AAS- atomic absorption spectroscopy .pdf

atomic absorption

1. Atomic
Absorption
andAtomic
Fluorescence
Spectrometry
I
nthischapterweconsidertwo(lpesofoptical
atomicspectrometricmethodsthatusesimilar
techniquesforsampleintroductionandatom-
ization.Thefirstisatomicabsorptionspectrometry
(.4AS),whichfornearlyhalfocenturyhasbeenthe
mostwidelyusedmethodfort~determinationof
singleelementsinarwlytjcalsamp!'!!.Thesecond
isatomicfluorescencespectrometry(4F'S),which
sincethemid-1960shasEeenstudiedextensil'e{v.
BycOr(trasttotheabsorp1ionfIleT/lOd,atomic
fluorescencehasnotgainedwidespreadgeneral
useforroutineelementalanalysis,however.Thus,
althoughseveralinstrumentmakershaveinrecent
yearsbeguntoofferspecial-purposeatomicfluores-
cencespectrometers,theoastmajorityofinstru-
mentsarestilloftheatomicabsorptiontype.
Becauseofthi.~differenceinuwge,wedel'ote
thebulkofthischaptertoAASandconfineour
descriptionofAFStoabriefsectionattheend.
r71Throughoutthischapter,thislogoindicates
lQ.janopportunityforonlineself-studyatwww
.thomsonedu.com/chemistry/skoog,linkingyouto
interactivetutorials,simulations,andexercises.
PriortodiscussingAASIindetaiLwefirstpresentan
overviewofthetypesofatomizersusedinbothAAS
andAFS
Ex.citation
Atomicions-.
loniz.ation
t
Excited
ire"-ersiblcl Ions
Atoms-.
Dissociation
I
Excited
(reversible)
atoms
Gaseuus
,\:,"-.
molecules
Thetwomostcommonmethodsofsampleatomization
encounteredinAASandAFS,flameatomizationand
electrothermalatomization,arefirstdescribed.We
thenturntothreespecializedatomizationprocedures
usedinbothtypesofspectrometry.
Volatilizationt....
.'~'--;~';.;.::
Solid/gas.::~.w':l.
aerosol '.(~.~.
Desohatlont".
Spray ')~:.'
Nebulizationt
~.:-
Inaflameatomizer,asolutionofthesampleisnebu-
lizedbyaflowofgaseousoxidant,mixedwithagaseous
fuel,andcarriedintoaflamewhereatomizationoccurs.
AsshowninFigure9-1,acomplexsetofinterconnected
processesthenoccurintheflame.Thefirstisdesalva-
tion,inwhichthesolventevaporatestoproduceafinely
dividedsolidmolecularaerosol.Theaerosolisthen
volatilizedtoformgaseousmolecules.Dissociationof
mostofthesemoleculesproducesanatomicgas.Some
oftheatomsinthegasionizetoformcationsandelec-
trons.Othermoleculesandatomsareproducedinthe
flameasaresultofinteractionsofthefuelwiththeoxi-
dantandwiththevariousspeciesinthesample.Asin-
dicatedinFigure9-1,afractionofthemolecules,atoms,
andionsarealsoexcitedbytheheatoftheflametoyield
atomic,ionic,andmolecularemissionspectra.Withso
manycomplexprocessesoccurring,itisnotsurprising
thatatomizationisthemostcriticalstepinflamespec-
troscopyandtheonethatlimitstheprecisionofsuch
methods.Becauseofthecriticalnatureofthe
atomizationstep,itisimportanttounderstandthe
characteristicsofflamesandthevariablesthataffect
thesecharacteristics.
Analytesolution
FIGURE9-1Processesoccurringduringatomization.
thattemperaturesof1700
0
eto2400·eoccurwiththe
variousfuelswhenairistheoxidant.Atthesetempera,
tures,.onlyeasilydecomposedsamplesareatomized,so
oxygen'ornitrousoxidemustheusedastheoxidantfor
more'refractorysamples.Theseoxidantsproducetem-
peraturesof2500
0
eto31oooewiththecommonfuels.
TypesofFlames
Table9-1liststhecommonfuelsandoxidantsusedin
flamespectroscopyandtheapproximaterangeoftem,
peraturesrealizedwitheachofthesemixtures.Note
Maximum
Burning
Temperature, Velocity,
Fuel Oxidant ·C cms-1
Naturalgas Air 1700-1900 39-43
Naturalgas Oxygen 2700-2800 370-390
Hydrogen Air 2000-2100 300-440
Hvdrogen Oxygen 2550-2700 900-1400
Acetylene Air 2100-2400 158-266
Acetylene Oxygen 3050-3150 1100-2480
Acetylene Nitrous 2600-2800 285
oxide
'GeneralreferencesonatomicahsorptionspectrometryincludeLH.J
L<ljunenandP.Pcramaki,SpectrochemicalAnalnisbyAwmieAbmrptiofl
andF.missio'l,2ndcd.,Camhridgc:RoyalSocietyofChemistry.2004:
J.A.CRroekaert,Anl/lyticalAromicSpf'CIfOmelryl~'ilhFlamesandPillS
mas,\Veinheim,Germany:Wiley-veH.2002;RMagyar.Guide-Linesto
PlanningAtomicSpectrometricAnalysis.NewYork:Elsevier,19HZ:1.0
IngkJr.andS.R.Crouch.SpectrochnnicalAnalysis.Chap.10.Engle
woodCliffs.NJ:PrenticeHaiL1'}88~M.SperlingandB.Welz.Atomic
AhsorptlonSpeetrometn.3rded..Nc\\:York:VCHPublishers.1999:N.H
Bill~s.A.Bogaerts.andJA.CHroekaert.Anal.Chern..2004.70..1113
Secondary
combustion
zone
Theburningvelocitieslistedinthefourthcolumnof
Table9-1areimportantbecauseflamesarestableonly
incertainrangesofgasflowrates.Ifthegasflowrate
doesnotexceedtheburningvelocity,theflameprop-
agatesbackintotheburner,givingflashback.Asthe
flowrateincreases.theflamerisesuntilitreachesa
pointahovetheburnerwheretheflowvelocityandthe
burningvelocityareequal.Thisregioniswherethe
flameisstable.Athigherflowrates,theflamerises
andeventuallyreachesapointwhereitblowsoffthe
burner.Withthesefactsinmind,itiseasytoseewhyit
issoimportanttocontroltheflowrateofthefuel-
oxidantmixture.Thisflowrateverymuchdependson
thetypeoffuelandoxidantbeingused.
FlameStructure
AsshowninFigure9·2.importantregionsofaflame
includetheprimarycombustionzone,theinterzonal
region,andthesecondarycombustionzone.Theap-
pearanceandrelativesizeoftheseregionsvarycon-
siderablywiththefuel-to-oxidantratioaswellaswith
thetypeoffuelandoxidant.Theprimarycombustion
zoneinahydrocarbonflameisrecognizablebyitsblue
luminescencearisingfromthebandemissionofC".
CH.andotherradicals.Thermalequilibriumisusually
notachievedinthisregion,anditis.therefore,rarely
usedforflamespectroscopy.
Theinterzonalarea.whichisrelativelynarrowin
stoichiometrichydrocarbon!lames,mayreachseveral
centimetersinheightinfuel-richacetylene-oxygenor

2. 1700
~ J7 50/--.
1600 '\~/ \
(SOO
§30
">
~
"
'0
~l.O
6
FIGURE9-3TemperatureprofilesindegreesCelsiusfor
anaturalgas-airflame.(FromB.LewisandG.vanElbe,
J.Chem.Phys.,1943,11,94.Withpermission.)
acetylene-nitrousoxidesources.Becausefreeatoms
areprevalentintheinterzonalregion,itisthemost
widelyusedpartoftheflameforspectroscopy.Inthe
secondaryreactionzone.theproductsoftheinner
coreareconvertedtostablemolecularoxidesthatare
thendispersedintothesurroundings.
Aflameprofileprovidesusefulinformationaoout
theprocessesthatgoonindifferentpartsofaflame;itis
acontourplotthatrevealsregionsoftheflamethathave
similarvaluesforav'ariableofinterest.Someofthese
variaolesincludetemperature.chemicalcomposition.
absorbance.andradiantorfluorescenceintensity.
TemperatureProfiles.Figure9-3showsatemperature
profileofatypicalflameforatomicspectroscopy.
Themaximumtemperatureislocatedintheflame
about2.5emaoovetheprimarycombustionzone.Itis
important-particularlyforemissionmethods(Sec-
tionJOe-I)-tofocusthesamepartoftheflameon
theentranceslitforallcalibrationsandanalytical
measurements.
FlameAbsorptionProfiles.Figure9-4showstypical
absorptionprofilesfmthreeelements.Magnesium
exhibitsamaximuminabsorbanceataooutthcmid-
dieoftheflamebecauseoftwoopposingeffects.The
initialincreaseinahsorhanccasthedistancefromthe
oaseincreasesresultsfromanincreasednumoerof
magnesiumatomsproducedbythelongerexposure
totheheatoftheflame.Asthesecondarycomoustion
zoneisapproached.howe\'er.appreciableoxidationof
themagnesiumbegins.Thisprocessevcntuallvleadsto
adecreaseinabsorbancebecausetheOXideparticles
formeddonotabsorbattheooservationwa\'e1ength.
Toachievemaximumanalyticalsensitivity.then.the
flamemustbeadjustedupanddownwithrespectto
thebeamuntiltheregionofmaximumabsorbanceis
located.
Thebehaviorofsilver.whichisnoteasilyoxidized.is
quitedifferent;asshowninFigure9-4.acontinuousin-
creaseinthenumberofatoms,andthustheaosorbance,
isooservedfromtheoasetotheperipheryoftheflame.
Bycontrast,chromium.whichformsverystaoleoxides,
showsacontinuousdecreaseinabsorbancebeginning
closetotheournertip;thisooservationsuggeststhat
oxideformationpredominatesfromthestart.These
ooservationssuggestthatadifferentportionofthe
flameshouldoeusedforthedcterminationofeachof
theseclements.Themoresophisticatedinstrumentsfor
flamespectroscopyareequippedwithmonochroma-
torsthatsampletheradiationfromarelativelysmallre-
gionoftheflame,andsoacriticalstepintheoptimiza-
tionofsignaloutputistheadjustmentofthepositionof
theflamewithrespecttotheentranceslit.
~Nebulizu
To\\'ask oxidanl
thatusesaconcentric-tubeneoulizer,suchasthat
showninFigure8-1la.Theaerosol.formedoytheflow
ofoxidant,ismixedwithfuelandpassesaseriesof
bafflesthatremoveallbutthefinestsolutiondroplets.
Thebafflescausemostofthesampletocollectinthe
bottomofthemixingchamberwhereitdrainstoawaste
container.Theaerosol,oxidant,andfuelarethen
burnedinaslottedournertoprovidea5-tollJ-cmhigh
flame.
Laminar-flowburnersproducearelativelyquiet
flameandalongpathlengthformaximizingabsorp-
tio~.Thesepropertiestendtoenhancesensitivityand
reproducioilityinAAS.Themixingchamberinthis
typ~ofburnercontainsapotentiallyexplosivemixture
thatcanflashoackifthetlowratesaretoolow.Note
thatthelaminar-flowburnerinFigure9-5isequipped
withpressurereliefventsforthisreason.Other
typesoflaminar-flowburnersandturbulent-flowburn-
ersarcavailaoleforatomicemissionspectrometry
andAFS.
FlameAtomizers
Flameatomizersarcusedforatomicabsorption.
fluorescence.andemissionspcctroscopy.Figure9-5isa
diagramofatypicalcommerciallaminar-flowburner
FuelandOxidantRegulators.Animportantvariaole
thatrequiresclosecontrolinflamespectroscopyisthe
flowrateofbothoxidantandfuel.Itisdesirabletobe
aoletovarveachoverabroadrangesothatoptimalat-
omizationconditionscanoedeterminedexperimen-
tally.Fuelandoxidantarcusuallycomoinedinapprox-
imatelystoichiometricamounts.Forthedetermination
ofmetalsthatformstableoxides,however,al1amethat
FIGURE9-5Alaminar-flowburner.
(CourtesyofPerkin-ElmerCorporation,
Norwalk,CT.)
containsanexcessoffuelisoftendesirablc.Flowrates
arcusuallycontrolledbymeansofdouble-diaphragm
pressureregulatorsfollowedoyneedlevalvesinthein-
strumenthousing.Awidelyuseddeviceformeasuring
flowratesistherotameter.whichconsistsofatapered,
graduated,transparenttubethatismountedvertically
withthesmallerenddown.Alightweightconicalor
sphericalfloatisliftedbythegasflow;itsverticalposi-
tionisdeterminedbythetlowrate.
PerformanceCharacteristicsofFlameAtomizers.
Flameatomizationisthemostreproducibleofall
liquid-sample-introductionmethodsthathavebeen
developedforatomicabsorptionandfluorescence
spectrometrytodatc.Thesamplingefficiencyofother
atomizationmethodsandthusthesensitivity.however,
aremarkedlybellerthaninflameatomization.There
aretwoprimaryreasonsforthelowersamplingeffi-
ciencyoftheflame.First,alargeportionofthesample
flowsdownthedrain.Second.theresidencetimeof
individualatomsintheopticalpathintheflameisbrief
(-10-·5).
Electrothermalatomizers,whichlicstappearedonthe
marketintheearlv1970s.generallyprovideenhanced
sensiti\'ityoecausetheentiresampleisatomiz.edina
shortperiod,andtheaverageresidencetimeofthe

3. atomsintheopticalpathisasecondormore.'Elec.
trothermalatomizersareusedforatomicabsorption
andatomicfluorescencemeasurementsbuthavenot
beengenerallyappliedfordirectproductionofemis.
sionspectra.Theyareusedforvaporizingsamples
ininductivelycoupledplasmaemissionspectroscopy.
however.
Inelectrothermalatomizers,afewmicrolitersof
sampleisfirstevaporatedatalowtemperatureandthen
ashedatasomewhathighertemperatureinanelec·
tricallyheatedgraphitetubesimilartotheoneinFig-
ure9·6orinagraphitecup.Afterashing,thecurrent
israpidlyincreasedtoseveralhundredamperes,which
causesthetemperaturetoriseto2000"Cto3000°C;at-
omizationofthesampleoccursinaperiodofafewmil·
lisecondstoseconds.Theabsorptionorfluorescenceof
theatomicvaporisthenmeasuredintheregionimme-
diatelyabovetheheatedsurface.
ElectrothermalAtomizers
Figure9·6aisacross-sectionalviewofacommercial
electrothermalatomizer.Inthisdevice,atomization
occursinacylindricalgraphitetubethatisopenat
bothendsandthathasacentralholeforintroduction
ofsamplebymeansofamicropipette.Thetubeis
about5emlongandhasaninternaldiameterofsome·
whatlessthan1em.Theinterchangeablegraphitetube
fitssnuglyintoapairofcylindricalgraphiteelectrical
contactslocatedatthetwoendsofthetube.These
contactsareheldinawater-cooledmetalhousing.Two
inertgasstreamsareprovided.Theexternalstream
preventsoutsideairfromenteringandincineratingthe
tube.Theinternalstreamflowsintothetwoendsofthe
tubeandoutthecentralsampleport.Thisstreamnot
onlyexcludesairbutalsoservestocarryawayvapors
generatedfromthesamplematrixduringthefirsttwo
heatingstages.
Figure9·6aillustratestheso-calledL'vovplatform,
whichisoftenusedingraphitefurnacessuchasthat
showninthefigure.Theplatformisalsomadeof
graphiteandislocatedbeneaththesampleentrance
port.Thesampleisevaporatedandashedonthisplat·
form.Whenthetubetemperatureisincreasedrapidly,
however.atomizationisdelayedbecausethesampleis
:F,lrdetaileddlSCU'isionsofd.:ctwlhL·rmalatomi.zCf<;.set'K.\\'.1J.cLon
ElectrmhcmUl!AromizmioflforAnalyticalAromicSpeCfrome[rv.Nt\.\,,'
York:Wiley.J999:A.Varma.eReHmllfhullkofFllrnal""Atm;licAh-
51!fpriofJSpcctroscOpLBocaRat,)n.FL:eRePr..:ss.19SY:D.JBUlch<.:r
and1.Sn.:'ddon,APraCtlc,l!(~'uidcIIIGraphileFU rrlilli'AtomicAhsorpfi()!l
S"earon/cay,NtC'wYork:\Viky.\yqS:K.\Vbckson.AnlllChem.,2000.
71.159.andprcYiousrcyi<.:\.\sintheseries
.~~~~
~/~~
(b) (cj
FIGURE9-6(a)Cross-sectionalviewofagraphite
furnacewithintegratedL'vovplatform.(b)Longitudinal
configurationofthegraphitefurnace.Notethetempera-
tureprofileshowninbluealongthepathofthefurnace.In
thelongitudinalconfiguration,thetemperaturevaries
continuouslyalongtbepath,reachingamaximumatthe
center.(c)Transverseconfigurationofthefurnace.The
temperatureprofileisrelativelyconstantalongthepath.
(CourtesyofPerkin-ElmerLifeandAnalyticalSciences,
Shelton,CT.)
nolongerdirectlyonthefurnacewall.Asaresult,at-
omizationoccursinanenvironmcntinwhichthetem-
peratureisnotchangingsorapidly,whichimprovesthe
reproducibilityofanalyticalsignals.
Figure9·6bandcshowsthetwowaysofheatingthe
graphitefurnacewhileitisheldintheopticalpath.Tra-
ditionally,thefurnacewasheatedinthelongitudinal
modeillustratedinFigure9-6b,whichprovidesacon-
tinuouslyvaryingtemperatureprofileasshowninthe
figure.Thetransversemode,showninFigure9-6e,gives
auniformtemperatureprolilealongtheentirelengthof
thetube.Thisarrangementprovidesoptimumcondi.
tionsfortheformationoffreeatomsthroughoutthe
tube.Recombinationofatomstomolecules,atomloss,
andcondensationonthecoolertubeendsexhibitedin
thelongitudinalmodearethusminimizedinthetrans-
verseheatingmode.
Experimentsshowthatreducingthenaturalporos·
itvofthegraphitetubeminimizessomesamplema-
t;ixeffectsandpoorreproducibilityassociatedwith
graphitefurnaceatomization.Duringatomization,
partoftheanalyteandmatrixapparentlydlfluseInto
thesurfaceofthetube,whichslowstheatomIzatIOn
process,thusgivingsmalleranalytesignals.Toover·
comethiseffect,mostgraphitesurfacesarecoatedWIth
athinlayerofpyrolyticcarbon,whichsealsthepores
ofthegraphitetube.Pyrolyticgraphiteisatypeof
graphitethatisdepositedlayerbylayerfromahighly
homogeneousenvironment.Itisformedbypassinga
mixtureofaninertgasandahydrocarbonsuchas
methanethroughthetubewhileitisheldatanelevated
temperature.
OutputSignal
Atawavelengthatwhichabsorbanceorfluorescence
occurs,thetransduceroutputrisestoamaximumafter
afewsecondsofignitionfollowedbyarapiddecay
backtozeroastheatomizationproductsescapeinto
thesurroundings.Thechangeisrapidenough(often
<1s)torequireamoderatelyfastdata-acquisitionsys-
tem.Quantitativedeterminationsareusuallybasedon
peakheight,althoughpeakareaisalsoused.
Figure9-7showstypicaloutputsignalsfroman
atomicabsorptionspectrophotometerequippedwith
anelectrothermalatomizer.Theseriesof{ourpeaks
ontherightshowtheabsorbanceatthewavelengthof
aleadpeakasafunctionoftimewhena2-flLsample
ofcannedorangejuicewasatomized,Duringdrying
andashing,threepeaksappearthatareprobablydue
toAlolecularevaporationproductsandparticulate
ignitio'nproducts.Thethreepeaksontheleftare{or
lead"standardsusedforcalibration.Thesamplepeak
onthefarrightindicatesaleadconcentrationofabout
0.05flg/mLofjuice.
PerformanceCharacteristics
ofElectrothermalAtomizers
Electrothermalatomizersoffertheadvantageofun-
usuallyhighsensitivityforsmallvolumesofsample.
Typically,samplevolumesbetween0.5and10flLarc
used;underthesecircumstances,absolutedetection
limitstypicallylieintherangeof10-
10
tolO-lJgof
analyte
3
Therelativeprecisionofelectrothermalmethodsis
generallyintherangeof5%to10%comparedwiththe
0.8
0.7
0.6
0.5
~
'"c
~..,
0.4
~..,
«
0.3
0.2
0.1
FIGURE 9-7Typicaloutputforthedeterminationoflead
fromaspectrophotometerequippedwithanelectro-
thermalatomizer.Thesamplewas2flLofcannedorange
juice.Thetimesfordryingandashingare20and60s,
respectively,(CourtesyofVarianInstrumentDivision,Palo
Alto,CA.)
1%orbetterthatcanbeexpectedforflameorplasma
atomization,Furthermore,becauseoftheheating-
coolingcycles,furnacemethodsareslow-typicallyre-
quiringseveralminutesperelement.Afinaldisadvan-
tageisthattheanalyticalrangeisrelativelynarrow,
usuallylessthantwoordersofmagnitude.Asaresult,
electrothermalatomizationisthemethodofchoice
whenflameorplasmaatomizationprovidesinadequate
detectionlimits.
AnalysisofSolidswithElectrothermalAtomizers
Inmostmethodsbasedonelectrothermalatomizers,
samplesareintroducedassolutions.Severalreports,
however,havedescribedtheuseofthistypeofatom-
izerforthedirectanalysisofsolidsamples.Onewayof
performingsuchmeasurementsistoweighthefinely
groundsampleintoagraphiteboatandinserttheboat
intothefurnacemanually.Asecondwayistoprepare
aslurryofthepowderedsamplebyultrasonicagita-
tioninanaqueousmedium.Theslurryisthenpipelted
intothefurnaceforatomization.'

4. }
LldH
------ ~h
-\.noJ\,.'P
j
FIGURE9-8(a)Crosssectionofacellforglow-discharge
atomizationofsolidsamples.(b)Cratersformedonsample
surfacebysixjetsofionizedargon.(TeledyneLeeman
Labs,Hudson,NH.)
Hvfar,th~mostcommon sample-introductionand
atomizationtechniquesforatomicabsorptionanaly-
s~sareflamesorelectrothermalvaporizers.Several
oth~ratomizationmethodstindoccasionaluse,how-
el·er.Threeofthesearcdescribedbri~flvinthissection.
Glow-DischargeAtomization
,\sdescribedinSection8C-2,aglow-dischargcdevice
producesanatomizedvaporthatcanbesweptintoacell
torabsorptionmeasurements. Figure9-8ashowsa
glow-dischargecellthatcanbeusedasanaccessory
tomostflameatomicabsorptionspectrometers.It
consistsofacylindricalcellabout17emlongwithacir-
cularholeabout2emindiametercutnearthemiddl~of
thecvlinder.AnO-ringsurroundsthehole.Th~sample
ispressedagainstthisholewithatorquescrewsothatit
sealsthetube.Sixfinestreamsofargongasfromtiny
no/zksarrangedinacircularpatternabovethesampk
impinge011the:-;ampksurfaceinahexagonalpattern.
The'argonisionizedbyacurrentbetweenananode
supportingthenozzlesandth~sample.whichactsasa
cathode.,\sar~sultofsputtering.sixcrat~rsquicklv
vaporpressureatambienttemperature.'Thedetermi-
nationofmerCUri'invarioustypesofsamplesisofvital
importancecurrentlybecauseofthetoxicityoforganic
mercurycompoundsandtheirwidespreaddistribution
intheenvironment.'Onepopularmethodforthisde-
terminationiscoldvaporizationfollowedbyatomic
absorptionspectrophotometry.Toperformadeter-
minationofthistype,mercuryisconvertedtoHg>by
treatmentofsampleswithanoxidizingmixtureofnitric
andsulfuricacids.followedbyreductionoftheHg"to
themetalwithSnCl,.Theelementalmercuryisthen
sweptintoalong-passabsorptiontubesimilartothe
oneshowninFigure9-9bybubblingastreamofinert
gasthroughthereactionmixture
9
Thedeterminationis
completedbymeasuringtheabsorbanceat253.7nO'.
Detectionlimitsintheparts-per-billionrangeare
achieved.Severalmanufacturersofferautomaticin-
strumentsforperformingthisdetermination.
FIGURE9-9Ahydridegenerationandatomization
systemforMS.
98ATOMICABSORPTION
INSTRUMENTATION
formonthesamplesurfac~asshowninFigure9-8h.The
sputteredatomsaredraWl!byavacuumtotheaxis
ofthecellwheretheyabsorbradiatiol!fromthespec-
trometersource.s
Forthistechniquetobeapplicable.thesamplemust
beanelectricalconductorormustbemadeintoapel-
letwithapowderedconductorsuchasfinelyground
graphiteorcopper.Solutionsampleshavealsoheen
analyzedhydepositionOl!agraphite,aluminum,or
coppercathode.Detectionlimitswiththistypeofde-
vicearereportedtoheinthelowparts-per-million
rangeforsolidsamples'"
InstrumentsforAASarcsimilaringeneraldesignto
thatshowninFigure7-1aandconsistofaradiation
source,asampleholder.awavelengthselector.adetec-
tor,andasignalprocessorandreadout
1O
Thesample
holderinatomicabsorptioninstrumentsistheatomizer
cellthatcontainsthegaseousatomizedsample.
98-1RadiationSources
Atomicahsorptionmethodsarepotentiallyhighly
specific;becauseatomicabsorptionlinesarcremark-
ablyllarrow(0.fXJ2to0.005nm)andbecauseelectronic
transitionenergiesareuniqueforeachelement.On
theotherhand,narrowlinewidthscreateaprohlem
thatdoesnotnormallyoccurinmolecularabsorption
HydrideAtomization
InSection8C-I.weconsideredmethodsforintroduc-
ingsolutionsampleshyhydridegeneration.Atomiza-
tionofthchydridesrequiresonlythattheybeheated
inaquartztube.asshowninFigure9-9.
'SeeL.H.J.LajunenandP.Peramakl.Specrrocht'fnicalAnillysuh}
Atomi,AbsorpliunandEmi.\Jion.2nded..p.lJJ,C<:Imhridg<.:Roy;.}1Sl lC i-
etvufChenllstfv,20(j4
kS'ec,forexample,D.A.Skoog.D.M.West.F1.Holler.andSR.Crouch,
FundllmenlaL~ofAnlll.l'ticalChcmislrr.Sthed.pp_86:'\-67,Bc:lrnonl,C--\."
Brooks/Cole,20ll-1.
~Foradiscussionoftheimp\lrtance\)fdeterminingmercury111thenn-i-
ronmcnl.refertothl'InstrumentalAnalysis111Aclln!lkatml'at,hI','nJ
ofSectIOn2
IORefcren.:ehooks1111;ltOO1I(ahsorpli,lOspcClfO<;COrVindudcL,Ii,J
LajuncnandP.Peramalo,Spn'lrtJchenliculAflillulsIIIAtoml<Absorption
andt:mi.uion.2ndtd..Camhndge.UK:RoyalSncidyll fChemistry.20114:
\f.Sperling.andBWdz.AI"miL'AhxurprillrlSPl:'crrll!1lr'tr\,;rdeJ."<:\1,
York:VCH.1999.
Cold-VaporAtomization
Thecold-vaportechniqueisanatomizationmdhodap-
plicableonlytothedeterminationofmercurvbecause
itistheonlymetallicdementthathasanappreciahle
'SeeE.II.Plepmeit'rinChili'DischargE'Spl'dro.I'Co{JII'S,pr61f-71.R.K.
\farcu~.cd.;..JewYurk:PlenumPre~s.19'-11
'Furafl:\-ieWtllpuls<.'Jglow-Jls..:;hargespe,:llllsCOr",St'C\\W,Harris,fn.
eYdllg.andF.Oxk\.An,ll('hem.200l.~.1.--psO.-- \
spectroscopy.InSection13B-2.weshowthatalin-
earrelationshipbetweentheanalyticalsignal(ab-
sorbance)andconcentration-thatis,forBeer'slaw
asgil"enhI"Equation6-3.•tobeobeyed-requiresa
narrowsourcebandwidthrelativetothewidthofan
ahsorptionlineorband.Evengood-qualitlmonochro-
mators.howcver,hal"eeffectivebandwidthssignifi-
cantlygreaterthanthewidthofatomicabsorption
lines.Asaresult.nonlinearcalibrationcurl"esarein-
evitablewhenatomicabsorbancemeasurements arc
madewithanordinaryspectrophotometer equipped
withacontinuumradiationsource.Furthermore.the
slopesofcalihrationcurvesohtainedintheseexperi-
mentsarcsmallbecauseonlyasmallfractionofthe
radiationfromthemonochromator slitisabsorbedby
thesample;theresultispoorsensitil"ity.Inrecent
Years,thedevelopmentofhigh-resolution(R>LO')
~ontinuum-source spectrometersbasedonthedouble
echellemonochromator coupledwitharraydetection
hascloudedthisissue,andsuchinstrumentsarebe-
ginningtocompetewithtraditionalspectrometers
equippedwithlinesources."
Theproblemcreatedbythelimitedwidthofatomic
absorptionlineshasbeensoll"edbytheuseofline
sourceswithbandwidthsevennarrowerthantheab-
sorptionlinewidth.Forexample,tousethe589.6-nm
lineofsodiumasthebasisfordeterminingtheele-
ment,asodiumemissionlineatthissamewavelength
isisolatedtoserveasthesource,Inthisinstance.a
sodiumI"aporlampinwhichsodiumatomsareexcited
byanelectricaldischargemayheusedtoproducethe
line.Theothersodiumlinesemittedfromthesource
areremovedwithfiltersorwitharelativelyinexpen-
sivemonochromator. Operatingconditionsforthe
sourcearechosensuchthatDopplerhroadeningofthe
emittedlinesislessthanthebroadeningoftheabsorp-
tionlinethatoccursintheflameorotheratomizer.
Thatis.thesourcetemperatureandpressurearckept
belo\\'thatoftheatomizer.Figure9-10illustratesthe
principleofthisprocedure.Figure9-lOashowsthe
emissionspectrumofatl"picalatomiclampsource,
whichconsistsoffournarrowlines.Withasuitahlefil-
terormonochromator. allhutoneoftheselinesarere-
moved.Figure9-10hshowstheabsorptionspectrum
fortheanalytehct\\"L.'cn\\'a\"ekngth~AIandA:.Note
B\\d/.11.B~ d,crRlh ,~.S,Flu[,:k.anJ\.H<.:Jlllldllll.f!Jgh-R,\olwi')IJ
C"rWflWWISource..lAYHohokL·ll.l'iJ:\\'l1cv-\·('H.2(n-,:HHC( --k<. 'l
R(lv;.S.F1,lrd;.RI'lsdlcllddrf.(iHSChlllL' cht'L!-r,'\(lliulI.·In,ll(·ft,'II)
199(1,15.;,'1~)

5. Monochromator
bandwidth
I I
fa)
Emission
spectrumof
source
(h)
Sample
absorption
spectrum
p
.4~log-J}
A
(e)
Emissionspectrum
afterpassage
throughsample
andmonochromator
A,
Wavelength
thatthebandwidthissignificantlygreaterthanthat
oftheemissionline.AsshowninFigure9-1Oc,passage
ofthelinefromthesourcethroughthe/lamereduces
itsintensityfromPotoP;theabsorbanceisthengiven
by10g(PoIP),whichislinearlyrelatedtotheconcen-
trationoftheanalyteinthesample.
Adisadvantageoftheprocedurejustdescribedis
thataseparatesourcelampisneededforeachelement
(orsometimesgroupofelements).
Hollow-CathodeLamps
Themostcommonsourceforatomicabsorptionmea-
surementsisthehollow-cathodelamp,suchastheone
showninFigure9-11.12Thistypeoflampconsistsofa
tungstenanodeandacylindricalcathodesealedina
glasstubefilledwithneonorargonatapressureofIto
5torr.Thecathodeisconstructedofthemetalwhose
spectrumisdesiredorservestosupportalayerofthat
metal.
Ionizationoftheinertgasoccurswhenapotential
differenceontheorderof300Visappliedacrossthe
electrodes,whichgeneratesacurrentofabout5to
15mAasionsandelectronsmigratetotheelectrodes.
Ifthevoltageissufficientlylarge,thegaseouscations
acquireenoughkineticenergytodislodgesomeofthe
metalatomsfromthecathodesurfaceandproducean
atomiccloudinaprocesscalledsputtering.Aportion
ofthesputteredmetalatomsareinexcitedstatesand
thusemittheircharacteristicradiationastheyreturn
tothegroundstate.Eventually,themetalatomsdif-
fusebacktothecathodesurfaceortotheglasswallsof
thetubeandareredeposited.
Thecylindricalconfigurationofthecathodetends
toconcentratetheradiationinalimitedregionofthe
metaltube;thisdesignalsoenhancestheprobability
thatredepositionwilloceuratthecathoderatherthan
ontheglasswalls.
Theefficieneyofthehollow-cathodelampdepends
onitsgeometryandtheoperatingvoltage.Highvolt-
ages,andthushighcurrents,leadtogreaterintensities.
Thisadvantageisoffsetsomewhatbyaninereasein
Dopplerbroadeningoftheemissionlinesfromthe
lamp.Furthermore,thegreatercurrentsprodueean
increasednumberofunexcitedatomsinthecloud.The
unexcitedatoms,inturn,areeapableofabsorbing
theradiationemittedbytheexcitedones.Thisself-
absorptionleadstoloweredintensities,particularlyat
thecenteroftheemissionband.
Hollow-cathodelampsareoftenusedassourcesin
AFS,asdiscussedinSection9E-1.Inthisapplication,
thelampsarepulsedwithadutycycleof1%to10%
andpeakcurrentof0.1to1A,whichincreasestheir
peakradiancebyafactorof10to100relativetothe
steady-stateradianceofdcoperation.
13
Avarietyofhollow-cathodelampsisavailablecom-
mercially.Thecathodesofsomeconsistofamixtureof
severalmetals;suchlampspermitthedeterminationof
morethanasingleelement.
ElectrodelessDischargeLamps
Electrodelessdischargelamps(EDLs)areuseful
sourcesofatomiclinespectraandprovideradiantin-
tensitiesusuallyonetotwoordersofmagnitudegreater
thanhollow-cathodelamps."Atypicallampiscon-
structedfromasealedquartztubecontainingafewtorr
ofaninertgassuehasargonandasmallquantityofthe
metal(oritssalt)whosespeetrumisofinterest.The
lampeontainsnoelectrodebutinsteadisenergizedby
anintensefieldofradio-frequencyormicrowaveradIa-
tion.Ionizationoftheargonoecurstogiveionsthatare
acceleratedbythehigh-frequencycomponentofthe
fielduntiltheygainsufficientenergytoexcitetheatoms
ofthemetalwhosespectrumissought.
EDLsareavailablecommerciallyforfifteenor
moreelements.Theirperformanceisnotasreliableas
thatofthehollow-cathodelamp,butforelementssuch
asSe,As,Cd,andSb,EDLsexhibitbetterdetection
limitsthandohollow-cathodelamps."Thisoccursbe-
causeEDLsfortheseelementsaremoreintensethan
thecorrespondinghollow-cathodelamps,andthus,
EDLsarequiteusefulindeterminingtheseelements.
Figure9-12isaschematicofacommercialEDL,which
ispoweredbva27-MHzradio-frequencysource.
Hollo\\"
Anode\ /cathode
1JbW~~~"rn
Glass Nco~Ar orPyrex
window
:i1.D.Inuit.:Jr.andS.R.Crouch.Spl'crfOchamwlAflahsis.p~to,Engle-
woodCliffs.:"JJ:Pr~ntic~HaiL1CJSX
I:S~CW.B.Barndt.1.W.,'u llmeLJ.nJS1\1.DcNuLW.A[Ahsorptiofl
Vew5lell"1976.15.33
"r:.Da\~nport.Am!'!"LllhSoo.}t:t99.31151.1\J1S
FIGURE9-11Schematiccrosssectionofahollow-
cathodelamp.
FIGURE9-12CutawayofanEDL.(FromW.B.Barnett,
J.W.Vollmer,andS.M.DeNuzzo,At.Absorption
Newsletter,1976,15,33,Withpermission.)
SourceModulation
Inthetypicalatomicabsorptioninstrument,itisneces-
sarytoeliminateinterferencescauscdbyemissionof
radiationbythe/lame,Muchofthisemittedradiation
is,ofcourse,removedbythemonochromator.Never-
theless,emittedradiationcorrespondinginwavelength
tothemonochromatorsettingisinevitablypresentin
thel1amebecauseofexcitationandemissionofanalyte
atomsand/lamegasspecies,Toeliminatetheeffectsof
l1ameemission,itisnecessarytomodulatetheoutputof
thesourcesothatitsintensity/luctuatesataconstant
frequency.Thedetectorthenreceivestwotypesofsig-
nal,analternatingonefromthesourceandacontinu-
ousonefromthe/lame.Thesesignalsareconvertedto
thecorrespondingtypesofelectricalresponse.A
simplehigh-passRCfilter(Section28-5)canthenbe
usedtoremovetheunmodulateddesignalandpassthe
acsignalforamplification.
Asimpleandentirelysatisfactorywayofmodulat-
ingthcemissionfromthesourceistointerposeacir-
cularmetaldisk,orchopper,inthebeambetweenthe
sourceandthe/lame.Rotatingdiskandrotatingvane
choppersarecommon(seeFigure5-7aandb),Rota-
tionofthediskorvaneataconstantknownratepro-
videsabeamthatischoppedtothedesiredfrequency.
Othertypesofelectromechanicalmodulatorsinclude
tuningforkswithvanesattachedtoalternatelyblock
andtransmitthebeam(seeFigure5-7c)anddevices
thatrotateavanethroughafixedarctoperformthe
samefunction."Asanotheralternative,thepower
supplyforthesourcecanbedesignedforintermittent
oracoperationsothatthesourceisswitchedonand
offatthedesiredconstantfrequency.

6. Instrumentsforatomicabsorptionmeasurementsare
offeredbynumerousmanufacturers;bothsingle-and
double-beamdesignsareavailable.Therang~ofso-
phisticationandcost(upwardofafewthousanddol-
lars)issubstantial.
Ingeneral,theinstrumentmustbecapableofpro-
vidingasufficientlynarrowbandwidthtoisolatethe
linechosenforthemeasurementfromotherlinesthat
mayinterferewithordiminishthesensitivityofthede-
termination.Aglassfiltersufficesforsomeofthealkali
metals,whichhaveonlyafewwidelyspacedresonance
linesinthevisibleregion.Aninstrumentequippedwith
easilyinterchangeableinterferencefiltersisavailable
commercially.Aseparatefilterandlightsourceare
usedforeachelement.Satisfactoryresultsforthede-
V1Exercise:Learnmoreaboutsingle-beamand
lQ.Jdouble-beamspectrophotometers.
terminationoftwenty-twometalsareclaimed.Most
instruments,however,incorporategood-qualityultra-
violet-visiblemonochromators,manvofwhichare
capableofachievingabandwidthoniheorderof1A.
Mostatomicabsorptioninstrumentsusephotomul-
tipliertubes,whichweredescribedinSection7E-2,as
transducers.Aspointedoutearlier,electronicsystems
thatarecapableofdiscriminatingbetweenthemodu-
latedsignalfromthesourceandthecontinuoussignal
fromtheflamearerequired.Mostinstrumentscur-
rentlyonthemarketareequippedwithcompulersys-
temsthatareusedtocontrolinstrumentparameters
andtocontrolandmanipulatedata,
pulsedpowersupply,anatomizer,andasimplegrating
spectrophotometerwithaphotomultipliertransducer.
Itisusedinthemannerdescribedonpage159.Thus,
thedarkcurrentisnulledwithashutlerinfrontofthe
transducer.The100%transmittanceadjustmentis
thenmadewhileablankisaspiratedintotheflameor
ignitedinanonflameatomizer.Finally,thetransmit-
tanceisobtainedwiththesamplereplacingtheblank.
Double-BeamInstruments
Figure9-13bisaschematicofatypicaldouble-beam-in-
timeinstrument.Thebeamfromthehollow-cathode
sourceissplitbyamirroredchopper,onehalfpassing
throughtheflameandtheotherhalfaroundit.Thetwo
beamsarethenrecombinedbyahalf-silveredmirror
andpassedintoaezerny-Turnergratingmonochroma-
tor;aphotomultipliertubeservesasthetransducer.
Theoutputfromthelatteristheinputtoalock-inam-
plifierthatissynchronizedwiththechopperdrive.The
ratiobetweenthereferenceandsamplesignalisthen
amplifiedandfedtothereadout,whichmaybeadigital
meteroracomputer.
Itshouldbenotedthatthereferencebeaminatomic
double-beaminstrumentsdoesnotpassthroughthe
flameandthusdoesnotcorrectforlossofradiantpower
duetoabsorptionorscatteringbytheflameitself.
Methodsofcorrectingfortheselossesarediscussedin
thenextsection.
Single-BeamInstruments
Atypicalsingle-beaminstrument,suchasthatshown
inFigure9-13a,consistsofseveralhollow-cathode
sources(onlyoneofwhichisshown),achopperora
9CINTERFERENCES INATOMIC
ABSORPTION SPECTROSCOPY
/lV
\I
\!
~
Grating
Intelferencesoftwotypesareencounteredinatomic
absorptionmethods.Spectralinterferencesarisewhen
the<!bsorptionoremissionofaninterIcringspeciesei-
theroverlapsorliessoclosetotheanalyteabsorption
oremissionthatresolutionbythemonochromatorbe-
comesimpossible.Chemicalinterferencesresultfrom
variouschemicalprocessesoccurringduringatomi-
zationthataltertheabsorptioncharacteristicsofthe
analyte.
Flame
~~""~I-
r'0..L...--.Yp••..---7
LampChopp~r~------:'~ _
Becausetheemissionlinesofhollow-cathodesources
aresoverynarrow,interferencebecauseofoverlap-
pinglinesisrare.Forsuchaninterferencetooccur.the
separationbetweenthetwolineswouldhavetobeless
thanabout0.1A.Forexample.avanadiumlineat
30R2.11Ainterferesinthedeterminationofaluminum
$
//Half-silvered
.'_.. mirror
r..firror-/_- Opell
FIGURE9-13Typicalflamespectrophotometers:(alsingle-beamdesignand
(b)double-beamdesign.
basedonitsabsorptionlineat3082.15A.Theinterfer-
enceiseasilyavoided,however,byobservingthealu-
minumlineat3092.7Ainstead.
Spectralinterferencesalsoresultfromthepresence
ofcombustionproductsthatexhibitbroadbandabsorp-
tionorparticulateproductsthatscatterradiation.Both
reducethepowerofthetransmittedbeamandleadto
positiveanalyticalerrors.Whenthesourceofthese
productsisthefuelandoxidantmixturealone,theana-
lyticaldatacanbecorrectedbymakingabsorption
measurementswhileablankisaspiratedintotheflame.
Notethatthiscorrectionmustbeusedwithboth
double-beamandsingle-beaminstrumentsbecausethe
referencebeamofadouble-beaminstrumentdoesnot
passthroughtheflame(seeFigure9-13b).
Amuchmoretroublesomeproblemoccurswhen
thesourceofabsorptionorscatteringoriginatesinthe
samplematrix.Inthisinstance,thepowerofthetrans-
mittedbeamPisreducedbythematrixcomponents,
buttheincidentbeampowerPoisnot;apositiveerror
inabsorbanceandthusconcentrationresults.Anex-
ampleofapotentialmatrixinterferencebecauseofab-
sorptionoccursinthedeterminationofbariumin
alkaline-earthmixtures.Asshownbythesolidline
inFigure8-8,thewavelengthofthebariumlineused
foratomicabsorptionanalysisappearsinthecenterof
abroadabsorptionbandforCaOH.Wetherefore
anticipatethatcalciumwillinterfereinbariumdeter-
minations,buttheeffectiseasilyeliminatedbysubsti-
tutingnitrousoxideforairastheoxidant.Thehigher
temperatureofthenitrousoxideflamedecomposes
theCaOHandeliminatestheabsorptionband.
Spectralinterferencebecauseofscatteringbyprod-
uctsofatomizationismostoftenencounteredwhen
concentratedsolutionscontainingelementssuchasTi,
Zr,andW-whichformrefractoryoxides-arcaspi-
ratedintotheflame.Metaloxideparticleswithdiame-
tersgreaterthanthewavelengthoflightappeartobe
formed,andscatteringoftheincidentbeamresults.
Interferencecausedbyscatteringmayalsobea
problemwhenthesamplecontainsorganicspeciesor
whenorganicsolventsareusedtodissolvethesample.
Here,incompletecombustionoftheorganicmatrix
leavescarbonaceousparticlesthatarecapableofscat-
teringlight.
Fortunatelv,withflameatomization,spectralinter-
ferencesbymatrixproductsarenotwidelyencoun-
teredandoftencanbeavoidedbyvariationsinthe
analyticalvariables,suchasflametemperatureand
fuel-to-oxidantratio.Alternatively.ifthesourceof

7. interferenceisknown,anexcessoftheinterfering
substancecanbeaddedtobothsampleandstandards.
Providedtheexcessaddedtothestandardsample
islargewithrespecttotheconcentrationfromthe
samplematrix,thecontributionfromthesamplema-
trixwillbecomeinsignificant.Theaddedsubstanceis
sometimescalledaradiationbuffer.Themethodof
standardadditionscanalsobeusedadvantageouslyin
somecases.
Intheearlydaysofelectrothermalatomization,
matrixinterferenceproblemsweresevcre.Asplatform
technology,newhigh-qualitygraphitematerials,fast
photometricinstrumentation,andZeemanback-
groundcorrectionhavcdeveloped,matrixinterferencc
problemshavedecreasedtothelevelencounteredwith
flames
17
Severalmethodshavebeendevelopedforcor-
rectingforspectralintcrferencescausedbymatrix
products.L8
TheTwo-LineCorrectionMethod
Thetwo-linecorrectionprocedureusesalincfromthe
sourceasarcference,Thislincshouldlieascloseas
possibletotheanalytelinebutmustnotbeabsorbed
bytheanalyte.Ifthcseconditionsaremet,itisas-
sumedthatanydecreaseinpowerofthereferenceline
fromthatobservedduringcalibrationarisesfromab-
sorptionorscattcringbythematrixproductsofthe
sample,Thisdecreaseinpoweristhenusedtocorrect
theabsorbanceoftheanalytelinc,
Thereferencelinemaybefromanimpurityinthe
hollowcathode,aneonorargonlincfromthegascon-
tainedinthelamp,oranonresonantcmissionlineof
theelementthatishcingdetermined,Unfortunatelv,a
suitablereferencelineisoftcnnotavailable, .
TheContinuum-SourceCorrectionMethod
Figurc9-14illustratesasecondmethodforback-
groundcorrectionsthatiswidelvused,Inthistech-
nique,adeuteriumlampprovide~asourccofcontin-
uumradiationthroughouttheultravioletregion,The
configurationofthechopperissuchthatradiation
fromthecontinuumsourceandthehollow-cathode
lamparcpasscdalternatelythroughtheelectrother-
malatomizer.Thcabsorbanceofthedeuteriumradia-
tionisthensubtractedfromthatoftheanalytehearn,
!'Scc\V.Sla\'in,Anal.Cht:m.1986.58.59VA.
I~Foracomparisonufthe\'ariousfilt:thodsf'l[hack£.founJcorrection,$..:e
D.JButcherandJSneddu[]..4PracticalCillitit:tnGraphj[1!Furnace
Ar,jmlcAhsorptiOflSpcctromt'lrr.Nev.York:\\'1Iey.II.J4Kpr.S~-89
r-r-JDeuterium
~ lamp
!
Analytehollow·
c~r:Rotaling
chopper
_~,,"--Tomono~
chromator
Electrothermal
,~ ~ C ~agnel n F=r1
~fi' szrq//~~~C==18
~-:"Vy/ /1 /
~/ //}////>/ //
D11/\/ /'1))'-1n_n_/
j\J;;(J+ EY J.LU,;U.LUL/
(J / \ F
Background Backgroundplus
only atomicabsorptionFIGURE9-14Schematicofacontinuum-sourceback-
groundcorrectionsystem,Notethatthechoppercarbe
eliminatedbyalternatelypulsingeachlamp,
FIGURE9-15Schematicofanelectrothermalatomicabsorptioninstrumentthatprovides
abackgroundcorrectionbasedontheZeemaneffect.(CourtesyofHitachiScientific
Instruments,MountainView,CA,)
Theslitwidthiskeptsufficientlywidesothatthefrac-
tionofthecontinuumsourcethatisabsorbedbythe
atomsofthcsamplcisnegligible,Therefore,thcatten-
uationofthecontinuumsourceasitpassesthroughthe
atomizedsamplereflectsonlythebroadbandabsorp-
tionorscattcringhythesamplematrixcomponents,
Abackgroundcorrectionisthusachieved,
Unfortunately,althoughmostinstrumentmanufac-
turersoffercontinuum-sourcebackgroundcorrection
systems,theperformanceofthcsedevicesisoftenless
thanideal,whichleadstoundercorrectioninsomesys-
temsandovercorrectioninothers,Oneofthesources
oferroristheinevitabledegradationofsignal-to-noise
ratiothataccompaniestheadditionofthecorrection
system,Anotheristhathotgaseousmediaareusually
highlyinhomogeneousbothinchemicalcomposition
andinparticulatedistribution;thusifthctwolamps
arenotinperfectalignment,anerroneouscorrection
willresultthatcancauseeitherapositiveoranegative
error.Finally,theradiantoutputofthedeuteriumlamp
inthcvisibleregionislowenoughtoprecludetheuseof
thiscorrectionprocedureforwavelcngthslongerthan
about350nm,
wereformed,Thisphenomenon,whichistermedthe
Zeemaneffect,i9isgeneralforallatomicspectra,Sev-
eralsplittingpatternsarisedependingonthetypeof
electronictransitionthatisinvolvedintheabsorption
process.Thesimplestsplittingpattern,whichisob-
servedwithsinglet(SectionSA-I)transitions,leads
toacentral,or7f,lineandtwoequallyspacedsatellite
alines,Thecentralline,whichisattheoriginalwave-
length,hasanabsorbancethatistwicethatofeach
aline,Formorecomplextransitions,furthersplitting
ofthe7fandalinesoccurs,
ApplicationoftheZcemaneffecttoatomicabsorp-
tioninstrumentsisbasedonthediffcringresponseof
thetwotypesofabsorptionlinestopolarizedradia-
tion,The7flineabsorbsonlythatradiationthatis
plane\polarizedinadirectionparalleltotheexternal
magneticfield;thcalines,incontrast,absorbonlyra-
diationpolarizedat90"tothefield,
Figure9-15showsdetailsofanelectrothermal
atomicabsorptioninstrument,whichuscstheZeeman
effectforbackgroundcorrection,Unpolarizedradiation
fromanordinaryhollow-cathodesourceAispasscd
througharotatingpolarizerB.whichseparatesthe
beamintotwocomponentsthatareplane-polarized
at90'tooneanotherC.Thesebeamspassintoatube-
typegraphitefurnacesimilartotheoneshowninFig-
BackgroundCorrectionBased
ontheZeemanEffect
Whenanatomicvaporisexposedtoastrongmagnetic
field(-10kG),asplittingofelectronicenergylevelsof
theatomstakesplacethatleadstoformationofseveral
absorptionlinesforeachelectronictransition,These
linesareseparatcdfromoneanotherbvabout0,01nm,
withthesumoftheabsorhancesforth~linesbeingex-
actlyequaltothatoftheoriginallinefromwhichthey
I)Foradetaileddiscussionofthe_applicatIOnoftheZeemaneffectto
atomiCahsorption,seeD.1.Butcherand1.Sneddon.APracticalGllideto
Graphltf'FurnuceAtomicAbsorptionSpectrometry.:\ewYork:Wiley,
19Yk,pro7,-8 -l:F1.Fernanda.S.A.Myers,and\\".Slavin.Anal.Chern..
1980,5':.7-lLS.D.Rr'Win.Anal.Chem.,1977.49\l-l),12fi9A
ure9-6a,ApermanentII-kGmagnetsurroundsthe
furnaceandsplitsthecnergylevelsintothethreeab-
sorptionpeaksshowninD,Notcthatthecentralpeak
absorbsonlyradiationthatisplanepolarizedwiththe
field,Duringthatpartofthecyclewhenthesourcera-
diationispolarizedsimilarly,absorptionofradiationby
theanalytetakesplace,Duringtheotherhalfcycle,no
analyteabsorptioncanoccur.Broadbandmolecular
absorptionandscatteringbythematrixproductsoccur
duringbothhalfcycles,whichleadstothecyclicalab-
sorbancepatternshowninF.Thedata-acquisitionsys-
temisprogrammedtosubtracttheabsorbanceduring
theperpendicularhalfcyclefromthatfortheparallel
halfcycle,thusgivingabackgroundcorrectedvalue,
AsecondtypeofZeemaneffectinstrumcnthas
beendesignedinwhichamagnetsurroundsthehol-
low-cathodesourcc,Here,itistheemissionspectrum
ofthesourcethatissplitratherthantheabsorption
spectrumofthesample,Thisinstrumentconfiguration
providesananalogouscorrection,Todate,mostin-
strumentsareofthetypeillustratedinFigure9-15,
Zeemaneffcctinstrumentsprovideamoreaccurate
correctionforhackgroundthanthemcthodsdescribed
earlier.Theseinstrumentsareparticularlyusefulfor
electrothermalatomizersandpermitthedirectdeter-
minationofelementsinsamplessuchasurincand
blood,Thedecompositionoforganicmaterialinthese
samplesleadstolargebackgroundcorrections(back-
groundA>I)and,asaresult.susccptibilitytosignifi-
canterror.
~ Tutorial:LearnmoreahouttheZeemaneffect.

8. BackgroundCorrectionBased
onSourceSelf-Reversal
Aremarkablysimplemeansofbackgroundcor-
rectionappearstooffermostoftheadvantagesof
aZeemaneffectinstrument.2<1Thismethod.which
issometimescalledtheSmith-Hieftjebackground
correctionmethod.isbasedontheself-reversalorself-
absorptionbehaviorofradiationemittedfromhollow-
cathodelampswhentheyareoperatedathighcur-
rents.Aswasmentionedearlier,highcurrentsproduce
largeconcentrationsofnonexcitedatoms,whichareca-
pableofabsorbingtheradiationproducedfromtheex-
citedspecies.Anadditionaleffectofhighcurrentsis
tosignificantlybroadentheemissionlineoftheex-
citedspecies.Theneteffectistoproducealinethat
hasaminimuminitscenter,whichcorrespondsex-
actlyinwavelengthtothatoftheabsorptionpeak(see
Figure9-16).
Toobtaincorrectedabsorbances,thelampis
programmedtorunalternatelyatlowandhighcur-
rents.Thetotalabsorbanceisobtainedduringthelow-
currentoperationandthebackgroundabsorbanceis
providedbymeasurementsduringthesecondpartof
thecyclewhenradiationattheabsorptionpeakisata
minimum.Thedata-acquisitionsystemthensubtracts
thebackgroundabsorbancefromthetotaltogiveacor-
rectedvalue.Recoveryofthesourcetoitslow-current
outputtakesplaceinmillisecondswhenthecurrent
isreduced.Themeasurementcyclecanberepeated
oftenenoughtogivesatisfactorysignal-to-noiseratios.
Equipmentforthistypeofcorrectionisavailablefrom
commercialsources.
Chemicalinterferencesaremorecommonthanspec-
tralinterferences.Theireffectscanfrequentlybemin-
imizedbyasuitablechoiceofoperatingconditions.
Boththeoreticalandexperimentalevidencesug-
gestthatmanyoftheprocessesoccurringinthemantle
ofaflamearcinapproximateequilibrium.Therefore.
itbecomespossibletoregardtheburnedgasesasa
solventmediumtowhichthermodynamiccalculations
canbeapplied.Theequilibriaofprincipalinterestin-
cludeformationofcompoundsoflowvolatility.disso-
ciationreactions,andionization. .
FIGURE9-16Emissionlineprofilesforahollow-cathoce
lampoperatedathighandlowcurrents.
FormationofCompoundsofLowVolatility
Perhapsthemostcommontypeofinterferenceisby
anionsthatformcompoundsoflowvolatilitywith
theanalyteandthusreducethefractionoftheanalyte
thatisatomized.Lowresultsaretheoutcome.Anex·
ampleisthedecreaseincalciumabsorbanceobserved
withincreasingconcentrationsofsulfateorphosphate.
Theseanionsformcompoundswithcalciumthatare
difficulttovolatilize.Forexample,atafixedcalcium
concentration,theabsorbancefallsoffnearlylinearly
withincreasingsulfateorphosphateconcentrations
untiltheanion-to-calciumratioisabout0.5;theab-
sorbancethenlevelsoffatabout30%to50%ofits
originalvalueandbecomesindependentofanion
concentration.
Cationinterferencehasbeenobservedaswell.For
example,aluminumcauseslowresultsinthedeter-
minationofmagnesium,apparentlyasaresultofthe
formationofaheat-stablealuminum-magnesiumcom-
pound(perhapsanoxide).
Interferencescausedbyformationofspeciesoflow
volatilitycanoftenbeeliminatedormoderatedbvuse
ofhighertemperatures.Alternatively,releasinga~ents,
whicharecationsthatreactpreferentiallywiththein-
terferantandpreventitsinteractionwiththeanalyte,
canbeused.Forexample,additionofanexcessof
strontiumorlanthanumionminimizestheinterference
ofphosphateinthedeterminationofcalcium.Thesame
twospecieshavealsobeenusedasreleasingagentsfor
thedeterminationofmagnesiuminthepresenceofalu-
minum.Inbotbinstances,thestrontiumorlanthanum
replacestheanalyteinthecompoundformedwiththe
interferingspecies.
Protectiveagentspreventinterferencebyforming
stablebutvolatilespecieswiththeanalytc.Three
commonreagentsforthispurposeareethylenedi-
aminetetraaceticacid(EDTA).8-hydroxyquinoline,
andAPCD,whichistheammoniumsaltofI-pyrro-
Iidinecarbodithioicacid.ThepresenceofEDTAhas
beenshowntoeliminatetheinterferenceofaluminum,
silicon,phosphate,andsulfateinthedeterminationof
calcium.Similarly.8-hydroxyquinolinesuppressesthe
interferenceofaluminuminthedeterminationof
calciumandmagnesium.
DissociationEquilibria
Inthehot,gaseousenvironmentofaflameorafur-
nace,numerousdissociationandassociationreactions
leadtoconversionofthemetallicconstituentstothe
elementalstate.Itseemsprobablethatatleastsomeof
thesereactionsarereversibleandcanbetreatedbythe
lawsofthermodynamics.Thus,itshouldbepossibleto
formulateequilibriasuchas
MO~M+O
M(OHh~M +20H
whereMistheanalytcatomandOHisthehydroxyl
radical.
Inpractice,notenoughisknownaboutthenature
ofthechemicalreactionsinaflametopermitaquan-
titativetreatmentsuchasthatforanaqueoussolution.
Instead,wemustrelyonempiricalobservations.
Dissociationreactionsinvolvingmetaloxidesand
hydroxidesplayanimportantpartindeterminingthe
natureoftheemissionorabsorptionspectraforanel-
em",,!.Forexample,thealkaline-earthoxidesarcrel-
ativelystable,withdissociationenergiesinexcessof
5eV.Molecularbandsarisingfromthepresenceof
metaloxidesorhydroxidesintheflamethusconstitute
aprominentfeatureoftheirspectra(seeFigure8-8).
Exceptatveryhightemperatures,thesebandsarc
moreintensethanthelinesfortheatomsorions.In
contrast,theoxidesandhydroxidesofthealkalimet-
alsaremuchmoreeasilydissociatedsothatlineinten-
sitiesfortheseelementsarehigh,evenatrelativelylow
temperatures.
Dissociationequilibriathatinvolveanionsother
thanoxygenmayalsoinfluenceflameemissionandab-
sorption.Forexample,thelineintensityforsodiumis
markedlydecreasedbythepresenceofHCI.Alikely
explanationisthemass-actioneffectontheequilibrium
ChlorineatomsformedfromtheaddedHCldecrease
theatomicsodiumconcentrationandtherebylower
thelineintensity.
Anotherexampleofthistypeofinterferencein-
volvestheenhancementofvanadiumabsorptionwhen
aluminumortitaniumispresent.Theinterferenceis
significantlymorepronouncedinfuel-richflamesthan
inleanflames.Theseeffectscanbeexplainedbyas-
sumingthatthethreemetalsinteractwithsuchspecies
as0andOH,whicharealwayspresentinflames.Ifthe
oxygen·bearingspeciesaregiventhegeneralformula
Ox,aseriesofequilibriumreactionscanbepostulated.
Thus,
VOx~V +Ox
AIOx~AI +Ox
TiOx~Ti +Ox
Infuel-richcombustionmixtures,theconcentrationof
Oxissufficientlysmallthatitsconcentrationislowered
significantlywhenaluminumortitaniumispresentin
thesample.Thisdecreasecausesthefirstequilibrium
toshifttotherightwithanaccompanyingincreasein
thevanadiumconcentrationandabsorbance.Inlean
mixtures,ontheotherhand,theconcentrationofOxis
apparentlyhighrelativetothetotalconcentrationof
metalatoms.Inthiscase,additionofaluminumorti-
taniumscarcelychangestheconcentrationofOx,and
thepositionofthefirstequilibriumisrelativelyundis-
turbed.Therefore,thepositionofthefirstequilibrium
isnotdisturbedsignificantly.
IonizationEquilibria
Ionizationofatomsandmoleculesissmallincombus-
tionmixturesthatinvolveairastheoxidant,andcan
oftenbeneglected.Inhighertemperatureflames
whereoxygenornitrousoxideservesastheoxidant.
however,ionizationbecomesimportant.andthereisa
significantconcentrationoffreeelectronsproducedby
theequilibrium
whereMrepresentsaneutralatomormoleculeand
M'isitsion.WewillfocusonequilibriainwhichMis
ametalatom.
TheequilibriumconstantKforthisreactiontakes
theform
[MlfeJ
[1\[1

9. FractionIonizedattheIndicatedPressureandTemperature
Ionization
p10-'aIm
I'=10-'atm
Element
Potential,eV 2000K
3500K
2000K
3500K
Cs
3.893
0.01
Rb
0.86
Oil
>0.994-176
0.004
0.74
K
004
>0.99
4339 0_003
066
Na
0.03
0.99
5.138
00003
0.26
0.003
Li 0.90
5.390
0.0001
018
Sa
O.OO!
0.82
5210
0.0006 0.41
OlX16
Sr 0.955_692.
0.0001
0.21
0001
Ca 0.876.111 3X10-5
0.11
0.0003
Mg
7.644 4X107
0.67
0.01
4X10'
0.09
DatafromA.L.Valleea
_ .ndR.E..Thlcrs,.tn.IreallseonAnalYllcaiChenllstr.r.IMKIlhffdPJ!
Interscience,196).Reprmtedwit hpermlSslonofJohn\\-"i1ey&Son s.Inc..,.()0an..~I\"ing,cds.,PariLVol .n,p.15C H:l,~cwYork:
Ifnoother'sourceofelectronsispresentintheflame
thisequationcanbewrittenintheform '
K=(-"'-)1'
1-"
where"isthefractionofMthatisionized,andI'isthe
partIalpressureofthemetalinthegaseoussolventbe-
foreionization_
Table9-2showsthecalculatedfractionionizedfor
severalcommonmetalsunderconditionsthatapprox-
Imatethoseusedinflameemissionspectroscopy.The
temperaturescorrespondroughlytoconditionsthat
eXIstinair-acetyleneandoxygen-acetyleneflames,
respectively.
.Itisimportanttoappreciatethattreatmentofthe
IO!llzationprocessasanequilibrium-withfreeelec-
tronsasoneoftheproducts-immediatelyimpliesthat
thedegreeofionizationofametalwillbestronglyinflu-
encedbythepresenceofotherionizablemetalsinthe
flame.Thus,ifthemediumcontainsnotonlyspeciesM
butspeciesUaswell,andifBionizesaccordingtothe
equation
thenthedegreeofionizationofMwillbedecreasedby
themass-aetlllllelfectoftheelectronsformedfromB.
Determillationofthedegreeofionizationunderthese
condItionsrequiresacalculationinvolvingthedissoci-
atIonconstantforBandthemass-balanCeexpression
FIGURE9-17Effectofpotassiumconcentrationonthe
calibrationcurveforstrontium.(Reprintedwithpermission
fromJ_A.BowmanandJ.B.Willis,Anal.Chem.,1967,39,
1220.Copyright1967AmericanChemicalSociety.)
.Atom-ionequilibriainflamescreateanumberof
Importantconsequencesinflamespectroscopy.Forex-
ample,mtensltlesofatomicemissionorabsorption
hnesforthealkahmetals,particularlypotassium,ru-
bidIUm,andcesium,areaffectedbytemperatureina
complexway.Increasedtemperaturescauseanin-
creaseinthepopulationofexcitedatoms.accordingto
theBoltzmannrelationship(Equation8-1).Counter-
actmgthiSeffect.however,isadecreaseinconcentra-
lionofatomsresultingfromionization_Thus,under
someCircumstances"dccreaseinemissionorabsorp-
tionmaybeobservcdinhotterflames.Itisforthisrea-
sonthatlowerexcitationtcmperaturesarcusuallv
specifiedforthedetcrminationofalkalimetals. .
Theeffectsofshiftsinionizationequilibriacanusu-
allybeeliminatedbyadditionofaniOfli~lltioflsuppres-
sor,whIchprOVidesarelatIvelyhighconcentrationof
electronstotheflame;suppressionofionizationofthe
analyteresults.Theeffectofasuppressorappearsinthe
cahbratlOncurvesforstrontiumshowninFigure9-17_
Notetheslglllficantincreaseinslopeofthesecurvcsas
slronli~mionizationisrepressedhytheincreasingcon-
centrationofpotassiumionsandelectrons.Notealso
theenhancedsensitivityproducedhyusingnitrousox-
Idemsteadofairastheoxidant.Thehighertempera-
tureachIevedwithnitrousoxideundoubtedlyenhances
thedegreeofdecompositionand\o!atilizailOnofthe
strontIumcompounds intheplasma.
90ATOMICABSORPTION
ANALYTICALTECHNIQUES
Thissectiondealswithsomeofthepracticaldetails
thatmustbeconsideredinflameorelectrothermal
atomicabsorptionanalysis.
Adisadvantageofflamespectroscopicmethodsisthe
reqlfir~mentthatthesamplebeintroducedintotheex-
citatio;'sourceintheformofasolution.mostcom-
moniyanaqueousone.Unfortunately,manymaterials
ofinterest,suchassoils,animaltissues.plants.petro-
leumproducts.andminer;]lsarcnotdirectlysoluhlein
commonsolvents,andcxtt:nsivepreliminarytrc:atmcnt
isoftenrequiredtoohtainasolutionoftheanalyteina
formreadyforatomization.Indeed,thedecomposition
andsolutionstepsarcoftenmoretime-consumingand
introducL:moreerrorthanthespectroscopicmeasure-
mentitself.
Decompositionofmaterialssuchasthosejustcited
usuallyrequirerigoroustreatmentofthesampleathigh
temperaturesaccompaniedhyariskoflusingthean3-
lytehyvolatilizationorasparticulatesinsmoke.Fur-
r?l.Tll[Oria.1:Learnmore<thoutatomicabsorption
l.Q.Jspectroscopy.
thcrmore.thereagentsusedindecomposingasample
oftenintroducethekindsofchemicalandspectral
interferencesthatwerediscussedearlier.Additionally.
theanalytemaybepresentinthesereagentsasanim-
purity.[nfact.unlessconsiderablecareistaken,itisnot
uncommon intraceanalysestofindthatreagentsarea
largersourceoftheanalytethanthesamples-asitua-
lionthatcanleadtoseriouserrorevenwithblank
corrections.
Someofthecommonmethodsusedfordecompos-
inganddissolvingsamplesforatomicabsorptionmeth-
odsincludetreatmentwithhotmineralacids;oxidation
withliquidreagents,suchassulfuric,nitric,orperchlo-
rieacids(wetashing);combustioninanoxygenbomb
orotherclosedcontainertoavoidlossofana[yte;ashing
atahightemperature;andhigh-temperaturefusion
withreagentssuchasboricoxide,sodiumcarhonate,
sodiumperoxide,orpotassiumpyrosulfate."
Oneoftheadvantagesofelectrothermalatomiza-
tionisthatsomematerialscanbeatomizeddirectly,
thusavoidingthesolutionstep_Forexample,liquid
samplessuchasblood,petroleumproducts,andor-
ganicsolventscanbepipelteddirectlyintothefurnace
forashingandatomization.Solidsamples,suchas
plantleaves,animaltissues,andsomeinorganicsub-
stances,canbeweigheddirectlyintocup-typeatomiz-
ersorintotantalumboatsforintroductionintotube-
typefurnaces_Calibrationis,however,usuallydifficult
andrequiresstandardsthatapproximatethesamplein
composition,
90-2SampleIntroduction
byFlowInjection
[nSection33B,wedescribethemethodsandin-
strumentationforfiowinjectionanalysis(FlA)_FlA
methodologyservesasanexcellentmeansofintro-
ducingsamplesintoaflameatomicabsorptionspec-
trometer.Alternatively,wemaythinkofanatomicab-
sorptionspectrometerasausefuldetectorforanFIA
system.Fromanyperspective.theperistalticpump
andvalvearrangementsofFlAdescribedinChapter
33arcaconvenientmeanstosampleanalytesolutions
reproducihlyandefficiently,especiallywhenitisim-
portanttoconservesample.Thecarrierstreamofthe
FIAsystemconsistingofdeionizedwaterordilute
electro[vteprovidescontinuousflushingoftheflame
'jn,Ko;'hh~'ku~inS ilfll('h 'f'rt'purd(J(JfITel·hniqllt'sillAn,ll\'llcaiChonisln
S.\litra.l'J.:\c'.••.YorK:Wiky.2u03:R.H<Kk.AlI<lndbookofIJCl"i1mr',
Illio n,\f!' lh l!d~m•.llla(I'l1<1l1Cheffmtry.N<.:wY'ork:\Viley.Ilj7Y.

10. atomizer,whichisparticularlyadvantageousfor
samplescontaininghighlevelsofsaltsorsuspended
solids.
Earlyinthedevelopmentofatomicabsorptionspec-
troscopyitwasrecognizedthatenhancedabsorbances
couldbeobtainedifthesolutionscontainedlow-
molecular-weightalcohols,esters,orketones.Theef-
fectoforganicsolventsislargelyattributabletoin-
creasednebulizerefficiency;thelowcrsurfacetension
ofsuchsolutionsresultsin-smallerdropsizesandare-
sultingincreaseintheamountofsamplethatreaches
theflame.Inaddition,morerapidsolventevaporation
mayalsocontributetotheeffect.Leanerfuel-oxidant
ratiosmustbeusedwithorganicsolventstooffsetthe
presenceoftheaddedorganicmaterial.Unfortunately,
however,theleanermixtureproduceslowerflame
temperaturesandanincreasedpotentialforchemical
interferences.
Amostimportantanalyticalapplicationoforganic
solventstoflamespectroscopyistheuseofimmiscible
solventssuchasmethylisobutylketonetoextract
chelatesofmetallicions.Theresultingextractisthen
nebulizeddirectlyintotheflame.Here,thesensitivity
isincreasednotonlybytheenhancementofabsorp-
tionlinesbecauseofthesolventbutalsobecausefor
manysystemsonlysmallvolumesoftheorganicliquid
arerequiredtoremovemetalionsquantitativelyfrom
relativelylargevolumesofaqueoussolution.Thispro-
cedurehastheaddedadvantagethatatIcastpartof
thematrixcomponentsarelikelytoremaininthe
aqueoussolvent;areductionininterferenceoftenre-
sults.Commonchelatingagentsincludeammonium
pyrrolidinedithiocarbamate,diphenyIthiocarbazone
(dithizone),8-hydroxyquinoline,andacetylacetone.
Intheory,atomicabsorptionshouldfollowBeer'slaw
(Equation6-34)withabsorbancebeingdirectlypro-
portionaltoconcentration.Unfortunately.calibration
curvesareoftennonlinear,soitiscounterproductive
toperformanatomicabsorptionanalysiswithoutex-
perimentallyconfirmingthelinearityoftheinstrument
response.Thus.acalibrationcurvethatcowrsthe
rangeofconcentrationsfoundinthesampkshouldbe
preparedperiodically.Inaddition,thenumberofun-
controlledvariablesinatomizationandabsorbance
measurementsissufficientlylargetowarrantmeasure,
mentofonestandardsolutioneachtimeananalysisis
performed.Itisevenbettertousetwostandardsthat
brackettheanalyteconcentration.Anydeviationof
thestandardfromtheoriginalcalibraiioncurvecan
thenbeusedtocorrecttheanalyticalresult.
TABLE9-3DetectionLimits(ng/mll'
forSelectedElements
-
AAS
Ele-AAS Electro-AES AES AFS
mentFlamethermal flame ICP Flame
Al 30 0.1 5 0.2 5
As 200
0.5- 2 15
CaI 0.25 0.1 (WOOl
0.4
Cd 1 om 2000 0.07 0.1
Cr 4 0.03 5 O.O~ 0.6
Cu 2 0.05 10
0.04 0.2
Fe 6 0.25 50 0.09 0.3
Hg 500 5 -- 5
Mg 0.2 0.002 5 0.003 0.3
Mn 2 0.01 - 0.01 1
Mo 5 0.5 100 0.2 8
Na 02 0.02 0.1 0.1 0.3
Ni 3
0.5 600 0.2
0.4
Pb 8 0.1 200 1 5
Sn 15 5 30n- 200
V 25 1 200 0.06 25
ZnI 0.005 5ססoo 0.1 0.1
Thestandard-additionmethod,whichwasdescribedin
SectionID-3,iswidelyusedinatomicabsorptionspec-
troscopytopartiallyorcompletelycompensateforthe
chemicalandspectralinterferencesintroducedbythe
samplematrix.
AASisasensitivemeansforthequantitativedeter-
minationofmorethansixtymetalsormetalloidele-
ments,Theresonancelinesforthenonmetallicele-
mentsaregenerallylocatedatwavelengthsshorter
than200nm,thuspreventingtheirdeterminationby
convenient,nonvacuumspectrophotometers,
From1.D.lngteJr.andS.R.Crouch,SpcClrochcmicalAllah's;s.
pr.250-51.300.321,EnglewoodCliffs,NJ:PrenticeHall.19RR.
,'lot e:PulsedhoHow-cathode-lampexcitationsourcewithfer
atomization.
a1ng/mL=10--3~g/mL=10Jppm.
AAS = =atomicabsorptionspectroscopy:AES=atomicemission
speclruscopy:AFS = =atomicfluorescencespectroscopy:ICP;=
inductivelycoupledplasma.
DetectionLimits
ColumnstwoandthreeofTable9-3presentdetection
limitsforanumberofcommonelementsbyflameand
electrothermalatomicabsorption.Forcomparison,
detectionlimitsforsomeoftheotheratomicproce-
duresarealsoincluded.Smalldifferencesamongthe
quotedvaluesarcnotsignificant.Thus,anorderof
magnitudeisprobablymeaningful,butafactorof2or
3certainlyisnot.
Formanyelements,detectionlimitsforatomicab-
sorptionspectroscopywithflameatomizationlieinthe
rangeofIto20ng/mL,or0.001to0.020ppm;forelec-
trothermalatomization,thecorrespondingfiguresare
(Ul02to0.01ng/mL.or2x10-
6
to1x10-'ppm.Ina
fewcases,detectionlimitswelloutsidetheseranges
areencountered.
9E"ATOMICFLUORESCENCE
.SPECTROSCOPY
Overthevears,significantresearchefforthasbeen
devotedt~thedevelopmentofanalyticalmethods
basedonatomicfluorescence."Thisworkhasdemon-
stratedclearlythatatomicfluorescencespectroscopy
Accuracy
Undertheusualconditions,therelati"eerrorassoci-
atedwithanameatomicabsorptionanalysisisofthe
orderofI%to2%.Withspecialprecautions,thisfig,
urecanbeloweredtoafewtenthsofapercent.Errors
encounteredwithelectrothermalatomizationusuallv
exceedthoseforflameatomizationbyafactor,;f
:itoto.
'-'-Forfurtherinformationonatomicfluorescencespectroscopy.seeLII
1.LaJuncnandP.Pcramaki,Sf'cclrIJclIl'miCll[Atwhsisb.I'AtomicAhsorp..
tionandEmission,2nded.,pp.276-85.Camhridg~:RoyalSoottv01
Ch cmis trv.20 0-1-:1-A,CBroeka.:rt.ArIah'(im[AtomicSpt'crromerr\'Il'ifh
Flamesm;dI'll1SffW5.pro290-96.We inhe im.Germany:Wiky-\'C IL:': OIJ: '::
D.1.ButcherinHandhookof!llJuumellw{Tt'chlllqllt'S(orA/lil/nil'al
Chemistr\'.FA.Settle,ed..tIpperSaddh:River.NJ.19\)7.pp,-I--ll-:'iK:
S.Greenfield.TrendsinA.na/FicalChemisT(\".1995.l-J.-1-35--1-2;JC.\'an
Loon.Anal.Chnu..1981.53,3-'2.-\.:D.Butchere : {aI.,1.Anal.A[o/11
Sl'furom..1988.3,\05Y
providesausefulandconvenientmeansforthequan-
titativedeterminationofareasonablylargenumberof
elements.Todate,however,theprocedurehasnot
foundwidespreadusebecauseoftheoverwhelming
successesofatomicemissionandespeciallyatomicab-
sorptionmethods,whichweredevelopedprior10
atomicfluorescencebymorethanadecade.Asmen-
tionedearlier,thesesuccesseshaveledtotheavail-
abilityofabsorptionandemissioninstrumentsfrom
numerouscommercialsources.[nrecentyears,anum-
berofmanufacturershaveintroducedatomicfluores-
cencespectrometersusefulfordeterminingelements
thatformvaporsandhydrides,suchasPb,Hg,Cd,Zn,
As,Sb,Bi,Ge,andSeD
Thelimiteduseofatomicfluorescencehasnotarisen
somuchfromanvinherentweaknessoftheprocedure
butratherbeca;setheadvantagesofatomicfluores-
cencehavebeensmallrelativetothewell-established
absorptionandemissionmethods.Thus,although
fluorescencemethods,particularlythosebasedonelec-
trothermalatomization,aresomewhatmoresensitive
forseveralelements,theprocedureisalsolesssensi-
tiveandappearstohaveasmallerusefulconcentra-
tionrangeforseveralothers,Furthermore,dispersive
fluorescenceinstrumentsaresomewhatmorecomplex
andmoreexpensivetopurchaseandmaintain,"These
disadvantageshavebeenlargelyovercomeinsomespe-
cial-purposededicatedinstrumentssuchastheonede-
scribedintheInstrumentalAnalysisinActionfeature
attheendofSection2.
Thecomponentsofinstrumentsforatomicfluores-
cencemeasurementsaregenerallyarrangedasshown
inFigure7-lb.Thesamplecontainerismostcom-
monl~aflamebutmavalsobeanelectrothermal
atomizationcell,aglo~discharge,oraninductively
coupledplasma,asdescribedinSectionlOA-I.Flow
cellsarcoftenusedinconjunctionwithvaporand
hydride-basedmethods.
Sources
Acontinuumsourcewouldbedesirableforatomic
fluorescencemeasurements,Unfortunately,however,
theoutputpowerofmostcontinuumsourcesovera
~;Exampksinclude:Teledyne/LeemanI.ahs(Hudson.NH)andAurora
Instruments.Ltd.t\'ancouveLBel
:lSee\V.[3,BarndtandH.L.Kahn.Alld/.Chon..197Z.-1-1.lB:'.

11. regIOnasnarrowasanatomicabsorptionlineistoolow
toprovidesufficientsensitivityforatomicfluorescence.
Intheearlyworkonatomicfluorescence,cOnY'en-
tionalhollow-cathodelampsoftenservedasexcitation
sources,Toenhancetheoutputintensitywithoutde-
stroyingthelamp,itwasnecessarytooperatethelamp
withshortpulsesofcurrentthatweregreaterthanthe
lampcouldtolerateforcontinuousoperation,The
detectorwasgatedtoobservethefluorescencesignal
onlvduringpulsesofsourceradiation,
Perhapsthemostwidelyusedsourcesforatomic
fluorescencehavebeentheEDLs(Section9B-I),
whichusuallyproduceradiantintensitiesgreaterthan
thoseofhollow-cathodelampsbyanorderofmagni-
tudeortwo.EDLshayebeenoperatedinboththecon-
tinuousandpulsedmodes,Unfortunately,thistypeof
lampisnotavailableformanyelements.
Lasers,withtheirhighintensitiesandnarrowband-
widths,wouldappeartobetheidealsourceforatomic
fluorescencemeasurements.Theirhighcostandoper-
ationalcomplexities,however,havediscouragedtheir
widespreadapplicationtoroutineatomicfluorescence
methods,
DispersiveInstruments
Adispersivesystemforatomicfluorescencemeasure-
mentsconsistsofamodulatedsource,anatomizer
(flameornonflame),amonochromatororaninterfer-
encefiltersystem,adetector,andasignalprocessor
andreadout.Withtheexceptionofthesource,mostof
thesecomponentsaresimilartothosediscussedinear-
herpartsofthischapter.
NondispersiveInstruments
Intheory,nomonochromatororfiltershouldbenec-
essaryforatomicfluorescencemeasurementswhenan
EDLorhollow-cathodelampservesastheexcitation
sourcebecausetheemittedradiationis,inprinciple,
thatofasingleelementandwillthusexciteonlyatoms
ofthatelement.Anondispersivesystemthencouldbe
madeupofonlyasource,anatomizer,andadetector.
Th.:reareseveraladvantagesofsuchasvstem:(I)sim-
plicityandlow-costinstrumentation.(2)adaptahilitv
tomultielementanalysis.(3)high-energythroughput
andthushighsensitivity,and(4)simultaneouscollec-
tionofenergyfrommultip!.:lines.whichalsoenhances
sensitivity,
Torealizetheseimportantadvantages.itisnec.:s-
sarythattheoutputofthesourcebefreeofcontami-
natinglinesfromotherelements;inaddition.the
atomizershouldemitnosignilicantbackgroundradia-
tion.Insomeinstanceswithelectrothermalatomizers,
backgroundradiationisminimal,butcertainly,itis
notwithtypicalflames.Toovercomethisproblem,fil-
ters,locatedbetweenthesourceanddetector,have
oftenbeenusedtoremovemostofthebackground
radiation.Alternatively,solar-blindphotomultipli-
ers,whichrespondonlytoradiationofwavelengths
shorterthan320nm.havebeenapplied.Forthesede-
vicestobeusedeffectively,analvteemissionmustbe
below320nm.
(g)sputtering.(h)self·absorption.(i)spectralinterference.(j)chemicalinter-
ference,(k)radiationbuffer.(I)Dopplerhroadenmg,
9.2Describetheeffectsthatareresponsihkforthethreedifferentabsorbancepro-
filesinFigure9-4andselectthrceadditionaldementsyouwouldexpecttohavc
similarproftks.
9-3Whyisanelectrothermalatomizermoresensitivethanaflameatomizcr"
9-4Describehowadeuteriumlampcanheusedtoprm'ideabackgroundcorrection
foranatomicabsorptionspectrum.
9-5Whyissourcemodulationusedinatomicabsorptionspectroscopy?
9-6ForthesamCconcentrationofnickel,theahsorbanceat352.4nmwasfoundtobe
about30%greaterforasolutionthatcontained50%ethanolthanforanaqueous
solution.Explain,
9-7Theemissionspectrumofahollow-cathodelampformolybdenumhasasharpline
at313.3nmaslongasthelampcurrentislessthan50mA.AthIghercurrents,
however,theemissionlinedevelopsacuplikecrateratItsmaximum.Explam.
9.8Ananalystattemptstodeterminestrontiumwithanatomicabsorptioninstru-.
mentequippedwithanitrousoxide-acetyleneburner,butthesensl,tl~ltyass~ct-
atedwiththe460.7-nmatomicresonanceImeISnotsatisfactory.Suggestatleast
threethingsthatmightbetriedtoincreasesensitivity,
9.9Whyisatomicemissionmoresensitivetoflameinstabilitythanatomicabsorption
orfluorescence"
9-10Figure9-1summarizesmanyoftheprocessesthattakeplaceinalaminar-flow
burner.WithspecificreferencetotheanalysisofanaqueousMgCl,solutIon,
describetheprocessesthatarelikelytooccur.
*9-11UseEquation7-13fortheresolvingpowerofagratingmonochromatortoesti-
matethetheoreticalminimumsizeofadiffractIongratmgthatwouldproYldea
profileofanatomicabsorptionhneat500nmhayingalinewidthof0.002nm.
AssumethatthegratingistobeusedinthefirstorderandthatIthasbeenruled
at2400grooves/mm.
*9-12Forthe!lameshowninFigure9-3,calculatetherelativeintensityofthe766.5-nm
emissionlineforpotassiumatthefollowingheightsabovetheflame(assumeno
ionization):
(a)2.0cm (b)3.0cm (c)4,0cm (d)50cm
9-13Inahydrogen-oxygenflame.theatomicabsorptionsignalforironwasfoundto
decre;seinthepresenceoflargeconcentrationsofsutrate1011 .
(a)Suggestanexplanationforthisobservation.. ...,,.
(b)Suggestthreepossihlemethodsforovercommgthepotenllalmterferenceof
.suli;teinaquantitativedeterminationofiron.
*9-14ForNaatomsandMg~ions.comparetheratiosofthenumberofparticlesinthe
31'excitedstatetothenumherinthegroundstakin
(a)anaturalgas-airflame(2100K).
(h)ahydrogen-oxvgenflame(2900K)
(c)aninductivelycoupledplasmasource(60(~)KI.
Interferencesencounteredinatomicfluorescence
spectroscopyaregenerallyofthesametypeandof
aboutthesamemagnitudeasthosefoundinatomicab-
sorptionspectroscopy."
Atomicfluorescencemethodshavebeenappliedtothe
determinationofmetalsinsuchmaterialsaslubricat-
ingoils,seawater,geologicalsamples,metallurgical
samples,clinicalsamples,environmentalsamples,and
agriculturalsamples,Table9-3listsdetectionlimitsfor
atomicfluorescenceprocedures.
*Answersarcprovidedattheendofthebookforproblemsmarkedwithanasterisk.
~Problemswiththisiconarebestsolvedusingspreadsheets.
9-1Definethefollowingterms:(a)releasingagent,(b)protectiveagent.(c)ionization
suppressor,(d)atomization.(e)pressurehroadentllg.IfIhollo,,-cathodelamp.