File:  [Nicadd] / ttbar / p20_taujets_note / plots / d0conf_ttbarljets_v19_rvk.dvi
Revision 1.1.1.1 (vendor branch): download - view: text, annotated - select for diffs
Wed May 18 21:30:44 2011 UTC (13 years, 2 months ago) by uid12904
Branches: tauNote, MAIN
CVS tags: v-00, HEAD
Initiall commit

; TeX output 2008.02.01:19012=PSfile="d0logo.eps" llx=206 lly=343 urx=406 ury=449 rwi=1000 o		cmr9DnoteT5234-CONFWۍg+(Ncmbx12Measurementof*gcmmi12n9'XQcmr12(pp	:,!",
cmsy10!URt$)6܍t>6)in+jetschannelPύ0K`y

cmr10The%D$CollabGoration
(S&j		cmti9URLhttp://www-d0.fnal.gov(Dated:pF:ebruaryT1,2008)mo..Thisnotepresen9tsanewmeasurementof5"		cmmi9pDZp		cmsy9!
tVۍ %tWX:proAductionatؽp1؉aHO|(s
O=
1:96T:eVusing350$pb-=q%cmsy6Aacmr61
nof|data|collectedwiththeDdetectorbAet9ween|2002and2005.SmW:efoAcusonthenalstate$where70the71WbAosonfromoneofthetopquarksdeca9ysinto70a?0leptonanditsassoAciatedneutrino,$whiletheotherdeca9ysintoaquark-antiquarkpair.e
W:eaimtoselectthoseev9entsinwhic9hthe$lepton#subsequen9tlydecays"tooneorthreecharged"hadrons,`zeroormoreneutralhadronsandatau$neutrino(thec9hargeconjugateproAcessesareimpliedinalloftheabAo9ve).MTheobserv|rablesignature$th9usconsistsofanarro9wcalorimetershowerwithassoAciatedtrack(s)characteristicofahadronic$taurdeca9y:,fourormorejets,ofwhic9htwoareinitiatedrbybquarksaccompanyingtherWH'sinthe$topNquarkNdeca9ys,]andalargenetmissingmomen9tuminthetransv9erseplaneduetotheenergetic$neutrino-an9tineutrinoTpairthatleaveUnotraceinthedetectormedia.pThepreliminaryresultforthe$measuredTcrosssectionis:fR(tSaHWtW)=5:1=#+4;cmmi6:38Q#3:5*(statw)=#+0:78Q#0:7(systƻ)1Ľ80:3#(lumij:0)pbOPreliminaryN<ResultsforFJalxl2006Conferences*22jt:		cmbx9I.$INTRCODUCTIONThetopquarkistheheaviestfundamentalparticleinnature.5PrecisemeasurementsofitsproGductionrateandotherpropGerties!allowustoperformprecisiontests!ofQCD predictions.!Inaddition,TanydeviationsfromtheStandardMoGdel%(SM)maysignalthe$presenceofnewphysics.U
InthisrespGect,bthedecaysofthetopquarkintothe
b>

cmmi10 leptonareespGeciallyinteresting,sincetheNistheheaviestlepton.0Anynon-standard
avor-andmass-depGendentcouplingscouldproGduceDaverysignicanteectinthischannel.[AninterestingexampleisthechargedHiggsbGoson,whichappGearsinextensionsoftheSMHiggssectorto2HDMs(Two-HiggsDoubletMoGdels),) andisrequiredinMSSM(MinimalSupGersymmetricStandardModel)[1].^SincetheHiggs-fermioncouplingisproportionaltothelatter'smass,Ҕdecaysto=the<heavyleptonwouldbGemuch<morefrequentthan<thosetothelightereand.~ThispromptsustosearchforGH^ٓRcmr7+"ղdecayingFtoleptons.DandCDF0pGerformedsuchsearchesinRunI/[2,3].ThemeasurementFoftheSMproGcess$presentedhere$isanimportantingredient$toextendingthereachof$suchsearchesinaddition$topGerformingacrucialUUtestoftheStandardMoGdel.TheoreticalGlcomputationsof[ٲ(pWp	b
!",

cmsy10!Zt'#btr)areconstantlyimproving.HThelatestpublishedNNLOG-crosssectionis6.80.4pb[4].ThedecaymoGdeto{в+Z	jetshasabranchingfractionof0.15,thesameastheeZ	+jetsandZ	+jetschannels.%However,n|secondaryite'sand'sfromiuleptonicdecaysofa;leptonarediculttodistinguishfromprompt(primary)onesatahadroncollider.	&Hence,Kqinthisanalysisweonlytrytoidentifytheeventsinwhichthesubsequentlydecaystooneorthreechargedhadrons,Szeroormoreneutralhadronsanda	0ercmmi7(thechargeconjugateproGcesses|!areimplied| inalloftheabove).*These|!account| foronly0.65oftheleptonbranchingratio,soweexpGecttheUUnumbGerofeventstobGelowerthaninthee8+jetsand+jetschannels.This2note1presentstherstmeasurement1ofthet'#btcrosssectioninthischannel,usingthedatacollectedduringRunIGIbytheDdetector.?Theintegratedluminosityofourdatasamplewas349[\23pb^O!cmsy71
y[5].?Aftercarefuloptimizationofjtriggers,o,ob8jectidenticationjandusinganarticialneuralnetworkforthenaljfull-eventpatternrecognition,o,weareabletoextractthefaintsignalfromtheverylargebackground,!whichisdominatedbymismeasuredQCDevents.!C%IKI.urTHEDDETECTORTheDRunIGIpdetector[6]iscomprisedofthefollowingmaincompGonents:thecentraltrackingsystem,#theliquid-argon/uraniumUUcalorimeterandthemuonspGectrometer.Thecentraltrackingsystemincludesasiliconmicrostriptracker(SMT)andacentralbGertracker(CFT),bGothloGcatedTinsidea2T!superconductingsolenoidmagnet..rTheSMT!isdesignedtoprovideecienttrackingandvertexingcapabilityatpseudorapiditiesofj[ٸj=ܵ<=۲3.GThesystemhasasix-barrellongitudinalstructure,heachwithasetoffourlayersarrangedaxiallyaroundthebGeampipe,andinterspersedwith16radialdisks.Atypicalpitchof50-80moftheB_SMTB"stripsallowsB^aprecisiondeterminationofthethree-dimensionaltrackimpactB^parameterwithrespGecttothe,primary-vertex,!whichis,thekeycompGonent-ofthelifetimebasedb-jettaggingalgorithms.MTheCFThaseightcoaxial#barrels,eachsuppGortingtwodoubletsofoverlappingscintillatingbGersof0.835mmdiameter,onedoubletbGeingUUparalleltothecollisionaxis,andtheotheralternatingby3^9relativetotheaxis.Thecalorimeterisdividedintoacentralsection(CC)providingcoverageouttoj[ٸj1andtwoendcalorimeters(EC) extending.coverage-toj[ٸjղ4,each-housedinseparatecryostats.
PScintillatorsplaced-bGetweenthe-CC!andECprovideUUsamplingofshowersat1:1<j[ٸj<1:4.Themuonsystem,\coveringj[ٸj<2,\residesbGeyondthecalorimetryandconsistsofthreelayersoftrackingdetectorsand>fscintillatingtriggercounters.j"Movingradially>goutwards,Btherstlayerisplacedinsidethe>g1.8T>`toroidmagnets,andUUthetwoUUfollowinglayersareloGcatedoutsidethemagnets.ThedDetriggerisathree-leveltriggersystem.@Level1isahardwaretrigger,whilelevelse2and3aresoftwaredlters.!C%RIKII.AOBJECTIDENTIFICATIONThe7?mostimpGortantob8jectsforthisanalysisarejets7@andhadronicYleptoncandidates.Jetsarereconstructedusing{an|iterativealgorithm,?Ewhichintegrates|theenergyobserved|inthecalorimeterinaconewithradiusRȲ=
P\n5u

cmex10pn5fe*f	ˍ`y
2+8`
2$8=0:5,where1y?isdetector2rapidityandistheangleintheplanetransversetothebGeamaxis.KThesejetsare=required=tobGeintheducialregionj[ٸj<2:4,Awith=pT	ε>20GeV/c.iThe=energyofthejetsafterreconstructioniscorrectedUUtorepresentthetruejetenergy[5].qAnequivqalentcorrectionisderivedforMonteCarloevents.	23jײA.
^IDP\1.	qTJauN<decaymodesThe%oleptonhasseveraldecaychannels,/=classiedbythenumbGerofchargedparticles(tracks)andEM{clustersassoGciatedUUwithit[7]:electronUUormuon(߸!eeKKTor!\5),BR=35%.singleUUchargedhadron(߸![ٟ^ڵ5),BR=12%.singleUUchargedhadron+1neutralparticle(i.e.,߸!^M!(n[ٟ^0,+8[ٟ^ڲ)5),BR=38%.3UUchargedhadrons+0neutralhadrons,BR=15%(so-called\3-prong"decays).!C%‰2.	qTJauN<IDvariablesAtD,leptonsareidentiedintheirhadronicmoGdesasnarrow(R)=20:3cone)jets,isolatedandmatchedtochargedUUtracks.qThe(mostimpGortant)discriminatingvqariablesare[8]:ProleUU-iErb}Zcmr51O
\cmmi5T+Erb}0Tɉfe$M.9#cmex7P
L@i<Eb}iύT"Q,whereE^ibT	istheETofthei^th`highestETtowerUUinthecluster.OIsolation,rdenedas&h8Eb}(0:5)E(0:3)8ʉfe3+Eb}(0:3);,whereE(RDz)istheenergycontainedinay,ofradiusRͲaroundthe
$calorimeterUUclustercentroid.T*rackisolation,0denedasscalarP=pT
Qqofnon-&tracksinay	εconeof0.5aroundthecalorimeterclustercentroid.Using"theseandothervqariables,,threeNeuralNetworks"(NNs)are"trainedtoidentifythreetypGesofDlepton([ٲ-typGe,-typGeUUand3-prong)Thet5outputt4oftheseNNsprovidesasetofthreevqariables(D<x

cmtt10Dnnout=1,2,3)tobGeusedtoselecttheleptonintheevent.huThetypGesroughlycorrespondtotheVleptondecaymodes.hvType1containsonlyonechargedtrack,typGe2hasHoneIchargedtrackIandoneormoreelectromagneticclustersandtypGe3hasmorethanonetrack(multi-prongdecays).p_HighQvqaluesofNNQcorrespGondtothephysicalrleptons,Qwhilelowonesshouldindicatejetsmisidentiedas!Dz'sUU(fakes).qF*ormoredetails,seeRef.[8].!C%~B.
^TheSecondaryVertexTaggingAlgorithmTheFµt'#bt
inalFstatecontainsFtwoFµb-jets,IwhilejetsinQCDFandW+jetseventsForiginatemostoftenfromlightquarksoregluons.:RequiringatdleastonejetintheeventetobGeb-taggedisthereforeaverypGowerfulemethoddofbackgroundrejection._Theb-taggingalgorithmusedinthismeasurementisasecondaryvertextaggingalgorithm(SVT),whichexplicitly4reconstructsverticesthataredisplacedfromtheprimaryvertex.ZdInthettingproGcedure, theSVT
usesonlygoGodqualitytracksthathaveanimpactparametersignicance>3.hAlso,&onlysecondaryverticeswhicharedisplacedJfromKtheprimaryvertexintheplanetransversetothebGeamlinebymorethan7standarddeviationsareconsideredUUinthismeasurement.E':

cmti10Ea.
8bE-taggingeciency
ItTisSknown[9]thattheb-tagging,TwhenapplieddirectlytoMonteCarlo,TgivesTaneciencythatgishigherthandata.%Inordertoaccountforg thisfactor,kSVTghasbGeenparameterizedont7ferYɍt0!侵D+jetsgMCgand+jetsWdatatoXcomputethecorrectionfactor,
whichscalestheMC-derivedeciency*.$Asaresult,
weobtaintheMCftaggingprobabilityandthedata-correctedone[9].HItcanbGenotedthatthedata-correctedeciencyisindeednoticeablyUU(30%)lowerUUthanwhatwewouldexpGectbyapplyingtheSVTdirectlytoMC.Eb.
8cE-taggingeciency
An`assumption`ismadethatthecorrectionfactorobtainedbydividingthesemi-leptonicb-taggingeciencyindatatotheoneinMCalsoiscorrectforcjets.H	HencetheMC-obtainedinclusivec-taggingeciencyUUismultipliedbythisfactorinordertoestimatethec-taggingprobability*.$24]YC9T:ABLETI:Bac9kgroundsources,relev|rantforthe7+8jetsanalysis.pThebranchingfractionintohadronicShasbAeenapplied0/37ffffffffYBac9kground<DescriptionfCrossTSectionΉffWW+8jjjj׽!ƻqjjjjôhasTiden9ticalsignaturetothesignal1Խ18Tpb
Z Z=
ƹ+8jjj׽!jjjbSleptonTisusuallyfoundasajet۠2.6TpbWHZ5!ƻqjjneedsTt9woextrajets(canbAegluonemission)۠0.2TpbWHWک!ƻqjjČneedsTt9woextrajet(canbAegluonemission)۠0.5TpbsingleTtopSsmallTcrosssection,buthasbjets۠0.5TpbQCDan9yT4-jeteventthatdoAesn'thavearealSleptoninitӝ>100Tn9bΉffffff(.Ec.
8Lightkjettaggingeciency
Thevb-tagvfakeratefromlightquarksiscomputedbymeasuringthenegativetagrate.vItisdenedbytherateofappGearanceofsecondaryverticeswithanegativedecaylengthsignicance.vItisassumedg~thatthelightquarkshaveanequalchancetoproGduceasecondaryvertexwithpGositiveandnegativedecaylengthƵsignicanceƶ(duetoniteresolutioneects)whiletheheavy
avorƶjetscanonlyproGduceaSVƘwithpGositivedecayRlengthRsignicance.pHowever,SFthisisRnotquitetrueandaspGecialscalingfactor(SFhf	ݧ)isintroGducedtocorrectforthefractionofheavy
avorsamongthejetswithanegativedecaylengthsignicance.hAnothercorrectionisfortheUUpresenceoflonglivedparticlesinlightjets(SFl
`l).qBothfactorsarederivedfromMonteCarlo.!C%ELIV.xBACCKGROUNDSTwomaindistinctivefeaturesofthesignaldeterminewhichbackgroundsareimpGortant.InordertobGerelevqant,theproGcessmusthaveahigh(>3)numbGerofjetsaswellasasizeable(>15GeV)6qȵETL.MTAllofthecandidateproGcessesareUUlistedinT*ableIV.qThecrosssectionslistedincludethebranchingfractionsintowleptons.W*e.canconcludethatthetwo.dominantbackgroundsourcesareQCD.{(\fake!Dz")andWc+4jets.dThesetwosourcesweretakenintoaccountinthisanalysis.wQCDisderivedfromthedata,-,whileMonteCarlosimulationisusedforWc+jets.This|simulation|utilizesevents|generatedatp!fe3荵s=1:96T*eV|withtheALPGEN1.2[10
]matrixelementgenerator,assumingatopmassof175GeV/c^2\andthepartondistributionfunctionsetCTEQ6.1M[11
].[NTheseeventsarethenproGcessedtthroughF-

cmcsc10FPythias6.2[12
].$Thispackqagetakestintoaccountthegluonradiationsandfragmentationeectsand)also*pGerformstheshortlivedparticledecays,eexceptforbhadronsandleptons.=EvtGen[13
]isusedtomoGdelthedecays'of'bhadrons.bIòleptonsaredecayed'usingT*auola[14
].bThegeneratedevents'arethenproGcessedthroughafullFgeantBj[15
]BksimulationoftheDdetectorprovidingtrackingBjhits,F3calorimetercellenergyandmuonhitinformation.Multiple	interactions	areaddedtoalleventsaccording	toaPoissondistributionwithameanthatisdeterminedfromtheaverageinstantaneousluminosity*.b]Thesamereconstructionandob8jectIDisappliedtodataandMonteCarloevents.qTheUUsignal(t'#btr)MonteCarloispreparedinthesameway*.V.>EVENTSELECTIONF*orthisanalysis,$Gweusedthedatacollectedwitha4-jettrigger,$Gwhichrequired4jetcandidateswithj[ٸj<3:6andETL>10GeV}usingasimpleconealgorithm.aThistriggerwasdesignedfortheall-hadronictopdecaymoGdeandworks!well!forourpurpGoses,,AsinceanenergetichadronicCòleptonisalways!foundasajetcandidate.`TheeciencyofthistriggerhadbGeenparametrizedandcross-checkedondataandisappliedasaweightingfactortotheMonteCarloevents.TheUUanalysisproGcedureinvolvedUUseveralstages:Preselectionqg(sectionqhVA).Atleast4jetsand6ET	signicance>3arerequired.653,727eventsqgareselectedintheUUdata,withapredictionof109.97.3t'#btDzevents,UUforaS:B1:6,000.IDzcutsz(sectionVB).nAtleastonegoGodzTleptonzcandidateandatleastonetightSVTztagarezrequired.oW*ealsorequired2jetswithj[ٸj<2:4andpT	ε>20GeV/c.K216eventsareselectedinthedata,~9.30.6t'#bt~amongthemUUareexpGected.qS:B1:58.T*opGologicalNN(sectionVD).5VAsequenceoftwofeed-forwardNN'shavebGeentrainedandapplied.5VTheoptimalcuto<onthesecondNNo5hasbGeenfoundtobe0.6.|Withthisnalcut,uweobtained13eventsindatawith4.90.3UUt'#btDzamongthemexpGected.qS:B1:2.5.8 25x@#FIG.T1:p6|rET:signicanceforQCD-dominateddata(blac9k),WW+8jets(blue)andtVۍ %t!ƻ7+jets(red).îT:ABLE]IAI:Preselectionresults.WSho9wnarethetotalacceptances(includingpreselection)andthenumbAerofeventsscaledto
349??23pb-=1
1(nosystematicuncertain9tiesexceptforthisluminosit9yerrorareincluded).^TheALPGEN?samplesgenerationcutsTaredescribAedin[16	?].pAnestimateofQCDbac9kgroundnotincluded.1~)(=ffffffffY(#Tpassed)/(total#)KR,Tpbl#TpassedscaledΉff!ndata653,727/17M̤Y653,727
􍍑tSaHWt!ƻ7+8jets:K6,141/10,878:z0.821T0.004
B109.93T7.26
WHbbjj׽!T+8bbjj:K2,321/11,576:z0.222T0.044ĭ@9.98T2.08ֻWHccjj׽!T+8ccjj:K2,289/10,995:z0.527T0.059]A24.77T3.22JWHcjjj׽!T+8cjjj:K2,169/10,435<$0.920T0.087]A42.23T4.87WHjjjj׽!T+8jjjj:K2,683/11,920?x14.14T1.3C720.33T81.48Ήffffff()|The$1WbackgroundhasbGeenmodeledusing$0ALPGEN#MonteCarlosimulation,Wwhile$0QCD#backgroundestimateswereextractedfromthedata	usingtheproGceduredescribedinsectionVC.]XInorderto	optimizetheselectionwealsousedxtheyt'#btMonteCarloandnormalizedittoknown(eithertheoreticalorcomputedbyALPGEN)bvqalues.T}IntheendweUUappliedthisoptimalselectiontomeasuretheactualt'#btDzcrosssectioninthedata.!8E԰A.
^PreselectionThee?totalnumbGere>ofe?eventsrecordedbye>ourtriggerinthis349pb^1!datasetis17million.Themaingoalofpreselectionwastoreducethisdataset,&whileimpGosingthemostobviousandstraightforwardrequirementsthatcharacterizeUUthesignalsignature.qSuchcharacteristicfeaturesincludethefollowing:@ۍMoGderateUU6qȵET	arisingfromboththeWvertexandwdecay*.fAtUUleast4jetsmustbGepresent.AtUUleast1wleptonand2bjetsarepresent.@ڍ6qȵETsignicanceK[17
]LisdenedasmeasureofthelikelihoGodofL6qȵET~arisingfromLphysicalsources,	Iratherthan
uc-tuationsHinGdetectormeasurements.*These
uctuationsarepredictedfrommeasuringandparametrizingtheob8jectresolutionsinthedata.AscanbGeobservedinFig.1,acutonthisvqariableprovestobGeaneectivewaytoreducetheUUdataset.qAcutof3wasusedforpreselection.Nowweneedtoscaletheoriginal10KeventsoftheMCsampleto349pb^1
t.<pThetotalt'#bt5Ucrosssectionis6.8pb[4].T*akingintoaccountthebranchingfractiontothehadronic~C+\|jetsmoGde,theeectivecrosssectioncomesouttoUUbGe:M|M/BqRDz(߸!hadrons%N:)8BRDz(t'#btc!Z+jets)[ٲ(t'#btr)=0:650:156:8=0:66UUpbTheGrelativeH
avorfractionsHoftheW
V+4jetsproGcessweretakenHfromALPGENsimulationasHratiosofthesimulatedcrossYsections.~ItwasYthennormalizedtothemeasuredtotalvqalueof4.52.2pb[18
].~T*ableIGIshowstheresultsoftheUUpreselectionforbGothdataandthebackgrounds.qB.
^ResultsoftheIDcutsThe&next'stepwastoapplytherequirementof!Dz-andb-tagging.F
T*ableIGII&showstheselectioncriteriathatweapplytodataandMCpandT*ableIVshowstheresultingselectioneciencies.ZTableVshowsthebreakdownoftheseeventsbGetweentype2and3V(type1isdisregardedduetoitssmallcontributiontothesignalandlargebackground).MuItcan
bGe
	notedthattheS:Batthisstageis1:58,~whichistoGosmall.YInsectionVDwewilldescribGethetopGologicalNNusedUUtoenhancethesignalcontent.AtUUthispGoint,weusetheseIDalgorithmstodene3subsamplesoutoftheoriginalpreselecteddatasample:@ۍThe\signal"sample-requireatleast1ϟleptonwithNNq>Z0:95andatleastoneSVTtag(asintableIGII).ThisUUisthemainsampleusedforthemeasurement.qItcontains268events.Q26T:ABLE	IAII:b-taggingandID.IntheMC,w9eusetheb-taggingcertiedparameterizationratherthanactualb-tagging,that
is,Uw9eappliedtheb-taggingweight.
W:ealsousedthetriggeringweightascomputedbythetriggereciencyparameterization. &iffffffffYdata
1taggingMC
kF1ȴwithj@Lj<2:4andpmTJ>20GeV/chF1ȴwithj@Lj<2:4andpmTJ>20GeV/c»F1SVT3Tr7ig@LWeight,bTagPr7obgF2jetswithj@Lj<2:4andpmTJ>20GeV/cdF2jetswithj@Lj<2:4andpmTJ>20GeV/cΉffffffRîT:ABLEIV:b-taggingandIDresults.GSho9wnarethetotalacceptances(includingpreselection)andthenumbAerofevents
scaledTto349pb-=1	.pAnestimateofQCDbac9kgroundisnotincluded.0)(=ffffffffYH(#Tpassed)/(total#)>3Acceptancel#TpassedscaledΉff!ndatam216/653,7279216
􍍑tSaHWt!ƻ7+8jetsٹ524.0/614140.0480T0.0020
B9.320T0.620
WHbbjj׽!T+8bbjj
ع54.5/232140.0150T0.0024
B0.012T0.002ֻWHccjj׽!T+8ccjj
ع13.3/228940.0039T0.0012
B0.034T0.005JWHcjjj׽!T+8cjjj]׹8.0/216940.0025T0.0010
B0.160T0.020WHjjjj׽!T+8jjjj]׹3.3/268340.0009T0.0006
B0.860T0.100Ήffffff()The\fvetosample"-Sameselection,butinsteadofNNo>o0:95,0o<NNo<o0:5wasrequiredforgleptoncandidates\andno\events\with\goGod"\(NN3>0:8)~rleptonswereallowed.Thissampleisused\forthetopGologicalNNUUtraining.qItcontains21,022events.TheD\bCveto"sample-atleast1
leptonwithNN>p0:95,ԿbutNO)SVT*tags.ThissampleistobGeusedfortheUUQCDprediction.qItcontains4,642events.+C%ZC.
^QCDmoKdelingThedierencebGetweenthetotalnumbGeroft'#bt]andW$eventsanddatahastobGeaccountedforbyQCD-initiatedevents,+wheregah.candidateisajet,mistakenlyidentiedhasa.lepton.Inordertoestimatethisbackgroundcontribution,UUthefollowingstrategywasemployed.W*estartedwiththe\bveto"sample.EThissampleisdominatedbymultijeteventsinwhichajetismisidentiedasJahadronic!Dz.R Sincethelcandidatesherearereallyjets,%wecanJsimplydividevspT
~distributionsofthecandidatesbythesamedistributionsforjetsbinbybintoparameterizethe<Mfakerate.^-Inordertoreducethiseectand'minimizethestatisticaluncertainty*,we'pGerformeda2Dttothis(distribution.=ThistwasthenusedfortheQCD,prediction.HTheTdistributionsinN-andUpT
?
haveTbGeenseparatelyttedwithA([ٲ)andBq(pTL).HThevqalidityofusingthenormalizedproGductofthesetwoeciencycurveshasbGeencheckedusingclosuretests.jTheresultofthisproGcedureWcanbeobservedinFig.@2.F*orWtypGeV3leptoncandidatesthe0:85:<j[ٸj:<1:1Wregionwasexcludedfromthe&tduetothepGoor&performanceoftheIDNeuralNetinthisregion.:Thenal2DparameterizationofthefakeUUrate(normalizedproGductof.andpT	ts)isshowninFig.q3.!C%1.	qComputingN<theQCDfractionW*eassumethatprobabilityforajettofakeaAdleptonissimplyFc([;pTL)=A()Bq(pTL)._ThentheprobabilitythatatUUleastoneofthejetsintheeventUUwillfakeawcanbGecomputedasthefollowing:kT:ABLE$V:$b-taggingand,ID$resultspAert9ypeafter$thewGcut(asexplainedinsectionVC).JJSho9wnarethenumbAer$ofevents
predictedTinsignalandobserv9edinthedata.pAnestimateofQCDbackgroundisnotincluded.!
nffffffffYJT9ypAeT2֜T9ypAeT3
!dataZb91@71
􍍑=tSaHWt!ƻ7+8jetsj5.61T0.372.81T0.18Wک!ƻ+8jetsj0.93T0.040.32T0.01Ήfffffff27)|{Ycmr8(a)	TÎyp<re2t-(b)	TÎyp<re3t8t1@FIG.T2:pFitofthegandpT:distributionsoftheSfak9erate.[
(a)	TÎyp<re22Dt
(b)	TÎyp<re32DFitG@oFIG.T3:pThe2Dcom9binedt(ingandpT)oftheSfakerate.W1Pev@Lentв=18QjU(1Fc(p1ɍjTL;[ٟ^j)):lByUUsummingupsuchprobabilitiesoverthetaggeddata,weobtaintheQCDbackgroundestimate.Using.theresults-describGedintheprevioussection,"#wegetNQC}D7=*71:131:56.forthetypGe2andNQC}D7=77:46>?0:80forthe(typGe3,3Gwhichagreeswiththeobserveddata(inT*ableV)fairlywell.rOnecanalsoobserve(seeWAppGendix)WthatthepredicteddistributionsofthemaintopGologicalvqariables(sectionVD)areinfairlygoGodagreementUUwithwhatisobservedinthedata. 粍D.
^TopKologicalNNSimilarlyGtoHtheall-jetsanalysis[5],wedene2neuralnetworksG(NN)+(theMultiLayerGPerceptron[19
]GprogramwasUUused).qThetwoNNscontain'1.3UUtopGological(aplanarity*,sphericityandcentrality)and2energy-based(HT	andpfe3荵sZ)vqariables.'2.theUUoutputoftherstNN,WandtmasslikelihoGods,UUpT	anddecaylengthsignicanceofthebjets.lTheUUkinematicandtopGologicalvqariablesusedare:HT	-UUthescalarsumofalljet'spT(andwleptoncandidates).SphericityأandآAplanarity[5]آ-thesevqariablesareformedfromtheeigenvqaluesofthenormalizedMomentumT*ensorpofothejetsintheevent./ThesearepexpGectedtobGehigherinthetoppaireventsthanpinatypicalQCDevent.Centrality*,UUdenedasHmT7feWQHmEha,whereHE
²issumofenergiesofthejets.)ˍT*op$and$W$masslikelihoGod-a$^2|s-likevqariable.aZL`텍	/M3j#mt	/ŭfemʟt*,`0ݱ26ֲ+T`텍	kM2j#MmW	kŭfe$gx̴mW.`4ݱ29m,-wheremtV;MW	;t;W<are$topandCW5masses(175GeVand80GeV4respGectively)Dandresolutionvqalues(45GeVand10GeVrespGectively)[5].M3jtandUUM2jareinvqariantmassescompGosedofthejetcombinations,soastominimizeL.pT	andUUlifetimesignicanceoftheleadingb-taggedjet.lMany~ofthesevqariables(forinstancethemasslikelihoGod~and~aplanarity)areonlydenedforeventswith2ormorejets.pThusQwerequire2jetswithpT	ε>20GeV/cQandj[ٸj<2:5.pPlotsofsomeofthesevqariables,RwhichalsoservesasanUUadditionalcheckUUoftheagreementbGetweenthedataandprediction,areincludedintheappGendix.The5resultofapplyingthese5NN5todataisshowninFig.g<4.The5maximumsignal5signicanceisusedtodeterminetheoptimalNN{cut.8oThesignalsignicanceisdenedasENmS7fe$Ti捍pɟW	 noNmS+NmB*R\,whereNS	bandNB
yaretheexpGectednumbersof
bgsignalandbackgroundeventsrespGectivelyandisshowninFig.=5.ThissignicancereachesitsmaximumatNN3>0:9forΐbGothtype2and3.DTherefore,thisselectionisusedfortheΏcrosssectionmeasurement.TheresultsaresummarizedinUUT*ableVI.|[28*:t
(a)	TÎyp<re2+2cmmi8lep-ton
(b)	TÎyp<re2lep-ton(zo<romed)P
(c)	TÎyp<re3lep-ton
(d)	TÎyp<re3lep-ton(zo<romed)\KFIG.y4:"ResultofapplyingthexNNWcut.޻tVۍ %tW,BW\andQCDVareplottedincremen9tallyinordertocomparewiththen9umbAeryof
ev9entsobservedindata.0Errorbarsincludeonlystatisticaluncertain9ties.0ܻR(tVۍ %tW)=5:54pbisassumed.0Therigh9tplotonlysho9wsTtheentriesathighv|raluesofNNcut.;E.̍
d(a)	dTÎyp<re2
yc(b)	ycTÎyp<re3,Z4lFIG.T5:ptVۍ %t!ƻ7+8jetssignalsignicance.U󍍒DVI.SYSTEMATICUNCERTAINTIESTheBmostAimpGortantsystematiceects(exceptfortheb-tagging,FwhichistreatedindepGendently)aresummarizedinT*able_VIGI.=uMostofthemareassociatedwithuncertaintiesonthecorrespondingquantities(Jet^EnergyScale,PrimaryV*ertex,'Branching
RatioandTriggersimulation).^TheQCDsystematicscomesmainlyfromtheerrorofthe2Dtofthe3nU5fakerate.fzTheW*!!duncertaintycomes3mfromtwosources:`theuncertaintyof3mthecrosssectionvqalueandtheconservqativeUUestimateoftheerrorassoGciatedwiththemethodthatweusedformodelingthisbackground. nDA.
^b-taggingTheeectsofuncertaintiesintaggingbjetsaretakenintoaccountbyvqaryingthesystematicandstatisticaluncertaintiesontheMC1taggingweights.V^TheresultingeectofalloftheerrorsourcesonthenalnumbGerissummarizedUUintableVIGII,UUalongwiththetotalbIDsystematicuncertaintyUU(quotedintableVIGI). neQVIKI.^{CRCOSSSECTIONTheUUcrosssectionisdenedas"=)&N,umberof&sig@LnalheventsK);feymJ"(tU)ȱJtV)BW=R	,(tU)ȱJt!+jgets)Luminosity~Ȳ.qTheresultsarethefollowing:;!Dz+jetsUUtypGe2(singleprong)crosssection:kpn@[ٲ(t7ferYɍtr)=3:63䍑0+4:7203:50 1(stat)䍑0+0:4900:48(syst;)80:240(lumiq)pb;҄gT:ABLEVI:ThenalresultsummaryaftertheNN>ƹ0:9cut.(tVۍ %tW)isthetotalsignalacceptance.Uncertain9tiesarestatistical
only:.,xMخffffffffnChannelNN-=obs{=BUzi		cmex9ŖLdt;pb-=1B{Bac9kgroundsQ)&h#ff^؉"(tVۍ %tW)T(%)sT(7pb)6s8+bΉffY
1+jetsTt9ypAe2VV5xK0.1~349.3RWک!ƻ0.60T0.03Yh1.57T0.01A3.83+0:468Q0:516.84+0:468Q0:51
Yfak9es2.41T0.09+jetsTt9ypAe3VV5xK0.1~349.3RWک!ƻ0.27T0.01Yh0.73T0.01A1.80+0:228Q0:234.39+0:228Q0:23Yfak9es2.33T0.09Ήffffff	ˠ29񍟜
}ӂT:ABLETVIAI:Systematicuncertain9tiesonR(tVۍ %tW)(inpb).6&kffffffffY-ChannelPr+jetsXtÎyp<re2Tr+jetsXtÎyp<re3
HJetXEnergyScale
:ޟq0:27;+0:30Dq0:69;+0:53!PrimaryXVJertex	7q+0:037;0:036q+0:095;0:093.H*MCXStat
:ޟq+0:25;0:22Dq+0:65;0:581TJriggerq0:020;+0:0025GIq0:069;+0:0056BrancÎhingXRatio	7q+0:074;0:071Dq+0:19;0:18QCDXfakÎerateParameterization
:ޟq+0:17;0:17Dq+0:34;0:34-Wک!ƻ
:ޟq+0:19;0:19Dq+0:19;0:19ΉffffffT:ABLETVIAII:b-taggingsystematicssources;?kffffffffY.Channel:r+jetsXtÎyp<re2Wr+jetsXtÎyp<re3
ub-taggingt0ncmsy50:13;+0:076qt0:26;+0:41bc-taggingTyt0:20;+0:16qt0:48;+0:60(=l-tagging9t0:0051;+0:0051
yt0:014;+0:014GlSFhft0:00036;+0:00036<9t0:00094;+0:0009412SFllt0:00036;+0:00036<9t0:00094;+0:00094Tb-tagging(data)Yt0:091;+0:094qt0:24;+0:25iTb-tagging(MC)Tyt+0:11;0:10qt+0:28;0:25Mtaggabilit9yYt0:048;+0:049qt0:13;+0:13Ήffffff2˵!Dz+jetsUUtypGe3(multi-prong)crosssection:l'?[ٲ(t7ferYɍtr)=9:39䍑0+10:1007:49$(stat)䍑0+1:2501:18 1(syst;)80:610(lumiq)pb:The1combinedcrosssectionwasestimatedbyminimizingthesumofthenegativelog-likelihoGod1functions(con-structedfromaPoissondistributionusingthenumbGerofobservedeventsand\hypGothetically"observedeventsasthePoissonmean)foreachchannel.UUThefunctionalformofthelikelihoGodfunctionwasthesameashasbGeenusedfortheeUUchannel[20
].qThecombinedcrosssectionyieldsn@[ٲ(t7ferYɍtr)=5:05䍑0+4:3103:46 1(stat)䍑0+0:6800:67(syst;)80:330(lumiq)pb:!C%JiVIKII.ߓSUMMARYThis-notepresents,themeasurementof[ٲ(t'#btr),bytheDcollabGoration,intheRun2aoftheF*ermilabTevqatron.WThedecayb9channelstudiedinvolvesonehadronicallydecayingb8lepton,er2bjets,2lightjetsand6CETL.sThetriggerandthecorrespGondingUU349823pb^1ɲdatasetweresharedwiththeall-jetschannel[5].The/mainchallengewastorejecttheoverwhelmingQCD
background,whileatthesametimepropGerlyhandlingthephysical	Wײ+I4jirreduciblebackground.SXInordertoachievethisgoal,KaNN-basedϲIDandSVTb-taggingalgorithmwasemployed.Inaddition,ItherelevqanttopGologicalvqariableswerecombinedintoaNNtrainedtodierentiatethesignal>fromQCD.=Thesingleandtripleprongedleptonchannelswere>treatedasindepGendentchannels>andthencombined.TheUUcrosssectionwasmeasuredtobGen@[ٲ(t7ferYɍtr)=5:05䍑0+4:3103:46 1(stat)䍑0+0:6800:67(syst;)80:330(lumiq)pb:Thisresultisinagreementwiththetheoreticallypredictedvqalueof6.80.4pb[4]aswellasotherDandCDFmeasurementsUU[5,16
UW].
Ǡ210j)@AcCknowledgmentsW*evthankvthestasatF*ermilabandcollabGoratinginstitutions,~andacknowledgevsuppGortfromtheDOEvandNSF(USA);1CEAandCNRS/IN2P3(F*rance);FASI,RosatomandRFBR(Russia);CAPES,0CNPq,hFAPERJ,FAPESPand/FUNDUNESP(Brazil);<DAEand0DST(India);Colciencias/(Colombia);CONACyT(Mexico);KRFand/KOSEF(Korea);CONICETandUBACyT(Argentina);FOM(TheNetherlands);PP*ARC(UnitedKingdom);MSMT(CzechRepublic);yCRCbProgram,سCFI,NSERCandW*estGridPro8ject(Canada);yBMBFcandDFG(Germany);xSFI(Ireland);ResearchUUCorpGoration,AlexandervonHumbGoldtF*oundation,andtheMarieCurieProgram.r.@)0  fz@=`򎷹[1]RAJ.TA.Coarasa,J.Guasc9h,andJ.Sola,RepAortNo.UAB-FT-451,1999,hep-ph/9903212.
[2]RAD.TAcostaet.N<al.,Ph9ys.Rev.D62,012004(2000).[3]RADirectTSearc9hforChargedHiggsBosonsintVۍ %tloevents,DNote4357.[4]RAN.TKidonakis,R.V:ogt,Ph9ys.Rev.D68,114014(2003).[5]RAMeasuremen9tTofthetVۍ %t!ƻal&9l4q8jetsproAductioncrosssection,usingSecondaryV:ertexTagging,DNote4879-CONF.[6]RAV.M.TAbazo9v,etN<al.(D0CollabAoration),\TheupgradedD0detector,"Nucl.Instrum.andMethodsA565,463(2006).[7]RAS.TEidelmanetN<al.(P9articleDataGroup),Phys.Lett.B592,1(2004)and2005upAdate.[8]RADTSIDcertication,DNote4773.[9]RADTSecondaryV:ertexb-taggercertication,DNote4796.[10]RAM.L.TManganoetN<al.,JHEP0307,001(2003),hep-ph/0206293.[11]RAJ.TPumplin,D.R.Stump,J.Huston,H.L.Lai,P:.Nadolsky,W.K.Tung,hep-ph/0201195.[12]RAT.TSj`ostrandetal.,LUTP01-21,hep-ph/0108264.[13]RAGߤN		cmtt9Ghttp://c9harm.physics.ucsb.edu/pAeople/lange/EvtGen/[14]RAS.TJadac9h,Z.W:as,R.Decker,J.H.Kuehn,Comp.Phys.Commun.,361(1993)(CERNTH-6793preprint).[15]RAR.TBrunandF.Carminati,CERNProgramLibraryLongW:riteupW50131993.[16]RAMeasuremen9tpNofthepOttbarcrosssectioninthelepton+jetsc9hannelatؽp%؉aHO|(s"=)g1:96T:eVp7(topAological),DNotes4667andRA4662.[17]RA6|rETkxSignicanceTAlgorithminRunIAIdata,DNote4254[18]RAMeasuremen9tJofWH-=
!?e-=Gބ+Inclusiven-jetCrossSectionswithCDFIDataatT:ev|ratronRunIAI,http://www-RAcdf.fnal.go9v/physics/new/qcd/wjets/WjetffgTyana/Wjetffgana.psT.[19]RAh9ttp://schwind.home.cern.ch/schwind/MLPt.htmlT.[20]RAC.Clemen9tVvetal.,fT:opquarkVuproAductioncrosssectioninthelepton+trac9kchannelusingVusecondaryvertexb-tagging,fDRANoteT5011.!C%ˮAppKendixA:KinematicDistributionsFIG.6:2ofthe5inputv|rariablesofthersttopAologicalNNbeforetheNNcut(t9ype2).kTheKolmogoro9v-Smirnov(KS)probabilitiesTaresho9wn,indicatingthequalityoftheagreement.$@h4*FIG.T7:pTheresultingoutputofthesecond(nal)NN(St9ypAe2).211x@h4*FIG.T8:pTheresultingoutputofthesecond(nal)NN(St9ypAe3).;
GߤN		cmtt9F-

cmcsc10E':

cmti10D<x

cmtt109#cmex7,!",
cmsy10+2cmmi8*gcmmi12)|{Ycmr8(Ncmbx12'XQcmr12&j		cmti9t:		cmbx9zi		cmex9q%cmsy6		cmsy9;cmmi65"		cmmi9Aacmr6o		cmr9
!",

cmsy10O!cmsy70ncmsy5
b>

cmmi10	0ercmmi7O
\cmmi5K`y

cmr10ٓRcmr7Zcmr5u

cmex10

FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>