HAL Id: in2p3-00014135
http://hal.in2p3.fr/in2p3-00014135
Submitted on 18 Nov 2003
HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.
Search for the radion using the ATLAS detector
G. Azuelos, D. Cavalli, H. Przysiezniak, L. Vacavant
To cite this version:
G. Azuelos, D. Cavalli, H. Przysiezniak, L. Vacavant. Search for the radion using the ATLAS detector.
EPJdirectC - Articles, 2002, 4, pp.1-13. �in2p3-00014135�
Search for the Radion using the ATLAS Detector
G.Azuelos 1
,D. Cavalli 2
,H. Przysiezniak 3
andL.Vacavant 4
1
LaboratoireRJALevesque,UniversitedeMontreal,Montreal,Quebec,H3C3J7,CANADA
2
DipartimentodiFisica,UniversitadiMilanoeINFNSezionediMilano,I-20133Milano,ITALY
3
LaboratoiredePhysiquedesParticules,IN2P3-CNRS,F-74019Annecy-le-Vieux,FRANCE
4
Lawrence BerkeleyNationalLaboratory,PhysicsDivision,BerkeleyCA94720-8167,USA
Received:date/Revisedversion:date
Abstract. Thepossibilityofobserving theradion usingthe ATLASdetectorattheLHCis investigated.
Studies on searches for the Standard Model Higgs with the ATLAS detector are re-interpreted to ob-
tainlimitsonradiondecay to and ZZ ()
. Theobservabilityof radiondecaysinto Higgspairs, which
subsequentlydecayinto+b
bor+b
bisthenestimated.
PACS. XX.XX.XX NoPACScodegiven
1Introduction
ThehierarchybetweentheelectroweakandPlanckscales
is oneof theprincipal puzzlesin modelsof unicationof
the interactions. The postulate of the existence of com-
pactiedlargeextradimensionsallowsforaPlanckscale
intheTeVrange,buttransferstheproblemtotheunnatu-
ralsizeoftheextradimension(s).InthemodelofRandall
andSundrum(RS)[1],two4-dsurfaces(branes)bounda
sliceof 5-dspace-time. TheSM elds areassumed to be
locatedononeofthebranes(theTeVbrane),whilegrav-
itylies in thebulk.The fthdimension isnot large,but
themetricisnon-factorizable,allowingforaresolutionof
thehierarchyproblem,givenanappropriatewarpfactor.
Thetheoryadmitstwotypesoffour-dimensionalmassless
excitations:theusualgravitonandagraviscalar,themod-
ulusorradion().Inordertostabilizethesizeoftheextra
dimensionwithoutnetuning ofparameters,Goldberger
and Wise [2] have proposed a mechanism by which the
radion acquiresa mass,expected to besmaller than the
J=2Kaluza-Kleinexcitations.Thepresenceoftheradion
is one of the important phenomenological consequences
of these theoriesof warped extradimensions[3{12].The
study of this scalartherefore constitutes acrucial probe
ofthemodel.
1.1Radion branchingratiosandwidth
The radioncouplingsto fermions and bosonsare similar
tothoseoftheStandardModel(SM)Higgs[3];onlytheir
relativestrengthschange.Theyareexpressedasafunction
ofthree parameters:thephysical massoftheradionm
,
thevacuum expectation valueofthe radionorscale,
,
Inthefollowingstudy,itisassumedthattheSMHiggs
hasbeendiscoveredandthatitsmasshasbeenmeasured.
The branching ratios of the radion are calculated using
thoseoftheSMHiggsascalculatedinHDECAY[15],and
using the ratio of the radion to Higgs branching ratios
given by [3]. For the mixing scenarios considered here,
( = 0 and = 1=6), the branching ratios of the light
HiggsareessentiallySM-like[14].
Figure1showstheprincipalbranchingratiosasafunc-
tionofscalarmassfordecaysoftheSMHiggs(topplots)
and ofthe radionwhenm
h
=125GeVand
=1 TeV,
for = 0 when there is no -h mixing (middle plots),
andfor=1=6whenandhareheavilymixed (bottom
plots).Thesegurespresentthefollowingfeatures:
{ BR ( !gg)is greatlyenhancedwith respectto the
Higgs and is close to unity for m
> 500 GeV and
=1=6,
{ theradiondecaysinto twoSMHiggsform
2m
h ,
{ BR (! )is enhancedfor =1=6andm
600
GeV,
{ for = 1=6 an interference is observed, producing a
strongsuppressionofdecaystovectorbosonsatapar-
ticularmassoftheradion.
Theradionhasaverynarrownaturalwidth.Figure2
showsthe totalwidth as afunction of mass, for theSM
Higgs and forthe radionwith =0and 1/6,for
=1
TeV.Thewidthisinverselyproportionaltothesquareof
.
Theaimofthepresentstudyistoinvestigatethepos-
sibilityofobservingaRSradionwiththeATLASdetector
throughthefollowingdecays:!,!ZZ ()
!4`,
! hh ! b
b and ! hh ! b
b +
. Only the di-
SN-ATLAS-2002-019 26 September 2002
10
-410
-310
-210
-11
50 100 150 200 10
-410
-310
-210
-11
200 400 600 800 1000
10
-410
-310
-210
-11
50 100 150 200 10
-410
-310
-210
-11
200 400 600 800 1000
10
-410
-310
-210
-11
50 100 150 200 10
-410
-310
-210
-11
200 400 600 800 1000
Fig. 1. Branching ratios of the SM Higgs (top), and of the
radionwhen=0(middle)and=1=6(bottom)asafunction
oftheirmass,for
=1TeV.TheHiggsmassinthefourlower
plotsissettom
h
=125GeV.
10 -2 10 -1 1 10 10 2 10 3
100 200 300 400 500 600 700 800 900 1000
Fig. 2. Width oftheSM Higgs andof theradion with =0
and1/6bothfor
=1TeV.
it is the main process at LHC and it benets from the
enhancementofthecouplinggg.
Othersignals,observableinSMHiggsdecay[16],such
asH !WW !``, shouldbeobservablein thecorre-
spondingradiondecaychannel[11].Themassreconstruc-
tionis diÆcultin thiscase, butobservationof thesignal
2! and !ZZ
()
!4`
Forthe(m
<160GeV)andZZ ()
(m
>100GeV )
decay channels, the radion signal signicance is deter-
minedfromtheSMHiggsresultsexpectedforATLAS[16],
for100fb 1
(one yearathigh luminosity10 34
cm 2
s 1
).
EvaluatingthesignalsignicanceasS=
p
B,whereS and
B are the numbers of signal and background events re-
spectively, theratiooftheradionsignalsignicanceover
that oftheSMHiggs,isgivenby[3]:
S=
p
B(;;ZZ)
S=
p
B(h;;ZZ)
=
!gg
BR (!;ZZ)
h!gg
BR (h!;ZZ) v
u
u
t
;ZZ
h
;ZZ
Accountingforexperimentalresolution,theZZresonance
width is given by ZZ
h;
= q
( h;
tot
=2:36) 2
+(0:02m
h;
) 2
,
where h;
tot
is thetotalintrinsic width of theHiggs or
resonance. In the energy range considered, the reso-
nance width is essentially independent of the negligible
intrinsic width of the resonance,
h;
= 0:10 p
m
h;
+
0:005m
h;
,andinthiscase,thelastfactorissimplyunity.
UsingtheATLAS SMHiggssignalsignicanceresults[16],
theradionsignalsignicanceisdeterminedandshownin
Figure 3asa function ofthe massof the radion,forthe
channel(top)andfortheZZ ()
channel(bottom),for
=1,10TeV,=0,1/6,andforanintegratedluminosity
of100fb 1
.
10
-210
-11 10
60 80 100 120 140 160 180 200
10
-210
-1
1 10 10
2100 200 300 400 500 600 700 800 900
Fig. 3. Signalsignicance versusthe massof theradion, for
the channel(top)and forthe ZZ ()
channel(bottom). In
both plots, the values for
=1,10 TeV and =0,1/6 are
shown,foranintegratedluminosityof100fb 1
.
3Radion production at the LHC
FromtheSMHiggsproductioncrosssection,evaluatedat
nextto leadingorder in [17],the crosssectionfor radion
productionwasestimatedaccordingto:
(gg!)=(gg!h) ()
(h)
BR (!gg)
BR (h!gg)
where isthetotalwidthoftheradionorHiggsreso-
nance,showninFig.2andthebranchingratiostoapair
ofgluonsareshownifFig.1.
Inthefollowingsections,forthepurposeofestimating
thelimitsof observation ofradiondecayto apairofSM
Higgsbosons,tworeferencevaluesaretakenforthemass
oftheradion:300GeVand600GeV.Theproductioncross
sectionsinthese casesare58pband8pbrespectively.
4!hh!b
b
As was mentioned in Section 1, the radion, similarly to
the heavy Higgs of the Minimal SuperSymmetric Model
(MSSM), candecay into Higgspairs withrelativelyhigh
branchingratio(seeFigure1).Thespecicdecaychannel
!hh!b
b oersaninterestingsignature,with two
high-p
T
isolatedphotonsandtwob-jets.Thebackground
rateisexpectedtobeverylowfortherelevantmassregion
m
h
>115GeVandm
>2m
h
.Inaddition,triggeringon
such eventsis easy andthe diphotonmass providesvery
goodkinematicalconstraintsforthereconstructionofm
.
The decay hh ! b
b was previously studied in the
contextoftheMSSMHiggs[18],althoughthemassranges
investigated were lower. The approach and the selection
usedin thisstudyareverysimilar.
4.1Signal
Signal events were generated with PYTHIA 6.158 [19].
The heavyHiggs H 0
production process viagluon-gluon
fusion(in theframework ofthe Minimal1-Higgsdoublet
Standard Model) was used to produce the radion. The
mass and the width of the H 0
were changed to reect
those oftheradionwhilethelightHiggsmasswasset to
m
h
=125GeV.
AsshowninFigure 2,thetotalwidthoftheradionis
afactorof10(100)smallerfor =0(1/6)thanthatofthe
Higgs,suchthatitiscompletelynegligiblewithrespectto
thereconstructedmassresolutions.
Twosamples of100kevents each were generated, for
m
=300GeVandform
=600GeV.
4.2Background
The backgroundsfor this channel are b
b (irreducible),
cc,bj,cjandjj(reduciblewithb-tagging).The
eventsweregeneratedwithPYTHIA6.158.Intheregion
processesaretheBorndiagramqq!andtheboxdia-
gramgg!.Theratesarethereforeverylow.However
large uncertaintiesapply to these backgroundssincethe
jets arise only from initial state radiation and not from
thehard-scattering.Generatingabackgroundsampleofa
sensible size turns outto be veryCPU time consuming,
andsomecutshad tobeappliedat theeventgeneration:
the sample was generated in seven dierent bins of p^
? ,
thetransversemomentumdenedintherestframeofthe
hard interaction. For each bin, ten million events were
generated.
Singlephoton production in the hard process j, ac-
companied by QCD orQEDradiation, and where either
the photon orjet is misidentied represents another re-
ducible background.This processwasstudied inthecon-
textoftheSMH!channel,andwasfoundtoincrease
thetotalbackgroundbyafactoroftwo.Inthecontextof
the radionwhere the backgroundsare negligible, this is
thereforenotexpectedtoaectthenalresults.
4.3Detectorsimulation
Thedetectoreects onthesignalandbackgroundevents
are simulated with a fast Monte-Carlo code, ATLFAST
2.53[20],basedonparametrizeddetectorresponse.While
mostparametersarechosenfromthestandardATLFAST
lowluminosityconditions(10 33
cm 2
s 1
),afewimprove-
mentswereapplied forthisstudy:
{ jets were recalibrated using a detailed parameteriza-
tion,
{ the photon reconstruction eÆciency was assumed to
be80%,
{ ap
T
-dependent b-tagging parameterizationwasused
with an average eÆciency of
b
=60% and rejection
factorsofapproximately93for light-quarkjetsand 7
forc-jets,respectively[16].
4.4Selection
Toextract thesignal,twoisolatedphotonswith p
T
>20
GeV and jj < 2:5, and two jets with p
T
> 15 GeV,
jj < 2:5 were required. At least one of the jets had
to be tagged asa b-jet. The diphoton and the dijet in-
variantmasses werethen formed. Figure 4showsthe re-
constructed invariant masses for m
= 300GeV, = 0
and
= 1 TeV. Two mass window cuts were applied:
m
=m
h
2GeVandm
bj
=m
h
20GeV.
Thephotonsandjetsfulllingtheserequirementswere
thencombinedtoformthem
bj
invariantmassasshown
inFigure5.Themassresolutionimprovesto5GeVwhen
constrainingthereconstructedmassesm
bj andm
tothe
lightHiggs massm
h
, asshownon theright-handplotof
Figure 5. The signal acceptances after the various cuts
describedabovearegiveninTable1.
Thesameanalysisprocedurewasappliedtotheback-
ground sample. Because of the uncertainties concerning
0 10 20 30 40 50 60
80 100 120 140 160
Mean RMS
125.0 1.859
mγγ (GeV/c2)
Events/1 GeV/30 fb-1
0 5 10 15 20 25 30 35 40
0 50 100 150 200 250
mjj (GeV/c2)Events/10 GeV/30 fb-1
bj bb
Fig.4.Diphoton(left)anddijet(right)invariantmassdistri-
butions, for m=300GeV, =0, =1TeVand 30fb 1
(threeyearsatlowluminosity10 33
cm 2
s 1
).Theright-hand
plotshowstheimpactofrequiringtwob-taggedjetsinsteadof
one.
0 5 10 15 20 25
200 250 300 350 400
mγγbj (GeV/c2)Events/8 GeV/30 fb-1
all window:
mh ± 2 mh ± 20
0 5 10 15 20 25 30
200 250 300 350 400
Mean RMS
301.0 10.16 7.589 / 35
Constant 25.47
Mean 300.4
Sigma 5.192
mγγbj (GeV/c2)
Events/4 GeV/30 fb-1
mass constraint
Fig. 5. Reconstructed bj invariant mass distribution, for
m
= 300 GeV, = 0,
= 1 TeV and for 30 fb 1
. The
plotsontheleft showallcombinationsandtheones fullling
themasswindowcuts(cf.text).Thedistributionontheright
isobtainedbyconstrainingthereconstructedmassesm
bj and
m
tothelightHiggsmassm
h
,afterthemasswindowcuts.
Cuts m
=300GeV m
=600GeV
photonskinematics 46% 51%
jetskinematics 36% 28%
b-tagging 76% 78%
m windowcut 83% 85%
mbj windowcut 49% 53%
total 5% 5%
Table1.Acceptanceforthesignal,for=0,
=1TeVand
forthetworadionmassesstudied.Foreachcuttheacceptance
isdenedwithrespecttothepreviousone.
themasswindowcutswereloosenedtokeepeventsfulll-
ing: m
=m
h
30 GeVandm
bj
=m
h
40GeV.The
backgroundlevelremainsneverthelessextremelylow,even
in thisconservativeapproach.
4.5Results
Thenalcandidateeventsare selectedin amasswindow
< m
bj
> 1:5
m
bj
for signal and background.The
m=300GeVm=600GeV
=0;
=1TeV 84:5 7:0
=0;
=10TeV 0:9 0:1
=1=6;
=1TeV 150:9 5:3
=1=6;
=10TeV 1:2 0:1
background 1:4210 4
0
Table 2. Numberofeventsselected for signal and for back-
ground, for m =300 and 600 GeV, for 30 fb 1
and for
mh=125GeV.
m
=300GeVm
=600GeV
=0;
=1TeV 4 (43)
=0;=10TeV (333) N=A
=1=6;=1TeV 2 (57)
=1=6;=10TeV (250) N=A
Table 3. Minimum integrated luminosity (fb 1
) needed for
discovery.N/AmeansthatthesignalisnotaccessibleatLHC.
Integratedluminositieslarger than30fb 1
areinparentheses
sincethefeasibilityoftheanalysisat highluminosity hasnot
beenstudied.
Sincethischannelispracticallybackgroundfree,asig-
naldiscoveryisdened hereasaminimumoftenevents.
The minimum integrated luminosities needed fordiscov-
eryarelistedin Table3.Afewfb 1
areneededfor=0
if
1TeV.
In the special case where = 0, the cross-section is
proportionalto 2
.Alowerlimiton
canthereforebe
derivedfromthisstudy.Itisobtainedusingtheprescrip-
tionof [21]:foraknownmeanbackgroundofzero,there
ismorethan95%chanceofobservingatleast10eventsif
the expected numberof signaleventsis greaterthan 18.
The correspondingreach in
is 2.2 TeV form
=300
GeV and 0.6 TeV for m
= 600GeV for an integrated
luminosityof30fb 1
.
5!hh !b
b +
Thechannel!hh !b
b +
providesanotherpoten-
tiallyinterestingsignalforradiondiscovery,althoughthe
background ishigher andthe reconstructed massresolu-
tionsarepoorerthaninthe!hh!b
bchannel.
Inordertoprovideatrigger,aleptonicdecayofthe
isrequired.Here,onlythecasewhenone decayslepton-
icallyandtheotherhadronicallyisconsidered.Asabove,
the events weregenerated byappropriately adaptingthe
process of MSSM decayof the heavy Higgs H 0
into two
light Higgs bosons (h) in Pythia 6.158 [19]. The eect
of the ATLAS detector on the resolution and eÆciency
of reconstruction of these eventswassimulated withthe
ATLASfastsimulationpackage(ATLFAST2.53).Theef-