Korean Physical Society 201226/10/2012

Study of top mass at CMS using Bi-Event

Subtraction TechniqueKorean Physical Society 2012

Junghwan Goh, Minsuk Kim, Intae Yu (SKKU),Tai Sakuma, Teruki Kamon (Texas A&M)

Korean Physical Society 2012

Outline Motivation of Bi-Event Subtraction Technique (BEST) BEST on tt̄ semi-leptonic channel Datasets and event selection Closure test with MC Results Summary and prespects

2

Korean Physical Society 2012

Motivation of BEST

3

How to find X→jj ?

‣ Choose a best dijet candidate among cascade decays of interest

‣ Or, allow combinations and estimate effect of random combinations

Challenge for background modeling

‣ Should be model independent & no kinematic preferences

‣ Data driven method

‣ More sensitivity for new physics search with multi jets

Korean Physical Society 2012

Combinatorial background with BEST Mixture of jet information from different events will give

random & uncorrelated pair of dijets Mass distributions from bi-event to subtract from the one with

all jet combinations - resulting dijet distribution with most of the combinatorial background removed

Published in the paper : Kamon et.al, PLB Vol 703, p475-478

4Mixture of two events

Korean Physical Society 2012

Top quark physics at CMS

5 [GeV]topm160 165 170 175 180 185-0.5

6.8

CMS combined result 0.9± 0.4 ±173.4 up to L= 5.0/fb syst.)± stat. ±(val.

CMS 2011 all-jets 1.3± 0.7 ±173.5 PAS-TOP-11-017 (L=3.54/fb) syst.)± stat. ±(val.

CMS 2011 lepton+jets 1.2± 0.3 ±173.5 CMS PAPER TOP-11-015 (L=5.0/fb) syst.)± stat. ±(val.

CMS 2011 dilepton 1.4± 0.4 ±172.5 CMS PAPER TOP-11-016 (L=5.0/fb) syst.)± stat. ±(val.

CMS 2010 lepton+jets 2.7± 2.1 ±173.1 )-1PAS-TOP-10-009 (L=36 pb syst.)± stat. ±(val.

CMS 2010 dilepton 4.6± 4.6 ±175.5 )-1JHEP 07 (2011) (L=36 pb syst.)± stat. ±(val.

CMS Preliminary

CMS combined result

CMS Preliminary

Available Top reconstruction methods :

Kinematic fitX2 fit to assumed mass constraints

MT2Applicable to dileptons

M3Invariant mass of 3 jet & dijet with highest pT

And so on...

Korean Physical Society 2012

Top quark study with BEST

6

Event topology with many jets

Jet combination problem can be solved by BEST

Strategy Start from semi-leptonic channel Extend to full-hadronic channel

Lepton+JetsDilepton Hadronic

Missing neutrino Jet combination

Clean event Full kinematics

Korean Physical Society 2012

Dataset and event selection

7

CM energy : 8TeVMC : tt̄ MadGraph + Pythia & tauolaReal data : 2012 summer (~5fb-1)

Event selectionMuon+3Jet triggered dataset1 isolated muon |η| < 2.1≥2 jets with pT > 45GeV≥4 jets with pT > 35GeV≥2 b tagged jets

High purity after event selection

Jet reconstruction and selectionParticle Flow, Anti-kT (R=0.5) with JESpT > 35GeV, |η| < 2.4b Tag algorithm based on SecVtx

W and top reconstructionNo overlap between jets (∆R > 0.4)2 untagged jet for W, +1 bjet for topNo lepton/MET for mass reconstruction

Korean Physical Society 2012

Counting combinations

8Same eventSame event Bi eventBi event

Pairs Counts Pairs Counts

W signal

biW

J+K

K+K

HadB+J(same top)

LepB+J

B+XB+X

J+J 1

J+J’ 4

J+K 2n J+K’, K+J’ 2(n+m)

K+K n(n-1)/2 K+K’ nm

J+b 2

J+B 2 J+B’, B+J’ 8

b+B 1 b+B’, B+B’ 4

b+K, B+K 2n b+K’, B+K’,... 2(n+m)

Objects in shorthand notationJ : Jets from Wb/HadB : b quark in hadronic topB/LepB : b quark in leptonic topk : jets not from top decay

Reducible by b tagging

W signalArtificial background

Korean Physical Society 2012

MC closure test

9

Same event8TeV

Bi event8TeV

Work in progress

Korean Physical Society 2012

Three-jet mass distributions

10

Same event8TeV

Bi event8TeV

Work in progress

Korean Physical Society 2012

First look on real data

11Work in progress

Same event8TeV

Same event8TeV

Bi event8TeV

Bi event8TeV

Korean Physical Society 2012

Bi event subtraction with fitting

12Work in progress

CMS work in progress, 5fb-1 at √s=8TeV

Korean Physical Society 2012

Possible applications Jet energy correction

A calibration tool with W jet resonance

Measurement of top quark properties Top quark mass and production cross section

without kinematic assumptions

Dijet and multi-jet resonance search Model independent, data driven background estimation Beyond Standard Model searches with multijets

13

Korean Physical Society 2012

Summary and prospects Combinatorial background with

Bi-Event Subtraction Technique Successfully applied BEST to tt̄ system

W and top mass peak extraction with CMS 2012 data Shown to be consistent with W and top quark mass Validation of technique with real data

Method is rapidly developing Aiming for independent measurement of top mass in 2012 Extension to BSM physics with Multi-jet

14