Surface modification of he pre exposed tungsten samples by he plasma impact in the divertor manipulator of asdex upgrade

ContentslistsavailableatScienceDirect Nuclear Materials and Energy journalhomepage:www.elsevier.com/locate/nme Surface modification of He pre-exposed tungsten samples by He plasma impact

ContentslistsavailableatScienceDirect

Nuclear Materials and Energy

journalhomepage:www.elsevier.com/locate/nme

Surface modification of He pre-exposed tungsten samples by He

plasma impact in the divertor manipulator of ASDEX Upgrade

S Brezinseka ,∗, A Hakolab , H Greunerc , M Baldenc , A Kallenbachc , M Oberkoflerc , G De

Temmermand , D Douaie , A Lahtinenb , B Böswirthc , D Bridac , R Caniellof , D Carraleroc ,

S Elgetic , K Kriegerc , H Mayerg , G Meislc , S Potzelc , V Rohdec , B Sieglinc , A Terraa ,

R Neuc , Ch Linsmeiera , the EUROfusion MST1 Team 1 ,2

a Forschungszentrum Jülich GmbH, Institut für Energie- und Klimaforschung - Plasmaphysik, Partner of the Trilateral Euregio Cluster, 52425 Jülich, Germany

b VTT Technical Research Centre of Finland Ltd., P.O Box 10 0 0, 02044 VTT, Finland

c Max-Planck-Institut für Plasmaphysik, D-85748 Garching, Germany

d ITER Organization, Route de Vinon-sur-Verdon - CS 90 046 - 13067 St Paul Lez Durance Cedex, France

e CEA, IRFM, F-13108 Saint-Paul-lez-Durance, France

f Istituto di Fisica del Plasma - CNR, Via R Cozzi 53, 20125 Milan, Italy

g Culham Centre for Fusion Energy, Culham Science Centre, Abingdon, OX14 3DB, UK

a r t i c l e i n f o

Article history:

Received 24 July 2016

Revised 27 September 2016

Accepted 1 November 2016

Available online xxx

Keywords:

PSI

ASDEX Upgrade

ITER

W divertor

W nanostructure

Helium

a b s t r a c t

Tungsten(W)willbeusedasmaterialforplasma-facingcomponents(PFCs)inthedivertorofITERand interactwithHelium(He)ionseitherfrominitialHeplasmaoperationorfromDeuterium-Tritium(DT) fusionreactionsintheactiveoperationphase.Laboratoryexperimentsreportedthatinaspecific opera-tionalwindowofimpactenergy,ionfluence,andsurfacetemperature(E in ≥ 20eV,φ≥ 1× 1024He+m−2,

T surf ≥ 1000K)amodificationofWsurfacesoccursresultingintheformationofHe-inducedW nanos-tructures.ExperimentsinASDEXUpgradeH-modeplasmas (B t =2.5T,I p =0.8MA,P aux ࣃ8.0MW)in

Hehavebeencarriedouttoinvestigateindetail(a)thepotentialgrowthofWnanostructureson pre-damaged W samples incorporating Henanobubbles, and (b) the potential ELM-inducederosionof W nanostructure.BothWsurfacemodificationsweregeneratedartificiallyintheGLADISfacilitybyHe bom-bardmentofWsamplesatE in =37keV(a)to φࣃ0.75× 1024He0m−2atT surf ࣃ1800Kand(b) φࣃ1×

1024He0m−2atT surf ࣃ2300KpriortoexposureinthedivertormanipulatorofASDEXUpgrade.Though

inpart(a)conditionsofWnanostructuregrowthwithatotalHeionfluenceofφࣃ1.6× 1024He+m−2 andpeakHeionimpactenergiesabove150eVweremet,nogrowthcouldbedetected.Inpart(b)lower densityplasmaswithmorepronouncedtypeIELMs,carryingenergeticHeionsinthekeVrange,were executedwiththestrike-linepositionedon2μmthickWnanostructureaccumulatingafluenceofφ ࣃ 0.8× 1024He+m−2.Post-mortemanalysisrevealedthatco-deposition bypredominantlyW,and Boron (B),erodedatthemain chamberwalland transportedintothedivertor, tookplaceonallWsamples ErosionofWnanostructureoritsformationwashinderedbythefactthattheouterdivertoratthe loca-tionofthesampleswasturnedundertheseHeplasmaconditionsintoanetdepositionzonebyW,Band Carbon(C)ions.Thesurfacemorphologywithlargeroughnessandeffectivesurfaceareaactasacatcher fortheimpingingimpurities.Thus, apartfromoperationintheexistencediagramofWnanostructure withrespecttoT surf ,φ,andE in ,alsotheimpingingimpurityfluxcontributionneedstobeconsideredin predictionsconcerningtheformationofWnanostructures

© 2016TheAuthors.PublishedbyElsevierLtd ThisisanopenaccessarticleundertheCCBY-NC-NDlicense (http://creativecommons.org/licenses/by-nc-nd/4.0/)

∗ Corresponding author

E-mail address: s.brezinsek@fz-juelich.de (S Brezinsek)

1 see http://www.euro-fusionscipub.org/mst1

2 This work has been carried out within the framework of the EUROfusion Con-

sortium and has received funding from the Euratom research and training pro-

gramme 2014–2018 under grant agreement No 633053 The views and opinions ex- pressed herein do not necessarily reflect those of the European Commission http://dx.doi.org/10.1016/j.nme.2016.11.002

2352-1791/© 2016 The Authors Published by Elsevier Ltd This is an open access article under the CC BY-NC-ND license ( http://creativecommons.org/licenses/by-nc-nd/4.0/ )

Please citethis articleas: S Brezinsek etal., Surface modification ofHe pre-exposed tungsten samplesby He plasmaimpact inthe

1 Introduction

Tungsten (W) willbe usedasmaterialforplasma-facing

com-ponents(PFCs)intheITERdivertor[1] duetoitsgoodpower

han-dlingproperties,highmeltingpoint aswellaslow fuel retention

andsputtering inplasmaswithhydrogenicfuel [2] These

advan-tageshavebeensuccessfullydemonstratedindeuterium(D)

plas-masinASDEXUpgradeandJETwithfullWdivertors[3] However,

operationinhelium(He)ormixedD/Heplasmashavebarelybeen

executedindeviceswithWPFCsandareofvitalimportancease.g

ITERplanstooperateinHeplasmasduringthenon-activestart-up

phaseandwill producefusion α-particles byDT reactions inthe

activephase

Laboratory experiments,i.e.in linearplasmadevicesand

neu-tralbeamfacilities, reported onmicrostructure modifications

un-der energeticHe ionor atom bombardmentof W PFCs resulting

inWnanostructures(WfuzzorW dendrilsorWcoral-like

struc-ture)[4,5] TherequiredWsurfaceandHeion(Heplasma)

condi-tionsfortheformationofWnanostructurehavebeensummarised

in[6] includingtheidentificationofHe nanobubblesinW as

po-tentialprecursor W nanostructures are formed when theimpact

HeionenergyE inexceeds20eV,thesurfacetemperatureT surf

ex-ceeds 1000 K, and the fluence φ is larger than approximately 1

× 1024He+m−2.Recentstudies[8] identified a threshold for their

formation at low fluence These W nanostructures could

poten-tiallycompromise power handling,impact the fuel retention [7] ,

andinduce finally dust formation fromW plasma-facing

compo-nents[2] A critical questionis ifsuch surfacemodifications can

beobserved inpresent-day tokamaksoperation inELMy H-mode

conditionsandiftheycompromisetheplasmaoperation

Pioneering experimental studies with this respect have been

carriedoutintheTEXTORdeviceunderL-modeconditionsin

plas-maswithHemajority[9] Anactivelypre-heatedlimiterequipped

withWstripes,pre-exposedintheNAGDIS-IfacilitytoHeplasmas

providingnanostructuredWononestripeandHenanobubblesin

Won anotherstripe, wasexposed to Hetokamakplasmas under

conditionswhichshallpromoteafurthergrowthofW

nanostruc-tureand its formation arising fromthe He nanobubbles,

respec-tively.TEXTOR, operating witha first wall madeof graphite and

thisactively heated, roof-like limiter equipped withthe different

Hepre-damagedWstripesshowednogrowthofWnanostructures

though all three criteriamentioned beforefor growth were met

Instead in the near scrape-off layer (SOL), W nanostructure

ero-sionbyimpingingC impurityionsoriginatingfromfirstwall

ero-sion[10] has beenobserved on thecorresponding limiter stripes

anddeposition by carbon wasdetected a few centimeters inside

the SOL on all W stripes Thus, in the case of L-mode plasmas

the impinging flux distribution of low-Z impurity ions is an

ad-ditionalparameterdeterminingifWnanostructureisgrowing.The

concentration oflow-Z impurities, inthe case ofTEXTOR C with

a flux ratioof C∼(0.03)×( He+D ) ,determines iferosion ofW

takesplaceordeposition bylow-Zimpurities.Comparablestudies

inavirtual-freelow-Zdevicewereexecutedin[11] intheall-metal

tokamakCMOD Here,indeed thethree abovementioned

param-eters required forW nanostructure formation were fulfilled at a

W samplelocated atthe noseof the outer divertorand clearW

nanostructuregrowthwasobservedwithin12discharges

accumu-latingin 15 the required He ionfluence However, the plasma

operationwasinanELM-freeH-mode,thuswithoutanyhigh

en-ergeticionsimpingingduringan ELMeventatthetargetplateas

itwillbethecaseinITER.Thus,inviewofITER,studiesintype-I

ELMy H-modedischarges are requiredto concludeif thebalance

between W nanostructure growth and erosion [12] is shifted in

favourofthenanostructureformationorifneterosionunderELM

impact takes place Studies in PISCES-B simulating ELM-like

be-haviourbyplasmabiasinginHedemonstratedaclearerosionafter singleevents[13]

Here, we report on experiments carried out in He ELMy H-modeplasmasinASDEX-Upgrade afullW-devicewithoccasional use of boronisation and low residual C content [3] The exposi-tion of W samples in the outer divertor of ASDEX Upgrade aim

to investigate indetail (a) the potential growth of W nanostruc-tures on pre-damaged W samples incorporating He nanobubbles and(b) the ELM-induced erosionof thick W nanostructures The pre-damage ofWsamples byenergetic Heatomswasdone prior

tothetokamakexposureintheneutralbeamfacilityGLADIS pro-vidinga set ofsamples exposed to differentHe fluence and sur-facetemperatureresultinginparticularinsampleswithcoral-like

W nanostructure andsamples withHe nanobubbles The GLADIS operational parameters were selected accordingto [5] where de-tailed characterisation of the surface morphology after the adia-baticloadingwasdonebypost-mortemanalysis.Subsequently, AS-DEXUpgradeplasmaandsurfaceconditionsareselectedtoinduce formation orgrowth of W nanostructures on thesepre-damaged

Wsamples.Theseexperimentalconditionsinthefullmetallic en-vironmentareclosetoconditionsexpectedinthedivertorofITER duringtheHestart-upphase[15]

2 Experiment

TheexperimentalapproachtostudyWnanostructurebehaviour

inHe-dominatedASDEX Upgradeplasmaspresentedhereis com-parable to the TEXTOR studies [9] employing three steps of W sampletreatment:(i)Hepre-exposureofWsamplesinlaboratory plasmas(here: GLADIS),(ii)Exposureofpre-damagedW samples

intokamakplasmas(here:ASDEXUpgrade),and(iii)post-mortem analysis of W samples andcomparison with the reference mor-phology (here: IBAand SEM) Fig 1 a showsthe set of polycrys-tallineW samples (T1-T6) afterHe expositionin GLADIS corre-spondingtopart(i).Fig 1 billustratesthearrangementofthesame

W samples(T1-T6)afterHe plasmaexpositioninASDEX Upgrade

-correspondingtopart(ii).Thepurposeofthethreeexperimental partscanbedescribedinthefollowingway:

• Pre-exposure of W samples to a set of different fluences of

Heparticles andsurfacetemperaturesincontrolled laboratory plasmaconditionsproviding pre-damagedW sampleswithHe nanobubbles atthe fluence onset to W nanostructure forma-tionuptofullydevelopedWnanostructuresofseveral microm-eterthickness.Here,GLADISwasusedfortheHepre-exposition

[5] in contrast to a linear plasma device in the case of the TEXTOR experiment The detailed surface morphology differs slightlybetweenGLADISoperatingathighHeatomimpact en-ergiesoftensofkeV inducinga coral-likestructureandlinear plasmadevicesoperatingatlowHeionimpactenergiesoftens

ofeVinducinga tendril-likestructure [11] Bothhowever pro-videarepresentativenanostructuredWsurfacewith compara-blepropertieslikehighporosity,reducedthermalconductivity etc

• Exposure of pre-damaged W samples installed in the DIver-torManipulator(DIM-II)[16] ofASDEXUpgradelocatedinthe outerdivertor.Thestrike-linepositionisusedtoexposein se-quence different He pre-damaged W samples under different plasmaconditions.Thisutilisationofthestrike-lineposition al-lowsto separate the two main questions: (a) Will W nanos-tructureformationbeinitiatedonthesample(T2)withalready pre-implanted He nanobubbles and initial thin nanostructure formed iftheoperational conditions arefulfilling the require-mentsforgrowth?(b)HowwillexistingthickWnanostructure

bemodifiedbytokamakplasmaimpact?Here,inparticularthe questionofELM-induced W erosionwill be addressedforthe

Fig 1 a) He-exposure arrangement of W samples in GLADIS b) He-exposure ar-

rangement of W samples in the divertor manipulator of ASDEX Upgrade (AUG) with

the three strike-line positions for the different experiments indicated

first time ina metallic tokamak environment The complexity

ofthelatterincludesalsothematerialmigrationfromthemain

chamberintothedivertorandwithinthedivertorwhereas

un-dernormaldeuteriumplasmaconditionstheouterdivertorleg

is representing a net-erosion zone [17] As reference for the

global migration behaviour, a second set of six Molybdenum

(Mo) samples of identical dimension is installed at the same

poloidalposition, buttoroidally shifted by 2 cm onthe same

WcoverageplateinDIM-II(Fig 1 b)

• In-situ information about the change of surface morphology

during tokamak plasma operation are very limited, i.e here

onlyavisiblecameraobservationisavailable(Fig 3 b)recording

the integral change of photon emission owing to variation in

surfacetemperature, reflectionproperties,He recyclingandW

erosion.Different ex-situanalysismethods likeScanning

Elec-tron Microscopy (SEM), Nuclear Reaction Analysis (NRA),

Sec-ondary IonMass Spectrometry (SIMS) are required to

charac-terisetheexpectedchangesinsurfacemorphology,inimpurity

coverage, W erosion,andHe content inW post mortem

Pre-characterisationofWsamplesisperformedbySEMafter

expo-sitioninGLADISwhichallowadirectcomparisonofWsamples

beforeandafter tokamakplasmaexposure asshowninFig 2

forsamplesT2andT3ontheleftandright-hand-side,

respec-tively The SEM methodology employing contrast for

compo-nentanalysisiscomparabletotheanalysisreportedin[19] on

WsamplesexposedtodeuteriumplasmasinASDEXUpgrade

Fig 2 Left: SEM images of samples T2 (just above the onset of W nanostructure

formation), T3 (W nanostructures formed), and T4 (W nanostructures formed) after

He bombardment in GLADIS Right: Comparable SEM images after exposure to He plasmas in ASDEX Upgrade (AUG) visualising changes in surface morphology

Table 1

Exposure conditions for the different tungsten samples in GLADIS

He flux [10 21 m −2 s −1 ] ∼ 1 7 ∼ 1 3 0 7

He fluence [10 24 m −2 ] ∼ 1 0 ∼ 0 75 ∼ 0 4 Integral loading time [s] ∼ 580 ∼ 580 ∼ 580 Peak surface temperature [K] ∼ 2300 ∼ 1800 ∼ 1300

2.1 Helium exposure in GLADIS

The He exposition conditions of W samples in GLADIS were comparabletothe experimental studies describedin[5] wherea comprehensiveandcompletecharacterisationofthesurface modi-ficationsoftheappliedpolycrystallineW,alsousedinthis experi-ment,isdone.SixWsamples(T1-T6:eachofdimension30× 12×

4mm3)wereembeddedinamolybdenum-alloy(TZM)targetplate andexposedtothebeamatE in=37keVofHe0.Fig 1 ashowsthe geometricalarrangement ofthe samplesaswell asimplicitly the heat andparticleload footprint onthe target plateby darkening withmaximum incident particle flux andsurfacetemperature in thecentreonsamplesT3andT4.ThoughtheGLADISbeamprofile

isnothomogeneousandtheloadingoccursadiabatically,itis pos-sibletodescribetheaveragedexposureconditionsforthedifferent samplesassummarisedinTable 1 Wnanostructureofabout2μm thicknessareformedonWsampleT3andT4withafluenceofφ

ࣃ1 × 1024 He0m−2 andatapeaksurfacetemperatureofT surf ࣃ

2300K.InparticularsampleT3isfurther usedforW nanostruc-tureerosionstudiesinASDEX Upgrade Thecorresponding values forT2andT5areφࣃ0.75× 1024He0m−2andT surfࣃ1800K.The strongestgradientintheimpingingfluxoccursonthesetwo sam-ples:thelowerpartofT2andtheupperpartofT5arejustatthe

Fig 3 a) Magnetic configuration in the divertor for the three phases of exposure

and the line-of-sight of the camera on the divertor manipulator location b) Image

of divertor manipulator plates during plasma exposure in the visible spectral range

thresholdforW nanostructureformation,butwithsignificant He

nanobubblesincorporatedintheWmatrixwhereastheupperpart

ofT2 andthe lower partof T5 alreadyfully develop the typical

coralnanostructure ofW.The centre ofsampleT2 hasbeen

tar-getedintheASDEXUpgradeexperimentbytheouter-strikelineto

studytheformationandgrowthofWnanostructuresasdescribed

below.However,duetothepresentflux andtemperatureprofiles

owingtothefootprint intheHepre-exposition, awider rangeof

Wnanostructureswithdifferentheight,density,andthicknesswill

result.RepresentativeSEMimagesofthemicrostructuresurfaceof

sampleT2,T3,andT4afterGLADISexposureareshowninFig 2 a

2.2 Exposure in ASDEX Upgrade He plasmas

H-mode experimentsinASDEX Upgradehavebeencarriedout

inlower singlenullconfiguration(Fig 3 a)ata toroidal magnetic

field of B t=2.5 T, a plasma current of I p=0.8 MA, and

auxil-iary power of about P aux=8.5 MW consisting of a mixture of

neutral-beaminjectioninHe andHofminimum2.1MW,ion

cy-clotron resonance heating (ICRH) of up to 3.9 MW, and central

electron-cyclotronresonanceheating ofabout2.6MW.The

oper-atinggas andmainplasma specieswasHe witha purityof

typ-ically n ( He )

n ( He )+n ( D )+n ( H ) >0.8 These He plasmas were executed

di-rectly after a fuel species exchange experiment from D into He

by ion-cyclotron wall conditioning (ICWC) plasmas It should be

notedthat the last boronisation on B2D6 in ASDEX Upgrade was

just2 experimental days with about 175 plasmaseconds before

theICWC change over in24 dischargestook place, thus thefirst

He plasmasmight still be affected by the boron coverage of the

firstwallanditsconditioningeffect[20] Thetypicalcoreplasma

conditionsoftheseHeplasmasduringtheH-modeflat-topphase

ofabout7 durationare:electrondensityn c

e9.5× 1019m−3and electrontemperatureT c

e 3.0keV

The goal of the exposure to predominately He plasmas is to

investigate(a) thepotential growthof W nanostructures on

pre-damagedW samplesincorporatingHe nanobubbles(type A

plas-mas/sampleT2),(b)theerosionofthick Wnanostructures(type

B plasmas / sample T3), and (c) the impact of Nitrogen (N) on

the W nanostructure (type C plasmas / sample T4) Thus, three

studiesinASDEXUpgradewereexecutedinoneexperimentalday

consistingof 25 diverted plasmas (#32642− 32466) by variation

oftheouter strike-linelocated onthreedifferentHe pre-exposed

W samplesinstalled inthe divertor manipulator atthe low field

side as shown in Fig 1 b The correlation between plasma type,

outerstrike-lineposition,andWsampleisdepictedinFig 4 The

coreplasmaconditionsaresimilarapartfromthestrike-line

posi-Fig 4 Schematic arrangement of the three exposure conditions, number of dis-

charges, and outer strike-line positions

Fig 5 Plasma conditions in the near scrape-off layer region measured by an array

of Langmuir probes for discharges in exposure conditions A: a) ion saturation cur- rent, b) electron temperature, and c) electron density as function of the change of the s-coordinate with respect to the magnetic outer strike-line position

tionvariedbetween =1.02m(typeA), =1.05m(typeB),and

s=1.08m(typeC)aswellasthelowerHegasinjectionrate dur-ingtypeBplasmasthoughcontrolledfuellingwascompromisedby

HeoutgassingfromtheWfirstwall.Thetwoplasmadischargesof typeChadadditionalN2blibs.Weassumethatthedifferent strike-pointpositionsandtheorderofexperimentsallowaseparationof thethreeexperimentalparts.Nosignificantimpactonthe plasma-exposedWsamplesurfacesisexpectedwhentheouterstrike-line

ismovingupwards,fromtype AtotypeCplasma, intothe direc-tionofSOLwhichleavesthemodifiedWsurfaceintheprivate-flux region(PFR)undernegligiblepowerandionicparticleload

• TypeA theWnanostructuregrowthregime.In thiscasethe strike-lineispositioned atthepoloidaldivertor coordinate =

1.02m and the correspondingtime averaged radial profiles for thelocalelectron temperatureT e,electrondensityn e,andthe ionsaturationcurrentdensityj sat areshowninFig 5 Assum-ingthat singlyionised He ionsare impingingthe outer diver-tortarget plate, thepeak impact ionenergycanbe estimated

tobeaboutE in=150eV,assumingk B T e=k B T i ,andthusclearly abovetherequired20eV.Thepeakionfluxdensityatthe loca-tionofthestrikelineamounts j sat=2.0× 1023m−2 −1,80%of whichcanbe attributedtoHe ions.In14discharges executed

inscenarioAan integralplasmaexposuretimeofabout100 hasbeenreached Thisleadsto an achievedHe ionfluenceof

φ=1.6× 1025m−2 andthusoneorderabovethethreshold re-quiredforW nanostructuregrowthfrom undamagedW sam-ples.Therehasbeennodirectmeasurementofthesurface tem-peratureatthe locationof theDIM-II,however,infrared mea-surements atdifferent toroidal location at normalW divertor target platessuggest atleast a temperatureof800 K of stan-dardW PFCs withgoodthermal conductivity.The W samples installedintheDIM-IIhaveapoorthermalcontacttothe base-plate.The surfacetemperatureisthereforeassumedtobe

sig-nificantlyabove1000K.Visiblecameraimages(Fig 3 b)indeed

observestrongthermalradiationatthepoloidallocationofthe

exposed W sample whereas the W covering plate shows no

such radiation though receiving the same power andparticle

loadat the samepoloidal location, butjust slightlytoroidally

shifted.It should be notedthat theH-mode performance was

overallpoorinthecaseofTypeAplasmas(pedestalconditions:

T e ped330 eV and n e ped6.15× 1019 m−3) with an ELM

fre-quencyinthekHzrangesuggestingtypeIIIlikeELMyH-mode

withmoderateintra-ELMHeionimpactenergiesandlow

ELM-induced W sputtering asobserved before in D plasmasusing

standardWPFCs[3]

Overall, the experimental conditions in the first part of the

experiment (type A) are fulfilling all the requirements for W

nanostructureformationandgrowthinASDEX Upgrade

How-ever,visible inspection ofthe W sampleT2 directlyafter end

oftheplasmaexposuresuggestedachangeinreflection

prop-ertiesfromablackenedtoashinyWsurfaceindicating

modifi-cationsinthesurfacemorphology,butnoclearsignatureofW

nanostructureformation

• Type B theW nanostructureerosion regime In this second

set of9 comparabledischarges, theaim is tostudy the

ELM-induced Wnanostructureerosion.The outerstrike-lineinthis

secondsetofdischargesispositionedat =1.05monW

sam-pleT3employinganabout2μmthickW nanostructureastop

surface The averaged edge plasmas conditionsT e,n e, andj sat

reflecting mainly theinter-ELM phase are comparableto type

AplasmasaswellasT surf.Thecorresponding fluenceφ inthe

about63plasmasecondsamounts to1.0× 1025 He+m−2 and

wouldstillpotentially allowW nanostructuregrowthbetween

ELMs.Thefuellingrateinthissecondsetofplasmasisreduced

to half of the value of type A plasmas in order to provide a

morepronouncedH-modewithlowerELMfrequency(120Hz)

andlargerenergydropperELM.Theimpactenergyof

presum-ably He2 + fromthe pedestal region arriving at the outer

tar-getplatein lessthan1 ms issubstantially above the

sputter-ing thresholdofW.Assumingarelationship betweenpedestal

energyandimpactenergyaccordingtothefree-streaming

ap-proach as recently observed in JET-ILW in [22] , the impact

energy in these He plasmas is assumed to be above 1 keV

ThiswouldindeedallowtheexpectedcompetitionbetweenW

nanostructuresputteringby ELMimpactunderconditions

oth-erwise favouring W nanostructure growth The visual

inspec-tionofsampleT3revealsalsoa shinyW surface,suggestinga

changeofthesurfacemorphologywithreducedsurface

rough-ness theoriginofwhichwillbediscussedinthenextsection

• TypeC theimpactofNonWnanostructure N2 wasinjected

into two He plasmaswith theouter strike lineat =1.08m

impactingonsampleT4withthickWnanostructure.The

inter-actionof Nwiththe W morphology willbe studied infuture

bypost-mortemanalysis

3 Discussion

ThefirstobservationafterexposureofthesixHepre-damaged

W samplestoASDEX Upgrademaybesummarised inthe

follow-ing way:change of surfacemorphology and roughnesson all W

samplesstartingatPFR,passingthethreeappliedstrike-line

posi-tions,andendingintheSOL(Fig 1 b).Differentpost-mortem

anal-ysistechniques,i.e.ionbeamanalysis(IBA) andscanningelectron

microscopy (SEM)combinedwithfocusedionbeam(FIB)cutting,

are applied to obtain physics information forinterpretation SEM

images ofsubareas (50μm × 40μm) onsamples T2, T3, andT4

before (left-handside)andafterASDEX Upgrade exposure

(right-hand side)are depictedin Fig 2 showingthe mentioned surface

modification withreduction ofsurfacerougheningandingeneral

Fig 6 a) Cross-section of a tungsten nanostructure sample after ASDEX Upgrade

plasma exposure b) Enlarged images to show detailed structure with sharp cuts from erosion as well as homogenous deposition of predominantly W c) Focused ion beam cut of a tungsten sample showing the deposition of a manifold of individual layers corresponding to the number of discharges

absence of the coral-like W nanostructure Cross-sectioning with FIBandSEMrevealsthattheobservedsurfacemodificationisdue

toanalmosthomogeneouscoveragebya0.5− 1.0μmthick depo-sitionlayerasshowne.g.inFig 6 aforsampleT3 Fig 6 cshowsa cutthroughthedepositionlayeroftheupperpartofsampleT2 in-dicatingmorethan20individualsubstructuressuggestingthat ef-fectivelyeachofthe25executed dischargesiscontributingtothe deposition.Indeedbothhigh-Z(W)andlow-Zspeciesarepresent

in thisdeposition layer accordingto contrast analysis with SEM Thechemicalcompositionanddistributionofthedeposition layer hasbeendeterminedbyIBAandidentifiedtobeamixtureofB,C,

Daswell asW;energydispersive X-rayspectroscopy(EDX) con-firmsthesignificantappearanceBandC.Fig 7 ashowsthepoloidal depositionprofiles ofB,C,DonthedifferentHe pre-damagedW samples and Fig 7 b the corresponding profiles on the polished

Moreference samples.In addition Fig 7 b includes alsoW depo-sition profile on the Mo samples Fig 7 a showsthe He ion flux distribution,thus the distribution ofimpinging fuel species mea-suredbyLangmuirProbesfortypeBplasmaswiththeouterstrike linelocated at1.05 m Common forthe deposition patternofall specieson the pre-damagedW andthe Mo referencesamples is the strong deposition deep inside the PFR region Indeed, there

isonly very limiteddeposition ofB on thepolished Mo samples above =1.02 mwhichisthelowest positionoftheouter strike linein theexecuted discharges.Thus, one can concludethat sig-nificant deposition is absent in the SOL in the case of polished surfaces with low surface roughness In contrast, in the case of sampleswithWnanostructure,whichhavefullyestablished coral-structure betweenapproximately x=20 mm andx=110 mm in poloidaldirection,astrongdepositionofBandCcanbemeasured without any significant D embedded in the deposit In the SOL, theB depositionprofile followstheimpinging Heionflux profile whereastheC depositionprofileisflat intherangeofthe estab-lishedW nanostructureincludingthree minorpeaksatthe inter-section of samplesT2/T3, T3/T4, and T4/T5 Re-erosion fromthis area ofcomplexsurfacemorphology by physicalsputteringis re-duced, whereas chemical erosion still can occur which might be thecauseforthedifferencebetweenBandC.Overalltheaveraged

BtoC ratiointhe depositisapproximatelytwo Unfortunately,it

isnotpossibletoprovidefromthismeasurementtheratioof

low-Zimpurities tothe alsopresentimpinging W ionflux whichhas beenmeasuredontheMoreferencesample

Wsamples withpre-formedW nanostructureswithlarge sur-faceroughness, large effectivesurfacearea,andhighporosity act

as a kind of catcher for the impinging ions from the plasma Quickly,after establishment of a thin deposition layer (B, C, W), theexistingWnanostructureiscoveredandprotectedfromfurther erosion.Every additionaldischarge isinducing another layer and contributingtotheobserveddepositionwithmorethan20 identi-fiedlayers.Contrary,inthecaseofthepolishedMoreference

sur-Fig 7 a) Poloidal distribution of B, C, and D along the six tungsten samples installed in the divertor The He ion flux distribution is shown accordingly as reference for the

impinging flux b) Deposition of W as well as B, C, and D on the reference poloidal sample stripe made of Mo

facearetheimpinginglow-Zparticleseitherre-erodedorreflected

fromthehigh-Zsurface similartoobservationsmadeinTEXTOR

withCon W [21] andno effectivedeposition layergrowthcan

occur.Onecanassumethatgeometricallyonbothpoloidalsample

stripes,theW andMoone,thesameimpurityflux distributionis

impinging, thus, the difference in the two deposition profiles on

thestripes provideseffectively thefluence ofimpinging impurity

ionsaccumulatedover all discharges.The nearSOLregion ofthe

outerdivertorlegisundernormalDplasmasconditionsinASDEX

Upgradeaclearnet-erosionsourceofW[17] ,butinthepresentHe

plasmasthebalancebetweenerosionanddeposition haschanged

towards deposition Indeed on the polished Mo samples it is

al-mostfullybalanced andinthecaseoftheW nanostructurewith

largersurface roughnessre-erosion is furtherreduced The outer

divertorislocally,atthelocationoftheWstripes,transferredinto

anet-depositionzone forCandB ThehighlevelofB inthe

de-posit is caused by the boronisation which took place about 175

plasma seconds before the actual experiments The protective B

layerinthemainchamberiscontinuouslyerodedbyplasma

bom-bardmentand thereleased B isionised and transportedintothe

divertor as previously observed in D [18] The B erosion in the

mainchamber islargerinthepresentHeplasmasthaninD

plas-masundersimilar plasmaconditionsdueto highermassand

ef-fectivechargeoftheimpingingion.ThelevelofCisinthenormal

rangeofASDEXUpgradeplasmas

The observedWinthedivertor isalsoaresultofmain

cham-ber limitersandheatshield erosion,butduetoELMimpactas

reportedforDplasmas before[14] The W source inthe present

Heplasmasis significantlyhigherthan inDduetoa)the higher

massoftheimpingingprojectileandb)thehigherionisationstage

ofimpingingHe incomparisonwithD, andfinallyduetotheuse

ofICRHantenna[14] Thetransport ofhigherionisationstagesof

Wfromthemainchamberintothedivertorthenresultsinthe

de-positiononboththe Wnanostructuresamplesaswell asthe Mo

referencesamples

FocusingagainontheinitialexperimentalgoalsoftheHe

plas-masoftypeAandtypeB,onecansummarisetheobservationsand

conclude:

• TypeA theW nanostructuregrowthregime.No signatureof

W nanostructureformationorgrowthunderthegiven

experi-mentalconditionsinASDEXUpgradethoughtheproper

condi-tionsforWnanostructureformationwithrespecttoionimpact

energy, fluence and surfaceconditions are met in these

plas-masof type Aandthe outer strikelinepositioned on sample

T2.SEMofcross-sectioncutsofthesamplerevealadeposition

layerconsistingofB,C,andWwithhighdensityonthe strike-lineareaonsampleT2.Strongestdepositionisobservedbelow theactualstrike-linepositioninthePFRandalsoonthe refer-enceMosamplesasvisibleinFig 1 b) Potentialreasonmight

be relatedto −→E ×−→B drifts in thedivertor assuggestedforD plasmasinsimilarconfigurationby[24] TheoverallDcontent

isnegligiblylowintheSOL,butsubstantial upto15%ofthe codeposit inthedepositionenrichedregioninthePFR

• TypeB the W nanostructureerosion regime Onlymoderate erosionofsampleswithWnanostructurewasobservedcloseto theouterstrike-lineregiononsampleT3asdepictedinFig 6 a

Fig 6 bshowsaSEMimageofacorrespondingcross-sectioncut

ofW sampleT3with initiallyabout2μm thick W nanostruc-ture The top part of the coral-like structure is flattened and overlaid by a homogeneous deposition layer There is no in-dication of melting of the top part of the nanostructure, but ratherlocalerosion ofindividual “W-corals” tookplace which effectively smoothed the microstructure while the deposition coveredthe whole surfacearea.The erosionis determined by ELMs carryingHe2+ withabout1keV impact energy fromthe pedestaltotheoutertargetplatesimilartoobservationsin JET-ILW[22] Evenifthesputtering yieldfornanostructuredW is slightlylowerthanforstandardW asreportedby PISCES-Bfor low impact energies [12] , the erosion yield dependence as a function of the impact energy can be assumed to be compa-rableatenergiesabovekeVtostandardW.AlthougheveryELM impactindeedfulfilsconditionsforWerosionbyphysical sput-tering,theerosion process wasinthe presentcaseingeneral overcompensatedbylocaldeposition predominantlyoccurring betweentheELMs It should benoted that theoriginal coral-likestructure onT3remains detectable bySEM ina very low fractionof surfacearea closeto the strike-lineposition Thus,

inthis case, theflux of ELM-induced sputteredparticles is in balancewiththefluxofdepositingparticles

The interplay between local erosion of W nanostructures at the strike-zone and deposition in the SOL observed here is in accordance withprevious TEXTOR W limiter experiments in He-dominated plasmas [9] though latter were executed in ELM-free conditionsandsolelyinfluencedbyCimpurities.Bothexperiments stresstheuniversalcharacterandtheimportanceofthelow-Z im-purityfluxinthelocalbalance equationbetweenerosionand de-positiononWnanostructuresurfaces

4 Conclusion

InHe ELMy H-modeplasmasin ASDEXUpgrade nogrowthof

Wnanostructurestakesplacedespitethefactthat theoperational

conditionsfulfiltherequirementsfortheir formationandgrowth

Thus, the formation of W nanostructures dependsin addition to

the known dependencieson thefluence of Heions, their impact

energy, andthesurfacetemperature[6,15] also onthe impinging

flux of impurities such as boron and carbon as well as W ions

onto the outer W target plate Neither W nor intrinsic impurity

ionshaveinthecaseofASDEXUpgrade theirorigininthe

diver-tor,butarearesultofmainchambererosion WPFCsandBfrom

boronisation andadjacentmaterialtransport towards the

diver-tor Inthe caseofELMy H-modeplasmas inASDEX Upgrade,the

plasmaoutfluxofB,C,andWionscompensatestheWerosionflux

andshiftsthelocalbalanceintheouter divertorfromnet-erosion

intonet-depositionthoughELM-inducedWsputteringclosetothe

strike-linetakes place.Moreover, the highestdeposition oflow-Z

andhigh-Z speciesislocatedinthe private-fluxregionindicating

eithera localtransport mechanismfromthescrape-off layerinto

the private-flux region or significant re-erosion atthe strike-line

location Thisgeneralerosionanddeposition patternisnotsolely

relatedtotheW surfacemodificationsasasimilar patterncanbe

observed onreferencepolished Mosamples atthesamepoloidal

position, buttothe globaltransport of impuritiesfromthe main

chamber into the divertor as seen before in [23] In the case of

the sampleswithW nanostructure, surfaceroughnessand

poros-ityincreasesstronglythedegreeofdepositionwithrespecttothe

polished Mo surfacesandvisualises effectivelythe impinging ion

fluence

It is evident that themain chamber W sourcein these

ICRH-heatedHeplasmasissignificantlyhigherthaninH-modeD

plas-masinordertocausetheshiftinthelocalWerosionand

deposi-tionpattern.Fastimpurityionsanddeuterons(i.e.intra-ELM)were

identified before[14] asthemain causeforW erosionatlimiters

andheatshieldsinthemainchamber.Theincrease ofthe

projec-tile massofthe mainfuel speciesfromD(m=2)to He (m=4),

thelower sputteringthreshold,andthehigherchargestate ofHe

ions hitting the wall are contributing to the increased W wall

source inHe.However, furtherpost-mortem analysisneedsto be

done to geta furtherinsight inthe interplaybetweenthe

differ-ent parametersforW growthandthe observednet-depositionof

WbyW,BandCalongthefulloutertargetplateincludingtheMo

referencesamples.Moreover,localspectroscopyatthelocationof

thedivertormanipulator wouldbe advisableinorderto measure in-situthedifferentfluxcontributions

TheobserveddepositioncontributionofBandCinASDEX Up-gradecan beseenasproxytointrinsicberyllium(Be) inHe plas-masinITER Theinitial formationof Wnanostructures atcritical areasintheWdivertorfulfillingthestandardcriteriaforWgrowth can be hinderedifthe local erosion/depositionflux balance is in favourofdeposition duetostrongerBemainchamber sources in

He incontrast to D Indeed the Beerosion source underHe im-pactis expectedto be atleast twice aslarge asin D Repetition

oftheexperimentatJET-ILWinHeplasmaswouldbeadvisableto investigateachangeintheprimary Besourceandmigration pat-tern[25] aswellastoprovideinputformodellingcodes,WallDYN

[26] fortheglobalmigrationpatternandthebalanceinthe diver-tor,andERO [27] forthelocalELM-induced W sputteringeffects Subsequently, more detailed studies on W nanostructure erosion duetoELMimpactneedtobeexecuted atITER-relevantELM im-pactenergiesandionfluxesbeforefurthermodellingoftheW sur-facemorphologychangescanbedone

References

[1] R A Pitts, et al., 2013 55th Annual Meeting of APS Division of Plasma Physics, Denver, USA, WE1.0 0 0 01

[2] V Philipps , J Nucl Mater 415 (2011) S2 [3] R Neu , et al , IEEE Trans Plasma Sci 42 (2014) 552 [4] M.J Baldwin , R.P Doerner , Nucl Fusion 48 (2008) 035001 [5] H Greuner , et al , J Nucl Mater 417 (2011) 495 [6] S Kajita , et al , Nucl Fusion 49 (2009) 095005 [7] M Baldwin , et al , Nucl Fusion 51 (2011) 103021 [8] T.J Petty , et al , Nucl Fusion 55 (2015) 093033 [9] Y Ueda , et al , J Nucl Mater 415 (2011) S92 [10] S Brezinsek , et al , Phys Scr T167 (2011) 14067 [11] G.M Wright , et al , Nucl Fusion 52 (2012) 042003 [12] D Nishijima , et al , J Nucl Mater 415 (2011) 230 [13] D Nishijima , et al , J Nucl Mater 434 (2013) 230 [14] R Dux , et al , J Nucl Mater 390–391 (2009) 858 [15] Y Ueda , et al , J Nucl Mater 442 (2013) S267 [16] A Hermann , et al , Fus Eng Design 98 (2015) 1496 [17] M Mayer , et al , Phys Scr T138 (2009) 014039 [18] A Hakola , et al , J Nucl Mater 415 (2010) S226 [19] M Balden , et al , J Nucl Mater 438 (2013) S220 [20] V Rohde , et al , J Nucl Mater 363365 (2007) 1369 [21] A Kreter , et al , Plasma Phys and Control Fusion 48 (2006) 1401 [22] C Guillemaut , et al , Plasma Phys Control Fusion 57 (2015) 085006 [23] A Hakola , et al , Phys Scr T167 (2016) 014026

[24] A Hakola, et al., - these proceedings

[25] S Brezinsek , et al , Nucl Fus 55 (2015) 063021 [26] K Schmid , et al , J Nucl Mater 463 (2015) 66 [27] A Kirschner , et al , J Nucl Mater 463 (2015) 116