Higgs WG meeting, 23 October 2007 R. Kinnunen, S. Lehti Analysis status for heavy charged Higgs with H + R.Kinnunen, S. Lehti Helsinki Institute of Physic, Helsinki Higgs WG Meeting CERN, 23 October, 2007
Jan 31, 2016
Higgs WG meeting, 23 October 2007R. Kinnunen, S. Lehti
Analysis status for heavycharged Higgs with H+
R.Kinnunen, S. LehtiHelsinki Institute of Physic, Helsinki
Higgs WG MeetingCERN, 23 October, 2007
Higgs WG meeting, 23 October 2007R. Kinnunen, S. Lehti
Contents:
- Workplan and status for fully hadronic gg->tbH+, H+
- Background analysis for 100 fb-1 with W-> in tt and W+jet events
- Background analysis for 100 fb-1 with QCD events
- ”Hard reconstruction” in the signal events
Higgs WG meeting, 23 October 2007R. Kinnunen, S. Lehti
Off-line event selection
• Veto on isolated electrons and muons, pT >~ 10-20 GeV/c
• MET > 100 GeV
• jet reconstruction, ET > 100 GeV
• isolation: no track with pT > 1 GeV/c in the isolation cone of 0.04 < R < 0.4 around the leading track
• Electron rejection: At least one HCAL cell with ET > 2 GeV in the jet
• One-prong selection:
– One reconstructed track in the signal cone
– ptrack / E jet > 0.8
• Three-prong selection:
– Three reconstructed tracks in the signal cone
– Matching between ET(CALO cluster) and pTtrack to select ->
– ptrack / E jet > 0.8 or ptrack / E jet < 0.4
• Tagging of one b jet and top mass reconstruction
• Transverse mass reconstruction from jet and MET
Higgs WG meeting, 23 October 2007R. Kinnunen, S. Lehti
Backgrounds for gg->tbH+, H+
Main background processes: tt, W+3jets, QCD multi-jet eventsIn ideal case, - signal at large mT( jet,MET), with an endpoit at mT( jet,MET) ~ mH+
- backgrounds from W-> (tt, W+jets) distributed to mT( jet,MET) < mW
Higgs WG meeting, 23 October 2007R. Kinnunen, S. Lehti
Sources background in the signal area
Background in the signal area (mT( jet,MET) > 100 GeV) can be devidedinto 3 gategories:
• Diffusion of background events to the signal area due to MET measurement - This background in tt and W+jets events can be measured with muons from W-> decays in tt and W+jets events - Background due to QCD multi-jet events can be measured in QCD events measuring fake probability in other processes
• Fake jets due to difficulty of the identification Could be measured from data exploiting + jet events as proposed in CMS NOTE-2006/028
• Fake leading tracks (pT > 80 GeV) Can be reduced with strong quality cuts, studies going on (PFlow group), L. Wendland for ’s in gg->tbH+, H+
Higgs WG meeting, 23 October 2007R. Kinnunen, S. Lehti
• Isolated electron veto to suppress W->e in tt events• One muon with pT
> 100 GeV• Muon isolation• ET
miss > 100 GeV• 3 jets, ET > 20 GeV• W+3jet selection: no b jets, suppression of tt • tt selection: 2 b jets, suppression of W+3jets• mT(,MET) reconstruction
Background analysis for 100 pb-1 in muon events
Event selection:1. Single muon trigger2. Off-line selection cuts:
Higgs WG meeting, 23 October 2007R. Kinnunen, S. Lehti
Background in the signal area can be estimates through:
Ntt-sel = (Ntt
-sel – NW3jcont) * (/) * (
-sel/ sel )* top rec.
-sel / b-tag
Nw3j-sel = (Nw3j
-sel – Nttcont) * (/) * (
-sel/-sel)* top rec.
-sel * b-tag / (b-antitag)
NW3jconand NW3j
conarethe W+3jet and tt contamination to tt and W+3jet selections
: to be estimated from MC not enough events from Z-> -> jet in this energy scale
MET( sel.): MET efficiency in signal (jet) selectionMET( sel.): MET efficiency in background (jet) selection
b-tag : to be measured in b enriched tt events
b-antitag : can be measured from +jet events, if fraction of +b events is below 1%
- top reconstruction: to be checked if affects the shape of the mT( jet, ET
miss) distribution
Higgs WG meeting, 23 October 2007R. Kinnunen, S. Lehti
Event selection:• Events from tau + MET trigger• At least one jet with pT
jet > 100 GeV• One of the jets with pT
jet > 100 GeV taken randomly as candidate• ET
miss > 100 GeV• 3 jets, ET > 20 GeV • one ”b” jet• top mass reconstruction• mT( candidate,MET) reconstruction
Background analysis for 100 pb-1 in QCD events
Events in the signal area:
NQCD(-sel) NQCD(QCD-sel) * miss-id)
miss-id) can be measured in +jet events
Higgs WG meeting, 23 October 2007R. Kinnunen, S. Lehti
Software version and data samples
- CMSSW version CMSSW_1_6_7 used
- Data samples used:
tt: tt0j_mT_70-alpgen, = 840 pbW+3 jets: W3jet_0ptw100-alpgen, 0 < pT
W < 100 GeV, = 588 pb W+4 jets: W4jet_0ptw100-alpgen, 0 < pT
W < 100 GeV, = 124 pb QCD: QCD_Pt_120_170, 120 < pT
jet < 170 GeV Signal: RelVal160pre8ChargedHiggsTausM200 mH+ = 200 GeV, 5000 events
Higgs WG meeting, 23 October 2007R. Kinnunen, S. Lehti
Muons from reco::MuonCollection - tracker isolation in a cone of R=0.3 and pT
cut = 1 GeV used
Muon off-line selection
Higgs WG meeting, 23 October 2007R. Kinnunen, S. Lehti
MET reconstruction
reco::CaloMET used including muons from reco::MuonCollectionNo MET corrections yet
Higgs WG meeting, 23 October 2007R. Kinnunen, S. Lehti
B tagging
JetTagCollection: trackCountingHighPurJetTags
2 tagged b jets, ET(jet) > 20 GeV, discriminator > 1.5 (not optimized yet)
Higgs WG meeting, 23 October 2007R. Kinnunen, S. Lehti
Expected behaviour from MC variables in tt events
Transverse mass mT(,MET) distributions
Higgs WG meeting, 23 October 2007R. Kinnunen, S. Lehti
Transverse mass mT(,MET) distributionsPreliminary results for tt, t1->b,t2->hadrons
Unnormalized distributions
Events in the signal area were wound to originate mainly from real neutrinosfrom W-> originating from the secod top-need to implement -jet veto in the muon selection and second veto in the signal selection
W2-> and W2->eremoved with MC info
tail due to MET resolution
Some events from W->
MET should originatefrom W->
Higgs WG meeting, 23 October 2007R. Kinnunen, S. Lehti
Transverse mass mT(,MET) distributionsPreliminary results for W+3jet, W->
All remaining events at large mT are signal-like with well measured MET
Originate from the BW tail of the W need to genetate the W mass in bins to study the effect
Higgs WG meeting, 23 October 2007R. Kinnunen, S. Lehti
- Select jets passing the the IsolatedTauTagInfoCollection with ET > 100 GeV Parameters for the IsolatedTauTagAlgo: pT
leading track > 20 GeV, R(match) =0.1, R(signal)=0.07, R(isol) = 0.45, pT(min) = 1 GeV
- Require matching with the MC jet from gg->tbH+, H+
- Definition of energy resolution: ETjet / ET
visible MC -1
- Event sample used: RelVal160pre9ChargedHiggsTausM200 - small statistics, 5000 events !
First tentative for a ”hard algorithm”
Higgs WG meeting, 23 October 2007R. Kinnunen, S. Lehti
Visible hadronic
mH+ = 200 GeV
ET(MC jet)
MC jet)
Higgs WG meeting, 23 October 2007R. Kinnunen, S. Lehti
Possible algorithm of reconstructing large ET one-prong ’s R. Kinnunen and A. Nikitenko
for good geometrical ECAL-HCAL-track matching do following: 1. If |ET
HCAL+ECAL - pTtrack| < 2 CALO: use ~ track p, corresponds to ±
2. If ETHCAL+ECAL – pT
track > 2 CALO 2.1 if |ET
HCAL - pTtrack| < 2 HCALuse ~ track p + ECAL cluster,
corresponds to ±n with charged pion not interacting in ECAL 2.2 if pT
track - ETHCAL > 2 HCAL ±interaction in ECAL, take the calo jet
2.3 if ETHCAL - pT
track > 2 HCAL hadronic jet, reject
- Optimization of the cones forenergy collection in ECAL and HCAL needed
2.1) 2.2)
±±
HCAL
ECAL
Higgs WG meeting, 23 October 2007R. Kinnunen, S. Lehti
Definition: ETcalo cluster / pT
track – 1 Calo cluster reconstruction from CaloRecHits withR(ECAL impact point, cell) < 0.1 in ECALR(ECAL impact point, cell) < 0.2 in HCAL
One prong decay mode without neutrals,-> h
Test CALO cluster/track matching for one-prong ’s
Almost all events withinETcalo cluster / pT
track – 1| < 0.5
Matching variable
Higgs WG meeting, 23 October 2007R. Kinnunen, S. Lehti
Calo cluster/track matching forn
Test here the HCAL cluster/track matching
pjet ~ ptrackBad track ?reject
Matching variables
III
Higgs WG meeting, 23 October 2007R. Kinnunen, S. Lehti
Loss of energy observed, try to minimize the effect with an asymmetric cut:
- 0.5 < ETcalo cluster / pT
track – 1 < 0.2
candidates with Calo Cluster/track matchingregion II
Energy resolution forn
Energy resolution forwith Calo Cluster/track matching
Resolutions
Higgs WG meeting, 23 October 2007R. Kinnunen, S. Lehti
HCAL cluster/track matching:add the ECAL cluster to include 0’s
±interaction in ECAL, take the calo jet
Hadronic jet: reject
HCAL cluster/track matching fornRegion III
Select large window |ET
HCAL cluster / pTtrack – 1 |< 0.8
due to energy loss in this area
Matching variables
Higgs WG meeting, 23 October 2007R. Kinnunen, S. Lehti
Comparison with reco::Jet
Resolutions
Higgs WG meeting, 23 October 2007R. Kinnunen, S. Lehti
All one-prong ’s selected with MC info
Higgs WG meeting, 23 October 2007R. Kinnunen, S. Lehti
Conclusions and plans
1) Background analysis with W-> in tt and W+3jet events and in QCD events in progress Large mass tail observed in mT(,MET): - mainly due to W-> in tt events - due to natural W width in W+3jet events
2) First results from a ”hard -tau” algorithm for one-prong ’s in gg->tbH+,
H+ with ET-jet > 100 GeV were presented
- Improvement of jet resolution possible with a simple algorithm
• New group members: Katri Lassila, fake probability from +jet events Matti Kortelainen, top mass reconstruction, Benedict Hegner should join
later Richard Cavanaugh, with PFlow, MET
Higgs WG meeting, 23 October 2007R. Kinnunen, S. Lehti
Backup slides
Higgs WG meeting, 23 October 2007R. Kinnunen, S. Lehti
const CaloJet* caloJet = dynamic_cast<const CaloJet*>(&(iTau->jet()));
vector<CaloTowerRef> towers = caloJet->getConstituents();
for(vector<CaloTowerRef>::const_iterator iTower = towers.begin(); iTower != towers.end(); iTower++){
// access CaloRecHits for(size_t j = 0; j < numRecHits; j++) { DetId recHitDetID = (**iTower).constituent(j);
DetId::Detector detNum=recHitDetID.det();
if( recHitDetID.det() == DetId::Ecal ){ if( recHitDetID.subdetId() == 1 ){ // Ecal Barrel
EBDetId ecalID = recHitDetID; EBRecHitCollection::const_iterator theRecHit = EBRecHits->find(ecalID);
if(theRecHit != EBRecHits->end()){ DetId id = theRecHit->detid();
const CaloCellGeometry* this_cell = EB->getGeometry(id); if(this_cell){
GlobalPoint hitPos = this_cell->getPosition(); double eta = hitPos.eta(); double phi = hitPos.phi();
double theta = hitPos.theta(); if(theta > 3.14159) theta = 2*3.14159 - theta;
// ecalCells.push_back(hitPos); double energy = theRecHit->energy();
double et = energy * sin(theta); Double_t px = energy * sin(theta)*cos(phi); Double_t py = energy * sin(theta)*sin(phi); Double_t pz = energy * cos(theta);
4. EBRecHit4. EBRecHit
Higgs WG meeting, 23 October 2007R. Kinnunen, S. Lehti
5. ECAL impact point
- using TransientTrack
const TransientTrack transientTk = (transientTrackBuilder->build(&(leadTk)));
- ecalHitPosition is taken as maxTowerEt position
TrajectoryStateClosestToPoint TSCP = transientTk.trajectoryStateClosestToPoint(ecalHitPosition);
GlobalPoint trackEcalHitPoint = TSCP.position();