Search for Higgs in the dilepton dineutrino final state with CMS

Post on 04-Feb-2016

55 views 0 download

description

Search for Higgs in the dilepton dineutrino final state with CMS. Frank Würthwein UCSD. Outline. Introduction H to WW to lν lν Background Suppression & Estimation Results H to ZZ to ll νν Background Suppression & Estimation Results Outlook. Higgs Production. Dominant Process. - PowerPoint PPT Presentation

Transcript of Search for Higgs in the dilepton dineutrino final state with CMS

04/22/23 1

Search for Higgs in the dilepton dineutrino final state

with CMS

Frank Würthwein

UCSD

GGI 2011

Outline

• Introduction• H to WW to lν lν

– Background Suppression & Estimation– Results

• H to ZZ to ll νν– Background Suppression & Estimation– Results

• Outlook

04/22/23 2GGI 2011

Higgs Production

04/22/23 3

Dominant Process

Subdominant Process

GGI 2011

Higgs Production x Decay

04/22/23 4

event yields/fb In this talk, I will focus ondilepton + MET .

WW -> lν lν

ZZ -> ll νν

There are crudely speaking3 experimental regimes.

“250”

Easy WW Easy ZZHard WW

GGI 2011

SM backgrounds• No striking signature to

distinguish higgs from bkg’s.– Counts & multi-dim. shapes

• Key selections:– Lepton pT > 20/10GeV with

tight Id & Isolation• Wjets, QCD

– Large MET, MT, and Z veto• Drell-Yan

– # of jet classification (pT>30GeV) and top veto

• Top

– Kinematics (mll, ΔΦ)

• WW

04/22/23 5

Higgs 130

In addition: WZ,ZZ,Wγ(*)

GGI 2011

H -> WW for low vs medium mass

04/22/23 6

Bkg from Wjets grows as lepton pT decreases.

Lepton pT decreases as W* becomes more virtual.

Plots from CDF out of S.-C. Hsu thesis

Bkg from DY grows for very small dilepton mass.

Dilepton mass is small forlow higgs mass.

GGI 2011

WW vs ZZ for high mass

• As higgs mass increases the boost of the Z in higgs to ZZ leads to significant MET and significant transverse mass of the ll+MET system.

• This makes H -> ZZ -> ll + MET an interesting channel in the mH ~ 300 – 500 GeV range already this year.

04/22/23 7GGI 2011

Outline

• Introduction• H to WW to lν lν

– Background Suppression & Estimation– Results

• H to ZZ to ll νν– Background Suppression & Estimation– Results

• Outlook

04/22/23 8GGI 2011

Summary of bkg estimation

• Bkg estimated from data– Evaluated at HWW selection

• WW (for mH < 200GeV)• Wjets• Drell-Yan

– Evaluated at WW level• Top

– rely on MC to extrapolate to HWW

• Bkg estimated from MC– Dibosons (WZ,ZZ,Wγ)– Z to ττ

04/22/23 9

0-jet

1-jet

GGI 2011

Cut based analysis

04/22/23 10

H160H130

WW selection WW selection

GGI 2011

Missing Energy and pile-up

04/22/23 11GGI 2011

Drell-Yan Estimation

04/22/23 12

• DY (mostly) in ee,μμ– Fake MET due to tails of detector

resolution => difficult to simulate

• Predict “out” from “in”– Measure same flavor “in”– Subtract non-res contribution via eμ– Subtract VZ based on MC– Multiply result by R

• Measured in MC and at higher MET in data.

GGI 2011

Wjets Estimation• Measure “FR” with QCD events

– FR = prob. for “fake” lepton that passes loose selection to also pass tight lepton selection.

– FR = function of (pT,η)

• Extrapolate Wjets applying FR onto 1 loose 1 tight after higgs selection

04/22/23 13

Validation of Method in same sign at WW selection:

GGI 2011

Top background

• Jet veto kills top• Remaining top can be tagged

– Soft b-jets

– Soft muons

• Top tagging eff. ~50% in 0-jet• Residual top estimated via:

• Measure ε in 1 b-jet events– There must be another b-quark

• Systematics ~20-30%

04/22/23 14GGI 2011

Yields at WW selection

04/22/23 15

Summary of yields for 1.6/fb in the 0-, 1-, 2-jet bins.

Yields agree with expectations.Measure WW cross section as a crosscheck

GGI 2011

WW background

• Irreducible background with no single striking discriminator• Kinematic shapes in multidimensional space

– low mass, ΔΦll , lepton momenta, transverse mass of higgs

04/22/23 16

Example: mH=130 GeV

GGI 2011

Cut based Higgs selection

04/22/23 17

Spin-0 physics

Example: mH=160 GeVGGI 2011

WW bkg estimate from data

04/22/23 18

• For mH < 200GeV there is very little higgs contribution above 100GeV in dilepton mass.

• We thus can use that region to determine the normalization of WW bkg, and extrapolate into the higgs selection region using MC.

GGI 2011

Outline

• Introduction• H to WW to lν lν

– Background Suppression & Estimation– Results

• H to ZZ to ll νν– Background Suppression & Estimation– Results

• Outlook

04/22/23 19GGI 2011

Yields after Higgs selection

04/22/23 20

Example: mH=140GeV

Good agreement between observed and expectedNo sign of Higgs

GGI 2011

Yields after Higgs selection

04/22/23 21

Example: mH=140GeV

Good agreement between observed and expectedNo sign of Higgs

0-jet has most of the sensitivity

GGI 2011

Limits for cut based analysis

04/22/23 22

Observed: exclude higgs in range [147,194] GeVExpected: exclude higgs in range [132, 201] GeV

GGI 2011

Outline

• Introduction• H to WW to lν lν

– Background Suppression & Estimation– Results

• H to ZZ to ll νν– Background Suppression & Estimation– Results

• Outlook

04/22/23 23GGI 2011

H to ZZ to ll νν

• MET resolution and its non-gaussian tails – Zjets• Estimating WW & top via eμ • MC used to estimate WZ/ZZ bkg

04/22/23 24

Key Issues:

GGI 2011

Analysis Strategy

• Two high pT isolated same flavor leptons (ee, μμ).• Tight dilepton mass window (+- 15GeV) around Z

mass.• Large MET to suppress Zjets• Veto events with:

– small Δϕ between MET and nearest jet• Suppress Zjets with MET from large undermeasurements of jets

– b-tagged jets => suppress top– third lepton => suppress WZ/ZZ

04/22/23 25GGI 2011

Selections

04/22/23 26GGI 2011

Bkg Estimation

• Data Driven:– MET tails measured in γ+jets to estimate Zjets.– Top, WW using eμ events

• From MC:– WZ, ZZ

04/22/23 27GGI 2011

MET from γ+jets reweighting

04/22/23 28

μμ eeGGI 2011

Higgs transverse mass

04/22/23 29

Transverse mass of Higgs larger than all backgroundsonly for very large higgs masses.

GGI 2011

Outline

• Introduction• H to WW to lν lν

– Background Suppression & Estimation– Results

• H to ZZ to ll νν– Background Suppression & Estimation– Results

• Outlook

04/22/23 30GGI 2011

Final Yields

04/22/23 31

Observed and predicted agree well.No sign of higgs.

GGI 2011

Final Yields

04/22/23 32

Bkg decreases tenfold from low to high mass.Sensitivity to higgs largely a matter of luminosity.

GGI 2011

Final Yields

04/22/23 33

ZZ dominates except for low mass.Zjets remains second largest bkg even at high mass.WZ bkg with lost 3rd lepton significant at all masses.

GGI 2011

Limits on H to ZZ to ll νν

04/22/23 34

Higgs excluded in range 340-375 GeV GGI 2011

Outline

• Introduction• H to WW to lν lν

– Background Suppression & Estimation– Results

• H to ZZ to ll νν– Background Suppression & Estimation– Results

• Outlook

04/22/23 35GGI 2011

Necessary Improvements

04/22/23 36

For higgs mass of 120 GeV the systematic error on Wjets is roughly the same as the expected higgs signal.

Progress requires innovation in addition to luminosity !!!

GGI 2011

Desirable Improvement

• There is information left inside the cuts• There is information in correlations of the kinematic

variables that is not fully exploited by square cuts.04/22/23 37

MVA output for eμ 0-jet and 130,140,160GeV higgs

GGI 2011

Conclusions

• No higgs found– H to WW to lν lν excludes mH from 147 – 194 GeV– H to ZZ to ll νν excludes mH from 340 – 375 GeV

• Pushing the sensitivity towards lower mass higgs requires innovation and luminosity.

• Pushing the sensitivity towards higher mass higgs requires mostly luminosity.

04/22/23 38GGI 2011