wereobtained by dumping the carbon material from a window through a fimnel into a plastic containerfor shipment. Vacuumed samples were collected with a vacuum/filter cartridge apparatus thatpenetrated the RF antenna port in Bay L. The apparatus consisted of a filter housing with a 0.2-p,mpore-size filter, a probe that reached into the W and a vacuum pump. Samples werecollected from three surfaces; the bottom row of bumper limiter tiles, the blank 12” port on theW floor, and inboard from the 12’’port on the W floor. Complete filter housings were receivedfor analyses.

The dust samples were transferred from the shipping containers by suspending the dust inethanol and transferring the ethanol/dust slurry into sample tubes used for surface-area and gas-content analyses. After the dust settled from suspension in the ethanol, most of the ethanol wasextracted horn the tube and the sample was dried by heating at 60°C for several hours and then at260”C for several more hours. Mass measurements for the dust samples were based on weighingempty sample tubes and tubes with completely dried samples.

.

20

.

,

>

.

.

.

Specific surface area measurements were made for the dried samples using a Micromeritics ModelASAP 2010 gas adsorption instrument with either I&or N2 as the adsorptive gas. Gasadsorption data were analyzed by the BET approach ~runauer, Emmett and Teller, J. ACS 60(1938) 309]. Specific surface area values that were measured for the TFTR dust samples rangedfrom 7 m2/g to 27 m2/g for MIRI window material, with the lowest values observed for materialcollected from windows on diagnostic pipes that penetrated the limiter. A value of 25 m2/g wasmeasured for a dust sample vacuumed from the blank port on the floor of the W.

Assay of the tritium and deuterium concentrations in the TFTR dust samples was derived fromtwo measurement approaches. First, the tritium concentration in the ethanol transfer fluid for asample was measured using liquid scintillation counting. After the specific surface areameasurements were completed,”thermal resorption spectroscopy (TDS) measurements weremade to assay the quantities of tritium and deuterium that were retained in the fully-dried dustsamples. These measurements were made with an integrated gas flow system that consistedsequentially of a tube furnace for heating the sample tube, an in-line getter to convert tritiumspecies to elemental form, an in-line ion chamber for measurement of the kinetic release oftritium, an on-line quadruple mass spectrometer (QMS) to measure kinetic variation in the gascomposition of the carrier gas, a heated, in-line copper-oxide unit to convert tritium species tooxide form, and ethyiene glycol traps to collect the tritiated water. High purity Ar with -0.5%H2carrier gas was used to transport gases evolving from a heated specimen through the system.Samples were heated from 20”C to 10OO”Con an approximately 30°C/min linear ramp, resultingin simultaneous TDS peaks for tritium and HD at about 780”C. Tritium quantification was basedon integrating the ion-chamber kinetic data and by assaying the tritium in the ethylene glycoltraps using liquid scintillation counting. Deuterium quantification was based on integration of thekinetic data corresponding to the QMS mass-3 peak for HD. A summary of the results of theseanalyses is given in Table 1.

TabIe 1. Summary of tritium and deuterium analyses for TFTR dust samples.

SampleID

MIRI-2-sm3

MIRI-1-sm7

MIRI-6-smll

MUU-8-sm15

MHU-10-sm19

NumberCarbon

@U?w3.67E21

5.17E21

9.37E21

6.16E21

6.19E21

NumberTritium(atoms)

6.01E17

9,68E17+

2.71E18

1.68E18

1.52E18

QMSIID(sCc)0.78

l.ll+

1.80

1,54

1.35

NumberD

(atoms)

2.1OE19

2.98E19

4.84E19

4.14E19

3.63E19

AtomRatio

DiT

34.9

30.8+

17.9

24.6

23.9

----1-Atom AtomRatio Ratio

D/C TIC

0.0057 0.00016

0.0058 0.00019+

0.0052 I 0.00029

0.0067 0.00027

0.0059 I0.00025

21

I

,,

MIRIaverage 26 0.0058 0.00023

(+/-25%) (+/-10%) (+,-24%)

TFTRfi1ter2 1.04E22 4.56E18 3.15 8.47E19 18.6 0.0081 0.00044

TFTlUilter3 7.68E20 4.56E16 .-.-- ----- .---- ----- 0.00006

+ Tritium value is based on estimate of liquid scintillation count value.

.

.

.

22

Mixed-Material Effects on P1asma-Facing Components

*

.

.

.

R. P. Doemer for the PISCES Experimental Team

Fusion Energy Research Program, University of California, San Diego.

When any plasma confinement device is fabricated ftom more than a single material which can

come into contact with either particle or heat flux, there is the potential for migration of one of

these materials to the locations of other materials. This combination of materials, or mixed

materials, can have substantially different properties than either of the original materials. The

PISCES-B lineal-plasma device is.examining the formation conditicms and properties of mixed-

material surface layers on plasma-facing components.

The PISCES-B device has been modified to incorporate an impurity gas (CD4) puffing system

in the target interaction region. It is, therefore, possible to control the fraction of impurities in the

incident plasma and to petiorm systematic tests on the conditions necessary to form mixed-

materials surface layers. Measurements of the rate of growth of the thickness of the mixed material

layer are performed. If the growth rate of the layer is determined by the arrival rate of the impurity

species at the surface, the thickness will be proportional to the exposure time. If the growth rate is

determined by the diffusion rate of the bulk species through the surface layer, the thickness will be

proportional to the square root of the exposure time. In the case of these PISCES exposures the

thickness increases linearly with time, indicating a deposition dominated regime. A simple erosion

model involving erosion, deposition and redeposition can adequately describe the growth rate of the

mixed-material layer in these PISCES experiments and should allow for predictions in other plasma

environments. This simple model can akio predict the threshold impurity concentration for film

formation. These results are shown in the attached figure.

It is also important to investigate the role of redeposition of metallic impurities in the formation

of mixed material layers, A beryllium evaporator has been independently installed upstream of the

target-interaction region to allow seeding of the incident plasma with beryllium. The presence of

beryllium on the sample surface is observed to reduce the chemical erosion of the graphite by more

23

than the depletion of the surface carbon concentration. The chemical erosion of the sample is

measured by monitoring the CD band emission from the plasma, after measuring the CD band

Auger Electron Spectroscopy is performed on the sample to determine the elemental composition

of the surface. These results are similar to the observed reduction of carbon chemical erosion by

doping of the graphite, however, these results indicate that the reduced chemical erosion will likely

happen naturally in the ITER divetior geometry.

And finally, the hydrogen isotope retention properties of carbon-containing layers on beryllium

could have serious implications for tritium accumulation in ITER. The retention properties of

coated beryllium samples horn PISCES-B indicates that the retention follows trends previously

I observed in bombarded graphite surfaces, namely that the retention tends to decrease as the

temperature of the sample during exposure in increased. By the time the exposure temperature is as

high as 500 C, there is effectively no difference between the retention in carbon coated and‘.

.

uncoated beryllium samples.

.

24

...-

.

. .

10

1

0.1

W aat~surface after plasma expcsur eEl a w surface after plasma exposure

- Partial surface coverage region

@ Thresholdcafculatedf rom erosion model

+

+

-1● T T1, ! 1 . . ..! . . ..’ . . ..> 1..!.

o 100 200 300 400 500 600

Sample Temperature (”~

Mixed-material layer formation conditions for PISCES. Coated surfaces indicate more than 50% carbon stu+ace

coverage, clean &aces Iess than 50°A surface carbon. The partial coverage region is indicated in the figure to gnide the

eye.

25

4“ International Workshop on Tritium Effects in Plasma Facing Components, Santa Fee, May 14-15,1998

Influence of Neutron Damage on the Tritium Retention ofITER representative Carbon Fibre Composites (CFC’S)

Forschungszentrum Karlsruhe,

R. Rolli, H. Werle

!NR, Postfach 3640, D-76021 Karlsruhe, Germany

C.H. WU

The NET Team, Max-Planck-lnstitut fur Plasmaphysik, Bol&mannstra13e 2, D-85748Garching, Germany

1. Introduction

Graphites, other carbon-based materials and beryllium are considered as attractivechoices for the protection of high heat flux components in existing and forthcomingtokamaks. As next generation devices, i.e. ITER, will produce intense neutron fluxes(ITER first wall: neutron fluence - 2. 1022/cm2 a 14 dpa in graphite per full poweryear [1]), the influence of neutron-induced damage on the microstructure and oncritical properties of these materials (e.g., thermal conductivity, swelling and tritiumretention) has to be known. A crucial problem is the observed increase of tritiumretention with neutron damage for carbon-based materials [2 - 6], because this couldrepresent a key safety problem. The quantitative dependence of tritium retention onneutron damage and especially the absolute values of tritium retention differremarkably for different carbon-based materiais. Therefore in this work three ITERrepresentative carbon fibre composites(CFC’s) have been studied.

26

2. Experiments

I

.

,

2.1 Samples and irradiation conditions

The investigated CFC’S are specified in table 1.

Table 1 Investigated CFC’S

Manufacturer Material type Identification Fibre characteristics

Carbon Loraine A05 R 2-d

SEP (i30rdeaux] NI12 T. 3-d

Japan CX2002U x 2-d

Samries from these materials were irradiated in the OSIRIS reactor at CEA Saclay(irradiation CERAM)are given in table 2.

and in the HFR Petten (irradiate MACIF). Irradiation conditions

Table 2 Irradiation conditions

Reactor

OSIRISSaclay

HFRPetten

Samplelrradi- Neutron lrrad. Temp.ation ~c) Identification Shape

CERAM 1.8 800 Xl, TI CylinderD21,7-17 1.3 840 XI, TI, RI Disk

1.05 800 X2, T2, R2 Cylinder

1.8 625 IR Disk

1.75 615 X3, IT Cylinder

1.25 640 IX, IT Disk

0.82 610 X4, 1R Cylinder

CERAM 1.8 1000 4X, 2T, R4 Cylinder

0217-19 1.3 1035 3X, 3T Disk

0.9 970 T4, 3R Cylinder

MACIF 0.85 430 T5, R5, X5 Cylinder0.36,0.39,0..44 385 T6, R6, X6 Cylinder

0.47,0.51,0.57 385 T8, X8, R8 Cylinder

27

2.2 Loading of the samples

The samples were prepared by cutting the disks radialiy in quarters and the cyiindersaxially in two haifs.

For loading severai sampies were inserted together in an alumina tube within afurnace. The tube was then evacuated and heated to the ioading temperature(1000 “C). After achieving the ioading temperature, the tube was flied with H, + 5ppm Tz at a pressure of 2.105 Pa (2 bar). After ioading under these conditions for 6h, the tube was again evacuated , the sampies were ailowed to cooi down and werethen transferred into the anneaiing apparatus.

2.3 Annealing of the samples

Amount and reiease kinetics of the ioaded tiritum were determined by purging withHe + 0.1 0/0 Hz and heating the sampies with -16 OC/min up to 1100 ‘C. Tritiumreieased as HTO is transformed to HT by a heated Zn-bed (- 390 “C) to avoidproblems with tritium water adsorption. The tritium activity of the purge gas ismeasured with a proportional counter. After correction for background, the count rateof the proportional counter is used to determine the rate (Bq/s) and the totai amount(Bq) of tritum reieased from the sampie (fig. 1).

As can be seen from fig. 1, the reiease rate is no zero after -4 h anneaiing at1100 ‘C, i.e. there is some residuai for tritium in the sampies. The residuai tritiumhas been determined for two sampies of each of the three Ci%’s by burning thesampies and {iquid-scintillation counting of the coiiected tritium. The residuai tritiumwas found to be approximately equai to the anneaied tritium. The totai retainedtritium is therefore about two times the anneaied tritium with an uncertainty of about* 50‘A.

3. Results

The vaiues of anneaied tritium (Bq/g) were transferred into appm (1 MBq/g = 0.0112appm for graphite) and multiplied by 2 . 105 to correct for the isotopic ratio of theioading gas. The data shown in fig. 2 refer therefore to a pure T2-atmosphere at 2 barand 1000 “C.

For aii three CFC’S the tritium retention increase with neutron damage in the range <0.3 dpa by about a factor seven. Above - 0.3 dpa, the retention seems to beessentially independent of the damage (fig. 2).

4

For aii dpa-vaiues tritium retention of NI 12 and CX2002U is comparable and about afactor three iess than that of A05 (fig. 2).

in tabie 3 the observed tritium retention data (twice the annelaed tritium) for A05 andNI 12 are compared with previous data [6]. ‘

28

Table 3 Comparison of tritium retention for A05 and NI 12

I

This work

Previous work[6]

Loading conditions

2 bar Tz, 1000 ‘C, 6 h

0.8 bar T2, 850 “C, 10 h

T’ritium retention (appm)A05 N’t12

Unirrad. Wrad. Unirrad.

-300 -30103 (0.5dpa) -140

-160 -900 (0.1 dpa) -140

Taking into account the higher loading pressure and temperature in this work, ourvalues are estimated to be within about a factor two in agreement with the previousdata, which is considered to be satisfactory.

4,

●

●

●

e

Conclusions

For all three investigated CFC’S (A05, NI 12 and CX2002U), tritium retentionincreases by about a factor seven with neutron damage in the ranges 0.3 dpa.

For all dpa-values, tritium retention of NI 12 and CX2002U is comparable andabout a factor three less than that of A05.

For identicai loading conditions, tritium retention values of this study for A05 andNI 12 are estimated to be within a factor two in agreement with previous data.

It seems that at irradiation temperatures above 800 ‘C annealing of tritium trapstake place. At 1300 ‘C abvout 70 % of the tritium traps are annealed within 2 h.

..+ ).,;7

References

[1]

[2]

[3]

[4]

[5]

[6]

C.H, Wu, J.P. Bonal and B.J. Kryger, J. Nucl. Mater. 208, 1 (1994).

W.D. Wampler et al., J. NUCL Mater. 176/177, 983 (1990). .

R.A. Causey et al., Fusion Technol. 19, 1585(1991).

H. Atsumi et al., J. Nucl. Mater. 191-194,368 (1992).

1-!.Kwast et al., J, Nuci. Mater. 212- 215~ 1472 (1994).

H. Kwast, H. Werle and C.H; Wu, Physics Scripts, Vol. T64, 41 (1996).

29

II

I

Tritium release of sample : Gr-Cd-lT.21 R1“.00E+05

1.00E+04

1.00E+03

1.00E+02

1.00E+OI

1.00E+OO

1.00E-01

1.00E-02

120011001000900800700600500400300200100

o 5000 10000 15000 20000 25000 30000 35000

Time (see)

E

Fig. 1 Annealing of loaded tritium

10000

1000

100

. . . ,...$,, . . . . . . . . . . ... ... 1

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 ‘1.3 1.4 1.5 1.6 1.7 +.8 f.9

Neutron damage (dpa)

Fig. 2 Annealed tritium as function of neutron darnage

(parameter irradiation temperature),

total retained tritiurn = 2 x annealed tritium

.

*

.

30

t1

I,.

I/

t

DEUTERIUM RETENTION AND DISTRIBUTION ON TUNGSTENTEST LIMITERS FROM TEXTOR

M. Ruben, V. Philipps2 and T. Tanabe3

1Aifk6n Laboratory, Royal Institute of Technology, S-1OO44 Stockholm, Sweden

21nst itmte of Plasma Physics, Forschungszentrum Jiilich, D-52425 Jiilich, Germany

3Center for Integrated Research in Science and Engineering, Flogoya University, Japan

1. Introduction and the aim of investigation

Experiments aimimg at the evaluation of tungsten M a material for PFC havebeen carried out at the TEXTOR tokamak since 1993. Test limiters of the mushroomshape - as shown in Fig. 1- and the area of the piamna facing surface of approx. 100cm2 were exposed to several tens of ohmic and auxiliary NBI heated d~charges. Theywere introduced either from the top or from the bottom of the torus and the top of thelimiters (tangency-point) was positioned at the”radius of 44 or 45 CMwhich was by 2 or1 cm, respectively, smaller than the minor radius (a = 46 cm) defined by the graphiteblades of the toroidal belt limiter. For the exposure, the W Iiiters were pre-heated to500 ‘C, i.e. they were kept above the ductile-to-brittle transition temperature in orderto avoid the material damage by thermal shocks during the plasma operation.

& 1Shape of test limiter used at TEXTC)R.

z?=:”

Following the exposure the limiters were aualysed with surface sensitive spectro-scopes and other methods in order to assess the changes in composition and structurecaused by the plasma. The emphasis was on the determination of deuterium and plasmaimpurity atoms, mainly carbon, deposited on surfaces. This contribution reports on sur-face studies of two typ& of limiters made OE a] sole tungsten; b) tungsten and graphite,i.e. twin limiter consisting of two halves with the same geometry but of different comp~sition. In the latter case the aim to compare the retention of D in two different materialsexposed under similar conditions.

2. Results and discussion

The inspection of the exposed tungsten surfaces revealed the existance of d~tinctdeposition zones at the ends of the limiters, i.e. at the linear d~tance of 40 to 70mm from the tangency point, as marked on Fig.1. These surfaces were covered with ablackish deposit and Fig. 2 shows the areal d~tribution of species deposited on.

In the case of the sole tungsten liiiter the width of the deposition zones is approx.17 mm on both sides. Deuterium is detected mainly in these areas which were notdirectly “immersed” in the plasma. The CD /CC ratio is 0.1 and 0.05 on the ion drift andthe electron drift side, respectively. The result indicates that the retention of deuteriumin the surface region is smociated with the presence of the carbon- containing codeposit,because outside these deposition zones the deuterium content does not exceed the leyel

31

of 1 x 1016 cm– 2. This may be explained by thermal release of deuterium implanted intungsten due to very high surface temperature, above lSOO ‘C, recorded in the centralpart of the limiter.

I I I I I I I I

E’

I I I

TUNGSTEN LIMITER I

lo” ~L

A DEUTERIUM

.= 10’7Eo q; 10’6 .0 ❑ CARBON

i=a 10’5 ~ x TUNGSTENa1-Z

0 SILICON

w[ I I I I I 1 I I

k 1 I I i i I I 1 I I 1 I 2

F TWIN LIMITER1- R d -1

10’7

1o“ .

10’6 g

TUNGSTENPART

nI f ) I I I I !- I I I ~ I 1 I

-60 “ 40 -20 0 20 40 60

DISTANCE FROM TANGENCY POINT (mm)

Fig. 2 Areal concentrations of species deposited on Iim”ters.

Sirdar results concerning the deposition and retention of deuterium are obtainedfor the tungsten part of the twin limiter. Distictly different situation is observed on thegraphite part exposed under very simikw operation conditions. Deuterium is detectedon the whole surfsce, even in hot areas located near the tangency point. In addition todeuterium, pronounced quantities of tungsten and silicon are measured. [Silicon orig-inates from the conditioning of the TEXTOR wall by siliconization and from formerexperiments involving the injection of SiH4 to the plasma edge. ]There are two strikingfeatures concerning the deposition of W and Si species: a) their greatest concentrationoccurs on top of the limiter; b) the deposition profiles are very simhr. The mech-anism deciding this deposition pattern on is still to be cIarified. However, one maytentatively suggest the formation” of the mixed W-C-Si compounds on the hot part ofthe graphite limiter surface. TM suggestion is partly justified by the shape of theRutherford backscattering spectra, recorded for the area under d~cussion, indicatingthe diffusion of Si and W into the carbon substrate. The formation of mixed compoundsmay probably explain the retention of deuterium on the hot limiter surface.

32

DEUTERIUM ANALYSIS IN THICK CODEPOSITED LAYERSPLASMA FACING COMPONENTS

M. Ruben, A. Vevecka - Priftaj2, J. von Seggern3 and V. Philipps3

lAlfv4n Laboratory, Royal Institute of Technology, 10044 Stockholm, Sweden

2Royal Institute of Technology, Stockholm, on leave from Polytechnic University,

ON

Tirana

31nst itute of Plasma Physics, Forschungszentrum - Jiilich, D-52425 Jiilich, Germany

1. Introduction

Formation of thick codeposited layers containing significant amounts of fuelsatoms is decisive for trit ium accumulation in D-T operated controlled fusion devices.Therefore, the codeposits’ properties - composition, chemical reactivity and structure -must be characterised in detail in order to enable the development of techniques for ef-ficient release of H isotopes and decomposition of the layers. This contribution reportson the microscopy and ion beam investigation of thick, flaking codeposits formed inTEXT OR on high heat flux components: the main poloidal limiter and the rf antennaprotection tile. The influence of oxidation at elevated temperature (300 and 550 ‘C) onthe state of codeposits was also studied.

2. Summary of results

The thickness of flaking codeposits was in the range 80-170 pm and this can

also be inferred from Fig. 2 which shows the codeposit-to-substrate boundary. The gapon the boundary indicates that the deposit, consisting of small agglomerates, is looselybound to the substrate. The layer is brittle and easily disintegrates forming small, 0.5-10 pm, dust particles. The structure of the flaking layer is exemplifiedin Fig. 2 a andb, whereas Fig. 3 shows the appearance of the graphite substrate after scraping-off theflakes.

The depth profiling with nuclear reaction analysis (NRA) revealed the deuteriumcontent to be in the range 0.9 -1.2 x 101s cm–2 and the CD/Cc ratio 0.04-0.06 in thesurface region of the flaking layer; see Fig. 4. The accessible depth for the NRA with a1500 keV 311e+ beam is only 4.5 pm, but this fairly small D content in the flakes wasalso confirmed with thermal resorption spectroscopy. As expected, greater amounts ofD (up to 8 x 1018 cm–2) were detected on side surfaces of the tiles, i.e. in the gapsexposed to lower thermal loads than the outer surfaces.

The oxidation in air at 300-320 ‘C (4 hours) resulted in the release of approx.65-70 % of deuterium. However, no differences were observed in the layer thicknessand structure prior and after the oxidation, giving no evidence on the decompositionof those thick codeposits. Some effects of the “powderisation” of flakes was observedduring the oxidation above 500 ‘C

33

.

Fig. I. Boundary between the deposit and

Fig. 4.

Fig. 2a. Structure of the layer with flakes.

J1.,

CO= 9 x 10’7cm-’, D/cmax= 0.043

I.

200

.“

150 .

.

100

J’

50 .. ~ y,%% .

1-

0 4I’10. 500 1000. 1500 2000 2S130 3000 3500 4000 4500

DEPTH (rim)

Deuterium depth profile in the surfwe region of thefiking deposit

.

34

4th InternationalWorkshop on TritiumEffects in PlasmaFacing Components - May 1998

Influence of Surface Roughness on the Deuteriumof ASDEX-UPGRADE Divertor Tiles

D. Hildebrandt, M. Akbi, B. Juttner*, W. SchneiderMax-Planck-Institut fir Plasmaphysik, Bereich Plasmadiagnostik,EURATOM Ass., D -10117 Berlin/Germany

Inventory

* Hurnboldt-Universit~tzu Berlin, Institut fir Physik, D - 10115 Berlin

The gas inventory of diverter target tiles used in ASDEX-UPGRADE has been analyzed by

Thermodesorption Spectrometry (TDS). Desorbed gaseous molecules have been measured by

heating up complete and cut divertor tiles.The largest samples (80X80mm2) could be heated up

to temperatures of 1500 K and smaller ones to even higher temperatures.

In addition, surface analysis techniques as Auger Electron Spectroscopy (AES), Secondary Ion

Mass Spectrometry (SIMS), Nuclear Reaction Analysis (NRA), electron microscopy and

optical surface profiling have been applied for investigating erosion and deposition phenomena.

The original plasma facing surfaces of the tiles were graphite (EK98) and plasma sprayed

tungsten, respectively. The graphite tiles”were used from 1991-1995 for about 1900 discharges

and the tungsten tiles in 1996 for about 800 discharges.

Significant differences in gas trapping behaviour and impurity deposition have been observed

for inner and outer divertor tiles. The plasma exposure also affected the surface morphology, in

particular, for graphite.

On inner divertor tiles a several ~m thick deposition layer was found. The mean surface

roughness of the graphite tiles was estimated to be about 3 pm.

Depth proftig and imaging by AES and SIMS showed ah uniform contamination at the surface

of the tungsten and graphite tiles consisting of carbon, boron, hydrogen and deuterium. The

deposited amount on inner divertor tiles was estimated to exceed 10~ C-atoms/m2 and 3x1023

B-atoms/m2. This contamination contained deuterium amounts up to 1.8 x2023 D-atoms/ m2

(graphite tiles) and 8 x 10 m22 D-atoms / m2 (tungsten tiles) as measured by thermodesorption.

In order to increase the local resolution in the TDS-measurernents some graphite tiles were cut.

Then a strong Iocal correlation between the released gas inventory and the amount of the

deposited impurity boron became evident. Both, the deuterium inventory and the thickness of

the deposition layer minimize in the contact region of the separatrix indicating erosion effects in

this restricted region with a width of about 20 mm. The results from cut tiles also demonstrated

35

that the deuterium amount released horn deposits in gaps on sides of the tiles can exceed that

from plasma exposed surfaces up to a factor of 4.

Outer divertor tiles showed less contamination and gas inventories in a region near the strike

point with a width of about 100 mm. Erosion processes like arcing and sputtering enhanced the

surface roughness, in particular, for graphite. Electron microscopy shows microcraters and

microcracks on the graphite surface after the plasma exposure. The mean surface roughness on

outer diverter tiles exceeded 10 pm.

For complete outer divertor tiles deuterium inventories up to levels of 3 x 1022 D-atoms/m2 for

graphite and about 1 x 1022 D-atoms/m2 for tungsten coatings were found in the contact region

of the separatrix. However, these values are much higher than deuterium inventones observed

after exposure to a pure deuterium plasma. In fact, exposure of virgin graphite to a deuterium

plasma (Te about 5 eV) representing the divertor plasma without sign~lcant hydrogen and

impurity contamination caused deuterium retention in the orderof1021 D-atoms/ m2.

AES- and SIMS- depth profiling and imaging on the target tiles did not show a connected

deposition layer at the surface. However, almost identical lateral distributions of deuterium and

boron indicate that codeposition of both species occurred on surface depressions of produced

microcraters, cracks and microcavities. At these locations the deposits are protected against re-

erosion. Boron contamination in surface depressions has been found in excess of several 1022

B-atoms / m2.

In order to investigate the depth distribution of the trapped deuterium more quantitatively

surface’ layers of graphite tiles were locally removed by milling. TDS measurements of

,: separated groove areas where layers between 5pm and 30 pm were milled off showed that the

dominant amount of the deuterium is trapped at the surface. In fact, samples where the first ten

micrometer with high surface roughness were removed contained deuterium amounts of about

5X1021 D-atoms/ m2. Only less than 1021 D-atoms/ m2 were found in samples where about

I 20 ~ were milled off. This has to be compared with 1.5 x1022 D-atoms / m2 found at the

neighbored unmilled area.

This demonstrates that the dominant amount of deuterium is trapped in the contamination layer,

consisting mainly of carbon, boron and the hydrogenic isotopes. In regions with prevailing

.

.

erosion this contamination is located in plasma produced surface depressions. In this way the

surface roughness has a significant influence on gas trapping in regions with prevailing erosion.

The contribution of diffusion effects along grain boundaries to the deuterium inventory in

~aphite “is estimated to be smaller than 10 % in eroded regions of the outer diverter and less.

than 1 % in deposition zones of the inner divertor.

36

.

.

.

Analysis of Invessel Components in TFTR Prior to DT Operations and in DIII-D

R. Wampkr

Org. 1111, MS 1056 Phone: 5058444114

SandiaNational Laboratories FAX: 5058447775

Albuquerque, NM 87185 email: wrwampl@sandia.gov

W R Wampler presented results from analysis of in-vessel components in TFTR prior to

DT operations and DIII-D. The D retained inside TFTR was estimated for each year of

operation between 1987 and 1991 by ion-beam analysis of limiter tiles and wall coupons

[refs 1,2]. D accumulation rates were 4 to 11 grams/year. Comparing with the quantity of

D used to fbel the plasmas, the average fraction of deuterium retained in the vessel was

44’%0A 17Y0.Of this, 41’Mowas on the wall and 44% on the plasma-facing surface of

the limiter and 15V0in the gaps between limiter tiles. In DIII-D about 1 gram of D was

retained on the lower divertor after 2000 plasmas, 40°/0of which was in the gaps between

tiles [ref 3]. The DIII-D DiMES probe showed net carbon deposition was highest on

surfaces shadowed from direct ion flux, suggesting a high flux of carbon or hydrocarbon

rieutrals from the plasma boundary onto the diverter surface. In both TFTR and DHI-D

co-deposition of D with C, with D/C -0.2-0.4, was found to be the main mechanism for

D retention. For this process the D accumulation is proportional to the net C deposition

which equals the net C erosion. Two potential ways to reduce retention would be to

reduce the net erosion rate e.g. with detached plasmas or a high Z material, or to use a

material which does not retain D when it redeposits.

[refj

1.

2.

3.

—

W R Wampler, B L Doyle, S R Lee et al. J. Vat. Sci. Technol. A6, 2111, (1988).

H. F. Dylla and K. L. Wilson editors, Tritium Retention in TFTR, Princeton Plasma

Physics Laboratory (PPPL-2523) and Sandia (SAND 88-8212) Report, April 1988.

D. S. Walsh, B. L. Doyle and G. L. Jackson, J. Vat. Sci. & Technol. A1O, 1174 (1992).

37

II

TRITIUM EXPERIENCE IN LARGE TOKAMAKS; APPLICATION TO ITER

C H Skinner, Princeton Plasma Physics Laboratory

“Tritium Experience in Large Tokamaks; application to ITER” was the

subject of an International Energy Agency workshop held at the Princeton

Plasma Physics Laboratory on March 16-18th, 1998. The workshop brought

together 90 scientists and engineers from ITER, TFTR, JET, Asdex, DIII-D,

Textor, Tore Supra; tritium laboratories, AEC(Canada), JAERI(Japan),

INEEL, LANL, SNL Westinghouse (US), Karlsruhe (Germany), Russian

Nuclear Center, and many other institutions. The participants shared

experience and expertise in tritium issues on large tokamaks with a view to

providing the optimal knowledge base for all aspects of tritium use in ITER.

Extensive discussion periods were devoted to exploring outstanding issues

and identifying potential R&D avenues to address them:

. First results of tritium fueled plasmas in TFT.R and JET and near

completion of ITER EDA has greatly expanded knowledge base of

tritium issues.I

● Long ITER pulse duration - > erosion effects dominant (first time in

tokamaks). Tritium removal likely to constrain plasma operational

time on ITER.

. Motivating fusion development by global warming (“green pricing”)

will draw more attention to tritium and safety issues.

. R&D urgently needed on in-vessel tritium & dust diagnostics, in-vessel

tritium & dust removal and behavior of mixed materials.

A summary of the workshop is being submitted to Nuclear Fusion. Preprints

are available as a PPPL Report 3300 on the PPPL report web site:

http: //www-local .pppl. gov/cgi-bin/pubrep,cpy req_l 998 .pl

A limited number of proceedings containing the viewgraphs from the

presentations are available. Institutional libraries and interested individuals

are invited to email requests for copies to: cskinner@pppl.gov.

For more information please see the workshop web site:

http: //www. FusionSei.ence. ORG/tritium98/

38

.

.

Dependence of surface oxidation on hydrogen absorption and resorptionproperties of Ti-6Al-4V alloy

Y. Hirohata, T.Nakamura Y.Aihara and T.Hino

Department of Nuclear Engineering, Hokkaido UniversityKita-13, Nishi-8, Kits-ku, Sapporo, 060-8628 Japan

The hydrogen absorption and resorption properties of oxidized and non-oxidized Ti-6Al-4V

alloys under a condition of a fusion reactor operation (pHz=O.03-100pz” T,b,oW.=323-872K)

were evaluated by thermal resorption spectroscopy(TDS). Oxidized samples were prepared

by an oxidation of degassed sample for 10 rnin at the temperatures of 323, 623 and 873 K in

the oxygen atmosphere with a pressure range from 5x10-3to 5 Pa (Oxygen exposure = (0.44-

16)x101902/cm2). Compared with the case of non-oxidized sample, in the oxidized samples,

the O/’Ti ratio at top surface varied from 0.5 to 2.5 with the oxidation temperature. In

particular, in the sample oxidized at 873 K, the thickness of oxide layer became about 10 nm.

The resorption peak (TP) of non-oxidized sample appeared at the temperature of 790 K and

was almost the same even if the absorption amount changed (Fig. 1). The resorption rate was

roughly symmetric with respect to the TP. The diffusion coefficient of hydrogen was

estimated as D=333 exp(- 1.55 [eV]/kT) [cm2/s] .On the contrary, in the case of oxidized

sample, the TP shifted to a higher temperature region with the increase of Oxygenexposure.

Hydrogen Absorption: 673K, lOPa, lhr10

8

6

4

2

(-)

I I I I I

-Oxidation Temperature:673K-Oxygen Exposure(02/cm2)

: 5.4X1018s : 1.1X1019

: 2.2X1019- i :1.lXl@

rWithout Oxidatio1-

’400 600 800 1000 1200 1400

Time (s)

1200

1000

400

200

Fig. 1 Thermal resorption spectra of hydrogen for Ti-6Al-4V alloy with and without oxidation

39

Figure 2 shows the hydrogen

concentration as a function of oxygen

exposure. Here, absorption experiments

were carried out at the temperature of 673 K

for lhr. The hydrogen concentration

exponentially decreased with an increase of

oxygen exposure in all oxidized samples.

The hydrogen concentration of the samples

oxidized at 673 K and 873 K was two

orders of magnitude smaller than that of

non-oxidized sample when the oxygen

exposure was higher than 1x10200ticm2.

Figure 3 shows the hydrogen

2(NJ

.000

800

600

400

200

(-1=-01234567 8

Oxygen Exposure (x1O2002/cm2)

Fig.2 Hydrogen concentration as a function of oxygen

exposure

Hydrogen Absorption: 10Pa, lhr

concentration as a function of T~b~OW.when ~ 1200“ ~ Non Oxidation

I I i I

PHand absorption time were kept 10 Pa and ~ 1000: ~ g:ko:x;;e~ger

lhr, respectively. Here, samples with thick.4 Boo-

and thin oxide layer were prepared by ~ ~oo::

oxidation at the temperatures of 873 K and ~u 400 ;

323 K for 10 rein, respectively. In the non- ~3 200:oxidized sample, it is seen that the %

absorption amount decreased with the ~ ~o? 500 600 700 800 900increase of absorption temperature in the Absorption Temperature (K)

range higher than 650 K. In this temperature l+g,3Hy&ogen concentration as a function of absorption

region, the absorption amount saturated temperature

within absorption time of lhr. This equilibrium concentration of hydrogen was

approximately a half of that of pure Ti. The Siverts’ constant, K,, was obtained by changing

the range of PH2 from 0.03 to 100 Pa at 820 K. This value was 170 (ppm/Pa]’2) which was

between those of c+Ti (140) and &Ti (340).

In the oxidized alloy with a thick oxide layer(about 10nm), the absorption amount

was two orders of magnitude smaller than that of non-oxidized alloy when Ta~Orp.waS

lower than 6!50 K. However, the absorption amount of oxidized alloy became comparable

with that of non-oxidized alloy when the absorption temperature was higher than 800

.

K, because the oxide layer disappeared in such the higher temperature region.

40

Co- and Counter-permeation of Hydrogen and Deuterium in Metals

Kaname KT.ZU*and Tetsuo TANABE**

*Department of Nuclear Engineering, Graduate School of Engineering,Nagoya University,

Furo-cho, Chikusa-ku, Nagoya 464-8603, Japan

**Center for Integrated Research in Science and Engineering, Nagoya University,Furo-cho, Chikusa-ku, Nagoya 464-8603, Japan

Experiments of co-permeation (H and D permeate same direction.) and counter-

perrneation (H and D permeate opposite directions) of deuterium and hydrogen through

a palladium membrane under surface limited condition were performed.

For co-permeation, permeation rates of H2, HD and D2 as a function of upstream

ED(fixed) and D2 pressure were measured. When total hydrogen concentration (H+D) on

the surface was small, the partition of permeated H2, HD and D2 was found to be the same

as that of upstream chamber. However, as increasing the total hydrogen concentration, D2

flux was decreased.

For counter-permeation, deuterium permeation rates as a function of upstream D2

pressure were measured keeping H2 gas pressure constant at the downstream side. It was

found that significant deuterium permeation occurred even when the deuterium upstream

pressure was much smaller than the hydrogen downstream pressure. The deuterium

permeation rate was gradually reduced by increasing the counter H permeation.

These results imply that adsorption (absorption) of D2 from the gas phase may

be inhibited and/or surface recombination of D may be blocked by H.

41

Release of Ion-Driven Hydrogen Isotopes from Molybdenum

K. Yamaguchil, K. 0hkoshi2, O. C)noue2 and M. Yarnawaki2

1. Nuclear Engineering Research Laboratory, University of Tokyo

2. Department of Quantum Engineering and Systems Science, University of Tokyo

The release of hydrogen isotopes, including trithnn, from a foil of molybdenum (Me, 0.1

nun in thickness) was investigated. While the permeation behavior of hydrogen [H;) or

deuterium (D;) ions, with an incident energy and flux of 3.0 kV and = 1017ions m-2 s-l,

was studied using a device installed with in-situ surface analyzers [1], while the implan-

tation of hydrogen isotopes ions containing tritkrn was performed using the tritium test

apparatus at the University of Tokyo [2], see Fig.1.

The ion-driven deuterium permeation behavior from Mo in the temperature range of 500

to 1000 K was observed to be strongly temperature dependent, see Fig. 2, which was quite

difFerent from the result on niobium (Nb) which was performed at similar experimental

conditions, see Fig.3 [3]. Moreover, the permeation rate of latter was quite sensitive to the

condition of downst~earn-side surface.

On the other hand, Some preliminary results on interaction between tritium and MO (as

well as Nl; not presented here), were obtained, where the thermal resorption spectra of I)T

and T2 from Mo pre-irradiated by tritium-containing ion beam were obtained and com-

pared with those of H2 or 1)2 [4]. However, the data are scarce, and prolonged irradiation

is required to reach the fluence level of H and D.

References

[1] V. Bandourko, K. Ohlmshi, K. Yamagd.i, M. Yamawaki, N. Koborov, V. Kurnaev, D.

Levchuk, Vacuum 47’ (1996) 947.

[2] M. Okada, R. Sate, K. Yamaguchi and M. Yamawaki, J. Nucl. Mater. 248 {1997) 72.

[3] M. Yamawaki, N. Chitose, V. Bandurko and K. Yarnaguchi, Fusion 13ng. Des. 28 (1995)

125.

[4] K. Yarnaguchi, M. Okada, 0. Onoue, F. Ono and M. Yamawaki, to be published in J.

Nucl. Mater.

.

.

.

42—

i..

It

hassle. ----- .-__ --__--_>, -------------------

‘b-------------------yRotaryPump

TfiumRemowUSys.--------------------------------

bn&&K%uge ] WI,*

, (

Flow IMeter ~

!--- -----

E?Oi?ow$Wrap

Fig.1: Schematic drawing of tritiurn beam test apparatus.

,,I I 1 I j 65x10’6

,o-~ ‘“1 1.2 1.4 1.6

lm [m

7m 1Niobium9

1000 KIT

Fig.2: Temperature dependence of ion-driven deu- l?ig.3: Temperature dependence

terium permeation rate in Mo. of ion-driven deuterium perme-

ation rate in Nb [3].,4

!-

43

STUDY AND DEVELOPMENT OF CARBON-BASED MATERIALS WITH

REDUCED HYDROGEN ISOTOPES RETENTION

T.A. Burtseval, I.V. Mazull, A.E. Gorodetsky2, A.P. Zakharov2, N.N. Shipkov3

1D.V. Efrernov Institute of Electrophysical Apparatus,

189631 Metallostroy, St. Petersburg, Russia

2Institute of Physical Chemistry, Russian Academy of Sciences,

117915 Moscow, Russia

3NH “Graphite”, 111524 Moscow, Russia

Tritium trapping by carbon materials remains one of the criticzil issues limiting the use

of these materials for cladding of the fiist wall and divertor plates for fusion reactor. The

tentative analysis of the results carried out by the authors allows for the assumption that it is

possible to reduce hydrogen isotopes permeability and sorption capacity in graphite by

introducing titanium dopes (2,, %), thus increasing the materiak density to 2.24-2.26 g/cm2 and

rising its crystallization degree, and by small (O,1-0.5 %) amounts of boron in the solid

substitution solution in the crystal lattice.

Doping with boron is shown to initiate the implantation of boron atoms into the crystal

graphite lattice and the formation therein of a solid substitution solution, resuhing in a

considerable reduction in thermal conductivity of material.

But the presence of boron in recrystallized graphite deteriorates the containment of

heavy hydrogen isotopes by the material, this being critical when selecting the protection layer

for a fusion reactor. As demonstrated in [1], deuterium trapping by recrystallized RG-Ti-B

graphite (lOW%Ti and 0.5ti% B in the initial mixture) is decreased twice as compared with the

graphite without boron additives.

The preliminary analysis of the results allows for assumption that a reduction in

deuterium trapping by boron-content graphite is attributable to the following facts:

● boron atoms can occupy the place in traps and prevent deuterium from occupying these

traps or they can reduce the energy of deuterium atom binding with the traps, thus

decreasing the number of traps occupied by deuteriurn,

● boron introduced into the graphite lattice retards deuterium diffusion therein, thus Iirniting

its access to the traps.

The foregoing does not explain uniquely the effect of boron state on the content of

hydrogen isotopes in material.

As is known [2], boron in RG-Ti-B graphite is in the state of implantation into the

crystal lattice of graphite, boron carbide and titanium diboride.

44

.

.

One of the goals of the study was to produce graphite samples with an equal total boron

content therein and different contents of boron in the form of a solid substitution solution of.

carbon atoms in the graphite lattice. It will apparently clarify the effect of boron state on the

value of deuterium trapping by graphite.

It was also interesting to find the ways to reduce the total concentration of boron in

material, with the composition of the solution with carbon substituted by boron being retained.

Thus, the main goal of the investigation was to fmd a new method for introduction of

boron into the initial mixture so as to change boron concentration in the substitution solution,

with the total concentration remaining constant.

This goal can be accomplished by introducing into the initial carbon mixture not a boron

powder but a thermally treated boron-doped carbon system containing boron carbide and boron

implanted into the graphite lattice.

With the aim of assessing the concentration of boron forming the solid substitution

solution in the crystal graphite lattice measured were the electron properties of samples

produced by doping the initial mixture with different amounts of boron.

Diamagnetic susceptibility and a EPR signal were measured on specially conditioned

samples. The analysis of the results of these measurements revealed a good agreement between

. the values of boron concentration in the crystal lattice assessed by both methods.

Conclusion

*

The method was found for introducing boron into the initial mixture of recrystallized

RG-Ti-B graphite making it possible to increase the concentration of boron solved in the crystal

graphite lattice with an equal content of boron in material.

Samples of RG-Ti-B graphite were obtained with an equal total content of boron and

different amounts of boron in the substitution solution in the crystal lattice.

The obtained materials were investigated on ion-, plasma- and electron- beam

accelerators in the regimes simulating normal operation conditions of the ITER divertor plates

and at plasma current disruptions. It is shown that titanium dopes increase essentially the

resistance of recrystallized graphites, as compared to non-doped ~aphites and C-C composites,

to ion and thermal erosion.

Boron is assumed to increase considerably the erosion resistance with increasing the

content of boron atoms in graphite in the form of a solid substitution solution and essentially

reducing deuterium retention [1, 3, 4]. The carbide phase of boron in these processes is

practically of minor importance.

45

References

[1] V.N.Chernikov, A.E.Gorodetsky et al., J,Nucl.Mater. 217 (1994) 250.

[2] C.A. Kupriyanov, A.C, Kotosonov, T.A. Burtseva et al., “Non Ferrous Metals” 10 (1994)

35 (in Russian).

[3] B. Khripunov, V. Shapkin, et al., Fusion Technology. 1 (1994) 243.

[4] A. Gorodetsky, A. Markin, V. Chernikov et al., Atomnaya Energiya, 82 (6) (1997) 448 (in

Russian).

.

46

Tritiurnworkshop,May 14-15,1998,SantaFe

Reconsideration of hydrogen recycling.

T. TANABE

process at the first wall

+. Center for Integrated Research in Science and Engineering,Nagoya University, Fare-who, Chikusa, 464-8603 Nagoya, Japan

Hydrogen recycling is one of the key issues to ensure steady or long pulsed

operations in DT fusion reactors. Particle valance among fueling, plasma, retention in the

first wall and evacuations is directly related to tritium inventory and the retention in the wall

influence ❑lplasma density control very much. They involve various processes such as

❑reflection, reernission, excitation, ionization, dissociation, recombination , gyration and so

on. Up to now most of the laboratory studies on the hydrogen recycling have been

concentrated to the particle balance between plasma and materials but energy and momentum

balance between plasma and materials has been scarcely considered. The present paper

demonstrates the significance of the energy distribution and/or energytl state of the released

hydrogen from the wall.&

Hydrogen particles escaping from plasma are impinging into the subsurface layer,

and subsequently remitting from, retaining in and permeating through the first wall.

Different from molecular hydrogen, energetic hydrogen particles escaping from plasma

directly impinge into the solid, deposit their energy by electron excitation, atomic

displacement and phonon excitation, and finally become thermal. All these deposited

energies cooperatively influence the subsequent migration of the therrnalized hydrogen in the

target through, for examples, producing excited or ionized atoms, defect trapping and

vibrational diffusion enhancement. As a result various anomalies are observed in remission,

diffusion and permeation of implanted hydrogen. In fusion environment, another energetic

particles like electrons and neutrons, and various radiation are loaded simultaneously, which

makes the behavior more complex. Thus hydrogen molecules reernitted andlor permeated

after the energetic hydrogen injection are often not in equilibrium with gas phase, taking

higher energy states than expected from the material temperature. And the hydrogen. recombination process may not be the rate limiting process. In addition to the reernitted

hydrogen molecules, various particles are emitted from the surface, which include reflected0

and sputtered particles, secondary electrons and photons as well. Those emitted particles are

not necessarily equilibrated thermodynamically with the surface temperature of the target but

47

I

are often in higher excited states and some are even ionized. If displaced interstitial atoms are

migrate to the top surface without completely losing their energy or momentum, they are

released from the materials surface with excess energy compared with thermalized particles.

Such an example is seen in radiation enhanced sublimation (RES) of graphite and carbon

based materials irradiated with energetic ion irrespective injected species. Chemical erosion

of carbon materials by implanted hydrogen producing methane is one of the most well known

phenomena. Recently even atomic hydrogen remission from carbon is observed at such a

low temperature of 1000IJ. Super

impinging of energetic hydrogen.

There is no doubt that the

❑permeation of hydrogen in metals is also caused by

release of hydrogen atoms dissolved in the materials

occurs through surface recombination except at elevated temperatures where the release of

atomic hydrogen can occur. And theoretical ❑estimations of recombination coefficient have

been proposed by Baskes, Pick and Sonneberg, and Richards. If reemitted molecules are,

however, not in thermal equilibrium, all models •lwhich are based on volubility and

diffusivity would be wrong. And data scattering in recombination coefficients should not be

simply surface contaminant and geometry.

k engineering aspect, recombination coefficient does not seem very important in

steady state recycling. Simply hydrogen remission rate is equivalent to the incoming flux

except the reflection

Some part of impinging hydrogen is directly reflected at the surface. Depending on

their velocities the reflected particles show various states in their electron configuration which

would results in a different chemical effect at the surface. Quite recently, even some of the

reemitted molecules are reported to be in excited states either or both vibrationally and

rotationally.

Thus ❑lthe energy distribution and energy state

fraction, rotational and vibrational states) of both reemitted

(electronic state including ion

and reflected hydrogen are very

important not only to investigate the surface kinetics occurring under energetic hydrogen

injection but also to understand the role of the reemitted hydrogen and impurities on the

plasma materials interactions because it directly correlates to their penetration length or

ionization length in the boundary plasma.

All these energetic hydrogen induced effect would originate from the super position

of electron excitation, atomic displacement, phonon excitation and thermal motion, all of

48

which have their own characteristic time. This indicates that time-resolved measurement with

using pulsed beam (either photon, electron or ion)

dynamic transport process of the implanted hydrogen.

Finally it should be pointed out that strong

must be very useful to examine the

electric and magnetic fields in fusion

environment may have some influence on these phenomena. But it has not been examined yet.

49

Depth distribution of deuterium atoms and molecules in beryllium oxide

implanted with D ionsl

V.Kh.Alimov, V. N. Chernikov, A.E. Gorodetsky, A.P. Zakharov

Institute of Physical Chemistry, Russian Academy of Sciences,

Leninsky prospect31, 117915 Moscow, Russian Federation

Abstract

lndepth concentration profiles of deuterium

implanted with 3 keV D ions to fluences,

atoms and molecules in beryllium oxide films

@, in the range from 2.2x1015 to 6.7x1018

D/cm2 at temperatures, T~, of 300 and 700 K have been determined using SIMS and RGA

(residual gas anzdysis) measurements in the course of surface sputtering. The

microstructure of implanted specimens was studied by TEM. BeO film thermally grown

onto S-65B 13esubstrates consist of small close-packed hexagonal microcrystals of size 25-

30 nm.

Implanted deuterium is found to be retained in BeO matrix in the form of both D

atoms and D2 molecules. The maximal concentration of D in the both states (as D atoms

and D2 molecules) in BeO irradiated with D ions at Tln = 300 and 700 K reaches value of

0.2 D/BeO and 0.08 D/BeO, respectively. The total amount of deuterium captured within

the BeO film of - 120 nm in thickness and adjacent metal layers as a result of the

implantation reaches 1.8x1017 at ~,, = 3~0 K and 0.7x1017 D/cm2 at T~ = 700 K.

Irradiation at 300 and 700 K leads to the formation of tiny D2 bubbles of 0.6-0.7 nm

in radius and of high volume density c~ ~ 4.5x 1018cm-3. These bubbles together with the

intercrystalline gaps are responsible for the accumulation of a molecular fraction of the

implanted deuteriwn.

At both irradiation temperatures the major part of implanted deuterium is present in

BeO matrix in the form of D atoms. The ratio of deuterium quantities retained in the form

of atoms and molecules, Q: @z , varies from 8:1 for Tim= 300 K to 2.5:1 for Z= =*

700 K. Most of D “atoms is trapped within the oxide at the lattice point defects and their

complexes which are beyond the resolution of a TEM.

.

1 Work supported by the United States Department of Energy, under Contract LF-7292 with Sandia NationalLaboratories.

50

I I

I: . .

*

a)

C3&0

CiiE!O

12

9

6

3

0

8

6

4

2

0

keV D+ —> Beo(l 20 nm)/f3e,300 K

0 100 200 300 y

o 100 200 300Depth (rim)

,.Concentration profiles of deuterium trapped as D atoms (a) and in the form of

Dz molecules (b) in 120 nm layer of BeO on Be substrate implanted with 3 keV

D ions at 300 K.

51

!I

‘!I

I

I

a)

n

b)m-

3 keV D+ —> Beo(f 20 ~m)/Be, 700 K

4

3

2

1

00 100 200 300

8

6

4

2

00 100 200 300

Depth (rim)

I/

Concentration profiles of deuterium trapped as D atoms (a) and in the form of

Dz molecules (b)in 120nmlayer of BeOon Besubstrate implanted with 3keV

D ions at 700 K.

52

Isotope Difference in Thermal Re-emission

of Hydrogen Implanted into WC Layers Depositael on Graphite

T.Horikawaand K.Morita

Department of Crystalline Materials Science and Department of Nuclear Engineering,

Graduate School of Engineering, Nagoya University,

Furo-cho, Chikusa-ku,Nagoya 464-8603, Japan

In future devices, high Z refractory metal and carbon are planned to be

used simultaneously as the divertor plate and the first wall armor,

respectively. The mutual redeposition of sputtered substance modifies

surface layers of the materials into metal carbides under plasma irradiation

at high temperatures for long term discharge.

Refractory metal carbides have a desired property of low sputtering yield.

There is a concern that the sputtering of high Z metal induces radiative

cooling of hot core plasma, but TiC and WC layers coated on graphite have

shown that the sputtering of metals is suppressed at high temperatures due

to self-sustaining coverage of the surface with segregated carbon layers from

graphite substrate, even under high heat flux plasm a irradiation.

Nevertheless, data on hydrogen behauior in high Z metal carbides are hardly

available . Therefore, it is quite important to investigate dynamic behavior

of hydrogen in refractory metal-carbon compsite materials.

In this paper, we report the experimental results on the themal re-

mission of protium implanted into the WC layers which are compared with

those of deuterium [1]. It is shown newly that the re-emission of retained

hydrogen isotopes due to thermal annealing takes place at three stages [2] :

large amounts of hydrogen are re-emitted from the two different traps in the

WC layers at lower temperatures and amount of remains re-emitted from

carbons in grain boundaries at higher temperatures which corresponds to

the previous result [1].

Isotropic graphite plates (IG- 110I3) of 0.5x5x35mms in size were used as

a substrate of the specimen. A W-film of 400 nm in thickness was deposited

the mirror-likely polished surface of graphite plate by electron beam

heating. The WC layers were prepared by direct current heating of the

graphite substrate at 1400~6 for 30 min. It was found that the average

atomic composition of WC layers by RBS was C/lV=O.96t0.04. Since the WC

layers were polycrystalline and the XRD analysis showed graphite peaks, a

small fraction of carbons were expected to exist in grain boundaries.

53

The implantation into the WC layers were done with 5 keV H2+and D2+

ion beams up to saturation at room temperature and at a flux of

2.4x101 acre-Z-l. The concentration of hydrogen in WC layers was measured

by means of the elastic recoil detection technique.

First of all, protium retention in the WC layers deposited on graphite

irradiated with 5 keV H2+ ion was measured as a function of irradiation

time. It was found, similarly to deuterium retention in the previous work [1]

that the retained number of H increases rapidly at an early stage of

irradiation and then saturates at a level which decreases by a factor of

about 15°A due to spontaneous re-emissionin several hours of terminating

the irradiation .

It is found on isochronal annealing for 10min that the re-emission takes

place at three stages: the first stage occurs in the retained fraction of

1.O&O.6, the second stage in the retained fractions of 0.6&0.2 and the third

stage below the retained fraction of 0.2, which is ascribed due to re-emission

from carbons segregated in the grain boundaries of the WC layers.

It is also found in isothermal annealing experiments that the

concentration of retained hydrogen decreases rapidly in the beginning and

hereafter very gradually with increasing the annealing tirne.The isothermal

re-emission curves have been analyzed by taking into account the two

trapping model and the mass balance equations with two kinds of thermaldetrappings (Zdl and zd2)> retrapping (St) and local molecular

recombination (IcI)between free hydrogen atoms .

The analytical solution of the mass balance equations has been fitted to

the experimental data on the isothermal re-emission profiles of protium and

deuterium, in order to determine the rate constants of the elementary

processes. It is found that the theoretical best fitting cures reproduce

excellently well the experimental re-emission profiles. In the fitting, It is

determined that the values of activation energy of the effective

recombination rate constants (kl/Co) (Xdl/Z~2 for the type 1 trap and (kI/co)

&2/st)2 for the typen 2 trap are 0.61eV and 0.73eV respectively for protium

and deuterium. It is also determined that tha values of activation energy of

the thermal detrapping rate constants for the type 1 and type 2 traps are

0.18eV and 0.24eV, respectively, for protium and deuterium which are

considerably lower than that in graphite (0.60eV). The two trapping sites

with such small trapping energies are proposed to be intrinsic interstitial

sites.

54

J

.

References

[1] T.Horikawa,B.Tsuchiya and K.Morita,J.NucL Mater. in press(1998).

[2] T.Horikawa,B.Tsuchiya and K.Morita, to be published.

.

55

List of Attendees

Russell DoemerRm. 460, EBU-11University of California, San Diego9500 Gilman Dr.La Jolla, CA. 92093-0417phone: (619) 534-7830fax: (619) 534-7716rdoerner@pisces.ucsd. edu

Yuko HirohataDepartment ofNuclear Engineering, Faculty of Engineering,Hokkaido UniversityKits-13, Nishi-8, Kits-ku, Sapporo 060-8628 JapanTEL +81-1 1-706-7108FAX +81-1 1-709-6413hirohata@qe.eng.hokudai.ac.jp

Chung H. WuThe NET TeamMax-Planck-Institut fier PlasmaphysikBoltzmannstrasse 2D-85748 Garching bei MuenchenGermanyTelephone No.: 49-89-32994232Fax No.: 49-89-32994198wuc@ipp.mpg.de

WolfgangJacobMax-Planck-Institut fuer P1asmaphysikBoltzmannstr. 2D-85748 Garching, GermanyPhone: +49 -89-3299-2618Fax: +49 -89-3299-1149jacob@ipp.mpg.de

Dieter HildebrandtMax-Planck-Institut fuer PlasmaphysikBereich Plasmadiagnostik, Mohrenstrasse41, D-10 117 Berlin, GermanyTel. (030) 20366164

56

Fax: (030) 20366111Hildebrandt@ipp.mpg.de

Heinz GroteMax Planck Institutfuer PlasmaphysikPostfach 1533D-85740 Graching bei Munchentel 498932992128fax 498932991212higfj?ipp.mpg.de

Tetsuo TanabeCenterfor IntegratedResearch in Science andEngineering,Nagoya University,Furo-cho, Chikusa, 464-8603 Nagoya, JapanTEL +81-52-789-5 177 or 5200FAX +81-52-789-5177 or 3791tanabe@cirse.nagoya-u. ac.jp

Jochen RothMax-Planck-Institut fuer PlasmaphysikBoltzmannstr. 2D-85748 Garching, Germanyjor@sat.ipp-garching. mpg.de

V. DoseMax-Planck-Institut tier PlasmaphysikBoltzmannstr. 2D-85748 Garching, Germany

Kaname KizuDepartmentofNuclear Engineering,GraduateSchool ofEngineering,Nagoya University,Furo-cho, Chikusa-ku, Nagoya 464-8603, JapanTel:+81-52-789-5481Fax:+81-52-789-5490Bh962402d@eds.ecip.nagoya-u.ac.jp

KenjlYamaguchiNuclearEngineering Research Laboratory,University of TokyoMailing address; Tokai-mura, Ibaraki-ken 319-1106, Japan

Tel; +81-29-287-8455, Fax; +81-29-287-$458yamag@tokai.t.u-tokyo. ac.jp

57

Keisliro OhkoshiDepartmentof QuantumEngineeringand Systems Science,GraduateSchool of Engineering, Universi@ of TokyoMailing address;c/o Nuclear Engineering andResearch Laborato~, University of TokyoTokai-mura, Ibaraki-ken 319-1106, JapanTel; +81-29-287-8434, Fax; +81-29-287-8488ohkoshi@tokai.t.u-tokyo.ac .jp

EgonVietzkeForschungszentrum JulichInstitut fuer Plasmaphysik D-52425 Julich, GermanyTel. 49-2461-61-3113Fax.49-2461-61 -2660E.Vietzke@fk-juelich.de

Charles H. SkinnerMS15Princeton Plasma Physics LaboratoryRoute 1 NorthP. O. Box 451Princeton NJ 08543phone:609-243 2214fax: 609-2432665cskinner@pppl. gov

Marek RubelPhysics DepartmentRoyal Institute of TechnologyAssociation EURATOM-NFRFrescativagen24S-104 05 Stockholm, SwedenTel. +46-8 161061FAX +46-8 1586’74rubelf?jmsi.se

R. A. AnderlIdaho National Engineering and Environmental LaboratoryIdaho Falls, ID 83415Tel. (208) 5334153FAX (208) 5265450raa@inel.gov

58

Glen LonghurstIdaho National Engineering and Environmental LaboratoryIdaho Fallx, ID 83415Tel. (208) 5269950FAX (208) 5260528gxl@inel.gov

Bill WamplerMS 1056SandiaNationalLaboratotiesPO Box 5800Albuquerque, NM 87185Tel. (505) 8444114wrwampl@sandia.gov

V. Kh. AlimovInstitute of Physical ChemistryRussian Academy of SciencesLeninsky Prospect 31117915 Mosclow, Russia Federationalimov@ipcras.msk.ru

Tatyana BurtsevaD.V. Efiemov Institute189631 St. Petersburg, Russia Federationburts@a14.niiefa.spb.su

Shigeki OhtsuDepartment of Quantum Engineering and Systems ScienceUniversity of Tokyoohtsu@q.t.u-tokyo. ac.jp

Kazuki KobayashiDepartment of Quantum Engineering and Systems ScienceUniversity of Tokyo

K. MoritaDepartment of Crystalline Material ScienceNagoya UniversityFuro-cho, Chikusa-KuNagoya 464-01, Japank-morita@nucl,nagoya-u.ac.jp

59

I

T. HorikawaDepartment of Crystalline Material ScienceNagoya UniversityFuro-cho, Chikusa-KuNagoya 464-01, Japan

Rion CauseyMS9161SandiaNational LaboratoriesLivermore, CA 94550Tel. (925) 2943326FAX (925) 2943231causey@ca.sandia.gov

Tom VenhausMS C348TA 21 Bldg. 209Los Alamos National LaboratoryLos Alamos, New Mexico 87545Tel. (505) 667-2138venhaus@titium. lanl.gov

.

60

DISTRIBUTION:

*200 MS 9161

>

3 MS 90181 MS 08991 MS 9021

1 MS 9021

R. Causey, 8716(for distribution to workshop participants)

Central Technical Files, 8940-2Technical Library, 4916Technical Communications Dept., 8815/Technical Library, MS 0899,4916Technical Communications Dept., 8815 for DOE/OSTI

61

1’I

Y

This page intentionally left blank

62