Search for a Standard Model Higgs boson decaying to b quarks and produced in association with Z/W bosons with the CMS detector

0.29 MB·English

by  Michele De Gruttola

Search for a Standard Model Higgs boson decaying to b quarks and
produced in association with Z/W bosons with the CMS detector
MicheleDeGruttola1,2,a
1 UniversityofFlorida,InstituteforHighEnergyPhysicsandAstrophysics,DepartmentofPhysics,Gainesville,FLUSA
2 FermiNationalAcceleratorLaboratory,Batavia,ILUSA
Abstract. A search for the standard model Higgs boson is performed in a data sample corresponding to an
2 integratedluminosityof1.1fb−1,recordedbytheCMSdetector[4]inproton-protoncollisionsattheLHCwitha
1
7TeVcenter-of-massenergy.Thefollowingmodesarestudied:W(µν)H,W(eν)H,Z(µµ)H,Z(ee)HandZ(νν)H,
0
withtheHiggsdecayingtobbpairs.95%C.L.upperlimitsontheVHproductioncrosssectionarederivedfora
2
Higgsmassbetween110and135GeV.Theexpected(observed)upperlimitat115GeVisfoundtobe5.7(8.3)

n timesthestandardmodelexpectation.
a
J

2 1 Introduction Anoptimizationoftheeventselection,thatdependson
2
the Higgs mass, is performed, and 95% C.L. upper limits

on the pp → VH production cross section are obtained
] The search for the Higgs boson [1] is currently one of
x forHiggsmassesbetween110-135GeV.Theselimitsare
the most important undertakings of experimental particle
e based on the observed event count and background esti-
- physics. mateinsignalregionsdefinedineithertheinvariantmass
p
e At the LHC the main Higgs production mechanism is distribution of H → bb¯ candidates (“M(jj) or cut-and-
h direct production through gluon fusion, with a cross sec- countanalysis”),orintheoutputdiscriminantofaboosted
[ tionof∼ 17·103 fbforaHiggsmassmH = 120GeV[2]. decisiontreealgorithm(“BDTanalysis”)[5].Thelatteren-
However,inthisproductionmode,thedetectionoftheH → hancesthestatisticalpoweroftheanalysisbymakingfull
1 bb¯ decayisrenderednearlyimpossibleduetooverwhelm- useofcorrelationsbetweendiscriminatingvariablesinsig-
v
1 ing QCD di-jet production. The same holds true for the nalandbackgroundevents.
1 nextmostcopiousproductionmode,throughvector-boson Forlackofspacewewillpresenthereonlytablesand
6 fusion,withacrosssectionof∼1,300fb.Insteadwecon- plotsforthe115GeVmasshypothesis,whileonlythefinal
4 sider processes in which the Higgs is produced in asso- limitsplotswillcontainsallmassrangesearch.
. ciation with a vector boson which have cross sections of
1
∼ 660and∼ 360fbforWH andZH respectively.Evenif
0
the resulting sensitivity of the H → bb¯ decay is less than
2 2 Event selection
1 otherfinalstatessuchas H → γγand H → ττforexam-
: ple,itisparamounttosearchfortheHiggsinthesemodes
v giventhat theobservation ofthe H → bb¯ decayis keyto CandidateW(→ (cid:96)ν)decaysareidentifiedbyrequiringthe
i presence of a single, isolated, lepton and additional miss-
X determine the nature of this particle, if and when discov-
ingtransverseenergy(MET).Muons(electrons)arerequired
r ered. tohavea p above20(30)GeV.CandidateZ →(cid:96)(cid:96)decays
a t
We summarize a search for the standard model Higgs
are reconstructed by combining isolated, opposite charge
boson in the pp → VH production mode with the CMS
pairsofelectronsandmuonsandrequiringthedileptonin-
detector.Theanalysisisperformedinadatasamplecorre-
variant mass to satisfy 75 <m < 105GeV. For Z candi-
spondingtoanintegratedluminosityof1.1fb−1,collected (cid:96)(cid:96)
datestheelectron p isloweredto20GeV.Theidentifica-
t
bytheCMSexperimentata7TeVcenter-of-massenergy.
tionofZ → ννdecaysrequiresMET> 160GeV(thehigh
Thefollowingfinalstatesareincluded:W(µν)H,W(eν)H,
thresholddictatedbythetrigger).
Z(µµ)H,Z(ee)HandZ(νν)H–allwiththeHiggsdecaying ThereconstructionoftheH →bb¯decayismadebyre-
tobb¯ pairs.
quiringthepresenceoftwocentral(|η|<2.5)jets,abovea
BackgroundsarisefromproductionofWandZbosons minimump thresholdandb-tagged.Ifmorethantwosuch
t
associatedwithjets(allflavors),singly(ST)andpair-produced jetsarefoundintheevent,thepairwiththehighestsumof
topquarks,anddi-bosons(VV).Simulatedsamplesofall theb-tagoutputsforthetwojetsischosen(exceptforthe
backgrounds are used to provide guidance in the analy- WHanalyses,inwhichthett¯backgroundislarger,where
sis optimization, and an initial evaluation of their contri- thepairofjetswithhighesttotal p ischosen).Thesecom-
t
butions in the search region. For the main backgrounds, binationsarefoundtoyieldhigherefficiencyandrejection
high-puritycontrolregionsareusedtoestimatetheircon- ofwrongcombinationsinsignalevents,asopposedtosim-
tributioninthesignalregion. ply selecting the two highest p jets in the event. After b-
t
taggingthefractionofH →bb¯ candidatesthatcontainthe
a e-mail:[email protected] twob-jetsfromtheHiggsdecayisnearunity.
EPJWebofConferences
After b-tagging, the background from V+jets and di- Table1.Purityandscalefactors(Data/MC)derivedfromback-
bosonsisreducedsignificantlyandbecomesdominatedby ground enriched control regions (CR), as described in the text.
the sub-processes where the two jets originate from real The scale factors for W → µνHand W → eνH were averaged
b-quarks. Events with additional jets (N ) or additional together,andthesamewasdoneforZ→µµHandZ→eeH.
aj
leptons (N ) are rejected to further reduce backgrounds
al
fromtt¯andWZ. W →(cid:96)νH Z→(cid:96)(cid:96)H
The topology of VH production is such that the W/Z CR Purity SF Purity SF
and the Higgs recoil away from each other with signifi- V+udscg 79.4% 0.84±0.10 92.8% 0.88±0.02
cant pt.Cutsontheazimuthalopeninganglebetweenthe tt¯ 85.8% 1.01±0.11 97.5% 0.99±0.05
vectorbosonandthereconstructedmomentaoftheHiggs V+bb¯ 20.2% 1.40±0.29 81.6% 1.16±0.08
candidate, ∆φ(V,H), on the p of the V-boson and on the
t
b-tagged dijet pair achieve significant rejection for most
backgroundprocessesandimprovetheanalysisreach. Table2.Z Purityandscalefactors(SF,Data/MC)derivedfrom
FortheZ →ννchannel,QCDbackgroundsarefurther inv
backgroundenrichedcontrolregions,asdescribedinthetext.
reducedbyafactorof∼ 30whenrequiringthattheMET
doesnotoriginatefrommismeasuredjets.
ThetrainingoftheBDTisdonewithsimulatedsamples CR Purity SF
for signal and background that pass a looser event selec- Z+udscg 92.4% 0.97±0.06
tion relativeto the M(jj) analyses.Several input variables W+udscg 94.1% 0.92±0.05
were chosen by iterative optimization. These include the Z→bb 44.4% 1.00±0.30
di-jetinvariantmassandmomentum:M(jj)andp ,theV tt¯ 89.9% 0.91±0.09
tjj
transverse momentum p (V) , the b-tag value for each of
T
thetwojets,theazimuthalanglebetweentheVandthedi-
jets,∆φ(V,H),andthepseudorapidityseparationbetween
tt¯) or vetoing (for W+jets) b-jets. Table 2 lists the con-
thetwojets,∆η(J1,J2).TheBDTanalysiswasexpectedto
trol regions and the corresponding purities and scale fac-
improve the sensitivity with respect to the M(jj) analysis
torsobtained.
byabout10%ineverychannel.
The QCD background in the signal region is also es-
timated from data using control regions of high and low
values of two uncorrelated variables with significant dis-
3 Control regions
criminating power towards QCD events. One is the angle
betweenthemissingenergyvectorandtheclosestjetinaz-
Appropriatecontrolregionsthatareorthogonaltothesig-
imuth,∆φ(pfMET,J) andtheotheristhesumoftheCSV
nal region are identified in data and used to adjust Monte
valuesofthetwob-taggedjets.Thesignalregionisathigh
Carlo estimates for the most important background pro-
cesses: W+jets and Z+jets (with light and heavy-flavor valuesofbothdiscriminants,whileQCDpopulatesregions
jets),tt¯andQCDmultijetandheavy-quarkproduction.Dif- withlowvaluesofeither.Themethodpredictsanegligible
ferent control regions are found for each of the different contaminationofthisbackground.
search channels by changing the event selection in a way
thatenrichesthecontentofeachspecificbackground.For
all cases, control regions with purity ranging from about 4 Systematics
20% to nearly 100% have been successfully found. Dis-
crepancies between the expected and observed yields in
Thefollowingsystematicuncertaintiesontheexpectedsig-
thedatainthesecontrolregionsareusedtoobtainascale nalandbackgroundyieldsaffecttheupperlimit.Theval-
factor bywhich the estimatesfrom the simulationare ad-
ueslistedareanapproximationofwhatisactuallyusedin
justed.Thebackgroundfromthesesourcesinthesignalre-
thelimitcalculation.
gionarethenestimatedfromtheadjustedsimulationsam-
The total uncertainty on the signal prediction is taken
ples, taking into account the associated systematic uncer-
to be 26% and 28% for ZH and WH production, respec-
tainty.Thepreciseconstructionofallthecontrolregionsis
tively. Background uncertainties range from 12% to 20%
involvedandoutsidethescopeofthissummary.Theproce-
dependingonmodeandmasspoint.
duresappliedinclude,forexample:reversingtheb-tagging Experimentalsourcesofsystematicsaretheb-tageffi-
requirements to enhance W+jets and Z+jets with light-
ciency(∼10%),thejetenergyresolution(∼10%)andscale
flavor jets; enforcing a tighter b-tag requirement and re-
(∼1%) uncertainty, the machine luminosity (∼4.5%), the
quiring extra jets to enhance tt¯and requiring low “boost” trigger efficiency (∼2%). The signal cross section is af-
inordertoenhanceV → bbovertt¯.Table1liststhecon-
fectedbyelectroweakcorrectionsforaboostof∼150GeV
trol regions and the corresponding purities and scale fac-
are5%forZHand10%forWH,andQCDcorrection,rel-
torsobtained.
evantinthecomparisonNNLOvs.NLO,whereanuncer-
TheZ →ννHchannelisuniqueamongthefivemodes
taintyof10%forbothZHandWHisestimated.
analyzed, in that it does not include charged leptons. An
importantcheckistocomparetheobservedpfMETdistri-
butionwiththepredicteddistributionfromsimulation.To
accomplish this, muons are removed from the Z → µµJ 5 Results
data sample. Reasonably pure samples of tt¯and W+jets
eventscanbeobtainedbyrequiringatleastoneadditional Thefinalpredictednumberofeventsinthesignalregions
isolated lepton in the event,and then either requiring (for oftheBDTandM(jj)analysesaredeterminedwithamixof
ConferenceTitle,tobefilled
data-drivenestimatesbasedonthecontrolregions,andex-
pectationsfromsimulation.Wesummarizethefinalsignal GeV) CMS Preliminary DVHat afi bb
ainngdtbhaecksygsrtoeumnadtiecstuimncaetretsaiinntibeosthsusmetmsoafriaznedalyinsetsh,einpcrleuvdi-- nts/(16 10 Ws( m=n 7)H T(ebVb,) L = 1.1 fb-1 WZWWS i+W n++ gj +eubleWtdbs TsZogp
ous section, and the expected and observed upper limits Eve tQtCD
using1.1fb−1ofintegratedluminosity.Wereportintables
1
3 and 4 and figures 1 and 2 the results for a single mass
point, 115 GeV. While the final limits plots include mass
pointsfrom110to135GeV. 10-1
Table 3. Predicted backgrounds, signal yields with total uncer- 0 50 100 150 20M0 [Ge2V5]0
jj
tainty,andtheobservednumberofeventsfor115masspointfor
tohneM5c(jhj)anfonreltsheM1(1jj5)amnaaslsyspios.inWteseraerpcohr.talsotheslidingwindows GeV) CMS Preliminary DVHat afi bb
s/(16 10 Ws( e=n 7)H T(ebVb,) L = 1.1 fb-1 WZWW +W ++ j +eubWtdbssZg
nt Single Top
Process W(→µν)H W(→eν)H Z(→µµ)H Z(→ee)H Z(→νν)H Eve tQtCD
M(jj)cut 100–130 100–130 95–125 95–125 100–130
W+udscg 0.081±0.038 0.01±0.004 - - 0.023±0.007 1
W→bb¯ 0.829±0.221 0.344±0.093 - - 0.310±0.084
Z+udscg - - 0.110±0.065 0.006±0.003 0.180±0.039
Z→bb¯ 0.184±0.131 0.204±0.146 2.050±0.396 1.545±0.254 1.890±0.578
tt¯ 1.109±0.287 0.543±0.136 0.090±0.036 0.133±0.073 1.470±0.504
ST 0.24±0.105 0.122±0.049 0.090±0.036 0.009±0.007 0.410±0.156 10-1
VV 0.153±0.064 0.065±0.026 0.160±0.064 0.189±0.074 0.460±0.174
Bexp 2.596±0.449 1.288±0.242 2.410±0.358 1.883±0.232 4.793±0.938
WH 0.354±0.099 0.296±0.083 - - 0.091±0.012 0 50 100 150 200 250
ZH 0.006±0.002 0.002±0.001 0.195±0.051 0.193±0.050 0.502±0.104 Mjj [GeV]
Nobs 4 4 3 2 5
V
e CMS Preliminary Data
s / 30 G102 Zs(m =-m 7+) THe(bV,b L) = 1.1 fb-1 ZZZZZ H+++( 1ubW1db5Zs +cGWgeWV)
nt tt + S.T.
ve 10
E
Table 4. Predicted backgrounds, signal yields with total uncer-
tainty,andtheobservednumberofeventsfor115masspointfor 1
the5channelsBDTanalysis.WereportalsotheBDTcutwechoose
forthesearch. 10-1
10-2
Process W(→µν)H W(→eν)H Z(→µµ)H Z(→ee)H Z(→νν)H 0 50 100 150 200 250 300
BDT >0.050 >0.040 >−0.145 >0.160 >−0.175 M [GeV]
W+udscg 0.667±0.192 0.155±0.063 - - - jj
ZWZ+→→udbsbbc¯b¯g 20..003056±±-00..503036 10..317988±±-00..317441 02..181508±±-00..066455 00..097074±±-00..013869 10..315098±±00..039379 GeV102 CMS Preliminary DZHat(a115 GeV)
VStTVt¯ 010...316575533±±±000...132406041 010...226955243±±±000...132116751 000...320396460±±±000...010311084 00..119759±±−00..017383 000...524795173±±±000...201177672 nts / 30 10 Zs(e =-e 7+) HTe(bVb, )L = 1.1 fb-1 ZZZttZ ++++ ubWSdb.ZTs.+cWgW
Bexp 4.889±0.806 3.930±0.658 4.773±0.641 1.354±0.240 2.901±0.572 ve
WH 0.587±0.164 0.477±0.134 - - 0.507±0.112 E 1
ZH 0.011±0.003 0.004±0.001 0.328±0.085 0.183±0.048 0.049±0.007
Nobs 7 9 4 2 1
10-1
10-2
6 Upper Limits 0 50 100 150 200 250 300
M [GeV]
jj
Preliminary 95% C.L. upper limits on the Higgs produc- V103
tioncrosssectionintheVHmodewithH → bb¯ wereob- 0 Ge CsM =S 7 P TreelVim, Lin =a r1y.1 fb-1 DVHat(a115) QZ+CuDdscg
tcaoirnreedspforonmdinbgotthoathneinBtDeTgraanteddMlu(mjj)inaonsailtyysoefs1fo.1rfabd−1a.taFsoert Events/ 311002 Z(n n )H(bb) VtsttV ZWW+++bubbdbscg
theexpectedandobservedlimits,andthe1-and2-σbands,
theCLsmethodcurrentlyrecommendedbytheLHCHiggs
1
CombinationGroupwasemployed[6].
The results of the five BDT analyses are combined to 10-1
produce limits on Higgs production in the bb¯ channel for
the assumed masses: 110 − 135 GeV. The identical pro- 10-2
cedurewasappliedtotheresultsofthe M(jj)analysis.Ta- 0 50 100 150 200 250 300
ble5summarizestheresulting,expectedandobserved,up- Mjj [GeV]
per95%C.L.crosssectionlimits,withrespecttothestan-
Fig. 1. Distributions of dijet invariant mass after all M(jj) se-
dard model cross section, for each of the mass points for lectioncriteriahavebeenappliedin(fromtoptobottom:W(→
µν)H,W(→eν)H,Z(→µµ)H,Z(→ee)H,Z(→νν)H).
EPJWebofConferences
s
Event11002 CWsM( m=Sn 7 )PH Tr(eeblVbim,) Lin =a r1y.1 fb-1 DVWZWWtSQt HiaC+W n++t Dgafij +eubleWt dbbs TsZbogp ss/95% C.L. Limit on SM2233405050 CVsHM =(bCCCCS 7bLLLL )TP,SSSS ec rOEEEVeo,xxxm blLipppsm b=eeeei nicccr1nvettt.eee1adeddd rdfyb ––-1 12 ss
15
1 10
5
10-1 0 110 115 120 125 130 135
Higgs Mass [GeV]
-0.6 -0.4 -0.2 0 0.2 0.4
BDT Output SM40 CMS Preliminary
Events102 CWsM( e=Sn 7 )PH Tr(eeblVbim,) Lin =a r1y.1 fb-1 DVWZWWtt Ha+W ++t afij +eubWt dbbssZbg ss/95% C.L. Limit on 22330505 VsH =(b 7bCCCC )TLLLL, ecSSSSVo ,OEEEm Lxxxb bppp=si neeee1ecccr.1dvttteee efdddbd- 1–– 12 ss
Single Top
10 QCD 15
10
1 5
0 110 115 120 125 130 135
Higgs Mass [GeV]
10-1
Fig.3.Expectedandobserved95%C.L.combinedupperlimits
on the ratio of VHbb production for the BDT (top) and M(jj)
-0.6 -0.4 -0.2 0 0.2 0.4
BDT Output (bottom)analyses.Themedianexpectedlimitandthe1-and2-σ
bandsareobtainedwiththeLHCCLsmethodasimplementedin
ents / 0.05110023 CZs(Mm =-m S7+) THPe(rbVe,b Ll)i m= 1in.1a frby-1 DZZZZZ Ha+++t( a1ubW1db5Zs +cGWgeWV) LandS,asaretheobservedlimitsateachmasspoint.
Ev tt + S.T. Table5.Expectedandobserved95%CLupperlimitsonthepro-
10
ductionofaSMHiggsbosoninassociationwithWandZbosons
1 anddecayingtobquarksrelativetotheexpectedcrosssection.
LimitsarelistedseparatelyfortheBDTandM(jj)analyses.
10-1
10-2
MH(GeV) BDTExpected BDTObserved M(jj)Expected M(jj)Observed
10-3 110 5.8 8.0 6.4 8.2
115 5.7 8.3 6.0 11.3
-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 120 7.7 9.5 8.1 11.4
BDT Output 125 9.6 15.3 8.6 11.6
130 11.0 16.3 12.1 14.0
s103 135 14.4 22.5 15.0 19.9
ent CMS Preliminary DZHat(a115 GeV)
Ev102 s = 7 TeV, L = 1.1 fb-1 ZZZ ++ uWdZs+cWgW
Z(e-e+)H(bb) Ztt ++ bSb.T.
10 theBDTandM(jj)analyses.Theresultsaredisplayedsepa-
ratelyinFig.3.TheprimaryresultistheonefromtheBDT
1
analysis.
10-1
10-2
References
10-3
-0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1. P.W.Higgs,“Brokensymmetryandthemassofgauge
BDT Output
vectormesons”,Phys.Rev.Lett.13(1964)508.
nts104 CMS Preliminary Data QCD 2. ALEPH,CDF,D0,DELPHI,L3,OPAL,SLDCollabo-
Eve103 Zs(n =n )7H T(bebV), L = 1.1 fb-1 VVVH(115) ZZ++ubdbscg rtraotinonEsl,etchteroLwEePakElWecotrrkoiwnegakGWroourpk,inagndGtrhoeupS,LthDeTEelveac--
102 tstt WW++ubdbsgc troweak and Heavy Flavour Groups, “Precision elec-
10 troweak measurements and constraints on the Standard
Model”,ArXiv:1012.2367.
1
3. CMS Collaboration, “Search for the Standard Model
10-1
HiggsBosonDecayingtoBottomQuarksandProduced
10-2 in Association with a W or a Z Boson”, CMS Physics
10-3 AnalysisSummary,HIG-11-012(2011).
-1 -0.8 -0.6 -0.4 -0.2 0 0.2 4. CMS Collaboration, “The CMS experiment at the
BDT Output CERNLHC”,JINST 3(2008)S08004.
5. Byron P. Roe, Hai-Jun Yang, Ji Zhu, Yong Liu, Ion
Fig.2.DistributionsofBDToutputfordata(pointswitherrors)
and all backgrounds, from top to bottom: W(→ µν)H, W(→ Stancu,GordonMcGregor,“BoostedDecisionTreesas
eν)H,Z(→µµ)H,Z(→ee)HandZ(→νν)H.
ConferenceTitle,tobefilled
an Alternative to Artificial Neural Networks for Parti-
cleIdentification”,Nucl.Instrum.Meth.A543(2005)577-
584.
6. CMS Collaboration,“Search for standard model Higgs
boson in pp collisions at ps = 7 TeV”, CMS Physics
AnalysisSummary,HIG-11-011(2011).