Latest Results from NA48 and NA481

M.W.SLATER

DepartmentofPhysics,HighEnergyPhysics,CavendishLaboratory,

MadingleyRoad,CambridgeCB30HE,England

ThefirstobservationsoftheraredecaysKS→π0e+e−andKS→π0µ+µ−havebeenmadebytheNA48/1collaborationattheCERNSPSaccelerator.FromhighintensityKSdatacollectedduringthe2002run,cleansignalsof7KS→π0e+e−eventsand6KS→π0µ+µ−

eventswereobserved,givingbranchingratiomeasurementsofBR(KS→π0e+e−)=5.8+2.9

10−9andBR(K2.9+1−S→π0µ+µ−)=.5−9

2.4×

−1.2×10

.TheseresultsconstraintheindirectCPviolatingcomponentofthecorrespondingKLdecays.OtherrecentresultsfromNA48arealsopresented.

1Introduction

TheNA48experimentwasoriginallydesignedtomeasuretheCPviolationparameterRe(ǫ

′

No. Events908070605040302010

2002 dataππ D (mee > 0.09 GeV)ππ + conversionsππ D + conversion kevlarππ D + conversion DCH1ππ(ee)0 0ππ DD0000000000mK /GeV0.650.60.550.50.450.40.35

0.022001 DATA+-KL → eeγγ candidates0.040.060.080.10.120.140.160

0.090.10.110.120.130.140.150.16mee/GeV

mγγ /GeV

Figure1:(a)meedistributionoftheDalitzdecayandconversionbackground.Reasonableagreementbetweendata(points)andMonteCarlo(curves)canbeseen.(b)2001datashowingthedistributionoftheKL→eeγγ

backgroundacrossthesignalregion.

2

PhysicsMotivationforKS→π0l+l−Searches

ThedecaymodeKL→π0νν¯isdirectCPviolatingandcanbeusedtodeterminetheparameterηintheWolfensteinparameterisationoftheCKMmatrix.Thischannelistheoreticallyverycleanbutisexperimentallychallengingduetothemissingenergyfromtheneutrinos.AsecondpossibilitytomeasureηisfromthedecayKL→π0l+l−.Thisdecayiseasiertomeasureexperimentallyasallthedecayproductsaredetectable,butismorecomplicatedtheoreticallyasithascontributionsfromCPconservingandbothdirectandindirectCPviolatingcomponents,whichalsointerfere.TheCPconservingcomponentcanbepredictedfromameasurementofKL→π0γγwhiletheindirectCPviolatingcomponentcanbepredictedfromameasurementofKS→π0l+l−.AnymeasurementofBR(KL→π0l+l−),andconsequently,η,requiresthemeasurementofBR(KS→π0l+l−)todisentanglethedirectfromtheindirectCPviolatingcomponents.3

AnalysisStrategyforKS→π0e+e−andKS→π0µ+µ−

BothKS→π0e+e−andKS→π0µ+µ−werepredictedtohaveverysmallbranchingratios,(∼10−9)andso,evenwithafluxof∼3×1010KSdecays,onlyahandfulofeventswereexpectedfromeitherchannel.Therefore,toavoidbiasingtheresults,ablindanalysisprocedurewasemployed.A2.5σmK×2.5σπ0signalregionanda6.0σmK×6.0σπ0controlregionweredefined.Bothregionswerekeptmaskedduringthebackgroundstudies.Onlyaftertheback-groundcontributionstothesignalhadbeenestimatedandthecutsfixedwasthecontrolregionunmasked.Finalselectionchangescouldthenbemadeifrequiredbeforethesignalregionwasunmasked.44.1

KS→π0e+e−SignalSelection

KS→π0e+e−candidateswereselectedwith400.95andformingagoodvertex,wererequired,aswellastwoclustersinthecalorimeterthatwerenotassociatedtotracks.

mK /GeVmK /GeV0.650.60.65

0.50.6

0.490.1340.1360.550.50.450.40.350.550.50.450.40.35

0.10.150.20.250.30.350.10.110.120.130.140.150.160.17mee /GeVmγγ /GeV

Figure2:The(a)mγγvs.meeand(b)mγγvs.mKmassplanesshowingKS→π0e+e−candidateevents.The

largebackgroundfromdalitzdecaysandphotonconversionscanbeseenatlowmee

4.2Background

BackgroundsfortheKS→π0e+e−decayweredividedintotwocategories:physicalbackground(arisingfromasinglekaondecay),andaccidentalbackground(arisingfromtwoseparate,over-lappingkaondecays).Threesignificantsourcesofphysicalbackgroundwereidentified.ThefirstcamefromDalitzdecays(π0→eeγ)andphotonconversioninthevariouspartsofthedetector.ThesebackgroundswerestudiedextensivelyusingMonteCarlosimulation(seefig.1a).Reason-ableagreementwithdatawasfoundinthemeedistributionaroundtheπ0mass.Aconservativecutofmee>0.165GeV/c2wasappliedtoremovethisbackground,withacorresponding48%lossinacceptance.

ThesecondsignificantphysicalbackgroundwasidentifiedasKL→eeγγ.Thisbackgroundspreadacrossthemγγsignalregionascanbeseeninfig.1b.UsingKLdatatakenin2001,the

.03

backgroundwasestimatedtobe0.08+0−0.02events.

Thefinalphysicalbackgroundcamefromvarioushyperondecaychannels.Thebeamcontainedasignificant(∼109)fluxofneutralhyperonsandthechannelsΞ0→Λ(pπ−)π0,Ξ0→Λ(pe−ν)π0andΞ0→Σ+(pπ0)e−νwereidentifiedaspossiblebackgroundsources.Acutonthemomentumasymmetryofthedecayproductsreducedthesebackgroundstoanegligiblelevel.

Theaccidentalbackgroundwasdominatedbyoverlappingfragmentsoftwodecays,e.g(KL→π±e∓ν)+(KS→π0π0).Thisbackgroundwassuppressedbycuttingonthespreadoftrackandclustertimes(|∆t|<3ns).Toestimatetheremainingaccidentalbackground,thetimingcutswererelaxedandthetimesidebandsinvestigated.Noeventswerefoundintheoutoftime(3ns<|∆t|<50ns)signalregion.Therefore,anextrapolationwasmadefromtheoutoftimecontrolregiontotheoutoftimesignalregion,takingintoaccounttheshapeofthebackgroundfromMonteCarlosimulation.Afurthertimeextrapolationcouldthenbeperformedtogetthefinalbackgroundestimatefortheintimeregion(|∆t|<3ns)of0.069events.4.3

Result

.05

Theoverallbackgroundestimatewas0.15+0−0.04events.Thecontrolregionwasunmaskedandnoeventswerefound.Nochangestotheselectionwereneededbeforethesignalregionwas

Events120100806040200DataKL → π0π+π– MC (no cτK)0+–KL → πππ MC (all cuts)mγγ (GeV)0.170.160.150.140.130.120.11

0.440.460.480.50.520.0.420.440.460.480.5mµµπ (GeV)

mµµπ(GeV)

Figure3:(a)ThekaonmassregionwiththeKL→π0π+π−background.Reasonableagreementbetweendataandmontecarlo(24timesthedata)isshown.(b)DistributionofoutoftimeeventsinthemKvs.mγγplane.

unmasked.Asignalof7eventswasfound,givingtheresult:2

.8−9

BR(KS→π0e+e−)=(5.8+2−2.3(stat)±0.8(syst))×10

(1)

wherethesystematicerrorisdominatedbyuncertaintiesarisingfromtheextrapolationtothe

fullmeeregion.Thecandidateeventsareshowninfig.2.55.1

KS→π0µ+µ−SignalSelection

KS→π0µ+µ−candidateswereselectedwith60Background

AswithKS→π0e+e−,thebackgroundsweredividedintophysicalandaccidentalsources.Threedominantsourcesofphysicalbackgroundwereidentified.ThefirstwasKL→π0π+π−wherethetwochargedpionshaddecayedinflight.ThisbackgroundwasstudiedextensivelyinMonteCarlo,(seefig.3a)andestimatedtobe≤0.019events.ThesecondsignificantbackgroundwasKL→µ+µ−γγ.This,again,wasstudiedwithMonteCarloandestimatedtocontribute0.04±0.04eventstotheoverallbackground.Thelastphysicalbackgroundcamefromhyperondecays.AswithKS→π0e+e−,acutonthemomentumasymmetryofthedecayproductswasusedtoreducethisbackgroundtoanegligiblelevel.

TheaccidentalbackgroundwastreatedsimilarlytoKS→π0e+e−,bylooseningthetimingcutsandextrapolatingintotheintime(|∆t|<1.5ns)region.6eventswerefoundintheoutof

.18

time(−115ns<∆t<60ns)signalregion,leadingtoabackgroundestimateof0.18+0−0.11events.Thedistributionoftheseeventsinthemassplaneisshowninfig.3b.5.3

Result

.19

Theoverallbackgroundestimatewas0.22+0−0.12events.Thecontrolregionwasunmaskedandnoeventswerefound.Nochangestotheselectionwereneeded,thesignalregionwasunmasked

mγγ (GeV)0.160.150.140.130.120.110.20.250.30.350.4mγγ (GeV)0.170.170.160.150.140.130.120.11

0.440.460.480.50.520.mµµ (GeV)mµµπ(GeV)

Figure4:The(a)mγγvs.meeand(b)mγγvs.mKmassplanesshowingKS→π0µ+µ−candidateevents.TheeventsatlargemγγareconsistentwithKL→π0π+π−backgroundandthesingleeventatlowmγγisconsistent

withaccidentalbackgroundand6eventswerefound.Thisledtotheresult:.5−9

BR(KS→π0µ+µ−)=(2.9+1−1.2(stat)±0.2(syst))×10

(2)

wherethesystematicerrorcomesfromuncertaintiesinthenormalisation,obtainedfromKS→

π+π−decays.Thecandidateeventsareshowninfig.4.6

InterpretationoftheKS→π0l+l−Measurements

TheformfactorforKS→π0l+l−decayscanbedescribedinChiralPerturbationTheorybyafirstorderpolynomial:

W(z)≃GFm2K(aS+bSz)

(3)

20+−wherez=m2ll/mK.ThisleadstothefollowingpredictionsfortheKS→πllbranching

ratios:3

2−10

BR(KS→π0e+e−)=[0.01−0.76aS−0.21bS+46.5a2S+12.9aSbS+1.44bS]×102−11BR(KS→π0µ+µ−)=[0.07−4.52aS−1.50bS+98.7a2S+57.7aSbS+8.95bS]×10

(4)(5)

TheVectorMesonDominance(VMD)modelpredictstheratiobS/aS=0.4,allowing|aS|tobeextractedforbothchannels:

BR(KS→π0e+e−)≃5.2×10−9a2SBR(KS→π0µ+µ−)≃1.2×10−9a2S

.26

⇒|aS|π0ee=1.06+0−0.21±0.07.38⇒|aS|π0µµ=1.55+0−0.32±0.05

(6)(7)

Theseresultsfor|aS|agreewithinerrors.

Bycombiningtheeeandµµresultsinalog-likelihoodfit,aSandbScanbedeterminedseparately.Ascanbeseenfromfig.5,theobservedKS→π0e+e−andKS→π0µ+µ−ratesarecompatiblewithbotheachotherandtheVMDmodel.

20

(b)

10

bS/aS=0.4

bS0

-10

-20-5

-4

-3

-2

-1

0

1

2

3

4

5

aS

Figure5:(a)AllowedregionsofaSandbSdeterminedfromtheobservednumberofKS→π0e+e−andKS→π0µ+µ−eventsseparately.Theregionbetweentheinnerandouterellipticalcontoursistheallowedregionat68%CL.(b)AllowedregionsofaSandbSfortheKS→π0e+e−andKS→π0µ+µ−channelscombined.Thecontoursdelimitthe1σand2σallowedregionsfromthecombinedlog-likelihood.Thedashedstraightlinein

bothplotscorrespondstobS=0.4aS,aspredictedbytheVMDmodel.

7

ImplicationsforKL→π0l+l−

TheCPviolatingcomponentoftheKL→π0l+l−branchingratioisgivenby:

BR(KL→πll)CPV×10

0+−

12

=CIND±CMIX

󰀂

Im(λt)

10−4

󰀃2

(8)

0+−whereCDIRisthedirect󰀁CPVcomponent,CIND∼BR(KS→πll)istheindirectCPV

component,CMIX∼

NA48 (PHOTOS MC)no systematic errorNA48 (FFS MC)no systematic errorNA48(MI MC)KTEV

NA31

FFSDonsel4ChPT O(p )

0.8750.90.9250.950.97511.0251.051.075

o

BR(K , 20 , 30 MeV)/BR(K )e3γe3

Figure6:ThemeasuredvalueofBR(K0→π±e∓νγ)/BR(K0→π±e∓ν)comparedwiththepublishedresult

fromtheKTeVcollaborationandtheoreticalpredictions.

Thisisingoodagreementwithrecenttheoreticalpredictions,(Fearing,FischbachandSmith(FFS)7,8,Doncel9andChiralPerturbationTheorycalculations10,11),ascanbeseenfromfig.6.IftheFFSmethodisusedtoapplytheradiativecorrections,asintheKTeVanalysis,theresultisfoundtobeinagreementwiththeKTeVmeasurement.9

KL→π0π±e∓ν

AhighprecisionmeasurementofthebranchingratioandformfactorsfortheKL→π0π±e∓νdecaywasmadeusingKLdatafrom2001.ThisdecayprovidesagoodtestofChiralPertur-bationTheory(CHPT)predictionsforlongdistancemesoninteractionsandtheformfactormeasurementsallowthedeterminationoftheCHPTparameter,L3.12,13

ThemainbackgroundtothisdecaywasKL→π0π+π−whereapionwasmisidentifiedasanelectron.Tominimisethisbackground,aneuralnetworkwasusedtodistinguishbetweenpionsandelectronsthattookintoaccountgeometriccharacteristicsoftheshowersandtracks.Thisreducedthebackgroundto1.13%.

Asignalofeventswasfoundwithabackgroundof62events.Thisgaveapreliminarybranchingratiomeasurementof:

BR(KL→π0π±e∓ν)=(5.21±0.07stat±0.09syst)×10−5

(13)

AssumingaV-Astructureforthematrixelement,theformfactorparameterswerethenmea-suredtobe:

¯fs=0.052±0.006stat±0.002syst¯fp=−0.051±0.011stat±0.005systλg=0.087±0.019stat±0.006syst¯h=−0.32±0.12stat±0.07syst

ThisledtothefollowingvalueforL3:

L3=(−4.1±0.2)×10−3

10

Ξ→Λγ

(14)

Usingdatafromashorttestrunfrom1999withahighintensityKSbeamalone,730Ξ→Λγeventswerefoundwithabackgroundof58.2±7.8events.Thisledtoabranchingratiomeasurementof:14

BR(Ξ→Λγ)=(1.16±0.05stat±0.06syst)×10−3(15)Thestatisticswerelargeenoughtoenableasignificantmeasurementofthedecayasymmetry

tobemade:

α(Ξ→Λγ)=−0.78±0.18stat±0.06syst(16)Thisisthefirstevidenceforanon-zeroasymmetryinthisdecaymode.Amuchlargersample

fromthe2002runiscurrentlyunderanalysis.Acknowledgments

Itisapleasuretothankthetechnicalstaffoftheparticipatinglaboratories,universitiesandaffiliatedcomputingcentresfortheireffortsintheconstructionoftheNA48detectorapparatus,intheoperationoftheexperiment,andintheprocessingofthedata.References1.2.3.4.5.6.7.8.9.10.11.12.13.14.

A.Laietal.,Eur.Phys.J.C22231(2001).J.R.Batleyetal.,Phys.Lett.B576,43(2003).

G.D’Ambrosio,G.Ecker,G.IsidoriandJ.Portoles,JHEP9808004(1998).

M.Battagliaetal.,“TheCKMmatrixandtheunitaritytriangle”,hep-ph/0304132TheKTeVCollaboration,Phys.Rev.D,112004(2001).

E.BarberioandZ.Was,Comput.Phys.Commun.79,291(1994).H.Fearingetal.,Phys.Rev.Lett.24,1(1970).H.Fearingetal.,Phys.Rev.D2,52(1970).M.G.Doncel,Phys.Lett.B32,623(1970).B.R.Holstein,Phys.Rev.D41,2829(1990).J.Bijnensetal.,Nucl.Phys.B369,81(1993).

G.Colangelo,J.Gasser,andH.Leutwyler,Phys.Rev.Lett.86,5008(2001).G.Amoros,J.BijnensandP.Talavera,Nucl.Phys.B585,293(2000).A.Laietal.,Phys.Lett.B584,251(2004).