Physics at LHC Prague, 6-12 July, 2003 R. Kinnunen Helsinki Institute of Physics A/H -> and H + -> in CMS R. Kinnunen Physics at LHC Prague July 6

Physics at LHCPrague, 6-12 July, 2003

R. KinnunenHelsinki Institute of Physics

A/H -> and H+ -> in CMS

R. Kinnunen

Physics at LHC Prague July 6 - 12, 2003

Helsinki Institute of PhysicsHelsinki, Finland



Contents

H/A -> H/A -> Cross sections and branching fractionsHadronic trigger-jet identification and hadronic jet suppression tagging with impact parameter measurementb-jet tagging in bbH/ASignal to background ratios and expected discovery reaches,tan mesurement from event rates

H+ ->

Cross sections and branching fractionsTrigger for H+ -> in fully hadronic events polarization in H+ -> and W+ -> Signal to background ratios and expected discovery reaches, tan mesurement from event rates



- Efficient background reduction with b tagging in bbH/A

Relative to SM, H/A -> ZZ, WW are strongly suppressed but

- Higgs boson mass reconstruction for H,A ->

Hbb, H, H couplings are enhanced at high tan

How to search for heavy neutral MSSM Higgs bosons

use bbH for production with H/A -> decay channels

- 3 final states for H/A -> (jet+jet, lepton+jet, lepton+lepton)

- H/A -> bb may be also viable (under study)



Production and decay of H, A

Production through gg -> H/A and gg -> bbH/A

gg->bbH/A dominates the production at large tan ~90% of the total production cros section

BR(H -> ) ~ 10% for tan > 10

For large mA, enhancement of BR(H->) for larger || parameter due suppression of H,A -> decays

HIGLU,HQQ from of M. Spira et al.

Pole mass for Yukawa coupling

g

gg

g

H

H

b

b



Final states investigated for H, A ->

H -> -> 2 jets BR ~ 42.0%

Backgrounds from QCD multijet events, Z,* ->, tt, Wt, W+jetsH, A -> -> lepton + jet BR ~ 45.6%H, A -> ->2 leptons BR ~ 12.4%

Backgrounds from Z,* ->, tt, bb, Wt, WW, W+jets

Simulation tools

PYTHIA for event generationHDECAY for normalization of cross sections and branching fractionsfor two- jet and two-lepton channelsFull simulation for trigger, selection, b tagging, mass reconstructionFast simulation for signal to background ratios



Hadronic Tau trigger

Level-1 output rate of 1-Tau and 2-Tau triggers: 3 kHz at low luminosity 6 kHz at high luminosityHLT output rate on tape for ’s: 4 (10) Hz at low (high) luminosity

93 (112) GeV for 1-Tau for low (high) luminosity 66 (76) GeV for 2-Tau for low (high) luminosity

Trigger requirements:

Efficiency: 78 (54)% for H->, mH = 200 GeV, 1 or 2 Tau 81 (72)% for H+->, mH+ = 200 GeV, 1 Tau

Level-1 1-Tau and 2-Tau triggers

on the calorimeter jet reconstructed in 12x12 trigger towers with maximum Et

in the central 4x4 towers and no significant activity in the neighbouring towers (trigger tower = HCAL cell + 5x5 ECAL cells, x = 0.087x0.087 in the barrel)

The required Level-1 rate can be achieved with the thresholds of



Level-2 Tau trigger

Reconstruction of a jet centered at the hardest Level-1 jetIsolation in the EM calorimeter:

Suppression of 3 for QCD background with Et

em < 5.6 GeV

Signal efficiency ~ 85% same for mH = 200 and 500 GeV

Efficiency (QCD vs H->1/3 prong jets)

as a function of Etcut-off

- sum of the Et deposits in ECAL within 0.13 < R(jet direction, cell) < 0.4

jet definition:Etem <

Etcut-off



Level-3 Pixel Tau trigger

1. Reconstruction of tracks around the Level-1 jet direction

2. Small signal cone (RS = 0.07) around the hardest track3. Larger isolation cone around jet direction

Accept tracks only in the signal cone

HLT efficiency for 1 or 3 tracks in the signal cone and for R = 0.35:

QCD suppression ~ 103

signal efficiency ~ 40%

using track counting in the Pixel (vertexing) detector: - good efficiency required, high pt accuracy not needed

Efficiency (QCD vs H->1/3 prong jets)as a function of the isolation cone size



Off-line jet identification

Exploits further the narrowness and the isolation of the jet in the full tracker

Define a narrow signal cone r = 0.03 (0.07 for HLT)

around the leading track direction

1. Leading track cut: Find the leading track in the L1 jet, set a cut pt > 40 GeV

jet isolation: No track, pt > 1 GeV, allowed within 0.03 < R < 0.4

3. Number of tracks in the signal cone: 1 or 3 tracks in the signal cone (pt

leading > 40 GeV)

4. Further reduction of hard QCD jets: Very hard QCD jets can be further suppressed with a cut in pt

leading / Etjet



Simulation of the QCD di-jet background

using (for the moment) a rejection factor as a function of Et

jet (initial QCD di-jet rate ~1012 events for 60 fb-1)

efficiency verified with full simulation and complet reconstruction for pt

gen < 170 GeV

Suppression of ~ 1000 per jetcan be obtained

Signal efficiency (per event) including Level-1 and HLT trigger from full simulation: mH = 200 GeV 0.8%mH = 500 GeV 8.9%

selection efficiency for hadronic QCD jets

from fast simulation



impact parameter tagging in H ->

The lifetime is small, c ~ 90 m, but can be still used to further supress the fake ’s from Z -> ll and from QCD multi-jet events using impact parameter measurement (1 or 3 prong ’s) and vertex reconstruction (3 prong ’s)

S. Le

hti

CMS full simulation and reconstruction

Best separation combining the measurementsin the two jets into one variable

sqrt(ip(1)2 + ip()2)

where ip(1) and ip() are significansiesof the impact parameter measurementsof the leading tracks in jets 1 and 2

Signal efficisiency ~ 60%QCD suppression factor of ~ 9



B jet tagging in gg -> bbH/ATagging of the associated b jets is the most efficient way to reduce

the Z,* -> (bbZ ~ 1-2 %)and to further reduce the QCD multijet eventsAssociated b jets in gg -> bbH/A are soft and uniformly distributed over || < 2.5:Efficiencies (Et threshold + tagging propability) relatively low

Significance of the signed transverse impact parameterCMS full simulation and complete reconstruction

Tagging algorithm:at least 2 tracks, pt > 1 GeV, ip > 2,inside the jet cone

Efficiency per jet: 32% non- jets in bbH~ 2 % light quark and gluon jets

non- jets in gg -> bbH/A

jets in QCD di-jet events



Higgs boson mass reconstruction in H ->

The neutrinos from H -> ( -> hadrons+, -> l+) are emitted closeto the directions of the visible ’s (jets or leptons):

neutrino reconstruction possible using the Etmiss

measurement in events with < 180o, efficiency ~ 50 % (for E E >0)

mH = sqrt(2 E E (1-cosjj

Higgs boson mass from full simulation for H -> -> two jets, mH = 500 GeV, tan = 20,with jetjet < 175o

Efficiency (cut E E>0) 36% fit 14.9%

Efficiency and resolution sensitive to the Et

miss measurement and to the cut

jet, e, jet, e,

jet

Etmiss

jj



Event selection for A,H->-> 2 jets

Basic event selection: - 2 jets passing the Level-1 and HLT triggers and the off-line selection (1 or 3 hard tracks, isolation)- tagging with impact parameters- Higgs boson mass reconstruction ( cut, E, E > 0)

i) Further selection with Etmiss:

- Etmiss > 40 GeV

-central jet veto beyond 30 GeV

ii) Further selection with b-jet tagging:- one b-tagged jet, Et > 20 GeV- central jet veto beyond 30 GeV

Two alternatives for further reduction:

larger staistics but poor S/B much improved S/B but lower statistics

Total background Total

backgroundTotal background



Leptonic final states, H/A -> -> e, ll Event selection:

- 2 isolated leptons, pt > 20 GeV - one tagged b jet, veto on second central jet beyond 30 GeV

- impact parameter tagging lepton+lepton final states can be used to double the statistics

lepton + jet final states, H/A -> ->l + jetEvent selection: - one isolated lepton (pt > 20 GeV), one jet (Et > 40 GeV) - one tagged b jet, second jet veto - Higgs boson mass reconstruction

Reach not yet optimized for large mA (> 200 GeV) in CMS

H/A -> -> emA = 200 GeVtan = 20

H/A -> -> ll

- Higgs boson mass reconstruction



Expected 5-discovery reach for H/A ->

Results for H/A -> from full simulation and complete reconstruction

Variation of BR(H -> ) due to H -> ij

decay modes ~ 40% at mH=500 GeV, tan=20,for -200 GeV < < 500 GeV

Higgs boson mass resolution ~ 2%



Measurement of tan in H -> from event rates

using the tan dependence BR ~ tan2 * x

- Luminosity uncertaintyL/L ~ 5% - Theoretical uncertainty on the cross section for gg -> bbH: dx/x ~ 30%- The gg -> bbH component is selected by b jet tagging: 1b or 2b tagging, less theoretical uncertainty and higher experimental purity with 2b tagging

tan/tan = ½ * sqrt((NS+NB) / NS2 + (L/L)2 + (x/x)2)

tan/tan 1b tagging 2b taggingH/A -> -> 2 jets, mA= 500 GeV, tan = 40, 60 fb-1 16%

19%

H/A -> -> lepton + jets, mA= 200 GeV, tan = 20, 30 fb-1 16%

H/A -> -> e, mA= 140 GeV, tan = 14, 30 fb-1 18%

dominated byrate uncertaintyup to tan ~ 30

Precision of Higgs boson mass measurement in H/A -> -> 2 jets mA= 500 GeV, tan = 40 , 60 fb-1 mH/mH = 1.5%mA= 200 GeV, tan = 40 , 60 fb-1 mH/mH = 1. 2%



How to search for charged Higgs bosons at LHC

If mH+ < mtop: Production through tt events, t -> bH+

accessible through the H+ ->t2-> lepton+qq final state If mH+ > mtop: Production through gg -> tbH+, gb -> tH+, qq’ -> H+, gg -> H+H-, gg -> W+H-

-gg -> tbH+, gb -> tH+ most promising processes, cross sections large enough and the associated top and b jets can be used for background reduction

H+

H+

- gg -> H+H-, gg -> W+H- have small production cross sections-Event rate sufficient for qq’ -> H+-> but suppression of the qq’ -> W-> background is difficult

tg

b

g

g

t

b



One tagged b jet enough to suppress the backgrounds, use gb -> tH+

Event generation with PYTHIANormalization of the production cross sections toT. Plehn, MADPH-02-1275

T. Plehn

Normalization of branching fractions to HDECAY

HDECAY

tan = 30

H+ ->



Trigger for H+ -> , -> jet +

-Cut on the leading track and isolation needed

Level-1: single Tau trigger (Et > 93 GeV, low luminosity)

High Level Trigger: cut on Et

miss in the calorimetry (possible due to off-line Etmiss > 100

GeV)

Level-3 Single Tau:

- Reconstruction of tracks in the full tracker within the L1 jet

Efficiency for ptleading track > 20 GeV and

isolation in 0.065 < R < 0.4: QCD rejection ~ 30Signal efficiency ~ 58%

QCD rejection ~ 100 for Etmiss > 65

GeVEfficiency (QCD vs H->1/3 prong jets)as a function of pt cut for the leading track



polarization in H+ -> and W+ ->

H+

+

+

+

H+ -> leads to harder pions from and from the longidutinalcomponents of and a1 than the corresponding decays in W+ ->

W+

+

+

+

TAUOLA interfaced to PYTHIA

->L

,a1

TL ,a1

L

Large suppression of W -> in tt, Wt, W+jet using the cut:

pleading track / E jet > 0.8:

Signal, mH+= 400 GeV 46%Signal, mH+= 200 GeV 22%tt background 1.8%

jet = calorimeter jet from -> hadrons +

tt background

Signal



Event selection for gb -> tH±, H± ->

Quasi two-body decay between the jet and Et

miss in fully hadronic events -> almost background-free situation in mT(-jet,Et

miss)

- Et jet > 80 – 100 GeV, containing a hard track with pt

track/Et jet > 0.8

- Etmiss > 100 GeV

- Veto on 5th jet, veto on second top quark

Cut on ( jet, Etmiss) ->

low mass background can be suppressed

( jet, Etmiss) > 20o

mT( jet, Etmiss)

- Reconstruction of associated hadronic top from two jetand one b-tagged jet

( jet, Etmiss)

tt background

Signal

Signal

Signal

background

background

30 fb-1

30 fb-1



Expected 5-discovery reach for charged Higgs boson

H+ -> tb and qq’ -> H+ -> are also promising

good background knowledge needed

Excess of ’s can be measured in tt, t -> bH+, H+ -> for mH+ < mtop

No sensitivity for intermediate tan with gb -> tH+ (with SM decay channels):H+ -> Wh, h -> bb accessible (in MSSM) only at small tan

tan measurement from event rates using ~ tan2 at high tanExpected uncertainty for tan >30 with 20% theoretical

uncertainty :tan/tan < 14% for mH+ = 200 GeVtan/tan < 20% for mH+ = 400 GeV



Conclusions

H/A -> accessible for 30 fb-1 with e and ll final states for tan > 14 at mA = 140 GeV jet+ lepton final states for tan > 10 at mA = 200 GeVand for 60 fb-1 with 2 jet final states for tan > 18 at mA = 200 GeV tan > 25 at mA = 500 GeV

tan determination with event rates: tan/tan = 16% for H/A -> -> 2 jets, mA = 500 GeV, tan = 40 16% for H/A -> -> lepton+jet, mA = 200 GeV, tan = 20 18% for H/A -> -> e, mA = 140 GeV, tan = 14 14% for gb -> tH+ ,H+ -> , mA = 200 GeV, tan = 30

H+ -> accessible in gb -> tH+ in fully hadronic final states with 30 fb-

1 for tan > 20 at mA = 200 GeV tan > 32 at mA = 400 GeV

Search for H/A and H+ can start early, with < 10 fb-1

With ~ 60 fb-1 masses in the 500 – 800 GeV range accessible

more specifically:

mH/mH in H/A -> -> 2 jets, tan = 40, 60 fb-1 : 1.2% for mA = 200 GeV 1.5% for mA = 500 GeV

Documents

Physics at LHC Prague, 6-12 July, 2003 R. Kinnunen Helsinki Institute of Physics A/H -> and H + -> in CMS R. Kinnunen Physics at LHC Prague July 6