Standard Model Higgs in HZZ and H

1
Standard Model Higgsin H?ZZ and H???
L.R. Flores-Castillo, Y. Fang, B. Mellado, W.
Quayle, T. Vickey, Sau Lan Wu University of
Wisconsin-Madison North American SM Higgs
Workshop 28.04.06
2
Outline

• H?ZZ
• H?ZZ?4l. Overview, Rome results and recent work
• VBF H?ZZ?lljj
• H???
• Overview
• Recent results
• Summary

3
H?ZZ?4l
4
Introduction

• Dominant after 2mZ
• Important for 130GeVltmHlt180GeV
• Narrow peak
• Low backgrounds
• Relatively clean signature
• No displaced vertices
• 4 isolated leptons

5
Backgrounds
H

• Backgrounds
• QCD ZZ production, each Z-gtee or ??
• two real Z bosons
• irreducible
• Zbb
• One real Z,
• leptons from b decays
• ?Non-isolated and displaced
• tt
• Non-resonant.
• Displaced leptons, non-isolated.

QCD
Reducible Zbb background before isolation and IP
cuts
Zbb
Zbb, tt
6
Signal reconstruction

• Muons tracks matched to MuId muons
• Electrons 3x7 clusters (recalibrated)
• Four lepton resolution at the level of TDR

H130 GeV?4?
H130 GeV?2e2?
H130 GeV?4e
s1.51 m129.9
s 1.43 m129.8
s 1.45 m129.9
MH GeV
MH GeV
MH GeV
TDR s 1.540.06 GeV
TDR s 1.510.06 GeV
TDR s 1.420.06 GeV

• DC1 1.6-1.7GeV

7
Signal and BG
Colors Signal QCD ZZ Zbb tt
H?4e, 30 fb-1, LO, 130GeV
2e2mu for other masses
180GeV
150GeV
200GeV
8
Significances (10fb-1)
NLO
LO
Colors NLO, LO Filled circles Now Hollow
circles Rome workshop
9
DC3 reconstruction

• Relevant differences between 11.0.41 and Rome
  production
• Muon reconstruction
• Fixed MuID combination
• ? Improved Pt reconstruction
• Higher efficiency for etagt2
• Electron reconstruction
• Track-matching efficiency drop in the crack
  region
• Slightly higher efficiency at low Pt

10
Vertexing information

• In signal (and QCD ZZ), all 4 leptons come from a
  single vertex.
• Not so for reducible backgrounds.
• Two variables can be built
• The ?2 of a 4-track vertex fit
• TDRs SUMDI variable
• Using these variables in 4?
• _at_95 efficiency, 50 higher rejection vs Zbb,
  100 higher vs tt.
• No change in 4e
• Overall, modest increase in significance, but
  theres still room for improvement.

(xi, yi) intersection points of pairs of lepton
tracks (in the transverse plane).
11
Calo-based muon id

• Hardware-related inefficiencies for turn-on
• Tracking efficiency 100
• Identify muon tracks using 4-2-0 topo clusters
• 100 efficient for muons
• Many samplings available
• Constructed a Likelihood Ratio formuon id,
  applied to non-muon tracks
• For now, only for etalt1.4
• Single ? efficiency 94.2 ? 97.7
• H?4? efficiency 15.7 higher
• Effect on backgrounds under study

StacoMuTag StacoMuTagCaloLR
12
Low energy e calibration
Colors 10, 20, 50, 100 GeV
Colors 10, 20, 50, 100 GeV

• Regular procedure LW obtained with 20, 50, 100
  GeV single e.
• Low energy electrons tend to be overcorrected
  (left plot)
• Using them also in the fit modifies the behavior
  of higher energies slightly, but keeps their
  linearity 0.3
• Applying this modified calibration to
  H130GeV?4e improves the Higgs resolution by 5

13
VBF H?ZZ?lljj
14
VBF H?ZZ?lljj

• VBF is a significant production mechanism for
  Higgs at LHC
• Forward jet tagging greatly helps to reduce
  backgrounds
• Based on ATL-COM-PHYS-2003-035
• Higher significance at 350GeV
• Extended to higher masses
• Current work
• Improve S/B for low masses
• Control samples to estimatebackground shape

H?ZZ
15
H???
16
Introduction

• Signal pp?HX, VBF
• Background
• Real ?? (jets)
• Fake ? ?jj???j, jjj???j
• Main experimental issues
• Photon calibration (energy scale and resolution)
• Separation of converted and unconverted photons
• Conversion identification with tracking
• Photon ID
• Achieve best rejection against jets
• Photon/?0 rejection
• Photon angle correction
• Photon angle with help of tracking vertex

17
Combined analysis

• Disjoint analyses have been assessed in ATLAS
• Inclusive H??? (TDR-like analysis) mostly with
  DC1
• (L.Carminati, M.Consonni, F.Derue, M.Escalier,
  B.Laforge, F.Tartarelli, G.Unal, Wisconsin, etc)
• H???1 jet (Zmushko, G.Unal, Wisconsin)
• H???2 jet (Japan, Wisconsin)

Highest significance combine H???0j and
H???jets analyses
18
Inclusive H??? with Rome layout

• Use full simulation (Rome or initial layout) to
  evaluate photon energy resolution (converted and
  unconverted)
• Use full simulation to evaluate the ?-jet
  rejection
• Optimize analysis by fixing ?-jet rejection
• Use NLO cross-sections

19
Combining ??0j, ??1j samples
Cross-sections for the H???0,1j analysis

• H? ??1j has lower statistics, but a much higher
  S/B ratio, which enhances the combined
  significance.

20
H???2jets

• Re-did analysis reported in ATL-PHYS-2003-036
• The following improvements have been applied
• Tune photon ID to get constant rejection
• ? Serious enhancement of photon efficiency
• Assume rejection of gluon initiated jets is a
  factor of 4 larger than quark initiated jets
• More realistic background generation
• Use ALPGEN to produce ??jj(j), ?jjj(j), jjjj(j)
  with Matrix Element Parton Shower matching
• Get much lower central jet veto survival
  probability for QCD processes
• Add more realistic contribution from gg?H???

21
Combined ??0,1,2j
22
Combining ??0j, ??1j and ??2j
H???1j
H???0j
H???2j
L 10 fb-1
Signal VBF Signal gg Fusion EWDPS ggjj QCD
ggjj gjjjjjjj
ATLFAST/DC1
23
(No Transcript)
24
11.0.4-1 Calibrationoverall mean vs ? after
calibration (topoEM 6.3.0)
Performance (of 11.0.4-1 calibration)similar to
Rome samples
Fixed window case is similar.
25
Calibrated H mass for at least one converted
photon from H?gg (11.0.4-1)
Vtx correction applied, g ID not applied
Topo 6.3.0
?1.49 GeV
?1.46GeV
M119.7 GeV
M120. GeV
RMS3.83 GeV
RMS3.90 GeV
Without separation between early converted and
late-non converted photons
With separation between early converted and
late-non converted photons
26
Summary
H?ZZ?4l H?ZZ?lljj H???

• Significant improvements since Rome
• Recent work
• Add low E electrons in LW calibration 5 smaller
  ?(4e)
• Vertexing 50 to 100 higher bg rejection in H?4?
• Calo-based muons increase signal efficiency by
  16
• Effect on background under study
• Improved significance for intermediate masses
• Extended to higher masses
• Re-evaluated H???2jet analysis
• Significance was seriously underestimated in the
  past. This may be a discovery channel for 30
  fb-1
• Rejection of quark/glue initiated jets under
  study
• Combination of H???0,1,2jets enhances the signal
  significance by at least 70 w.r.t. inclusive
  analysis
• Systematic errors to be investigated

27
Backup slides
28
H?ZZ?4l
29
Single electron reconstruction

• 3x7 clusters
• Recalibrated longitudinal weights
• Requirements
• Track match
• IsEM calorimeter cuts
• Ptgt7GeV, etalt2.5
• For the 4 momenta
• Energy from clusters, angles from tracking

Fractional deviation within 0.3 in the barrel
30
Single muon reconstruction

• Identified using the combined muon block
• Requirements
• Good fit (91)
• Track match (100)
• Ptgt7GeV, etalt2.5
• 4-momenta from matched tracks (dRlt0.05)
• Better single ? resolution

Combined muons Matched ID tracks
31
H?4l selection

• All leptons Ptgt7GeV, ?lt2.5
• Two leptons Ptgt20GeV
• Defining
• M12 ? mass(ll-) closest to MZ mass (on-shell Z)
• M34 ? second closest mass (off-shell Z for low
  mH)
• the following cuts are applied
• M12-MZ lt ?M12
• M34gtM34min
• ?M12 and M34min depend on mH defined as in
  ATL-COM-2004-042 (DC1) (parameterization wrt mH,
  values close to TDR).

32
H?4l. Rejection of Reducible BG

• Variables considered
• For each lepton,
• SEt,calo Sum of Et depositions in (EMTileHec)
  with ?Rlt0.3
• For electrons, the electrons Et is subtracted
  from SEt,calo
• SPt,trk Sum of Pt for tracks within ?R0.2
• Ptmx Largest Pt of tracks in a 0.2 R-cone.
• IP significance of the 4 leptons.
• Following plots one entry per Higgs candidate.
  For each candidate, the maximum value from its
  four leptons is used.

33
H?4l. Isolation
Colors Signal tt Zbb
34
H?4mu adding 1 CaloLR muon
StacoMuTag
StacoMuTagCaloLR
s 1.590.04 GeV
s 1.590.04 GeV

• LR cut 0.99, adding CaloLR candidates only up to
  eta1.4
• Only considering candidates with 1 or 0 caloLR
  muons
• Efficiency gain (over StacoMuTag) 15.7
• No loss of resolution
• Similar tails (85)

35
VBF H?ZZ?lljj
36
Introduction
H?ZZ

• VBF is a significant production mechanism for
  Higgs at LHC
• Forward jet tagging greatly help to reduce
  backgrounds
• Above 200GeV, largest fraction decays to WW and
  ZZ
• H?ZZ?lljj rate is six-times smaller than
  H?WW?l?jj
• Still, H?ZZ will provide information on the Higgs
  couplings
• Backgrounds
• ZNjet production, N4, Z?ee, ??
• ttbar production
• Largest contribution from di-lepton events
• Selection ATL-COM-PHYS-2003-035 plus some
  additions

37

• Lab-frame ??(lljj) versus mlljj

For Large Higgs Masses
For Lighter Higgs
ZNjet Background

• Cutting on ??(ll,jj) greatly improves S/B for
  light Higgs
• S/B from 0.67 to 1.33 for mH200 GeV.
• Significance from 3.61 to 3.83 for mH200GeV and
  a 10 Syst. Unc.

38
Background Control Sample

• From data, the shape can be predicted to within
  10
• To determine the best control sample region, we
  examine bin-by-bin ratios (control region over
  signal-like) for different regions of the
  ??jj-mjj plane (tagging jets)

AGREE TO WITHIN 10
39
H???
40
H???. Signal
pp?HX (MC_at_NLO)
VBF (PYTHIA)
41
H???. Backgrounds
Real ???(Jet) (MADGRAPH RESBOS)
Fake ? ?jj???j, jjj???j (MADGRAPH)
42
Rejection versus Efficiency for different ltPTgt
of jets normalized with ATLFAST
B
A
C
Three possible scenarios
43
(No Transcript)
44
Cross-sections for the H???2j analysis
Significance (10fb-1) the H???0,1,2j analysis