Rare Decays in LHCb DiegoMart´ınezSantos,a,onbehalfoftheLHCbCollaboration EuropeanOrganisationforNuclearResearch(CERN),Geneva,Switzerland. Abstract. TherareBdecaysB0 →µ+µ−,B0→K∗0µ+µ− andB0→φγ arestudiedusingupto∼0.41 fb−1of (s) s √ ppcollisionsat s=7TeVcollectedbytheLHCbexperimentin2010and2011.AsearchforthedecaysB0 → (s) 2 µ+µ− isperformedwith0.41 fb−1.TheabsenceofsignificantsignalleadstoB(B0s → µ+µ−)< 1.4×10−8 and 1 B(B0→µ+µ−)<3.2×10−9at95%confidencelevel.Theforward-backwardasymmetry,fractionoflongitudinal 0 polarizationanddifferentialbranchingfractionof B0→ K∗0µ+µ− ,asafunctionofdimuoninvariantmass,are 2 measuredin0.31 fb−1.TheratioofbranchingratiosoftheradiativeBdecaysB0→K∗0γ andB0→φγ hasbeen s measuredusing0.34 fb−1.Theobtainedvaluefortheratiois1.52±0.14(stat)±0.10(syst)±0.12(f /f ).Using n s d theHFAGvalueforB(B0→K∗0γ),B(B0→φγ)hasbeenfoundtobe(2.8±0.5)×10−5. a s J 5 2 1 Introduction for the background. Specific vetoes are used in order to eliminatenoncombinatorialbackground. ] The LHCb experiment [1] has provided preliminary re- Thetrigger,reconstructionandofflineselectioncanall x sults in the measurement of the forward-backward asym- bias the measured angular distribution of B0 → K∗0µ+µ− e - metry,fractionoflongitudinalpolarizationanddifferential candidates. The detection acceptance is accounted for by p branchingfractionofB0→ K∗0µ+µ− [2]andthemeasure- weighting events when fitting for AFB, FL and dBF/dq2 e mentoftheB(B0→ φγ)[3].LHCbhasalsoprovidedup- (whereq2isthedi-muonmasssquared).Eventweightsare h s perlimitsinB(B0 →µ+µ−)andB(B0 →µ+µ−)[4].Sect.2 calculatedonaper-eventbasisinasmallphasespacewin- [ s sumarizestheanalysisandresultsobtainedbyLHCbinthe dow around each candidate, using fully simulated Monte 1 study of B0→ K∗0µ+µ− . Sect. 3 sumarizes the measure- Carlo (MC) simulation events. Simulated events are re- v mentofB(B0→φγ)/B(B0→ K∗0γ)andSect.4sumarizes weightedtoaccountforknowndata-MCdifferencesinPID 9 theanalysisasndresultsofB0 →µ+µ−. performance, impact parameter resolution, tracking effi- 5 (s) ciencyandtrackmultiplicity. 3 ThefitresultsforA ,F anddBF/dq2,andtheircom- 5 FB L 1. 2 B0→ K∗0µ+µ− parisonwiththeoreticalpredictions[7],areshowninFig.1. The systematic error on A , F and dBF/dq2 is typ- 0 FB L ically ∼ 30% of the statistical error. In the high-q2 re- 2 Theraredecay B0→ K∗0µ+µ− isab → s,flavourchang- 1 gion, the dominant contribution to the systematic uncer- ing neutral current decay, mediated by electroweak box : tainty comes from the overall uncertainty on the accep- v andpenguindiagramsintheStandardModel(SM).Inmod- tancecorrectionwhichisdictatedbythelimitedsimulation i els beyond the SM, new particles can enter in competing X statistics.Thiscanclearlybeimprovedforfutureanalyses. loop-orderdiagramsresultinginlargedeviationsfromSM Throughout,asub-dominantcontributioncomesfromthe r predictions(seeforexampleRefs.[5,6]). a data-derived performance corrections. In particular, from B0→ K∗0µ+µ− candidatesareselectedbyfirstapply- knowledgeofthePIDperformanceandtrackingefficiency ingaloosepre-selectionbasedonthe B0 lifetime,daugh- in data. This is again statistically limited and can also be terimpactparametersandarequirementthattheB0points improved with larger datasets. When fitting for A and FB backtooneoftheprimaryverticesintheevent.Atighter F thesignalandbackgroundmassmodelandtheangular L multivariate selection, based on a boosted decision tree modelforthebackgroundhavebeenvariedandyieldcor- (BDT), is then applied to select a clean sample of B0 → rectionsatthelevelof10-20%ofthestatisticaluncertainty. K∗0µ+µ− candidates,withasignal-to-backgroundratioin The uncertainty on the differential branching fraction in- a 100MeV/c2 window around the reconstructed B0 mass cludes the ∼ 4% uncertainty coming from the measured of about three-to-one. The BDT is based on the B0 kine- B0 → J/ψK∗0 and J/ψ → µ+µ− branching fractions [8]. matics, B0 vertex quality, daughter track quality, impact Thesemeasurementsarecurrentworldbest,anddon’tcon- parameterandkaon,pionandmuonparticleidentification. firmprevioushintsofanon-SMvalueofA atlowq2. Theofflineselectioncriteriaareexplicitlychosentomin- FB imise angular acceptance effects. The multivariate selec- tionwastrainedusing B0 → J/ψK∗0 candidatesfromthe 2010 data as a proxy fodr the signal and B0 → K∗0µ+µ− 3 B0s→ φγ candidatesfromtheuppermasssidebandofthe2010data IntheSM,theamplitudeofthese→sγpenguintransitions ¯ a e-mail:[email protected] isdominatedbyavirtualintermediatetopquarkcoupling EPJWebofConferences to a W boson. Extensions of the SM predict new heavy 5 Conclusions particlesthatmaypropagatevirtuallywithintheloopand modifythedynamicsofthetransition.Therefore,thesera- AscanbeseeninFig.1,thereisgoodagreementbetween diativemodesarepromisinglaboratoriesthatcouldreveal recent SM predictions and LHCb’s measurement of A , FB the presence of new phenomena beyond the SM with the F anddBF/dq2 inthesixq2 bins.Ina1 < q2 < 6GeV2 L precisemeasurementofthebranchingratios,asymmetries bin,LHCbmeasuresA =−0.10+0.14±0.05,F =0.57+0.11± FB −0.14 L −0.10 or angular distributions. The offline selection of both the 0.03 and dBF/dq2 = 0.39±0.06±0.02, to be compared B0 → K∗0γ and B0 → φγ decays is performed with with theoretical predictions of A = −0.04+0.03, F = FB −0.03 L the strategy of maximizing the cancellation of systematic 0.74+0.06 and dBF/dq2 = (0.50+0.11)×10−7 respectively. uncertainties when performing the ratio. The analysis of −0.07 −0.10 The experimental uncertainties are presently statistically ∼341pb−1ofLHCbdatagives: dominated,andwillimprovewithalargerdataset.Sucha datasetwouldalsoenableLHCbtoexploreawiderange B(B0→ K∗0γ) =1.52±0.14(stat)±0.10(syst)±0.12(f /f ) ofnewobservables [14]. B(B0→φγ) s d In340pb−1of ppcollisionsatacentreofmassenergy s √ (1) of s = 7TeVthemostprecisemeasurementofB(B0→ Where fd(fs)aretheprobabilitiesofthebquarktohadronize φγ)hasbeenperformed,giving: intoB0(B0).Thisresultsiscompatiblewithin1.6standard s deviationswiththetheoryprediction. B(B0→ K∗0γ) =1.52±0.14(stat)±0.10(syst)±0.12(f /f ) B(B0→φγ) s d Combining the ratio of branching fractions in 1 with s (3) the World Average measurement for the B(B0 → K∗0γ ) The B(B0 → µ+µ−) and B(B0 → µ+µ−) upper limits s from[9],weobtain, obtainedbyLHCbare: B(B0s→φγ)=(2.8±0.5)×10−5 (2) B(B0→µ+µ−)<1.2(1.4)×10−8at90%(95%)CL, s which agrees within 1.6 standard deviations with the pre- B(B0→µ+µ−)<2.6(3.2)×10−9at90%(95%)CL. vious experimental measuremen, and wich correspond to In Fig. 2 the luminosity needed for a 3σ evidence as themostprecisemeasurementofthisBRtodate. a function of B(B0 → µ+µ−) is shown. Approximately s ∼ 2fb−1 are needed in the case that the value is equal to 4 B0 → µ+µ− the SM prediction, but statistical fluctuations can make it (s) possiblewith∼1fb−1.Fig.2alsoshowsthatexclusionsof B(B0 → µ+µ−)downtothe(2×)SMlevelwouldimpose s TheSMpredictionsforthebranchingfractionsoftheFCNC importantconstraintsinregionaroundthecurrentNUHM1 decays B0s → µ+µ− and B0 → µ+µ− areB(B0s → µ+µ−)= bestfitpoint [15].Alltheresultspresentedherearecurrent (3.2±0.2)×10−9 and B(B0 → µ+µ−) = (0.10±0.01)× worldbest. 10−9 [10]. However, contributions from new processes or newheavyparticlescansignificantlyenhancethesevalues. For example, within Minimal Supersymmetric extensions References of the SM (MSSM), in the large tanβ regime, B(B0 → s µ+µ−) receives contributions proportional to tan6β [11], 1. LHCbcollaboration,A.A.Alvesetal.,“TheLHCbde- wheretanβistheratioofthevacuumexpectationvaluesof tector at the LHC”, JINST 3 (2008) S08005, and refer- thetwoneutralCP-evenHiggsfields,andcandiffersignifi- encestherein. cantlyfromtheSMprediction.TheLHCbanalysisisdone byclasifying B0 → µ+µ− candidatesinbinsofa2Dpa- 2. LHCbcollaboration,R.Aaijetal.,“Angularanalysisof (s) B0 → K∗0µ+µ−”,LHCb-CONF-2011-038(2011). rameterspacemadebytheinvariantmassandamultivari- 3. LHCb collaboration, R. Aaij et al., “Measurement of ateclasifierwhichcondensatesgeometricalandkinemati- the ratio of branching fractions B(B0→ K∗0γ)/B(B0→ calinformationoftheevent.Thesignalexpectationineach φγ)withtheLHCbexperimentat √s = 7TeV”,LHsCb- biniscalculatedusingdatafromcontrolchannelssuchas CONF-2011-055(2011). B0 → h+h(cid:48)− and B+ → J/ψK+. The background expec- (s) 4. R. Aaij et al. [LHCb Collaboration], “Search for tationiscalculatedbyinterpolatingfrommasssidebands. the rare decays B0 → µ+µ− and B0 → µ+µ−” The B0 → h+h(cid:48)− peakingbackgroundyieldiscalculated s (s) arXiv:1112.1600[hep-ex]. using π → µ and K → µ misidentification probabilities 5. A. Ali, T. Mannel and T. Morozumi, “Forward back- obtained from data using decays such as Λ → pπ− and ward asymmetry of dilepton angular distribution in the D0 → K+π−.Thesignalandbackgroundexpectationsare decayb→ sl+l−”Phys.Lett.B273(1991)505. comparedwiththedistributionofobservedevents,andthe 6. W. Altmannshofer, P. Ball, A. Bharucha, A. J. Buras, limitsaresetusingtheCLs method [12,13].TheB(B0s → D.M.StraubandM.Wick,“SymmetriesandAsymme- µ+µ−)andB(B0 →µ+µ−)upperlimitsobtainedare: triesofB→ K∗µ+µ− DecaysintheStandardModeland Beyond,”JHEP0901(2009)019[arXiv:0811.1214[hep- B(B0→µ+µ−)<1.2(1.4)×10−8at90%(95%)CL, s ph]]. B(B0→µ+µ−)<2.6(3.2)×10−9at90%(95%)CL. 7. C. Bobeth, G. Hiller and D. van Dyk, “More Benefits ofSemileptonicRareBDecaysatLowRecoil:CPVio- Fig.2showstheluminosityneededtoimposestronger lation,” JHEP 1107 (2011) 067 [arXiv:1105.0376 [hep- limitsortoachievea3σevidenceofB0 →µ+µ−. ph]]. s HadronColliderPhysicsSymposium2011,November14-18,Paris,France 8. K. Nakamura et al. [Particle Data Group Collabora- Theory Binned theory LHCb tion],J.Phys.GG37(2010)075021. B F A 9. The Heavy Flavor Averaging Group, D. Asner, “Av- erages of b-hadron, c-hadron, and τ-lepton Proper- 0.5 ties”, arXiv:1010.1589 [hep-ex]. Updates available at http://www.slac.stanford.edu/xorg/hfag/osc/end 2009. 0 10. A.J. Buras, G. Isidori and P. Paradisi, “EDMs versus CPV in B mixing in two Higgs doublet models s,d -0.5 LHCb with MFV”, Phys. Lett. B 694 (2011) 402–409, Preliminary [http://arxiv.org/abs/1007.5291 arXiv:1007.5291]; A. J. Buras, “Minimal flavour violation and be- 0 5 10 15 20 q2 [GeV2/c4] yond: towards a flavour code for short distance dynamics”, Acta Phys. Polon B 41 (2010) 2487, Theory Binned theory LHCb [http://arxiv.org/abs/1012.1447arXiv:1012.1447]. L F 1 LHCb 11. L. J. Hall, R. Rattazzi and U. Sarid, “The top quark Preliminary massinsupersymmetricSO(10)unification”Phys.Rev. D50(1994)7048; C. Hamzaoui, M. Pospelov and M. Toharia, “Higgs- 0.5 mediated FCNC in supersymmetric models with large tanβ”,Phys.Rev.D59(1999)095005; K.S. Babu and C.F. Kolda, “Higgs-mediated Bs,d → 0 µ+µ− in minimal supersymmetry”, Phys. Rev. Lett. 84 (2000)228. 0 5 10 15 20 q2 [GeV2/c4] 12. T.Junk,“ConfidenceLevelComputationforCombin- ingSearcheswithSmallStatistics”,Nucl.Instrum.Meth. Theory Binned theory LHCb A434(1999)435,hep-ex/9902006. 2]V 1.5 LHCb 13. A. Read, “Presentation of Search Results: The CLs Ge Preliminary Technique”,J.Phys.G28(2002)2693. 4/c 1 14W. .UR.eeEcgee,d“eN,eTw. oHbsuerrtvha,blJe.sMinatthiaes,deMca.yRmaomdeonB0an→d ·-7 10 K∗0µ+µ− ” JHEP 0811, 032 (2008) [arXiv:0807.2589 2 [q [hep-ph]]. /dF 0.5 B 15. O. Buchmueller et al., “Supersymmetry in Light of d 1/fbofLHCData,”arXiv:1110.3568[hep-ph]. 0 0 5 10 15 20 q2 [GeV2/c4] Fig.1. A ,F andthedifferentialbranchingfractionasafunc- FB L tion of q2 in the six Belle q2 bins. The theory predictions are describedfromRef.[7]. EPJWebofConferences BB((BB00 fifi mm ++ mm --)) 33 ss ddiissccoovveerryy [[1100--88]] ss 1.6 LHCb 1.4 Projection from 370 pb-1 1.2 1 0.8 0.6 0.4 SM 0.2 0.5 1 1.5 2 2.5 3 3.5 4 Luminosity [fb-1] BB((BB00 fifi mm ++ mm --)) UUppppeerr LLiimmiitt aatt 9955%% CC..LL.. iiff SSMM [[1100--88]] ss 1.8 LHCb 1.6 Projection from 370 pb-1 1.4 1.2 1 0.8 0.6 0.4 SM 0.5 1 1.5 2 2.5 3 3.5 4 Luminosity [fb-1] Fig. 2. Luminosity needed in order to get a B0 → µ+µ− 3σ s evidence (top) or a 95%CL exclusion in the presence of a SM signal (center). The bottom plot shows how upper limits in the 10−9 level would constraint the region around the minimum of theNUHM1fitfrom [15].