Photoelectron Spectroscopy - FHI

1 PhotoelectronSpectroscopyModern Methods inHeterogeneous CatalysisResearchAxel Knop Spectroscopy : General PrincipleSurface sensitivityInstrumentationBackground subtractionPES peaks, loss featuresPeak fittingChemical stateBinding energy calibrationHigh pressure XPSXPS at the liquid-solid interfaceexamplesProblem: What is the (chemical) composition of a surfaceMe0Me2+ Photoelectron Spectroscopy : General PrincipleExcitation of PhotoelectronGeneral Principle- Detection of photoelectrons fromthe valence band region and corelevels- Detection of Auger electrons andX-Ray FluorescenceBinding Energy out of.

2 H = Evsp+ Ekinh = Ekin+ Evsa+ sp- sah = Ekin+ EF+ spEvsaEvspX-ray twin anodeParallel plate mirror analyserHalbkugelanalysatorElektronen werden zu-n chst auf einen bestimmtenEnergiebetragabgebremstNur Elektronen mit be-stimmter"Passenergie"k nnen den DetektorerreichenJe kleiner die "Passener-gie" desto besser dieEnergieaufl sungDetector: ChanneltronXP spectra: two different anode materialsWhere do the electrons come from?Distance electron can travelin solids depends on: Material Electron kinetic energy Measure attenuation ofelectrons by covering surfacewith known thickness ofelementSampling DepthSampling depth:For normal takeoff angle, cos =1: When d= : -ln(I/I0)= , from within When d=2 : -ln(I/I0)= , from within 2 When d=3 : -ln(I/I0)= , from within 3 Disregarding elastic scattering.

3 Mean Free Path of Electrons in Solids IMFP is average distance between inelastic collisions Minimum of 5-10 for KE ~ 50-100 eV Maximum surface sensitivityOrigin of backgroundBackground of scattered electrons due to limited IMFP Higher for low KEBackground CorrectionBackground substraction Choose suitable energy range for substraction Baseline on high KE sideBackground CorrectionSeveral ways to substract background:- Stepwise- Linear- Method of Tougaard Most common:Method of Shirley et al.:Always use same backgroundsubstraction method for all peaks!

4 Shirley, Phys. Rev. B, 5, S. 4709, features: PE PeaksOrbital peaks used to identifydifferent elements in sampleObservation: s orbitals are not spin-orbit splitted singlet in XPS p, d, Orbitals are spin-orbit splitted doublets in XPSS pectral features: PE PeaksSpectral features: Auger PeaksAuger peaks:Result from excess energy of atom during relaxation (after corehole creation) always accompany XPS broader and more complex structure than PES peaksKE energy independent of incident h Spectral features: Loss PeaksAfter Emission of PE the remaining electronsrearrange (Secondary peaks): Relaxation by excitation of valence electron tohigher state (shake up) loss of kinetic energy of PE leads to new peaksshifted in BE with respect to main peak Relaxation by ionisation of valence electron(shake off) broad shoulder to main peak PlasmonlossesSpectral features: Loss PeaksLoss peakscan be usedto identify chemicalcompositionProblem.

5 References needed(theoretical orexperimental)Relationship between the degree of core levelasymmetry and the density of states at the Fermilevel (BE=0)Identification of FE:Easy for high density of states near EF, butmay be ambigeous for samples with lowDOS near EF, band gap or charged samplesAnalyzing the data: Calibration of Binding EnergiesLine profile modification by chargingMo oxide on silicareal catalyst is powdersample after impregnationand the data: Calibration of Binding EnergiesIdentification of FE:In case of low DOS near EF, band gap or charged samples (spectrum willmove in this case!)

6 Look for reference peaks with known BE (be sure about that!!!) Works for any PE or Auger peak, peaks originating from substrateChemical ShiftBE of core electrons depends on the electron density of theatom (screening of the core electrons), affected by theelectronegativity of neighboring atomsFor unknown BE: Calculation of Pauling s Charge (easy, but not always true) More elaborate calculations (theorists, also not always true)References: Use of data bases (NIST,Handbook of PhotoelectronSpectroscopy etc.*) Clean materials (puremetals, well definedsubstrates, untreated samples etc.)

7 Literature about similartopicsAnalyzing the data: Chemical States* Handbook of Photoelectron Spectroscopy , Perkin-Elmer,1992G. Ertl, J. K ppers; Low Energy Electrons and Surface Chemistry ; VCHV erlag, 1985S. H fner; Photoelectronspectroscopy ; Springer, Briggs, Seah; Practical Surface Analysis , Vol. 1; Wiley + Sons, the data: Peak FittingSeveral influences on peakshape: Broadening mainly due toexcitating light (natural),structural and thermal effects Asymmetry due to final stateeffects Species with similar BEThese effects have to be considered when fitting by reasonablechemical/physical model of the sample!

8 !!Analyzing the data: Peak FittingAsymmetric Peakshapes modeled byDoniach-Sunjicfunctions(convolutedwith Gaussian profiles)Most fitting programs provide useful tools:Levenberg-Marquardtalgorithm to minimizethe 2M: measured spectrumN: energy valuesS: synthesized spectrumP: parameter values : peak parameterM: mixing ratio : asymmetry parameterh: peak height 0: Lorentzian profileConvolution ofLorentz (or )Gaussian profilessuited best!! Voigt functionStrategy: Use of references forasymmetry andbroadening Reasoning of mostlikely chemicalcompositionAnalyzing the data: Peak FittingPeak broadening andasymmetryshouldbethe same for samecomponent in differentspectrapure AugraphiticCNTdisorderedBut: both may differ for different components discrepancy for different peaks should always be based on physical reasonsQuantitative AnalysisMe0Me2+Quantitative Analysis: First StepsIn general:IX= B tot T nXB:all instrumental contributions.

9 Ionisation cross section for given photon energy tot: total escape depthT: transmission through surfacenX: atomic density of analyzed species in sample = tot f(X, )with f(X, ) = 1+( (X)/4) (1-3cos2 )* tot: total ionisation cross sectionf: form function accounting for asymmetry of peak : asymmetry parameter : angle between photon beam and emitted electron(different for standard x-ray source and synchrotron)*Yeh and Lindau, Atomic data nucl. data tables32(1985)1 tot= E(X) 1/a[ln(E(X)-b)*E: KE of electrona, b: parameters dependent on dielectric function andconcentration of host*Penn, And Rel.]

10 Phen. 9(1976)29 Cross section changes with energyof incident beam!!!Quantitative Analysis: Cross SectionCalculated changes of Cross section beam energy(Yeh, Lindau; Atomic data and nuclear data tables 32 (1985) 1)Quantitative AnalysisIn most cases the sample surface is quite heterogeneous Need ofmodelsto extract an accurate approximation ofcomposition out of spectral intensitiesRuO2 RuxOyRu 3d5/2-map of the chemical states on Ru(0001)after pre-treatment with Operformed with Scanning PhotoelectronMicroscopy (ELETTRA):Quantitative Analysis: Useful Exampleswith the formulasabove followsandaA: radius of A Mol fractionA.