• Nie Znaleziono Wyników

"(Classical) binary encounter model in molecular ionization. Restricted Hartee-Fock calculations and target polarizability" 

N/A
N/A
Protected

Academic year: 2021

Share ""(Classical) binary encounter model in molecular ionization. Restricted Hartee-Fock calculations and target polarizability" "

Copied!
6
0
0

Pełen tekst

(1)

ContentslistsavailableatScienceDirect

International Journal of Mass Spectrometry

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

Electron-impact ionization of fluoromethanes – Review of experiments and binary-encounter models

Grzegorz P. Karwasz

a,∗

, Paweł Mo ˙zejko

b

, Mi-Young Song

c

aFacultyofPhysics,AstronomyandAppliedInformatics,NicolausCopernicusUniversity,Grudzi ˛adzka5,87-100Toru´n,Poland

bFacultyofAppliedPhysicsandMathematics,Gda´nskUniversityofTechnology,Narutowicza11/12,80-233Gda´nsk,Poland

cPlasmaTechnologyResearchCenter,NationalFusionResearchInstitute,814-2Osikdo-dong,573-540Gunsan-si,RepublicofKorea

a r t i c l e i n f o

Articlehistory:

Received17November2013 Receivedinrevisedform 27December2013 Accepted14January2014 Availableonline25January2014

Keywords:

Ionization Electronimpact Fluoromethanes

a b s t r a c t

Experimentsandrecommendeddataonelectron-impactionizationofmethaneandfluoromethanes (CH3F,CH2F2,CHF3,CF4)arereviewedandcomparedwithbinary-encountermodels(Gryzi ´nski’s,Deutsch andMärk’s,andKimandRudd’s).Agoodagreementbetweenrecentexperimentsandthetwolatter classical-likemodelsisshown.KimandRudd’smodel(calculatedpresentlyintherestrictedHartree- Fock6-31**Gorbitalbasis)predictswelltotalionizationcrosssectionsforallfivemoleculesconsidered.

However,counting-ionizationcrosssectionshavetobeextractedfromexperimentaldatatoshowthis agreement.TheadditivitymodelofDeutschandMärkperformsequallywell,onceallmolecularorbitals aretakenintoaccount.Themaximaoftotal(counting)ionizationcrosssectionscalculatedpresentlyin KimandRudd’sbinary-encounterapproximationcorrelatelinearlywiththemolecularpolarizability.

©2014ElsevierB.V.Allrightsreserved.

1. Introduction

OnehundredyearsfromBorn’sformulationoftheatomicmodel [1]theamountofexperimentaldataonelectron-atomscattering seemstobequitesatisfactory,seefore.g.reviewsbyMärkand Dunn[2],Karwaszetal.[3],LindsayandMangan[4].However,a morecarefulinsightshowsthatsomesimplesystematicsarestill missing,seefore.g.[5,6].Inthecaseofelectron-impactionization neitherthetotalionizationcrosssectionnorpartitioningintospe- cificionicchannelshavebeencorrelatedtosomeotheratomicor molecularfeatureslikeelectronaffinity,ionizationpotential,polar- izability.Numericalfittingofcrosssectionsforparticularprocesses likeionizationforavastchoiceoftargets(e.g.Refs.[7–9])and/or reviewsofcrosssectionsetsforseveralpossibleprocessesbutfor selectedtargetsonly,likemethane[10]orfluoromethanes[11,12]

wereundertaken,butnoconclusivepictureshavebeenobtained sofar.

Inthepresentworkweperformananalysisofrecentexperi- mentaldataonfluoromethanes(andmethane)andcomparethem withtwobinary-encounter models:DeutschandMärk[13] and KimandRudd[14].

∗ Correspondingauthor.Tel.:+48566112407.

E-mailaddress:karwasz@fizyka.umk.pl(G.P.Karwasz).

2. Deutsch–Märkandbinary-encounterBethe’smodels AsresumedbyMargreiteretal.[15],Thomson[16]wasthefirst toapplyclassicalmechanicstoderiveaformulafortheelectron impactionizationcrosssection

=



n

4a20n



R

In



2t−1

t , (1)

wherengoesthroughatomsubshells,nisthenumberofelectrons onthenthsubshell,Inistheionizationenergyofthenthsubshell, tisthenormalizedkineticenergyoftheincidentelectront=E/In, RtheRydberg’sconstantanda0 istheBohr’sradius.Thompson assumedthatthevelocitiesoftheelectronofthetargetatomsare smallcomparedtothevelocityoftheincidentelectron.

Gryzi ´nski[17]includedexplicitlythecontinuousdistribution ofthevelocitiesoftheelectronsinthetargetatomderivingthe followingexpressionfor

=



n

4a20n



R

In



2

1 t



t

1 t+1



3/2



1+2 3



1 1 2t



ln[2.7+(t1)1/2]



(2)

DeutschandMärk[13]proposedamodificationofGryzi ´nski’sfor- mula,substitutingBohr’sradiusa0withradiirnofthenthsubshells andthefactor4inEq.(2)byweightingfactorsgn obtainedfrom

1387-3806/$seefrontmatter©2014ElsevierB.V.Allrightsreserved.

http://dx.doi.org/10.1016/j.ijms.2014.01.010

(2)

Fig.1. DifferentapplicationsofDeutsch–Märkmodelformolecules:H2DMmodel [15];forCH4“DM1”and“DM2”modelsfrom[15]usingtwoalternativeMulliken populationsofmolecularorbitalsandorbitalsradii;forCF4“DM”from[15]and

“modifiedDM”from[24].Experiments:CF4,Nishimuraetal.[42]andCH4presently suggestedtotalcountingcrosssections,seeFig.3.Lowersymbolsareexperimental recommended[4]totalgrosscrosssectionsofH2andF2.

empiricaldataforelectronscatteringonnoblegases[18,19]

=



n

gnrnn1 t



t1

t+1



3/2



1+2 3



1− 1 2t



ln[2.7+(t−1)1/2]



(3) ThissocalledDeutsch–Märk(DM)approachwasextended [15]

tomolecularionizationassuminganadditivityrule,i.e.thatthe molecularcrosssection(i.e.AB)isasumofthecrosssectionsof constituentatoms(AandB)withappropriateweightingfactors

Aandb).

AB=AA+BB (4)

Theouterelectrons(i.e.thosefromthemolecularorbitals)bring themaincontributiontotheionizationcrosssection.Inorderto includethedetailsofthemolecularstructureintotheadditivity ruleEq.(4),theweightingfactorsnaremadeequaltothepartial contributionsoftheconstituentatoms(i.e.AandB)tothespe- cificmolecularorbitals.Letusrecallagainthatradiirn inEq.(3) andinconsequenceinEq.(4)aretakenfromquantummechani- calcalculationsforconstituentatomsandthefactorsgnfromthe semiempiricalanalysis[18,19].Forexample,forH2theweighting factorisgH=3,rH=a0(andH=2)[15].AcomparisonbetweenDM resultsfromRef.[15]andpresentlyrecommendedexperimental valuesforH2,CH4andCF4areshowninFig.1.TheDMmodelforCH4 predictsthemaximumofthetotalcrosssectionwithintheexisting spreadforrecentexperimentaldata.ForCF4theagreementofthe DMmodelwithrecentexperimentsisalsoquitegood.

For the DM additivity model for molecules, the knowledge of partial occupancies of molecular orbitals by electrons from constituentatoms(Mullikenpopulations)isneeded.Further,the molecularcrosssectionsdependontheionizationenergiesofthe constituentatoms.Obviously,thisisanindirectwaytoaccountfor thestructureofthemolecule.TheDMmodelpredictsreasonably wellthemagnitudeoftotalionizationcrosssections,inparticu- larintheregionbetweenthethresholdand themaximaofthe crosssection,i.e.intherangewhichismostimportantforpractical applicationsofplasmas.

TheDMmodelwasalsosuccessfullyappliedtomultipleioniza- tion[20]andinner-shellionizationofatoms[21].Amorerefined analysisoftheweightingfactorsiinformula(5),basedonabroad experimentalbasewasincorporatedintoa modifiedDMmodel

[22–24].ThemodifiedDMmodelpredictsquitewellamplitudes ofthetotalionizationcrosssectionsinmolecules,seeFig.1forCF4. NeitherGryzi ´nski’snorDeutschandMärk’smodeltakesinto accountexplicitlythevelocitiesoftheelectronsinthedifferent orbitalsandtheirbindingenergies(i.e.therealionizationenergies ofthemolecule).Afurtherstepinthesemiempiricalunderstand- ingoftheelectron-impactionizationwasdonebyKim[25]and Rudd[26]whoproposednewformulaenotonlyfortotalbutalso fordifferentialcrosssections(vs.angle,vs.collisionenergyand vs.energy-loss).Subsequently,usingMott’s[27]approximationfor lowandBethe’s[28]approximationforhigh-energycollisions,Kim andRudd[14,29]proposedthefollowingapproximationfortotal ionizationcrosssections

=



n

4a20n



R

In



2 1 t+un+1



1−1 t +

lnt 2



1− 1 t2



− ln t t+1



(5) whereunisanormalizedkineticenergyofanelectrononthenth orbital,un=Un/In.Asthismodelincludesthebinary-encounterand Bethe’sideasitiscommonlycalledBEB.

InBEBbothionizationenergiesofelectronsonnthorbital(i.e.

theirbindingenergies)aswellastheiraveragekineticenergies havetobeevaluatedfromquantummechanicalcodesformolec- ularstructures.Asaconsequence,theBEBresultsdependtosome extentonthemolecularorbitalbasissetschosen.TheBEBmodel wassuccessfullyappliedtolightatoms(He),(Ne),molecules(H2, H2O),ions(Li++)[14],metals[30]andtargetslikeN2,O2,CO,CO2, NH3,C3H8,butrarelytoCH4[31].

TheDMandBEBmodelsoffluoromethanes(CH3F,CH2F2,CHF3 andCF4)wereextensivelydiscussedbyTorresetal.[32].Inpar- ticulartheyexploitedtheoriginalDMapproachwiththeadditivity rule[15]andwiththemodifiedDMadditivityrule[23]calculat- ingMulliken’smolecular-orbitalpopulationsandweightingfactors foratomicelectrons(gn)directlyfromquantum-mechanicnumer- icalcodes (GAUSSIAN98W).Torreset al.[32] concludedfor all thefourgasesconsideredthattheD-MmodelwithHartree-Fock STO-3Gtendstooverestimatetotalionizationcrosssectionswhile withothermolecularbasesconsidered(HF/6-311G,HF/6-311**G, MP2/6-311G,MP4/6-311G,CC/6-311G,CISD/6-311G)theresults differ inalmostundistinguishableway (within 1%)and tend to underestimateslightlytheexperiments,seetheirFig.3in[32].

IntheBEBmodeltheinfluenceofthemolecularbasischosen issimilarlyinsignificant(withdifferenceswithin2%),apartfrom theHF/STO-3Gbasiswhichoverestimatestheexperiments[32].

InpresentworkwecalculatedBEBcrosssectionsusingrestricted Hartree-Fock6-31**Gorbitalbasissetfortheentireseries,CH4, CH3F,CH2F2,CHF3,CF4andcomparedthemwithexperimentaldata inFigs.3–5.

3. Discussionofexperimentaldata

3.1. Methane

Methaneisoneofthemostextensivelystudiedmoleculartar- gets–itisimportantforglobalwarmingbalance,fortokomakedge plasmas[5] and for hydrogen production viaelectric-discharge pyrolysis[33].The BEBmodel was appliedtoCH4 byKim and collaborators, [34] and [31], using vertical (14.25eV) and adia- batic(12.6eV) ionizationpotentials,respectively. Obviously,the useoflowerionizationpotentialrisesthetotalcrosssection,bring- ingthemaximumtoabout4.3×10−20m2,i.e.abovethehighest experimentaldata[35,36]andshiftingthemaximumtowardlower energies.

Theresultsofdifferentmeasurementsoftotalionizationcross sectionsofCH4areshowninFig.2.At100eVthedifferentdata

(3)

Fig.2.ReviewofexperimentalgrossionizationcrosssectionsforCH4.Rappand Englander-Golden[37],Schrametal.[79],Chathametal.[80],OrientandSrivastava [35],NishimuraandTawara[39],Vallanceetal.[61],Straubetal.[40],TianandVidal [36],thedataofGluchetal.[41]havebeenre-normalizedtothemeasurementsof RappandEnglander-Golden[37].“L-B”withlinedrawnaseye-guidestandsfor recommendedvaluesofRef.[4].

showaspreadofabout20%whatmakesthecomparisonwithBEB modelsdifficult.Somevalidationofexperimentaltechniquesused indifferentmeasurementsisneeded.RappandEnglander-Golden [37]normalizedtheirdatatoownmeasurementsofH2;Tianand Vidal[36]normalizedtheirdatatoAr+measurementsbyStraub etal.[38].Nishimura andTawara[39]measuredabsolutecross sections;theirdataalmostcoincidewiththosebyStraubetal.[40].

LindsayandMorgan[4]compareddifferentexperimentsincluding thatfromtheirlaboratory[40]andproducedasetofrecommended values;thissetdiffersby−9%at100eVandby−2%at1000eVfrom theexperimentbyStraubetal.[40].

Gluchetal.[41]normalizedthesumoftheirpartialcrosssec- tionstothetotalvalueofRappandEnglander-Golden[37].InFig.2 wehavere-normalizedthedataofGluchetal.byafactorofa02 totheoriginalvalueofRappandEnglander-Goldenat100eV.The mostrecentdata[39–41]andtherecommendedvalues[4]agree

Fig.3. SearchfortotalcountingionizationcrosssectionsforCH4.Experimentand

“LBtotal”,seeFig.2.“Wang”singleionization[50].“LBcounting”hasbeenobtained fromrecommendedgrosstotal“LBtotal”[4]bysubtractinghalfoftheH+yield(“LB:

H+ion”[4]),seetextforthediscussion.BEBKim[34]“Ward:double”isthesumof H+,CH+,C+,H2+,CH3+,CH2+yieldscomingfromdoubleionizationprocesses.“BEB 2a1ispresentcalculationforionizationfromthe2a1molecularorbital.

within10%,seeFig.3.ThepresentBEBmodelremainslowerthan theseexperiments.

Theusualexplanation[42]ofthisdifferenceisthatBEBmodels includealsothedissociationintoneutrals.This,inturndisagrees withtheverybasisoftheclassicalmodels,Eq.(1)–(5):theirkine- maticsassumeanelectronleavingthemolecule.Measurementsof dissociationintoneutralsexistforonlyfewmolecules.Winters[43]

measuredthetotalyieldofdissociatedfragments(ions+neutrals) byadsorptionontitaniumgetterobtainingacrosssectionofabout 4.0×10−20m2atthemaximumat80eV,comparabletothemax- imumin thegrossionizationcrosssection.At80eVtheparent ionization (i.e. into CH4+)is 1.55×10−20m2 [4] so BEB would underestimatethe sumof theionization and dissociation-into- neutralscrosssections.

ThedifferencebetweenpresentBEBcalculationsand experi- mentscanbeexplainedintermsofmultipleionization,andmore preciselyintermsofmultiplecountingionscomingfromthesame ionizationevent.Methodsinwhichthetotalchargeismeasured giveso-calledgrosstotalcrosssection.Thisquantitydiffersfrom thecountingionizationcross sectionsas multiplychargedions arecountedwithweightsoftheircharge.Forargonat100eVthe Ar2+/Ar+ratioamountsto6.6%,see[3].Inmolecules,dissociative ionizationisacompetingchanneltomultipleionization.Gener- allythedoubleionizationinmoleculesislow:inCF4at180eVthe summedcontributionfromCF22+andCF32+isabout1–2%ofthe totalgrossionizationcrosssection[44,45].

However,inmoleculesconsideredinthispaper,particularlyin CH4 and CF4 careful attentionmustbepaid todiscriminatefor doubleioncounts,like(CH2++H+)etc.Formethaneseveralrecent experimentalpapers[46–48]showedthatthedicationCH42+inits groundandexcitedstatesdissociatesimmediatelyanddecaysinto twoionizedfragments.Inparticular,usingthecoincidencetech- nique,Wardetal.[46]havereportedcrosssectionsforformation ofpairsofdissociatedionsandattributedthemtoaspecificpre- cursor(monocation,dication,trication);theirdataarerelativeto theCH4+yield.ThistechniquewasalreadyusedbyLindsayetal.

[49],who reportedcross sectionsforproduction of(CH2++H+), (CH++H+),(C++H+)ionpairsbuttheirresultsareprobablyunder- estimated,comparewith[46].Intotal,at200eVasmuchas58%

ofthegrossionizationcrosssectioncomesfromdissociativeion- ization(i.e.fromotherchannelsthanformationoftheCH4+ion) [4];obviouslyinpartoftheseeventsthesecondfragmentcanbea neutral.

AlreadyWangandVidal[50]noticedarelativelyhighamount ofmultiplecountingsofionscomingfromdissociativechannels that contributes to the gross total cross section. They evalu- ated the cross section for the double ionization at 200eV as 0.23×10−20m2 (vs.2.89×10−20m2forsingleionization).Wang and Vidal[50] estimatedroughly that theH+ ion is formed in abouthalfofthedissociativeionizationevents.AccordingtoWard et al. [46], at 200eV as much as 52% of H+ ionsresult from a doubleionization(and1.6%fromtripleionization).Theabsolute valuefortheH+overallyieldat200eVis0.328×10−20m2accord- ing to therecommended values in [4] and 0.30×10−20m2 for thesummed-upyieldofsingle,doubleandtripleionizationchan- nelsreportedbyWardetal.[46](andnormalizationthemtothe CH4+ yield from[4]).Thesum ofall cross sectionsfor appear- anceofionpairsreachesamaximumof0.37×10−20m2at100eV [46].

Foralltheseexperimentalevidencesitisreasonabletoassume thatthecountingcrosssection(i.e.withtheexclusionofdouble countingionpairs)ofCH4canberoughlyevaluatedbysubtracting halfofthecrosssectionfortheformationofH+ion[4]fromthegross totalcrosssection[4].WedidsoinFig.3–weconsiderfullpointson thisfigureassemi-empiricallysuggestedvaluesforthetotalcount- ingionizationcrosssectioninCH4.NowtheBEBmodelcoincides

(4)

Fig.4.Ionizationcrosssectionsinfluoromethanes.CH3F(shiftedby−1):grosstotal Vallanceetal.[61],Torresetal.[32],Rejoubetal.[58,4];H++F+yieldfrom[32].

CH2F2:Torresetal.[57].CHF3(shiftedby+1):BeranandHevan[60],Jiaoetal.[59], Torresetal.[45].BEBispresentRHF6-31**Gcalculation.

inthe40–1000eVrangewiththeexperimentwithinthecombined uncertaintyoftheexperimentalandevaluationprocedures.

Tovalidatefurtherourestimate,inFig.3weshow:(i)thediffer- encebetweentheL-Brecommended[4]grosstotalionizationcross sectionandthepresentBEBresults,(ii)theH+ ionyieldaccord- ingtotheL-Breview[4],(iii)thesumofH+,CH+,C+,H2+,CH2+, CH3+yieldscomingfromdoubleionizationprocesses(tableIIfrom Wardsetal.[46])normalizedpresentlytotherecommendedCH4+ yield[4].Thesethreesetsshowasimilarrangeofamplitudes.

Thedifferencebetweentherecommendedgrossandpresent BEB cross section remains unexplained only around 30eV, i.e.

wheredissociationsintoneutrals(CH2andCH3signals)reachtheir maxima[51].However,wearenotabletojudgeontheamplitude ofpossiblecontributiontothegrossionizationcrosssectioncom- ingfrompossibledissociationintoaneutral+ionizedfragmentvia anexcitationtosomehigherelectronicstates.Existingexperiments [51–53]arefragmentaryanddisagreewiththeories[54].Addition- ally,theBEBmodelwiththeverticalionizationpotentialprobably underestimates crosssectionsin thenear-thresholdregion, see theBEBcalculationwiththeadiabaticpotential[31].Experimen- talchecks[55,56]of theenergydependencesin thisregionare thereforeprecious.

3.2. Fluoromethanes(CH3F,CH2F2,CHF3)

Torresetal.[32,45,57]performedrecentlyaseriesofmeasure- mentsof partialand total ionizationcross sectionsupto85eV withadeclared uncertaintyless than10%.Morgan andLindsay [4] gave recommendedcross sectionsfor CH3F coinciding with themeasurementsofRejoubetal.[58].Forallthreemixedflu- oromethanes,CH3F,CH2F2andCHF3theagreementbetweenthe presentBEBmodelandrecentexperiments[25,45,57–60]isvery good,seeFig.4.ForCH3FweshowalsothesumofH+andF+ion production[57],withamaximumof0.4×10−20m2,butit does notmeannecessarilythattheseionscomefromdoubleionization events.

3.3. Tetrafluoromethane(CF4)

Tetrafluoromethaneisthegasmostcommonlyusedforplasma etching in semiconductor industries [62], therefore numerous experimentshavebeenperformed,startingfromthedetermina- tionofthedissociation(ionizationplusneutral-dissociation)cross

Fig.5.IonizationcrosssectionsinCF4.Experimentalgrosstotal:Polletal.[24], BruceandBonham[64],Nishimuraetal.[42],Torresetal.[45],Sieglaffetal.[65,4];

Bonhamcounting[70];Sieglaffdouble(CF3++F+andCF2++F+)from[65];BEBKim iscomplete-active-spacecalculation;BEBpresentisRHF6-31**Gcalculation.

sectionof5.5×10−20m2atitsmaximum[63].Allionizationevents in CF4 lead tothe dissociation of the molecule, with the CF3+

ionpredominant(3.3×10−20m2atitsmaximum[4]).Someearly experimentssufferedfromincompletecollectionoflightfragment ions but newer results[24,42,45,64] forma congruent dataset, agreeingwithin±15%,seeFig.5.Outofseveralrecentexperimen- taldata,theresultsfromInnsbrucklaboratory[24]thataccount for thecorrectionof iontrajectoriesin theion sourceformthe highest set,givea maximum of 6.0×10−20m2 at120eVwhile theresultofNishimuraetal.[42]is5.3×10−20m2at125–175eV.

ThecoincidencemeasurementsbySieglaffetal.[65]doneinabso- lutewaywith±5%uncertaintygaveatmaximumalowervalue, 4.95×10−20m2.

ThepresentBEBmodelagreeswellwiththerecentexperiments [42,45,64]andcoincideswiththeBEBcalculationdoneinslightly higher molecularorbitalbasis, RHF 6-311+G(d)[42], seeFig.5.

Kimandcollaborators[42]triedtoevaluatetheeffectofmultiple ionizationperformingBEBcalculationswiththeuseofacomplete- active-space wave functionsset (CAS, in Fig. 5).Such a model overestimatesthegrossionizationmeasurements[42,45,64],see Fig.5.

WhilstallionizationprocessesinCF4leadtothedissociationof themolecule,thedataofSieglaffetal.[64]showthatthedou- ble ionization(CF2++F+ andCF++F+)islowerthan forCH4 and reachesamaximumof0.23×10−20m2at200eV,seeFig.5.This canbeexplainedbythehighelectronegativityoffluorineandits componentswhatmakestheformationoftwopositiveionsless probableinCF4thaninCH4(i.e.thesecondfragmentinthedisso- ciativeionizationofCF4shouldpredominantlybeaneutral).This isalsoindirectlyconfirmedbystudiesofdissociationintoneutrals [52,63].ThecrosssectionoftheformationofanFatominionizing andneutral-dissociationeventsisabout7×10−20m2atitsmaxi- mum[52]comparedto0.74×10−20m2fortheformationoftheF+ ion[4].

Note,however,thatthedistinctionbetweengrossandcounting ionizationcrosssectionsofCF4 isfarlessdecisivethanforCH4. Awholeseriesofre-measurementsandre-analysiscomingfrom severalgroups[24,44,66–70]testifiesthis.

4. Towardsystematicsoftotalionizationcrosssections Presentanalysisofexperimentaluncertaintiesandcorrections bringcredibilitytothesystematictendenciesobservedbothinBEB

(5)

Table1

Lineardependenceofthemaximumcrosssectionsmax[10−20m2]inpresentBEB modelvs.moleculardipolepolarizability˛[10−30m3],experimentalvaluesfrom Ref.[73].

Molecule ˛[×10−30m3] max[×10−20m2] max

CH4 2.593 3.571 1.377

CH3F 2.97 4.02 1.354

CH2F2 3.27a 4.5 1.376

CHF3 3.57 4.92 1.378

CF4 3.838 5.273 1.374

aValueinterpolatedbetweenCH3FandCHF3.

(presentand[32])andinDM[42]estimatesofionizationcrosssec- tions.Inparticular,wereputethattheBEBmodelwith6-31**G molecularorbitalbasis reproducesthetotal countingionization cross sectionsfor the wholeseries CH4 – CF4 within5% accu- racy.Inthisseriesthemaximumofthecrosssectionrisesfrom 3.57×10−20m2forCH4to5.27×10−20m2forCF4.Thepositionof thecrosssectionmaximumrisesinasimilarmanner,from75eV inCH4to140eVinCF4.Thisratherreflectstheincreaseinenergy depthoforbitallevelsthanthechangesinthresholdvalues,14.13, 13.63,14.01,15.47and17.08eVforCH4,CH3F,CH2F2,CHF3 and CF4,respectively.

Inseveralworkssystematicdependencesofionizationcrosssec- tionswereexamined.Harlandetal.[71]studiedthecorrelationof atomicandmolecularmaximumionizationcrosssectionsmaxon thedipolepolarizability␣,checkingbothmax∝˛andmax∝√˛ proportionalities.Kimandcollaborators[42]postulatedforfluoro- carbonsCF4,C2F6,C3F8amax∝√

˛ZdependencewithZbeingthe totalnumberofelectronsinthetarget.Asimilardependenceonthe polarizabilityintherangeof50–100eVandtheapplicationofthe additivityrulewasalsonoticedfortotal-scatteringcrosssections [72].

MaximaoftheBEBtotalionizationcrosssectionsmaxvs.the dipolepolarizabilities[73] of thefivemolecules consideredare resumedinTable1.Themaxrises proportionallytothedipole polarizability of the target with the proportionality coefficient 1.37(±0.01),max=1.37˛,ifmaxisexpressedin10−20m2and˛ in10−30m3.Thecorrelationofthelinearfitis0.999.

Thisrathersimpledependencedeservessomecomments.All fourclassical-likeformulaeconsideredhereindicatethatthenum- berofelectronsonmolecularorbitalsdeterminestheamplitude oftheionizationcrosssection. Ontheother hand,themolecu- lar polarizabilitycan beconsideredas a sumof polarizabilities fromseparatemolecularbonds.Secondly,fromclassicalelectro- dynamics,the polarizability is themeasure of the deformation oftheelectroniccloudintheexternalelectricfield.Fromquan- tummechanicalpoint ofview,thepolarizabilityisthemeasure forthesumofallvirtualexcitationsofthetarget:electronic[74]

andvibrational[75].Again,thevibrationalexcitationsreflectprop- ertiesofthemolecularbonds;forelectronicexcitationsthesum of oscillator strengths is equal to the number of valence elec- tronsevenifthedetailedelectronic-excitationspectraofmolecules canbequitecomplex[76].Therefore, presentlyobservedinter- dependenceoftheionizationcrosssectionandthepolarizability canreflectthesame,somewhatsyntheticmolecularfeature:the numberofvalenceelectronsconvolutedwiththeirbinding(and kinetic)energies.

Weareawarethattheseargumentsarenotconclusivesofur- theranalysisareneeded.Notealsothatforsakeofcomparisonson anequalbasisweusedpolarizabilitiesfromthesametypeofopti- calmeasurements[73].Theories,includingthepresentHFmethod, tendtounderestimatethemolecularpolarizabilities,evenifthe generaltrendfortheseriesofmoleculesispreserved.Thequality ofcalculationdependsmuchonthechoiceofthemolecularorbital basis,comparefore.g.[77,78].

5. Conclusions

Binary-encountermodels[13,14,17]provetobequitesuccess- ful in predictingionization cross sections. Whatbecomes clear fromthecaseofmethane,theseareratherexperimentalsingle- ionization(i.e.counting)crosssectionsthatagreewithKimand Rudd’s[14] BEBmodel.Outofnumerous experimentsonlyfew allowtoestimatethesingle-ionizationcrosssectionwhichisthe outputofthebinary-encountermodels.

Formethane,Wardetal.[46]showedthatat200eVmoreH+ ionscomefromthedoubleionizationprocessthanfromthesingle ones.Subtractingtheestimateforthedoubleionizationfromthe recommended[4]grossionizationoneleadstoanalmostperfect agreementwiththepresentBEBmodel(6-31**Gorbitalsandthe verticalionizationenergyof14.128eV).

MoreuncertaintyexistsonsuchadistinctionforCF4andother fluoromethanes.Quantumchemistrycalculationsoftheenergetics ofspecificdissociationchannels–intoneutrals,intoneutralsand ions,andintoionpairs,areimportantinthiscontext[32].

Fromexperimentalside,singlecoincidence[46] andkinetic- energyreleasemeasurements[41]areimportantforunderstanding possibleionization channels.Such a knowledge is essential for projectingand diagnosisofplasmaprocessesin manytechnical applications.

DeutschandMärk’smodel[13,15,24]underestimatestheposi- tionoftheionizationmaximumbutthisshouldberatherattributed tothesimplifiedapplicationsofthemethod,whereonlythelowest ionizationthresholdsareconsidered,asinFig.1.AsshownbyTor- resetal.[32]whoincorporatedamorecompletesetofquantum levels,includinginnerorbitals,intotheircalculation,theagreement oftheDMmodelwithexperimentsisnotworsethanforBEB.

AnextensionoftheBEBmodeltodoubleanddissociativeion- ization wouldbe desirable.Already Gryzi ´nski [17] showedtwo possiblemechanismofdoubleionization:thedirectprocess,i.e.

thesecondionizationbythe(scattered)incomingelectron,andthe recoilprocess,i.e.ionizationbythe(leaving)secondaryelectron.It isnottobeexcludedthattheinformationonthedissociativeion- izationcanbealreadyextractedfromthestandardBEBcalculation.

InFig.3wecomparetheexperimentalH+yield[4]withtheioniza- tionfromthe2a1molecularorbitalinthepresentBEBmodel:the twocurvesseemtocoincide.

Presentcalculationsdonottakeintoaccountthepossibilityof ionizingtransitionscomingfromhighervibrationalstates.Inother words,weassumethatthevibrationaltemperatureofthetarget moleculesis T=0.Asalready noticedbyKim andcollaborators, thevibrationalstructureoftheelectroniclevelswouldmodifythe ionizationcrosssectionsinthenear-to-thresholdregion.Further calculationswouldbeneededtoexplorethisquestionindetail.

Acknowledgments

ApartofthisworkhasbeendoneinQuantumOpticsCenter (COK),UMKToru ´n.NumericalcalculationsintheGAUSSIANcode havebeenperformedattheAcademicComputerCenterinGda ´nsk (TASK).WethankdrA.Karbowskiforthepreparatoryworkonthe analyticalfitsforCH4crosssections.GKthanksforthehospitality atNFRI,Gunsan.

References

[1]N.Bohr,Phil.Mag.25(1913)10;

N.Bohr,Phil.Mag.30(1913)581.

[2]T.D.Märk,G.H.Dunn(Eds.),ElectronImpactIonization,Springer-Verlag,Wien, NewYork,1985.

[3]G.P.Karwasz,R.S.Brusa,A.Zecca,Riv.NuovoCim.24(1)(2001)1–118;

G.P.Karwasz,R.S.Brusa,A.Zecca,Riv.NuovoCim.24(4)(2001)1–101;

G.P.Karwasz,R.S.Brusa,A.Zecca,Riv.NuovoCim.19(3)(1996)1–146.

(6)

[4]B.G.Lindsay,M.A.Mangan,Ionization,in:Landolt-Börnstein(Ed.),Numeri- calDataandFunctionalRelationshipsinScienceandTechnology.NewSeries.

GroupI:ElementaryParticles,NucleiandAtoms,vol.17:PhotonandElec- tronInteractionswithAtoms,MoleculesandIons.SubvolumeC.Interactions ofPhotonsandElectronswithMolecules,2003,p.5001.

[5]G.Karwasz,K.Fedus,FusionSci.Technol.63(2013)338.

[6]A.E.D.Heylen,Proc.R.Soc.Lond.456(2000)3005.

[7]R.K.Janev,D.Reiter,Phys.Plasmas9(2002)4071.

[8]V.Dose,P.Pecher,R.Preuss,J.Phys.Chem.Ref.Data29(2000)1157.

[9]T.Shirai,T.Tabata,H.Tawara,Y.Itikawa,Atom.DataNucl.DataTables80(2002) 147.

[10]M.C.Fuss,A.Mu ´noz,J.C.Oller,F.Blanco,M.-J.Hubin-Franskin,D.Almeida,P.

Lim˘ao-Vieira,G.García,Chem.Phys.Lett.486(2010)110.

[11]L.G.Christophorou,K.J.Olthoff,J.Phys.Chem.Ref.Data28(1999)967.

[12]R.A.Bonham,Jpn.J.Appl.Phys.PartI33(1994)4157.

[13]H.Deutsch,T.D.Märk,Int.J.MassSpectrom.IonProcess.79(1987)R1.

[14]Y.-K.Kim,M.E.Rudd,Phys.Rev.A50(1994)3954.

[15]D.Margreiter,H.Deutsch,M.Schmidt,T.D.Märk,Int.J.MassSpectrom.Ion Process.100(1900)157.

[16]J.J.Thomson,Phil.Mag.23(1912)449.

[17]M.Gryzi ´nski,Phys.Rev.107(1957)1471;

M.Gryzi ´nski,Phys.Rev.115(1959)374.

[18]K.Stephan,H.Helm,T.D.Märk,J.Chem.Phys.73(1980)3763.

[19]K.Stephan,T.D.Märk,J.Chem.Phys.81(1984)3116.

[20]H.Deutsch,K.Becker,T.D.Märk,J.Phys.B29(1996)L497.

[21]H.Deutsch,D.Margreiter,T.D.Märk,Z.Phys.D29(1994)31.

[22]H.Deutsch,T.D.Märk,V.Tarnovsky,K.Becker,C.Comelisse,L.Cespiva,V.

Bonacic-Koutecky,Int.J.MassSpectrom.IonProcess.137(1994)77.

[23]H.Deutsch,K.Becker,T.D.Märk,Int.J.MassSpectrom.IonProcess.167(168) (1997)503.

[24]H.U.Poll,C.Winkler,D.Margreiter,V.Grill,T.D.Märk,Int.J.MassSpectrom.

IonProcess.112(1992)1.

[25]Y.-K.Kim,Phys.Rev.28(1983)656.

[26]M.E.Rudd,Phys.Rev.A44(1991)1644.

[27]N.F.Mott,Proc.R.Soc.Lond.Ser.A126(1930)259.

[28]H.A.Bethe,Ann.Phys.5(1930)325.

[29]Y.-K.Kim,M.E.Rudd,CommentsAtom.Mol.Phys.34(1999)293.

[30]Y.-K.Kim,J.Migdałek,W.Siegel,J.Biero ´n,Phys.Rev.A57(1998)246.

[31]W.Hwang,Y.-K.Kim,M.E.Rudd,J.Chem.Phys.104(1996)2956.

[32]I.Torres,R.Martínez,M.N.SánchezRayo,F.Casta ˜no,J.Chem.Phys.115(2001) 4041.

[33]M.Jasi ´nski,M.Dors,H.Nowakowska,G.V.Nichipor,J.Mizeraczyk,J.Phys.D:

Appl.Phys.44(2011)194002.

[34]Y.K. Kim, W. Hwang, N.M. Weinberger, M.A. Ali, M.E. Rudd, J. Chem.

Phys. 106 (1997) 1026, Seealso Kim etal. NIST,http://physics.nist.gov/

cgi-bin/Ionization/graphnew.pl?element=CH4.0 [35]O.J.Orient,S.K.Srivastava,J.Phys.B20(1987)3923.

[36]C.C.Tian,C.R.Vidal,J.Phys.B31(1998)895.

[37]D.Rapp,P.Englander-Golden,J.Chem.Phys.43(1965)1464.

[38]H.C.Straub,P.Renault,B.G.Lindsay,K.A.Smith,R.F.Stebbings,Phys.Rev.A52 (1995)1135.

[39]H.Nishimura,H.Tawara,J.Phys.B27(1994)2063.

[40]C.Straub,D.Lin,B.G.Lindsay,K.A.Smith,R.F.Stebbings,J.Chem.Phys.106 (1997)4430.

[41]K.Gluch,P.Scheier,W.Schustereder,T.Tepnual,L.Feketeova,C.Mair,S.Matt- Leubner,A.Stamatovic,T.D.Mark,Int.J.MassSpectrom.228(2003)307.

[42]H.Nishimura,W.M.Huo,M.A.Ali,Y.-K.Kim,J.Chem.Phys.110(1999)3811.

[43]H.F.Winters,J.Chem.Phys.63(1975)3462.

[44]K.Stephan,H.Deutsch,T.D.Märk,J.Chem.Phys.83(1985)5712.

[45]I.Torres,R.Martínez,F.Casta ˜no,J.Phys.B35(2002)2423.

[46]M.D.Ward,S.J.King,S.D.Price,J.Chem.Phys.134(2011)024398.

[47]H.Luna,E.G.Cavalcanti,J.Nickles,G.M.Sigaud,E.C.Montenegro,J.Phys.B36 (2003)4717.

[48]R.Flammini,M.Satta,E.Fainelli,G.Alberti,F.Maracci,L.Avaldi,NewJ.Phys.

11(2009)083006.

[49]B.G.Lindsay,R.Rejoub,R.F.Stebbings,J.Chem.Phys.114(2001)10225.

[50]P.Q.Wang,C.R.Vidal,Chem.Phys.280(2002)309.

[51]T.Nakano,H.Toyoda,H.Sugai,Jpn.J.Appl.Phys.30(1991)2908–2912.

[52]S.Motlagh,J.H.Moore,J.Chem.Phys.109(1998)432.

[53]C.Makochekanwa,K.Oguri,R.Suzuki,T.Ishihara,M.Hoshino,M.Kimura,H.

Tanaka,Phys.Rev.A74(2006)042704.

[54]M.Ziółkowski,A.Vikár,M.L.Mayes,Á.Bencsura,G.Lendvay,G.C.Schatz,J.

Chem.Phys.137(2012)22A510.

[55]T.Fiegele,G.Hanel,I.Torres,M.Lezius,T.D.Märk,J.Phys.B33(2000)4263.

[56]A.N.Zavilopulo,M.I.Mykyta,O.R.Shpenik,Tech.Phys.Lett.38(2012)947.

[57]I.Torres,R.Martínez,M.N.SánchezRayo,F.Casta ˜no,J.Phys.B33(2000)3615.

[58]R.Rejoub,B.G.Lindsay,R.F.Stebbings,J.Chem.Phys.117(2002)6450.

[59]C.Q.Jiao,R.Nagpal,P.D.Haaland,Chem.Phys.Lett.269(1997)117.

[60]J.A.Beran,L.Hevan,J.Phys.Chem.73(1969)3866.

[61]C.Vallance,S.A.Harris,J.E.Hudson,P.W.Harland,J.Phys.B30(1997)2465.

[62]M.J.Kuschner,J.Phys.D:Appl.Phys.42(2009)194013.

[63]H.Winters,M.Inokuti,Phys.Rev.A45(1982)2777.

[64]M.R.Bruce,R.A.Bonham,Int.J.MassSpectrom.IonProcess.123(1993)97.

[65]D.R.Sieglaff,R.Rejoub,B.G.Lindsay,R.F.Stebbings,J.Phys.B34(2001)799.

[66]CeMa,M.R.Bruce,R.A.Bonham,Phys.Rev.A45(1992)6932;

CeMa,M.R.Bruce,R.A.Bonham,Phys.Rev.A44(1991)2912.

[67]M.R.Bruce,R.A.Bonham,Chem.Phys.Lett.190(1992)285.

[68]M.R.Bruce,L.Mi,C.R.Sporleder,R.A.Bonham,J.Phys.B27(1994)5773.

[69]M.R.Bruce,R.A.Bonham,J.Mol.Struct.352/353(1995)235.

[70]R.A.Bonham,M.R.Bruce,Aust.J.Phys.45(2002)317.

[71]P.W.Harland,C.Vallance,Int.J.MassSpectrom.IonProcess.171(1997)173.

[72]G.P.Karwasz,R.S.Brusa,A.Piazza,A.Zecca,Phys.Rev.A59(1999)315;

G.P.Karwasz,R.S.Brusa,L.DelLongo,A.Zecca,Phys.Rev.A61(2000)024701.

[73]D.R.Lide(Ed.),CRCHandbookofChemistryandPhysics,71sted.,CRC,Boca Raton,1990.

[74]J.Mitroy,M.S.Safronova,C.W.Clark,J.Phys.B43(2010)202001.

[75]D.M.Bishop,L.M.Cheung,J.Phys.Chem.Ref.Data11(1982)119.

[76]A.Zecca,G.P.Karwasz,R.S.Brusa,T.Wróblewski,Int.J.MassSpectrom.223 (224)(2003)205.

[77]H.N.Varambhia,J.J.Munro,J.Tennyson,Int.J.MassSpectrom.271(2008)1.

[78]P.ThvanDuijnen,M.Swart,J.Phys.Chem.A102(1998)2399.

[79]B.L.Schram,M.J.vanderWiel,F.J.deHeer,H.R.Moustafa,J.Chem.Phys.44 (1966)49.

[80]H.Chatham,D.Hills,R.Robertson,A.Gallagher,J.Chem.Phys.81(1984)1770.

Cytaty

Powiązane dokumenty

Present data in benzene down to 2 eV agree very well with the results of Kimura et al [2] obtained with the apparatus with a retarding field analyser but without corrections for

ing on chloromethanes in units of 10 220 m 2.. calculation of atomic total cross sections for H, C, F, Si, S, and Cl. They used Hartree-Fock atomic functions and atomic

They are compared with other recent absolute measurements in figure 20 and with selected theoretical calculations in figure 2(b). Our set of results agree within

In the present experiment we have measured total absolute cross sections for electron scattering on NH,, OCS and N 2 0 using a non-magnetic linear transmission

There are also many modifications of the saddle point method (see e.g. The final density expression obtained from such a procedure is relatively simple and it allows

• Rotational excitation, even if (maybe) not important for plasma temperature, can be decisive in electron slowing down, and (in biological tissues), in „brougthing” them to

Electron scattering on molecules - partial (and total) cross sections:.. search for uncertainties and errors in

Synopsis The total cross sections for the single electron-impact ionization of 5-chloropyridine (5-C 5 H 4 ClN) and 5-bromopyridine (5-C 5 H 4 BrN) molecules have been calculated