DRAFTBRUKERXRFSPECTROSCOPYUSERGUIDE: SPECTRALINTERPRETATIONANDSOURCESOFINTERFERENCE

TABLEOFCONTENTS
TABLEOFCONTENTS ABSTRACT XRFTHEORY INSTRUMENTATION 1 3 4 6

EDXRFEQUIPMENT TRACER SIPINDIODEDETECTORPARAMETERS ARTAX SI(LI)SDDDETECTORPARAMETERS

SPECTRALINTERPRETATION INTERACTIONSINTHEDETECTOR

6 8 8 9 9
9 11

SUMPEAKS ESCAPEPEAKS ESCAPEPEAKS(CONTINUED) HETEROGENEITY HETEROGENEITY(CONTINUED)

INTERFERENCEWITHINSTRUMENTATION

11 12 ERROR!BOOKMARKNOTDEFINED. 15 16
17

EQUIPMENTANDINSTRUMENTCONTRIBUTION EQUIPMENTANDINSTRUMENTCONTRIBUTION(CONTINUED) THINFILMANALYSIS(BACKGROUNDCONTRIBUTION)

1

17 18 19

THINFILMANALYSIS(BACKGROUNDCONTRIBUTION)
PHENOMENAINTHESAMPLE

ERROR!BOOKMARKNOTDEFINED.
20

RAYLEIGH(ELASTIC)SCATTERING RAYLEIGH(ELASTIC)SCATTERING(CONTINUED) COMPTON(INELASTIC)SCATTERING MATRIXEFFECTS BRAGGSCATTERING BRAGGSCATTERING(CONTINUED) BRAGGSCATTERING(NISTC1122EXAMPLES) BRAGGSCATTERING(GEMSTONEEXAMPLES) BRAGGSCATTERING(GEMSTONEEXAMPLES)

SIXFIELDAPPLICATIONS

20 21 22 23 24 ERROR!BOOKMARKNOTDEFINED. 26 ERROR!BOOKMARKNOTDEFINED. ERROR!BOOKMARKNOTDEFINED.

28

IDENTIFYINGGENUINEARTIFACTS(CHELSEABULLFINCHEXAMPLE) IDENTIFYINGTRUEORIGINS(STONEWAREEXAMPLE) MEASURINGCHLORINEWITHTHETRACERANDOTHERFIELDAPPLICATIONS

APPENDICES

28 29 30
41

APPENDIXA APPENDIXB APPENDIXC APPENDIXD APPENDIXE

ByDrBruceKaiserandAlexWright November11,2008

43 48 49 50 51

 ABSTRACT
WhileperformingXRFspectroscopy,threemainfactorscontributetotheanalyticalspectrum:interactions inthedetector,interferencewiththeinstrumentation,andphenomenainthesample.ThisuserguidetoEDXRF providesabasicoutliningofthephysicsinvolvedinXRFspectroscopy,anoverviewofthemaincomponentsofthe BrukerTRACeRandARTAXunits,aswellasadelineation,explanation,andresolutionoftheseveralphenomena includedinperformingXRFspectroscopy.Eachsectionincludesatextualexplanationofwhyitoccursandhowit affectsaspectrum,aswellasexamplespectrathatclearlyidentifythephenomena.Severalfieldapplicationsand examplesareprovided,aswellasanappendixwithadditionalinformation

 XRFTHEORY
GeneralConceptBehindXRayFluorescenceSpectroscopy Every element has a characteristic electron structure. When inner shell electrons are ejected from an atom,electronsfromshellswithlessbindingenergyfilltheholesandmayreleasexrayradiationequivalenttothe difference in energy between the level the electrons came from to that which they went. The x ray radiation releasedduringthesetransitionsischaracteristictotheelementandhasaspecificenergy(2eV)dependingon thetransitionmadewithintheatom.Bybombardingasamplewithradiationthatexceedsthebindingenergyof theelectronsintheatomsofwhichthematerialiscomposedofanddetectingtheenergyandnumberofresultant characteristic x rays emitted from each element, it is possible to determine the composition and proportional concentrationsofthoseelements. Two common methods of Xray spectroscopy exist: Wavelength Dispersive XRF (WDXRF) and Energy Dispersive XRF(EDXRF).Themaindifferencebetweenthetwomethodsishowtheemittedxraysaremeasured;WDXRF usesananalyzingcrystaltodiffractthedifferentxraywavelengthsanddetectorsareplacedatthevariousangles tomeasurethenumberxraysdiffractedateachangle.Asingledetectormaybeusedtomeasureallthevarious energies if one moves the detector to cover all the angles, because each energy comes out of the crystal at a differentangle. Energydispersivexrayfluorescence(EDXRF)usesadetectorthatcollectsxraysofallenergiesandsortsouteach xrayenergybytheamountofelectronseachxrayknocksfreeinthedetectorlattice,typicallysilicon.Thenumber ofelectronsknockedfreedependsontheincomingxrayenergyandtheparticularinteractionthatthatxrayhas withthemateriallattice.Toaccuratelydeterminethexrayenergyalltheelectronsfromeacheventthatoccursin thedetectormustallbecollectedandconvertedultimatelytoadigitalsignal.Thusthedetectormeasuresonex rayatatime. BremsstrahlungRadiation In most xrf systems the beam of x rays incident on the sample are produced with a vacuum tube and createdbybombardingatarget(suchasRh,W,Cu,orMo)withhighlyacceleratedelectrons.ShowninFigure1,as theelectronspenetratethetargetatoms,theymayhavetheirdirectionchangedastheypassnearthenucleusof the target atoms causing a sudden deceleration and loss of kinetic energy. In this loss of kinetic energy the electron may emit an x ray with energy relatedtotheamountofenergylost.Asa Atomofthe Outboundelectron targetmaterial (deceleratedanddiverted) resultabroadspectrumofxrayenergies, known as a Bremsstrahlung continuum, is

Fastinbound electron

Figure1:DiagramoftheBremsstrahlungeffect

Figure2:Diagramofpossibleexcitationroutes

emittedfromthexraytubetarget.Thiscontinuumcanbeadjustedbytubehighvoltagesettings,beamfiltering andsecondarytargetstoallowonetofocusondetectionofspecificelementswithinthesample.Thiscapitalizes onthedifferentabsorptionedgesofeachelement.Theacceleratedelectronsalsocausethetargettofluoresce. Thesetargetcharacteristicxraysarealsoincidentonthesample,andmustbeconsideredduringdataanalysis ThePhotoelectricEffectandInnerShellIonization When an x ray interacts with an atom several different reactions can occur depending on the x rays energy.Lowenergyphotons,suchasalightwaves,canexciteandejectoutershellelectronsfromtheatomand exciteinnershellelectronstohigherenergylevels.Xrayscanejectinnershellelectronsfromtheatom,creatinga vacancy in the inner shell and put the atom in an unstable condition, providing the that incoming x ray has an energywhichexceedsthebindingenergyoftheelectronitinteractswith.Whenavacancyiscreated,theatom quickly relaxes (less than 107 s) by transitioning an electron from a higher shell to the vacancy. During this transitionanxraywiththeenergyequaltotheenergydifferenceofthetransitionmaybereleased.Thisxrayis knownasacharacteristicxrayandisspecifictothetransitionandelementinwhichitoccurs.Figure2illustrates the different interactions that can occur when x rays interact with the bound electrons of an atom. Once the characteristicxraysarecreated,theymayescapetheatomandthematrixmaterialatrandomanglesandavery smallfractionenterthedetector.Ifgeometry,densityandotherfactorsareknown,thenumberofxraysentering
Table1:Possibleelectronlocationsinanatom

the detector can be related to the type and numberofatomspresentinthesample. CharacteristicLines Theenergyofthexrayreleasedbytherelaxation of an ionized atom is dependent on the element from which it came, the location of the vacancy, and the electron that fills the vacancy. Electrons existinquantizedenergylevels;theyarenotfree to roam anywhere around the nucleus of the atom. A given electron is located in a given electronshellusingthequantumnumbersn,l,m, s,wherenindicatestheshell,lindicatesthesub shell,mlindicatestheenergyshiftwithinthesub
Table2:ExampleofpossibletransitionnotationsforaBariumatom.

shell,andmsindicatesthespinoftheelectron.Table1listssomeofthepossibilitiesforelectronlocationsinan atom. There are a couple ways of describing an x ray of a certaintransition. It canbe written inLine notation, Siegbahnnotation,ordescribedbythecharacteristicenergyorwavelengthconnectedthatxray.Thesenotations areshownbelowinTable2withanexampleofBariumelectrontransitions. IntheSiegbahnnotation,theGreek subscriptdenotestheprobabilityofthetransition(intensity),proceedingfromthemosttoleast(,,,etc.).

INSTRUMENTATION
TheTRACeRandARTAXarebothEDXRFunitswithSiliconbaseddetectors.TheTRACeRisahandheld unitcommerciallyofferedbyBrukerAXSandprovidesforquickandeasyqualitativeanalysisandchemistriesfor elements as low as Mg. The Tracer handheld XRF analyzer provides spectral analysis through PXRF analytical software. The instruments high sensitivity allows the user to identify the elements in a sample matrix, with concentrations as low as ppm. The PXRF software program provides qualitative and quantitative analysis, in addition to the voltage andcurrentcontrol of the Xraytube, which makes possiblea wider rangeof elemental analysis. The ARTAX is the first commercially available, portable microXRF spectrometer designed to meet the requirements for a spectroscopic analysis of unique and valuable objects on site, i.e. in archeometry and art history.ThesystemperformsasimultaneousmultielementanalysisintheelementrangefromNa(11)toU(92) and reaches a spatial resolution of down to 30 m. Both instrumentsallow one to utilize filters and secondary targettoadjusttheincidentxraybeaminbothenergydistributionandintensity.

EDXRFEQUIPMENT
TheARTAXandTRACeRsystemshaveseveralseparatecomponentsthatallservetheirownfunctionin the process of recording Xray fluorescence. The main components in terms of functionality are the Xray tube system, collimators, filters, detector and signal processing hardware and software. The ARTAX and the TRACeR Turbo are very similar in functionality. Both units employ energy dispersive technology and a Silicon based detector.Beingahandheldunit,theTRACeRisbatteryoperated,moreconvenient,buthasabeamspotsizeof3 by4mm(muchlargerthantheArtax).TheARTAXisportable;however,itisnotgenerallyusedinthefieldlikethe battery operated handheld TRACeR unit. Both units can use a variety of changeable filters, tube voltage and current settings making them uniquely capable of being configured to maximize their sensitivity to specific elementsofinterest. Figure4:Diagramofanendwindowxraytube XrayTube:Electronsaregeneratedandacceleratedtohigh speedsandthenbombardedatargetusuallycomposedof a pure metal (e.g. W, Mo, Cr, or Rh). Upon reaching the target the electrons either interact and ionize the target, creating characteristic xrays, or are decelerated upon nearing the nuclei, creating a Bremsstrahlung continuum. Figure 4 illustrates the production of the XRay beam.
Figure 5: Overlapping emission spectra of W and Mo target Xray tubes run at differentvoltages

Figure6:EmissionspectraofanRhtargetXraytube 6 runatthreedifferentvoltages

Filter:Afiltercanbeplacedbetweenthetubeandthesampletoremoveundesirablebackgroundradiationbelow a certain voltage. The level of radiation filtered out is dependent on the filter element composition and its thickness.Table3suggestssomefiltertypesforcertainapplicatonsandAppendixAprovidesalistofothersfor theTRACeRandARTAXunits,aswellasotherfilterinformation.Notebothunitsallowtheusertofabricateany filterorsecondarytargettheythinkisbestfortheirapplication.

Filter Nofilter Cellulose Thinaluminum Thickaluminum Thinanodeelement Thickanodeelement Copper

Thickness N/A Singlesheet 2575m 75200m 2575m 100150m 200500m

kVrange 450 510 812 1020 2540 4050 50

Elements All,NaCa SiTi SV CaCu CaMo CuMo >Fe

Table3:AvailablefiltersthatcanbeusedwiththeARTAXandTRACeRunits

Collimator:Collimatorsareusuallycircularoraslitandrestrictthesizeorshapeofthesourcebeamforexciting smallareas.Collimatorsizesrangefrom12micronstoseveralmm.Figure7illustratesthegeneralfunctionofa collimator. Detector: The detector is used to convert incoming xrays into proportionally sized analog pulses that are then converted by a digitalpulseprocessingsystemtoinformationthatcanbereadby a computer and displayed on a spectrum. The resolution of the detector depends on the type and quality of the detector (see Figure8).
Figure7:Xraypassingthroughacollimator Figure 8: Graph comparing the resolution of several differenttypesofdetectors

TRACER
The TRACeR is a handheld EDXRF unit used for instant nondestructive elemental analysis anywhere, anytime. It can be used in a wide range of applications including elemental analysis in material research, archeological digs, museum artifact analysis, conservation and restoration, electric utility industry, engine assembly,airframeassembly,scrapindustry,metalproducers,foundries,andmaintenanceassessment,andmany otherapplications. Xraytube(typicallyAg,RhorRe Target) Upto45kVXrays 170eVSiPINor145eVSDD 13BeDetectorWindow IRSafetySensor Vacuumwindow Userselectablefilter/target Upto200kcps(SDD)
Figure10:SketchofTRACeRmodule

Figure9:FrontviewoftheTRACeRunit

VacuumPort

iPAQPDAuserinterface PDAPinwith Lock XRAYSWITCH

SIPINDIODE/SDD DETECTORPARAMETERS

Table4:SiPIN/SDDcharacteristics (seeAppendixBforexplanations)

XrayTube Filter VoltageSelection CurrentSelection ScanLength OptimalPulseDensity Environment

DetectorChannels

Ag,Rh,orRe Selectable(SeeAppendixA) Variable,045kV Variable,060A Selectable 15,000 max cps (PIN), 150,000 (maxcps(SDD) AirorVacuum
1023PIN/2048SDD

Table5:TRACeRoperatingparameters

ARTAX
The ARTAX unit is a semiportable open beam EDXRF machine used for nondestructive elemental analysisofsurfacesandthespectralmappingofsurfaceareaswithinminutes.Itmaybeusedinmanyapplications such as in archeometry, art history, restoration, forensic sciences, process related quality control, and material sciences.AppendixCincludesacompletetableoftechnicalparametersfortheARTAXunit.

Mo, W, Rh, or Cr, Cu, TiXraytube Upto50kV Less than 155 eV resolution Si(Li) SDD Detector Collimator ChangeableFilter Morethan100kcps

Figure11:ARTAXmodelwithlabeledparts

SI(LI)SDDDETECTORPARAMETERS
XrayTube Mo,W,Rh,Cr,orCu Selectable(SeeAppendixA) 050kV 01000A Variable <50,000cps AirorHeliumFlush
4096

Filter VoltageSelection CurrentSelection ScanTime OptimalPulseDensity Environment DetectorChannels

Table6:TableofSi(Li)detectorparameters

Table7:ARTAXoperatingparameters

SPECTRALINTERPRETATION
AlthoughthegoalofXRFspectroscopyisgenerallytoelementallyanalyzethesample,severalphenomena inherentfromxrayphysicsinvolvedcontributetothespectra.Theseinfluencesrequireinterpretationinorderto correctlyunderstandthedata(seeFigure12).Threemaininfluencescontributetotheoutputspectraofasample: interactionsinthedetector,xrayscontributedbytheanalysissystem,andxrayinteractionsinthesample.These interactionsaredelineatedbelowanddiscussedindetaillaterinthedocument. InteractionsintheDetector o SumpeaksInterpretationoftwoormorepulsesasone

o o o

EscapepeaksPartialLossofenergyduetofluorescenceinmatrix(Si)detector ComptonscatteringPartiallossofxrayenergyenteringdetector HeterogeneityOverlappingpeaksduetodetectorresolution

xrayscontributedbyinteractionsintheanalysissystem o XraytubetargetcharacteristiclinesRayleighscatteredintothedetector o DetectorcanlinesIronandNickeltracepeaks o WindowlinesCalciumtracepeaks o CollimatorandinstrumentstructurelinesAluminumtracepeaks o ThinFilmAnalysisDetectionofsurfacebelowsample PhenomenaintheSample o Rayleighscattering(elasticcollisions)Nolossofxrayenergy o Comptonscattering(inelasticcollisions)Partiallossofxrayenergyinthesample o MatrixeffectsMisrepresentationduetosecondaryabsorption/excitation,densityeffects o BraggscatteringConstructiveinterferenceofXraysinlatticestructures

TheATRAXandTRACeRunitshavetheoptiontochangethevoltage,current,andfilterselection,andthe use of a secondary target to specify the most efficient parameters for a given sample. By selecting the correct combination of these parameters, the above phenomena can be isolated, identified, and/or corrected yielding valuableinformationaboutthesamplescharacter.

Figure12:AbstractviewofPhysicsinvolvedinXraySpectroscopy

10

 INTERACTIONSINTHEDETECTOR
SUMPEAKS
Whentwoormorexraysenterthedetectorattheexactsametimetheyarereadandconvertedintoone pulsewithenergy(e.g.amplitude)equaltothetwopulsescombined.Sumpeaksappearonaspectrumwhenthis occursenoughtimestocreateavisiblepeak,asseeninFigure13andFigure14.Intheory,sumpeakscanappear inanycombinationofcharacteristicenergies,buttheyaremostcommonlyfoundasdoubleKK,KKandKK becausethehigherrateofocurranceofthesexraysleadstoahigherprobabilityofasumeventbeingrecorded. Althoughsumpeaksaresmall,theymaybemistakenastraceelementsandcausespectralinterferencewithother characteristicpeaks. ARTAXunit 25kV 500A MoTube Nofilter 600seconds 49574cps

CuK1 peak(8.904keV) CuK1peak(8.047keV)

CuK K sumpeak (8.047+8.904=16.951keV) CuK K sumpeak (8.047+8.047=16.094keV)

MoKpeak(17.48keV)

Figure13:NISTstandardC1122(seeAppendixDforcomposition)spectrumonalogarithmicscaleshowingcopperKandKbpeaksandtheir ARTAXunit CuK1 peak(8.904keV) sumpeaks

CuK1peak(8.047keV)

CuK K sumpeak (8.047+8.047=16.094keV)

CuK K sumpeak (8.047+8.904=16.951keV)

25kV 500A MoTube Nofilter 600seconds 49574cps

11

Figure14:Linearscaleofabovespectrum

ESCAPEPEAKS
While most characteristic xrays entering the detectorareconvertedintopulseswhichareprocessedby the digital pulse processor, an incoming xray can excite andcausefluorescenceinanatominthedetector.Ifthe xrayenteringthedetectorhasanenergygreaterthanthe absorption edge of an element in the detector (for the ARTAX and the TRACeR: Silicon), then fluorescence in the detector may occur. Figure 15 shows the typical relationshipbetween incoming x rayenergy and resulting Si escape peak counts. The inbound x ray will lose the amountofenergyrequiredtofluorescethedetectoratom, leaving the x ray with an energy E=E inbound E Characteristic energyofdetector,thuscausingthedetectortoreadthexrayas havinganenergyofE.

Figure15:Typicalrelationshipbetweenescapepeakcountand parent energy peak. Parent peak energies closer to the absorptionedgeofsiliconwillcreatemoreescapecounts

Escapepeaksaremuchlessintensethanthecharacteristicpeaksfromwhichtheyarederived.Several escapepeakscanoccurinonespectrum,giventhatallcharacteristicenergiesabovetheabsorptionedgeofthe detector are capable of causing fluorescence. In the case of a Si based detector, escape peaks will appear approximately 1.74 keV lower than a characteristic peak because silicon has a K absorbtion edge. Error! Referencesourcenotfound.showstheescapepeakfromtheCuKpeak.Thisfigurealsoshowsdetectoredge effect,whichoccursatapproximately60%ofthetotalKparentpeakenergy.Escapepeakscanbeautomatically correctedbycomputeralgorithmsandsoftwarethatcalculatesandoutputsacorrecteddatacurve(seeFigure17 andFigure18).

12


SiEscapepeak (8.041.73=6.26keV) Detectoredgeeffects (8.04x.6=4.82keV) Cusumpeaks Cu K parent peak (8.04keV) Cu K parent peak (8.90keV)

Figure16:SpectrafromaTRACeRmodelshowingboththeSiescapepeakoffofaCuparentpeakandthedetectoredge effects.

13

Figure18:LinearscalespectrumofaNISTstandardC1122(seeAppendixD)withoutthecorrecteddatacurveshowstheappearanceofthe Cuescapepeak.

Figure17:Byapplyingthecorrecteddatacurve,theescapepeakisremovedandthetruecharacterofthesampleisshown.

14

HETEROGENEITY
After radiation enters the detector and converts to pulses, discrepancies between peaks due to similar energy levels may occur. Heterogeneity can occur with any combination of lines, including different elements characteristicpeaks,sumpeaks,satellitepeaks,escapepeaks,etc.Theresolutionofthedetectordeterminesthe amountofoverlapbetweensimilarpeaks.
x 1E3 Pulses 150

ARTAXunit 40kV 998A WTube 315mAlfilter 600seconds 11940cps

Pb Fe Cu Zn Pb

100

Zn Fe Cu

50

0 0 5 10 - keV 15 20 25

Figure19:Linearscalespectrumofabronzeingotshowingtheoverlapofcopperandzinccharacteristicpeaks

Zn K peaks overlap with the Cu K peak to form a shoulder on the Cu peak insteadoftwoseparatepeaks

15

HETEROGENEITY(CONTINUED)
x 1E3 Pulses 150

100

ARTAXunit 40kV 998A Wtube 315mAlfilter 600seconds 11813cps

Zn Fe Cu

Pb

Fe

Cu Zn

Pb

50

0 0 5 10 - keV x 1E3 Pulses

15

20

25

Figure 20: Linear scale spectrum of a bronze sample showingtheeffectofoverlappinglines

4.0

The Si K escape peak off of the Cu K peak (6.31 keV) can be easily confused withtheIronKpeak(6.40keV).Figures 3133showtheaffectoverlappingpeaks canhaveonapparentsamplecharacter.
0.0 x 1E3 Pulses 2 4 6 8 - keV 10 12 14 3.0

Pb

Sn

Fe

Ni

Cu

Zn

Pb

2.0

1.0

4.0

Figure 21: Closeup linear scale spectrum without the corrected datacurve.TheCuescapepeakappearstobeaFepeak

3.0

Pb Sn Fe Ni Cu Zn Pb 2.0

The downside of using a corrected data curve is the possibility of removingordiminishingtruesample lines.

1.0

0.0 2 4 6 8 10 12 14 Figure22:CloseuplinearscalespectrumwiththeARTAXsoftwarecorrecteddatacurve.TheFe - keV peakisdrasticallylowerthanwiththeescapepeakontop

16

INTERFERENCEWITHINSTRUMENTATION
EQUIPMENTANDINSTRUMENTCONTRIBUTION
Astheincidentradiationtravelsfromthesourcetothesample,itmaycausefluorescenceinmaterialsin themachinewhichmaybedetectedandshownonthespectrum.Thetargetelementmaybedetected(seeFigure 23)inadditiontoiron,zinc,copper,andnickelinthetube,collimators,lens,etc(seeFigure24).Byaddingafilter inbetweenthetubeandthesample,muchofthisunwantedradiationcanberemovedfromthespectrum(see Figure26).
Pulses

100

WL1

80

ARTAXunit 40kV 998A WTube 315mAlfilter 120seconds 1489cps WL2

60

WL1
W

WL1
40

WL3
20

0 0 5 10 - keV 15 20

Figure23:LinearscalespectrumofanAl203refractorywithtungstenpeaksfromxraytube

SourcesofContribution XraytubetargetKandLlines(e.g.Cu,Rh,Mo,W,etc.) StainlessSteelDetectorCanLines(e.g.Fe,Co,Ni,onlyappearwhentestinglowZelements) Windowlines(e.g.Ca) Collimatorandinstrumentstructure(e.g.Al) Ifusingathinfilmsample,elementsinsurfacebelowthesample

17

EQUIPMENTANDINSTRUMENTCONTRIBUTION(CONTINUED)
Pulses

600

400 Ca Zr Fe Zr W Pb Ca Fe W Pb Zr

ARTAXunit 40kV 998A Wtube 315mAlfilter 600seconds 620cps WL1 WL1

200

Fe K and KPeaks

WL3

0 0 5 10 15 20 - keV Figure24:LinearscalespectrumofaFefreemicroscopeslide.BothWandFepeaksappearonthespectrumbecauseofionizationofthe instrumentation

Figure25:Linearscalespectrumshowingtheappearanceofpeaksofelementsfoundinthecanandotherinstrumentation(Ni,Fe,Cu, Cr). By adding a filter much of the unwanted radiation is removed from the spectrum. (Taken from Bruker AXS presentation IntroductiontoXraySpectrometry)

WDXRF

WL1 WL2

18

THINFILMANALYSIS(BACKGROUNDCONTRIBUTION)
X rays with high energy have the ability to partially penetrate through the surface of a sample. This phenomenon is also found in the use of filters, where a thin layer of metal or substance is used to attenuate certain energies from the exciting x ray beam that is incident on thesample of radiation. Ifthe sample being testedisthinenoughfortheradiationtoentirelypenetratethrough,elementsinthesurfacebelowthesample maybefluorescedanddetected.Thefiguresbelow(seeFigure26,Figure27,andError!Referencesourcenot found.)showstheeffectoflayeringthinfilmsamplesandthedetectionoflowerlayers.
Figure26:Measuredseparately,samples1,2,and3shows11565,21077,and50522countsofPb,respectively.Whenthesamples arelayeredintheorder321thetotalcountsincreasesto55768(around5000abovethatofjustsample3),indicatingthepartial detectionofthelowerlayersofPb.Whenthesamplesarelayered123thetotalcountsincreasesto22757(nearly2xthatofjust sample1),indicatingthedetectionoflowerlayersisdiminishedbutstillaffectsthespectra

Figure27:Spectrausedinfigure26.NotethevariationoftheintensityshiftofthePbLalphaandPbLbetadependingwherethehighest concentrationofPbisinthesample.HighestclosetothesurfaceshowsPbLalphaabout15%moreintensethatPbLbeta.Lowest concentrationclosetothesurfaceshowsPbLbetatobeabout3%higherthanPbalpha.

19

PHENOMENAINTHESAMPLE
RAYLEIGH(ELASTIC)SCATTERING
Incidentradiationfromthetubethatreachesthesampleiseitherabsorbedinthephotoelectriceffector reflectedandscattered.Whenanxrayreflectsofftheatomsofthesamplewithoutlosinganyenergyitiscalled Rayleigh(orelastic)scattering.Theenergyoftheoutboundxraywillbeequaltotheenergyoftheinboundxray, thus being detected as a source peak with the energy of the inbound x ray. The Rayleigh scatter peaks visible correspondwiththecharacteristicenergiesofthexraytubetargetelement.Figure29andFigure30showthe appearanceofRayleighpeaksinaspectrum.Rayleighscatterpeaksarecharacterizedbysharpshapesthatarethe sameasthexrayflourescencepeaksbecausetheyareproducedinthedetectorbysingleenergyxrays.
x 1E3 Pulses

2.5

2.0

0.5 1.5

The W L peak (8.398 keV), W L peaks (9.5259.951keV),andWLpeaks(11.29 11.68keV)allappearonthespectrumdue toRayleighscatter

Fe

Co

Ni

Cu

ARTAXunit 40kV 998A CuKandKpeaks WTube 315mAlFilter 600seconds 16115cps

1.0

SiKescapepeak

0.0 4 6 8 - keV 10 12

Figure28:CloseuplinearscalespectrumofaNISTstandardC1251a(seeAppendixDforcomposition)showingthe appearanceofthetungstentargetRayleighscatterinspectra

RhLlinesfromrhodiumtargetin TRACeR tube appear in spectra duetoRayleighscattering.

TRACeRunit

Figure29:CloseuplinearscalespectrumshowingtheappearanceofRhcharacteristiclinesinaspectrumduetoRayleighscatter

20

 RAYLEIGH(ELASTIC)SCATTERING(CONTINUED)
InadditiontoRayleighscatterpeaksatthecharacteristiclinesofthetargetinthetube,aBremsstrahlung radiationcurvemayappearinthebackgroundofaspectrumduetoRayleighandComptonscatterofallincidentX rays(Figure31). ARTAXunit 40kV 998A WTube 315mAlfilter 600seconds 1486cps
Figure30:LinearscalespectrumofahighpurityAl2O3 refractory withawelldefinedBremsstrahlungcontinuum

21

COMPTON(INELASTIC)SCATTERING
Incident radiation with sufficient energy to ionize an inner shell electron in an atom does not always cause fluorescence, but insteadcausesanexcitationwithoutlosingallofitsenergy(seeFigure 32).Intheseinteractions,calledComptonscattering,axraystrikesan atom and loses energy, causing the excitement of an innershell electron.Becausenovacancyiscreatedintheatom,nocharacteristic energy is released; however, the x ray will lose energy and be scattered in all directions (noted by the formula in Figure 32). Comptonscatterxraysappearasabroadpeakdefinedbytheangle between the incident beam and the detector for the target characteristicxraylowerinenergythanRayleighscatterpeaksbecause Figure 31: Compton scattering causes a shift in theyonlyloseasmallamountofenergyintheexcitationofanelectron. wavelengthoftheincidentxray Theyaregenerallybroadduetotheareaofthedetectorandtheareaof theexcitingbeam,andtheyoccurmoreofteninlowZelements.ComptonscatteringcanbeseeninFigure33.

RhK RayleighScatter (20.21keV) ComptonScatter

RhK RayleighScatter (22.72keV)

22

Figure32:LinearscalespectrumshowingComptonscatteringofrhodiumparentpeaksonaTRACeRmodel.

MATRIXEFFECTS
Absorption: Any element that can absorb or scatter the incident xrays is capable of reabsorbing characteristic xrays of other elements (see Figure 34). After an atom undergoesthephotoelectriceffectandemitsa characteristic x ray, the x ray may be reabsorbed by another atom in the sample. When this happens, it causes a misrepresentation of the counts of elements detectedfromthesamplebyfailingtocounta x ray for the initial element. The expected number of x ray counts recorded by the detector will be lower than expected because someareabsorbedwithinthesample.

Figure33:Exampleofsecondaryabsorption.

SecondaryExcitation:Whenthecharacteristic radiation from one atom is reabsorbed by another atom and has sufficient energy to ionize the atom, it will cause fluorescence in the second atom, producing only the characteristic radiation of the second atom (see Figure 35). This can lead to a misrepresentation of elements by enhancing the appearance of elements through secondaryexcitation.
Figure34:Exampleofsecondaryexcitation

23

BRAGGSCATTERING
Many samples are composed of a periodic arrangement of atoms or molecules that create a crystal lattice.Whenamaterialexhibitsalatticestructure,severaldifferentlatticeplanescanexistorientedindifferent directions. All planes that are parallel to a given lattice plane are a set distance away from one another, as establishedbythecrystalstructure. Whentwoparallelincidentxraysstrikeapairofparallellatticeplanes,theraysarereflectedandcan interferewitheachother.ThisphenomenonisknownasBraggscattering,orthescatteringofincidentxraysdue tothecrystallatticeofasample.Braggslawstatesthattwowavesinterfereconstructivelywhenn=2dsin where:nisthereflectionorder,isthewavelengthoftheincidentxray,disthedistancebetweenthelattice planes,andistheangleofreflectionuponthecrystallattice.AccordingtoBraggslaw,agivenanglehasa specificsetofwavelengths(orenergies)thatcancauseconstructiveinterference.Sampleswithuniformcrystal structurewillexhibitnarrowpeaks,whileasamplewithanonuniformcrystallatticewillbebroader.Allthedata usedintheexamplesbelowaretakenwiththeArtaxsystem.ThesharperlinesinFigure37suggestthatthe sampleisofhigherquality(moreuniformcrystal)thanthatinFigure38.

0 Degrees 10 Degrees 25Degrees

As the angle is changed, certain peaks due to Bragg scattering appearanddisappear.

ARTAXunit 40kV 998A WTube NoFilter 60seconds 3,343cps

Figure35:LinearscalespectrumofabluegemstoneshowingtheeffectofincidentangleonBraggpeaks

Ifthesamplehasapreferredcrystalorientation,onlyasmallangleisrequiredforBraggscattering(<1), soBraggscatteringcanbeidentifiedandpreventedbychangingtheangleofincidenceofthexraytothesample. Indoingso,apeakinonegraphwillnotappearonanotherbecausetheanglethatsatisfiesBraggslawhasbeen changed.ToensurethatBraggpeaksarenotconfusedwithcharacteristiclinesofelements,theangleofincidence betweentheincidentxraybeamandthesamplecanbealtered,thusdisconnectingtheconstructiveinterference whichcreatesBraggpeaks.

24

Whenthesampledoesnothaveastronglypreferredcrystallineorientationasinthecaseofagemstone, byaddingafilteritispossibletoeliminateBraggpeaksbyremovingthespecificwavelengthsthatcreate constructiveinterference.However,sometimesBraggpeaksareimpossibletoremovewithoutinterferingwith theregionofinterest.

0 Degrees 10 Degrees 25Degrees

MultipleBraggpeaks

ARTAXunit 40kV 998A Wtube 315mAlfilter 60seconds 1242cps

Figure36:Linearscalespectrumofagreengemstoneatangles0,10,and25,showingthechangesinBraggpeaksandtheireffectonother characteristicpeaks.

Bragg Scattering can interfere with peaks and causeamisrepresentation ofelementsinthesample. In this example, the peak is greatly enhanced by Bragg scattering in the green spectrum, whereas the pink spectrum peak is not exaggerated because theanglewaschanged.

25

 BRAGGSCATTERING(NISTC1122EXAMPLES)
Angle 1 Angle 2 Angle3

ARTAXunit 25kV 498A Motube NoFilter 600s 40360cps

Figure37:LinearscalespectrumofaNISTstandardC1122(seeAppendixD)samplewithnofilterexhibitingtheBraggpeaks

26


BRAGGPEAKS

BraggPeaks

Angle 1 Angle 2 Angle3

Figure38:Braggpeaksappearindifferentintensitiesandpositionsbasedonthecrystalorientationofthesample.

BraggScattering(Gemstoneexamples)

Angle 1 Angle 2 Angle 3 Angle4

BraggPeaks

Figure39:CertainBraggpeaksareamplifiedatspecificangles,whilediminishedatothers.

27

Angle 1 Angle 2 Angle 3 Angle4

BraggPeaks

Figure40:Braggpeaksmayappearontopofotherpeaks,asseeninthered.

BraggPeaks

Angle 1 Angle 2 Angle 3 Angle4

BraggPeaks

Figure41:TheredspectrumdisplaysnoBraggpeaks,whilethegreenandpinkhaveseveraldifferentpeaks

28

FIELDAPPLICATIONS
IDENTIFYINGGENUINEARTIFACTS(CHELSEABULLFINCHEXAMPLE)
WiththeincreasedportabilityandeaseofuseintheTRACeRandARTAXunits,severalapplicationscanbemadeinseveraldifferentareasofstudy. ThefollowingareafewexamplestakenfromapresentationbyDr.BruceKaiseronEDXRFapplications.

Figure42:OverlappingspectraofthetwodifferentpaintsprovethattheceramicBullfinchwasrestoredfromitsnaturalcondition.

29

IDENTIFYINGTRUEORIGINS(STONEWAREEXAMPLE)
ByUsingtheTRACeRandXRFspectroscopytechnology,historians can easily identify the origins of otherwise unknown artifacts. Here, English stoneware (blue spectrum) contains significantly moreironthanGermanstoneware(redspectrum).Byusingthis knowledge,historianscanidentifytwovisuallyidenticalpots.

Figure43:TrueoriginsofartifactscanbefoundusingelementalanalysisthroughXRFtechnology

30

Measuring Chlorine with the Tracer Measurement of Cl is very important as it is often involved in corrosion and degradation of artifacts in marine environments. Or in some cases is a key constituent of pigments or other coatings, or an issue in paper conservation. The following slides first depict how to make up thin film standards to determine the Cl surface content in micro grams per square centimeter. And then how to set the Bruker handheld xrf instrument up to measure levels as low as 10 micro grams per square and shows 2 applications. It should be noted the Cl analysis is very much a SURFACE ANALYSIS when using xrf, as the Cl atom emits only a 2.7 keV x ray. This low an energy x ray is not able to escape the sample unless the atom is very near the surface.

31

CreationofveryverythinfilmChlorineStandards

1.83gmsofZirconiumdichlorideoxide(ZrOCL2.8H2O)wasaddedto100mlofdistilled water. Thenvariousamountswerepipettedontolightweightpapercircles8.2cmindiameter Thepaperwassaturatedwiththesolutionineachcasetoassurethatthesolution distributeduniformlyovertheentiresurface Eachpaperwasthenletdryonaplasticsheetfor1hour Theresultingmicrogram/cmvaluesforZrandCl

6.347 12.693 9.866 4.933 Zr-ug/sq-cm 0.000 1.269 Cl-ug/sq-cm 0.000 0.987

32

Cl senstivity (ugm/square cm)

350 300 # of x-rays 250 200 150 100 50 2 2.2 2.4 2.6 2.8 keV 3 3.2 3.4

C
Pure iron 0 .0 Cl 0.99 Cl 4.93 Cl 9.87 Cl

Each standard was analyzed for 3minutes at 2 different voltage and currentsettings. A 0.001TitaniumfoilwasusedinbothcasestoeliminatetheRhLlinesandgenerateTix raystoexciteClefficiently.ThethinpaperstandardswerebackedbypureFetomimicCl corrosiononFe.Theaboveisaplotofthemeasurementsthatweretakenat8kVand35 micro amps. The peak at 2.6 keV is the Cl K xray. The peak at 2.95 keV is a constant amplitudeandisaresultofArKxraywhichisintheairinthepaper.Itisclearthesystem issensitivetoCldowntolevelsaslowas1microgram/squarecm
33

BrukerArtaxXrfScanofHunleyRivet
Theoperatingparametersare:
Spotsize0.070mm2(microfocustube) Samplinggrid0.070mm2 15kVtubexraytubevoltage Motubetarget 300microampsxraytubefilamentcurrent 60secondanalysistimeperpointwithHeliumflush Beamarmwaspointeddownbutcanbeorientedin anydirectionforanysizedobject. Systemtripodisonwheelsandcanbemovedquickly. AnalyticalsoftwarerunseasilyonanyWindowsXPsystem

34

Micrograms/cm2 of Chlorine

26-28 24-26 22-24 20-22 18-20 16-18 14-16 12-14 10-12 8-10 6-8 4-6 2-4 0-2

Cl Distribution on Rivet Side at Machined Boundary

0.56 0.49 0.42 0.35 0.28 0.21 0.14 0.07 0.00 0.00 0.07 0.14 0.21 0.28 0.35 0.42 0.49 0.56 0.63 0.70 0.77 0.84

Start scan

mm
Machined area

mm

un Machined area

35

Fe Si escape peak

Scan56showingtypicalinstrumentresponse
36

SourcingObsidian ppmsensitivitytokeyelements Keytechniquetodeterminehumanmovementandactivity BrukerTRACeRxrfsystemsfoundtobeveryaccurateforthisapplication

BrukerTRACeROperatingParameters
40kVand10microamps .006Cu,.001Ti,.012Alfilters 180secdataacquisition

ThefollowingdataisanexampleofTraceranalysisdonebyJeffSpeakmanoftheSmithsonian

37

 38

6000

5000

4000

3000

Fe
2000

alca Chivay CRG 0002 Ixtepeque MLZ 1019 Mono Glass MTN Otumba Pico de Orizaba Quispisisa Sierra de Pachuca Ucareo UNL-050_1 XMC 020 Yellowstone

Zr

1000

Zr
0 5 6

Rb F7 C
8

Nb Sr
14

10

11

12

13

15

16

17

18

19

20

39

1200 1100 1000 900 800 700

alca Chivay CRG 0002 Ixtepeque MLZ 1019 Mono Glass MTN Otumba Pico de Orizaba Quispisisa
Rb Zr

Zr

Nb

600 500 400 300 200 100

Sierra de Pachuca Ucareo UNL-050_1 XMC 020 Yellowstone

Sr

C
0 13 13.5 14 14.5 15 15.5 16 16.5 17 17.5 18
40

MeasurementofToningAgentsonPhotographs
Michele

Use.006Cu,.001Ti,.012Alfilter Analyzeatphotograph

1970

40kV 6microamps(whatisavailable) Novacuum 5to10minin

Whiteareawherethereinnotoner Areastonevariesgreytoblack Takethedifference(tonedwhite)seebelow Thedifferencewillgiveyouaverygoodcleanspectrumofthetoningagent.And thegreytoblackvariationwillgiveyouanestimateoftheamountofagent.The reasonthisworkssowellare Thetoningmaterialsareverythinandhaveverylittleeffectonthespectrum fromthepaper

Thewhiteareaisjustthepaperandmountingmaterials Thefilterusedremovesmostofthebackscatteredxrays
41

1500

Ba L

michele dark 1970 michele white1970 Difference

Sr

Note the red difference spectrum clearly shows that the image forming agent is only Ag
Ag

1000

Pd K

500

Ag L
Rh K Rh K Pd K inelastic

Ag Ba

Fe K

Cu

Sr

Elastic backscatter backscatte

0
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

X ray Energy (keV)

42

APPENDICES
APPENDIXA
FILTERSAVAILABLEFORTHEARTAXANDTRACERUNITS(NOTEUSERCANMAKEUPANYFILTERORSECONDARY
TARGETHECHOSESDEPENDINGONHISPARTICULARNEEDANDAPPLICATION )

SomeoftheFiltersavailablefortheARTAXunitfromBruker o 315mAl o 25mNi o 12.5mNi o 12.5mMo o 100mAl o 200mAl SomeoftheFiltersavailablefortheTRACeRunitfromBruker o Bluefilter(1milCu) o Yellowfilter(12milAl+1milTi) o Redfilter(12milAl+1milTi+1milCu)

EFFECTOFAFILTERONANXRAYTUBESPECTRUM

Figure44:Scatteredexcitationspectrumprovidedbyasilvertargetxraytubeoperatedat15kV.Theplotsshowtheeffectoftwo thicknessesofanalprimarybeamfilter:(1)unfiltered;(2)thinalfilter;(3)thickalfilter.

43

EXAMPLEOFANATTENUATIONGRAPHFORAFILTERCOMPOSEDOFVARIOUSTHICKNESSESOFFE

Figure45:Toattenuateahigherpercentageofradiation,aswellasahigherenergyradiation,thefilterthicknessmustincrease.

44

Figure46:Thisgraphismadetodeterminethefilterthicknessrequiredforoptimalresultsintestingacoppersample.

TYPICALFILTER,VOLTAGEANDCURRENTSELECTION FOROPTIMUMXRFELEMENTAL GROUPANALYSISUSINGTHETRACER

ScreeningforallElements(LabRatmode): Nofilter 40kV 3to5microamps(fornonmetallicsamples) 0.6to1.4microamps(formetallicsamples) Utilizethevacuum. Thesesettingsallowallthexraysfrom1keVto40keVtoreachthesamplethusexcitingalltheelementsforMgto Pu. TooptimizeforparticularelementalgroupsonewantstousefiltersandsettingsthatpositiontheXrayenergy impactingthesamplejustabovetheabsorptionedgesoftheelement(s)ofinterest.Examplesofhowtogoabout thisisgivenbelow.Noteaswellthatthedepthofanalysisisalsoverymuchafunctionofboththexrayenergy usedtoprobethematerialandtheelementthatisbeingexcited,bothareexponentialfunctionsdependentonthe matrixofelementsthatthematerialiscomposed. 1. 2. 3. 4. 5.

45

MeasurementofObsidianforhigherZelements(Rb,Sr,Y,Zr,andNb): 0.006Cu,.001Ti,.012AlFilter 40kV 4to8microamps Novacuum Thesesettingsallowallthexraysfrom17keVto40keVtoreachthesamplethusefficientlyexcitingtheelements fromFetoMo.Thesearesomeofthekeyelementstoidentifyingtheoriginoftheobsidianandmanyothernatural occurringmaterialsusedbyearlyman.ThereislittleornosensitivitytoelementsbelowFewiththesesettings. 1. 2. 3. 4.

MeasurementofMg,Al,SiandPtoCu(andanyLandMlinesfortheelementsthatfallbetween1.2and8keV) Nofilter 15kV 15microamps Vacuum Thesesettingsallowallthexraysfromthetubeupto15keV.InparticularthisallowstheRhL(2.5to3keV)lines fromthetubetoreachthesample.Theseareparticularlyeffectiveatexcitingtheelementswiththeirabsorption edgebelow2.3keV.NotethissetupisnotgoodforClandSdetection,asthescatteredRhLlinesinterferewiththe xrayscomingfromtheseelements. 1. 2. 3. 4.

46

 MeasurementofMg,Al,Si,P,Cl,S,K,Ca,V,Cr,andFe(andanyLandMlinesfortheelementsthatfallbetween 1.2and6.5keV) Tifilter 15to20kV 15to20microamps Vacuum Thesesettingsallowxraysfrom3to12keVtoreachthesample.InparticularthisdoesnotallowtheRhLlines fromthetubetoreachthesample.TheseRhLxrayswouldinterferewithClandSanalysis.Forexample,thisisa verygoodsetupformeasuringClonthesurfaceofFe. Measurementofmetals(TitoAgKlinesandtheWtoBiLines): 0.001Ti,.012AlFilter(yellow) 40kV 1.2to2.6microamps Novacuum Thesesettingsallowallthexraysfrom12keVto40keVtoreachthesamplethusefficientlyexcitingtheelements notedabove.Thesearethesettingsusedtocalibratethesystemforallmodernalloysofthoseelementsofthose listedinthetitleofthissection.ThereislittleornosensitivitytoelementsbelowCawiththesesettings. 1. 2. 3. 4. 1. 2. 3. 4.

47

 MeasurementofPoisons(higherZelementsHg,Pb,Br,As): 0.001Cu,.001Ti,.012AlFilter 40kV 4to8microamps Novacuum Thesesettingsallowallthexraysfrom14keVto40keVtoreachthesamplethusefficientlyexcitingtheelements Hg,Pb,Br,As.Thesearesomeofthekeyelementsthatwereusedtopreserveorganicbasedartifacts.Thereislittle ornosensitivitytoelementsbelowCawiththesesettings. 1. 2. 3. 4.

48

APPENDIXB

Figure47:ExplanationofSiPINcharacteristics

49

APPENDIXC

50

APPENDIXD
BerylliumCopper C1122 PhosphorizedCu C1251a

Elem Sb Sn Ag Bi Pb Se Zn Cu Ni Co Fe Mn Cr Al P Si Mg S Be As Te Au

% .01 .005 .003 .01 97.45 .01 .22 .16 .004 .002 .17 .004 .17 1.75

Elem Sb Sn Ag Bi Pb Se Zn Cu Ni Co Fe Mn Cr Al P Si Mg S Cd As Te Au

% .0014 .0016 .0080 .00037 .00235 .0011 .0024 99.89 .00236 .00132 .0285 .00046 .0003 <.002 .0420 <.005 <.002 .0035 <.0003 .0016 .0016 .00155

Table 8: Elemental composition of NIST standards C 1122 andC1251a

51

APPENDIXE

Figure48:PeriodictableincludingtheKandLseriescharacteristicenergiesofalloftheelements

52

53