Capabilities of Angle Resolved Time of Flight electron spectroscopy with the 60 degrees wide angle acceptance lens

JournalofElectronSpectroscopyandRelatedPhenomena224(2018)45–50

ContentslistsavailableatScienceDirect

Journal

of

Electron

Spectroscopy

and

Related

Phenomena

j o ur na l ho me p a g e :w w w . e l s e v i e r . c o m / l o c a t e / e l s p e c

Capabilities

of

Angle

Resolved

Time

of

Flight

electron

spectroscopy

with

the

60

◦

wide

angle

acceptance

lens

Danilo

Kühn

a,∗,1

_,

_Florian

_Sorgenfrei

a,1

_,

_Erika

_{Giangrisostomi}

b

_,

_Raphael

_Jay

a

_,

Abdurrahman

Musazay

b

_,

_Ruslan

_Ovsyannikov

b

_,

_Christian

_Stråhlman

b,2

_,

Svante

Svensson

c

_,

_Nils

_Mårtensson

c

_,

_Alexander

_Föhlisch

a,b

a_{InstitutfürPhysikundAstronomie,UniversitätPotsdam,Karl-Liebknecht-Str.24/25,D-14476Potsdam,Germany}

b_{Helmholtz-ZentrumBerlinfürMaterialienundEnergieGmbH,Albert-Einstein-Straße15,12489Berlin,Germany}

c_{DepartmentofPhysicsandAstronomy,Box516,75120Uppsala,Sweden}

a

r

t

i

c

l

e

i

n

f

o

Articlehistory:

Received19December2016

Accepted27June2017

Availableonline11July2017

Keywords: Artof Electronspectroscopy Wideangle Timeofflight Energyresolution Synchrotron

a

b

s

t

r

a

c

t

Thesimultaneousdetectionofenergy,momentumandtemporalinformationinelectronspectroscopyis

thekeyaspecttoenhancethedetectionefficiencyinordertobroadentherangeofscientificapplications.

Employinganovel60◦_{wideangleacceptancelenssystem,basedonanadditionalacceleratingelectron}

opticalelement,leadstoasignificantenhancementintransmissionoverthepreviouslyemployed30◦

electronlenses.Duetotheperformancegain,optimizedcapabilitiesfortimeresolvedelectron

spec-troscopyandotherhightransmissionapplicationswithpulsedionizingradiationhavebeenobtained.

TheenergyresolutionandtransmissionhavebeendeterminedexperimentallyutilizingBESSYIIasa

photonsource.Fourdifferentandcomplementarylensmodeshavebeencharacterized.

©2017TheAuthors.PublishedbyElsevierB.V.ThisisanopenaccessarticleundertheCCBY-NC-ND

license(http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction

Electronspectroscopyhasbeenapowerfulexperimental tech-niqueforinvestigatingmatterinallphasessinceitsdevelopment inthe1960sbyKaiSiegbahn.Advancingsimultaneousdetection ofbothelectronkineticenergiesandtheirangularormomentum distribution hasbeen the pathway toenhance the experimen-talperformance.Inthisrespect,thecombinationofanadvanced electron opticallens with anhemispherical deflectionanalyzer (HDA)[1]establisheda workhorse that linkshighenergy reso-lution,highangularresolutionand flexibilityinoperationmodi spanningfromhighesttransmissionandenergyresolutionto com-binedhighangularandenergyresolvingoperation.Inordertoavoid thegeometricconstraintsoftheHDAentranceslit,theenergy dis-persivehemispherecanbereplacedbyahightransmissionTime ofFlightenergydispersingunit.Thishasledtothedevelopmentof theAngleResolvedTimeofFlightAnalyzer(ARTOF)[2].Operating

∗ Correspondingauthor.

E-mailaddress:danilo.kuehn@helmholtz-berlin.de(D.Kühn).

1 _{Previous address: Helmholtz-Zentrum Berlin für Materialien und Energie}

GmbH,Albert-Einstein-Straße15,12489Berlin,Germany.

2 _{Presentaddress:MalmöUniversity,20506Malmö,Sweden.}

atanacceptanceangleof30◦theapproachwasestablishedinproof ofprincipleexperiments,utilizingthesinglebunchtimestructure ofsynchrotronradiationandlaboratorysources[3–6].Sincethe electronflighttimewithintheARTOFsetsanupperlimitforthe suitablerepetitionrateforthephotonsourceintherangeoffew MHz,areductionoftheGHzrepetitionratesofmultibunch syn-chrotronradiation,i.e.500MHzatBESSYII,iscrucialforawider applicationof this advancedelectron spectrometer.Thus BESSY IIhasdevelopedPulsePickingbyResonantExcitation(PPRE)[7] andmechanicalpulseselectionbyaMHzChopper[8]aswellas electronicdetectorgating[9].

Combining theseadvances, we have now taken the step to enhancetheacceptanceangleoftheARTOFlenssystemthrough anadditionalacceleratingelectronopticalelementwhichallowsto widentheangularacceptanceto60◦keepingbothenergyand angu-larresolution.UtilizingtheBESSYIIFEMTOSPEXinstallations[10] inthisworkwepresenttheworkingprincipleandtheexperimental parameterspaceofthe60◦ARTOFsystem.

2. Electronanalyzerandsetup

The60◦ wideangleacceptanceARTOFisthecoreofthenew endstationforsurfaceandsolidstatedynamicsattheBESSYIIsoft X-Rayfemtoslicingbeamline.

https://doi.org/10.1016/j.elspec.2017.06.008

0368-2048/©2017TheAuthors.PublishedbyElsevierB.V.ThisisanopenaccessarticleundertheCCBY-NC-NDlicense(http://creativecommons.org/licenses/by-nc-nd/4.

Fig.1.a)Schematicdrawingofthe60◦_{wideangleacceptanceARTOF.Electronsemittedfromthesampleunderhighangleswithrespecttothesurfacenormal(redand}

bluelines),aredeflectedstronglytowardstheopticalaxis(greenline)whenenteringthespectrometerentranceduetotheretardationmesh.Depictedelectronsclosetothe

detectormarkequalflighttimes.b)InsightoftheARTOFentrancec)Photoemissiongeometry.

Fig.2.Left:ContourplotofthesimulatedconversionmatricesfortheAng564-lensmodeforEcen=10eV.TheverticallinesmarkequivalentenergiesineVandthehorizontal

linesmarkequivalentangle␽indeg.Right:Schematicenergyvs.time-of-flightdependencyforϑ=0(top)anditsderivative∂E

∂t (bottom).

TheARTOFisaTimeofFlightelectronspectrometerwhich con-sistsofa1.3mlongdrifttubewithanadvancedelectrostaticlens systemofmultiplelenselementsandatwo-dimensionaldelayline detectorbyRoentdekGmbHincludingaMCPelectronmultiplier whichenableseventdetectionofarrivaltimeandpositionofevery analyzedelectron.Thewideanglelensis amodularupgradeof theARTOFlenssystemwithanelectronretardationmeshatthe ARTOFentrancewith80%transmissionprovidingadeflectionofall electrontrajectorieswithinthehighacceptanceangleintothedrift tube.Fig.1showsaschematicdrawingoftheARTOFwith trajec-toriesofelectronsemittedunderdifferentangles␽,␸afteroptical excitationofthesample.Electronsemittedunderhigherangles␽ withrespecttotheopticalaxisofthespectrometerhavelonger flightpathsand thus longerflighttimes comparedtoelectrons emittedonaxis (ϑ=0).Theelectronarrivaltimetandhitting posi-tionx,yatthedetectordependonthekineticelectronenergyEand thetake-offangle␽inanon-linearwayandcanbetransformedto energyEandemissionangles␽,␸bytwotransformationmatrices incylindricalcoordinatesE (t,r) andϑ (t,r) whichareobtainedby simulations[2](seeFig.2).The␸-matrixistrivialduetocylindrical symmetry.

TheARTOFisoperatedintheconstantretardationratio(CRR) modewhichmeansthatallvoltagesofthelenssystemareraised proportionallytothecenterenergyEcenoftheanalyzedenergy

win-dow.Conclusivelytheelectrontrajectoriesareindependentfrom Ecenandonlyonesetofconversionmatricesisnecessarytocover

thetransformationforthefullenergyrange.Theenergywindow isaconstantfractionofEcenandtheenergyresolutionisscaling

withE3/2_{accordingtoEq.(1)forhighenergies.Inthisoperation}

modeonecandefinealensmodebyspecifyingalllenselement voltagesforasingleenergye.g.Ecen=10eV.Thesizeoftheenergy

windowinwhichallelectronscanbeanalyzedsimultaneouslyand theangleacceptancecanbederivedfromsimulations.Thestandard nomenclatureofalensmodeis‘Ang_␪ew’where_␪isthefullangle acceptanceand‘ew’isthesizeoftheenergywindowinpercentage e.g.‘Ang564’.

AllmeasurementsinthisworkhavebeenperformedattheUE 56/1PGMsoftX-RaybeamlineattheBESSYIIinstandard multi-bunchmode.WhilecommonlyoperatingwithPPRE[7]wehere usedtheunexcitedCamshaftbunchinthemiddleoftheionclearing gaptoavoidanypossibledecreaseinbeamlineenergyresolution which might arise frombunch excitation. We have chosenthe

D.Kühnetal./JournalofElectronSpectroscopyandRelatedPhenomena224(2018)45–50 47

1200l/mmmonochromatorgratingandthe50␮mexitslitfor opti-malbeamlineenergyresolutionandsmallelectronspotsizeatthe samplerespectively.Thephotonbeamisverticallypolarizedand thephotonenergyrangesbetween260and970eVwithaphoton pulselength<70psFWHM.TheARTOFopticalaxisisparalleltothe samplesurfacenormalandliesinthebeamplane.Thephotonbeam ishittingthesampleunder50◦withrespecttothesurfacenormal. Thebasepressureoftheexperimentalchamberis2·10−10_mbarand

allmeasurementswereperformedatroomtemperature.

3. Results

Oneofthemostimportantcharacteristicsofanelectron spec-trometer is its energy resolution Eand the related resolving powerE/E.We haveanalyzed thisproperty inthesoft X-Ray regimefor kineticelectron energies between90eVand 810eV. IntheCRRmodetheenergyresolutionisgivenbythefollowing equation:

E=



˛·E3/2_·t



2₊



_ˇ·E·d



2 ₍₁₎

Thefirstpartofthisequationcontainsthecontributionfromthe temporalresolutiont,thesecondpartdescribestheeffectofthe spatialdetectorresolutiond.Theparameters␣and␤describethe temporalenergydispersionandspatialenergydispersion, respec-tively,anddependonthespecificlensmode.Theyaregenerallya functionofdetectorhitpositionandflighttime(examplesaregiven below).Furthercontributionsmaycomefromtheelectronsource diameterandarotatedsamplesurfacethatisnotperpendicularto theopticalaxisofthespectrometer.Howeverforhighkinetic ener-giesEkin>50eVandsmallspotsizes≈100␮mastypicallyexistat

asynchrotronradiationbeamline,thetemporalresolutionisthe dominatingcontributiontotheintrinsicenergyresolution[2]and Eq.(1)simplifiestoEq.(2):

E=˛·E3/2·t (2)

Thetotaltemporalresolutiontisgivenbytheconvolution oftheexcitingphotonpulselengthtph=70ps FWHMandthe

temporalresolutionofthedetectortdet

t=



t2

ph+t2det (3)

andcanbedeterminedbymeasuringonlyphotonsatthe detec-tor,whichisachievedbyapplyingalargenegativebias.Herewe determinet=216±2psFWHMfortheglobaltemporalresolution byapplyingaGaussianfittothephotonsignalintegratedoverthe wholedetectorareaandt=166±13psFWHMforthedetector positionadaptedtemporalresolution,asprovidedbythephoton peakfittoolavailablewithintheARTOFLOADERanalysispackage.3

Fortheanalysisoftheenergyresolutionwedistinguishbetween localenergyresolutionElocwheretheelectronspectraare

deter-minedonlyfromasmalldetectorareaaroundtheoriginandthe globalenergyresolutionEglobwherethewholeangleacceptance

is usedto derivethe spectra.Besides thefundamental limit of theenergyresolutiongivenbythetemporalresolution,theglobal energyresolutioncanbefurtheraffectedbyresidualmisalignments andsignalpropagationtimedifferencesofthedetectorwhereas thelocalenergyresolutionisonlylittleaffectedbythose. There-foredeviationsbetweenElocandEglobforagivensettingcan

givehintsof‘howfarthespectrometerisawayfromoptimal per-formance’.

InthisworkweusetheS2pphotoemissionlinesof semicon-ducting2H-MoS2forthecharacterizationoftheenergyresolution.

3_{ARTOFLOADERv.2.5.−156,seeref.[4]forfurtherinformations.}

The sample is preparedby cleaving, which gives reliably clean surfacesleadingtoawelldefinedelectronsourcegeometry.The S2p3/2 andS2p1/2 peakswithEbind=162.7eVandEbind=163.9eV

[11]respectivelyhaveanarrowsymmetriclineshapewitha nat-urallinewidthof0.054eV4_{andgiveagoodsignaltobackground}

ratio.WeobserveanadditionalconstantGaussianlinebroadening

ofEsample=0.19eVFWHMwhichisdominating thelinewidth

atkineticenergiesbelow100eV.Thiswasalsoobservedby Mat-tilaetal.andisattributedtoterraces,stepsandpointdefectsatthe samplesurface[11].InFig.3differentrepresentationsofthespectra fortwodifferentkineticenergiesareshownfortheAng564-lens mode. Theenergy-angle-cuts(topand middlepanels)showthe intensityoftheS2plinesoverthewholeangleacceptance reveal-ingonlyminordistortions.Theenergyspectrawhichareintegrated overasmalldetectorareaandthewholedetectorarearespectively, haveverysimilarlineshapesandshowonlya minoradditional broadeninginthelattercase.Thecountratesofthedifferentlens modesgivequalitativeproofoftheincreasedtransmissionforthe 60◦ instrument.TheintensityratiooftheS2plinesofAng60and Ang50lensmodesisabout1.4,inaccordancewiththeratioofthe solidangles.However,thequantitativedifferencesbetweenAng56, Ang58andAng60aredifficulttodeterminepreciselyduetothe inhomogeneousangulardistributionoftheemittedelectrons.

ForquantifyingtheenergyresolutionwehavemeasuredS2p photoelectron spectrafor five differentkinetic energies in four differentlensmodesandappliedafitconsistingoftwoVoigt pro-filepeakswithafixedspinorbitsplittingof1.19eV,aLorentzian widthof0.054eVconvolvedwithaGaussianbroadeninganda lin-earbackground(seee.g.Fig.4a).Thecenterenergyissettothe kineticenergyoftheS2plinesforeachmeasurement.The spec-trometerenergyresolutioncannowbecalculatedbydeconvolving theGaussianlinewidthfromthesamplebroadeningandthe beam-lineresolutionEph.Inanextstepwehavedeterminedthescaling

parameter␣·tinEq.(2)byapplyingafit(Eq.(4))totheglobal Gaussianlinewidth(seeFig.4b)andthelocalGaussianlinewidth, respectively(seeTable1).

E=



˛·t·E_kin3/2



2 −



Eph



2 −



Esample



2 (4) Using␣loc·tand␣glob·tonecancalculatethelocalandglobal

energyresolution(seeFig.5a)andtheresolvingpower(seeFig.5b). Theparameters␣and␤fromEq.(1)canalsobedetermined fromthesimulatedenergyconversionmatricesE (t,r) bynumerical differentiationwithrespecttothetimeofflight(t)orpositionon thedetector(radiusr).

˛matr(t,r) =−

∂E (t,

r)

∂t

·E −3/2 cen (5) ˇmatr(t,r) =

∂E (t,

r)

∂r

·E −1 cen

Theparametersdependonrandtitisthereforenotpossibleto giveasingleenergyresolutionforagivenlensmode.Foralllens modesonecanobservearathernarrowdistributionofalpha val-uesaround˛matr(r=0,E=Ecen)(seeTable1).Therelativewidth

ofthedistribution␣matr/␣matrisapproximately±15%.Thevalue

forˇmatrisrather homogeneouslydistributedbetweenˇmatr=0

forr=0andˇmatr=1.5m−1forhigherradii.Thecontributionsof

spatialandtemporalresolutionontheenergyresolutionareabout equalbetweenEcen=10eVand30eVandthetemporalresolution

becomesincreasinglydominantforEcen≥50eV.

Fig.3.Angle/energy2Dmapsfortwodifferentkineticenergies(left:Eph=260eV,right:Eph=790eV)recordedintheAng564-lensmoderepresentedinbipolarcoordinates.

Onthetoppanels,intensityisdisplayedvs.energyandtheverticalanglealphaandintegratedoverthehorizontalanglebeta.Inthemiddlepanels,intensityisdisplayedvs.

energyandbetaandintegratedoveralpha.Atthebottompanel,angleintegratedenergyspectraareplottedfortwodifferentintegrationareas:Theredspectraaretaken

fromanareaaroundthedetectororigin(markedbyredboxesin2Dplots)andthebluespectraaretakenfromfullangleacceptance.Thesmallareaspectrashowrealelectron

yield(counts)whilethefullangleacceptancespectraarenormalizedtosolidangle(scalingfactoramountsto30.8).Notethatonlyafractionoftherecordedenergywindow

isdisplayedforbothkineticenergies.

Table1

ListofexperimentalandsimulatedlensmodeparametersdeterminingtheenergyresolutionaccordingtoEq.(2)andtheangularresolution.

Energywindow[%] ˛



loc·t(local)

10−5eV−0.5

˛glob·t(global)



10−5eV−0.5

˛loc ˛glob ˛matr



105_eV−0.5_s−1

_{Angularresolution[deg]}

ang50 7 5.34±0.04 6.95±0.02 0.77 3.43 0.13±0.02

ang56 4 3.06±0.03 3.67±0.01 0.83 1.76 0.18±0.04

ang58 4 3.41±0.02 4.18±0.01 0.82 1.93 0.20±0.04

ang60 2 4.17±0.03 4.88±0.01 0.85 2.54 0.15±0.02

Theangularresolutioncanbeestimatedfromtheangular

dis-persion of the lens system in the image planeat the detector

bycalculatingthederivativeofthethetaconversionmatrix.For

sourcesizesdsource≤100␮m,witha spatialdetector resolution

ofddet=100␮mandneglectingtheminorinfluenceoftemporal

resolution,theangularresolutionisgivenby:

ϑ (r) ≈

∂

ϑ

∂

r (r) ·ddet (6)

Table1showstheangularresolutionaveragedoverthe accep-tance angle for all lens modes. The uncertainties are given as standarddeviations.

4. Discussion

ThelocalenergyresolutiondeterminedfromtheS2ppeakfits can bedirectly compared withthe expected energy resolution fromtheenergyconversionmatriceswiththetemporalresolution

D.Kühnetal./JournalofElectronSpectroscopyandRelatedPhenomena224(2018)45–50 49

Fig.4.Left:Angleintegrated(±28◦_{)energyspectrumofS2pphotoelectrons(bluedots)withEph=260eVrecordedintheAng564-lensmode.Afitcontaining2Voigtprofiles}

andalinearbackground(redline)isappliedtothedata.Right:GaussianlinewidthcontributioninS2pspectraexperimentallydeterminedforfivedifferentkineticenergies

(dots)fordifferentlensmodes.Thefitstothedataaccountforsamplebroadening(lightbluecurve)andbeamlineresolution(greycurve).

Fig.5.Left:Globalenergyresolution(solidlines)calculatedwith␣globandlocalenergyresolution(dashedlines)calculatedwith␣locRight:Resolvingpowerofthe

spectrometercalculatedwith␣globfromtheenergyresolutionfit.Samecolourcodeasinleft.

obtainedfromthephotonpeakfittool,forexampleforAng56 4-lensmode:

˛matr·tloc=1.76·105eV−0.5·s−1·166ps=

2.92·10−5eV−0.5≈3.06·10−5eV−0.5=˛loc·t

Wefindthatthekineticenergydependenceoftheresolutionfor allfourlensmodesareinverygoodagreementwiththeexpected energyresolutioninthelocalcase,seeFig.4.Thisalreadyshows clearly that thetiming system of the spectrometer is working locallyasexpected.However,itshouldbeemphasizedthatthefits aremadeoverawideenergyrange.Althoughthefitisinverygood agreementwiththeoreticalexpectationsandpointtoaresolving powerasshowninFig.5,theintrinsicsampleenergybroadening istoohightoprovetheresolvingpoweratlowenergies.Further experimentsareneededtodeterminetheenergyresolutioninthe UPSregime.

Onecanfurthercomparethelocalandtheglobalenergy resolu-tion.Wefindthattheglobalenergyresolutionisdecreasedwith respecttothelocalenergy resolution byabout20%for alllens modes.Possiblereasonsareresidualmisalignments,different sig-nalpropagationtimesofdifferentdetectorareasand increasing ␣(fitparameter)athigheremissionangle.Weareconfidentthat carefulalignmentandadetectorpositionadaptedtimecorrection, whichwillbeavailableinfuture,canfurtherimprovetheglobal energyresolution.

Apparentlythedegradationoftheglobalenergyresolutionis sensitivetothesizeoftheenergywindowwhichcouldbeexplained bythefactthattheelectronflight timesneedtobemore com-pressedforlargerenergywindowsandarethereforemorestrongly influencedbysignalpropagationdelaysbetweendifferentdetector positions.

Comparingthefourlens modesonecanclearlyseethe gen-eraltrendthatlargersizeoftheenergywindowandlargerangle acceptancearetothedisadvantageoftheenergyresolution.

There-foretheoptimumchoiceofthelensmodedependsonthespecific experimentaldemands.

5. Conclusion

Inthisworkwehaveoperatedthe60◦wideangleacceptance ARTOFatthefemtoslicingbeamlineatBESSYII.Fourdifferentlens modeshavebeeninvestigatedinthesoftX-Rayregimeforkinetic energiesoftheelectronsrangingfrom90eVto810eV.Alllens modesperformasexpectedinthefullenergyrange,particularlythe resolvingpowerandangleacceptanceareinverygoodagreement withsimulations.Ahighresolvingpowerof1000isdemonstrated evenforhighenergiesandsuggestsaresolvingpowerof3000for 90eVkineticenergy.Electronspectroscopictechniques,drivenby pulsedlightsources,thatrelyonhightransmission,highenergy resolution,wideangleacceptanceorhighangularresolutione.g. ARPES,XPD,pump-probespectroscopymightstronglybenefitfrom thenewwideanglelens.

Acknowledgment

TechnicalsupportbyMikeSperlingisgratefullyacknowledged. PålPalmgrenandPatrikKarlssonfromScientaOmicronGmbHare acknowledgedforfruitfuldiscussionsabouttheARTOF.

D.K.,F.S,R.JandA.F.acknowledgefundingfromthe ERC-ADG-2014 − Advanced Investigator Grant No. 669531 EDAX under the Horizon 2020EU Framework Programme for Research and Innovation.NMandSSacknowledgefundingfromtheEuropean ResearchCouncilundertheEuropeanUnion’sSeventhFramework Programme(FP7/2007-2013)/ERC grant agreementn◦ [321319], and support from the Carl Tryggers foundation for scientific

reseach/CTH),andtheSwedishResearchCouncil(VR) TheauthorsthankHZBfortheallocationofbeamtime.

References

[1]N.Mårtensson,P.Baltzer,P.Brühwiler,J.-O.Forsell,A.Nilsson,A.Stenborg,B. Wannberg,J.ElectronSpectrosc.Relat.Phenom.70(no.2)(1994)117–128.

[2]B.Wannberg,Nucl.Instrum.MethodsPhys.Res.,Sect.A601(no.1–2)(2009) 182–194(SpecialissueinhonourofProf.KaiSiegbahn.).

[3]G. ¨Ohrwall,P.Karlsson,M.Wirde,M.Lundqvist,P.Andersson,D.Ceolin,B. Wannberg,T.Kachel,H.D ¨urr,W.Eberhardt,S.Svensson,J.ElectronSpectrosc. Relat.Phenom.183(no.1–3)(2011)125–131(ElectronSpectroscopyKai SiegbahnMemorialVolume.).

[4]R.Ovsyannikov,P.Karlsson,M.Lundqvist,C.Lupulescu,W.Eberhardt,A. F ¨ohlisch,S.Svensson,N.Mårtensson,JournalofElectronSpectroscopyand RelatedPhenomena191(2013)92–103.

[5]A.Vollmer,R.Ovsyannikov,M.Gorgoi,S.Krause,M.Oehzelt,A.Lindblad,N. Mårtensson,S.Svensson,P.Karlsson,M.Lundvuist,T.Schmeiler,J.Pflaum,N. Koch,J.ElectronSpectrosc.Relat.Phenom.185(no.3–4)(2012)55–60.

[6]U.B.Cappel,S.Plogmaker,J.A.Terschlusen,T.Leitner,E.M.J.Johansson,T. Edvinsson,A.Sandell,O.Karis,H.Siegbahn,S.Svensson,N.Mårtensson,H. Rensmo,J.Soderstrom,Phys.Chem.Chem.Phys.18(2016)21921–21929.

[7]K.Holldack,R.Ovsyannikov,P.Kuske,R.M ¨uller,A.Sch ¨alicke,M.Scheer,M. Gorgoi,D.K ¨uhn,T.Leitner,S.Svensson,N.Mårtensson,A.F ¨ohlisch,Nat. Commun.5(2014)05.

[8]D.F.F ¨orster,B.Lindenau,M.Leyendecker,F.Janssen,C.Winkler,F.O. Schumann,J.Kirschner,K.Holldack,A.F ¨ohlisch,Opt.Lett.40(2015) 2265–2268.

[9]C.Strahlman,Time-of-FlightIonandElectronSpectroscopy:Applicationsand ChallengesatStorageRingLightSourcesDoctoralThesis,MAXIVLaboratory, LundUniversity,Sweden,2016.

[10]K.Holldack,J.Bahrdt,A.Balzer,U.Bovensiepen,M.Brzhezinskaya,A.Erko,A. Eschenlohr,R.Follath,A.Firsov,W.Frentrup,L.LeGuyader,T.Kachel,P. Kuske,R.Mitzner,R.M ¨uller,N.Pontius,T.Quast,I.Radu,J.-S.Schmidt,C. Sch ¨ußler-Langeheine,M.Sperling,C.Stamm,C.Trabant,A.F ¨ohlisch,J. SynchrotronRadiat.21(2014)1090–1104.

[11]S.Mattila,J.Leiro,M.Heinonen,T.Laiho,Surf.Sci.600(no.24)(2006) 5168–5175.