The study Higgs decaying into tau tau with CMS EPS HEP 2013, - - PowerPoint PPT Presentation

▶
the study higgs decaying into tau tau with cms
SMART_READER_LITE
LIVE PREVIEW

The study Higgs decaying into tau tau with CMS EPS HEP 2013, - - PowerPoint PPT Presentation

Feb 02, 2023 255 likes •589 views

The study Higgs decaying into tau tau with CMS EPS HEP 2013, Stockholm, 18-24 July Micha Bluj LLR/cole Polytechnique CNRS/IN2P3 on behalf of the CMS Collaboration Documentation: CMS PAS HIG-13-004, CMS PAS HIG-12-053 Outline

slide-1
SLIDE 1

The study Higgs decaying into tau tau with CMS

EPS HEP 2013, Stockholm, 18-24 July

Michał Bluj

LLR/École Polytechnique – CNRS/IN2P3

  • n behalf of the CMS Collaboration

Documentation: CMS PAS HIG-13-004, CMS PAS HIG-12-053

slide-2
SLIDE 2

2 EPS-HEP 2013, Stockholm

Outline

* Not discussed here, see A. Nayak's talk

◎ Observation of a new boson with mass around 125 GeV consistent

with the standard model scalar driven by the bosonic decay modes: H→ZZ→4l, H→γγ

◎ Observation of the new boson decaying into tau tau is

an important test of compatibility with the standard model

◎ Decays into tau leptons interesting also in frame of extended

models (e.g. MSSM)*

◎ This talk: overview of the H→ττ search

with CMS

○ Search strategy strategy ○ Analysis tools ○ Discussion of results

125

slide-3
SLIDE 3

3 EPS-HEP 2013, Stockholm

Analysis overview

Z→ττ

Embedding: Z→μμ data with μ replaced by simulated τ

W+jets

  • Jet→τh fakes
  • Suppressed

by topological cuts (e.g. mT)

  • Simulation

normalized to data in sideband

QCD

  • Jet→l/τh fakes
  • Suppressed by isolation
  • Same-sign data

(corrected for OS/SS ratio)

Z→ll

  • l/jet→τh fakes
  • Shape from

simulation corrected for yield in visible mass region

◎ 5 channels for ggH and VBF

○ μτh, eτh, τhτh eμ, μμ

◎ 3 channels for VH

(l=e/μ, L=e/μ/τh)

○ W(l)H(lτh/τhτh), Z(ll)H(LL)

◎ Analysis strategy

○ Require well identified and

isolated leptons, taus (τh)

○ Topological cut to suppress

background (e.g. mT against W+jets in eτh/μτh)

○ Categorisation based on jet

multiplicity and pT of τh/μ

○ Simultaneous max-likelihood

fit of di-tau mass (mττ) shape (binned) in all channels and categories

slide-4
SLIDE 4

4 EPS-HEP 2013, Stockholm

Event categories

0-jet, low-pT

  • Large background
  • No fit for signal,

constrain uncertainties

1-jet, low-pT

  • Enhancement due to jet

requirement

  • Better mass resolution

1-jet

  • pT(ττ) > 140 GeV
  • Better mass resolution
  • QCD suppressed

2-jet, VBF

  • pT(ττ) > 110 GeV
  • Mjj > 250 GeV
  • Δηjj > 2.5
  • Central jet veto

0-jet, high-pT

  • Large background
  • No fit for signal,

constrain uncertainties

1-jet, high-pT

  • Enhancement due to jet

requirement

  • Better mass resolution
  • Z→ττ suppressed by

high-pT(tau)

2-jet, VBF

  • ≥ 2 jets
  • Mjj > 500 GeV
  • Δηjj > 3.5
  • Central jet veto
  • VBF H signal

enhanced

τhτh eτh, μτh, eμ, μμ

pT ( τh / μ )

  • No. of jets
slide-5
SLIDE 5

5 EPS-HEP 2013, Stockholm

Event categories

0-jet, low-pT

  • Large background
  • No fit for signal,

constrain uncertainties

1-jet, low-pT

  • Enhancement due to jet

requirement

  • Better mass resolution

eτh, μτh, eμ, μμ

1-jet

  • pT(ττ) > 140 GeV
  • Better mass resolution
  • QCD suppressed

2-jet, VBF

  • pT(ττ) > 110 GeV
  • Mjj > 250 GeV
  • Δηjj > 2.5
  • Central jet veto

0-jet, high-pT

  • Large background
  • No fit for signal,

constrain uncertainties

1-jet, high-pT

  • Enhancement due to jet

requirement

  • Better mass resolution
  • Z→ττ suppressed by

high-pT(tau)

2-jet, VBF

  • ≥ 2 jets
  • Mjj > 500 GeV
  • Δηjj > 3.5
  • Central jet veto
  • VBF H signal

enhanced

τhτh

pT ( τh / μ )

  • No. of jets

Constrain uncertainties

No fit for signal

slide-6
SLIDE 6

6 EPS-HEP 2013, Stockholm

Identification of τh

◎ Reconstruction of individual decay modes

(Hadron-plus-strip algorithm. HPS)

○ Particles by Particle Flow algorithm used ○ 1-prong, 1-prong+π0's, 3-prongs

◎ Isolation

○ Multivariate discriminant based on ΣpT of

particles in rings around τh

○ Pileup mitigated with FastJet rho used by

isolation MVA

◎ Additional selection to reject leptons

slide-7
SLIDE 7

7 EPS-HEP 2013, Stockholm

Di-tau mass estimation

visible mττ full mττ

◎ Neutrinos present in tau decays ⇒ invisible

component

◎ Use maximum likelihood based full di-tau

mass (mττ) estimate

○ Computed on event-by-event basis using

four-momenta of visible decay products, Ex

miss, Ey miss, expected ET miss resolution

◎ 15-20% resolution on reconstructed mττ

○ better Z/H separation than with the visible mass

slide-8
SLIDE 8

8 EPS-HEP 2013, Stockholm

Mass distributions

  • 2-jet VBF
  • Best S/B category
  • VBF H signal

enhanced

slide-9
SLIDE 9

9 EPS-HEP 2013, Stockholm

Combined mass distribution

  • Combine channels and

categories weighting each by S/B

  • S – expected signal
  • B – fitted background
  • Both S and B in

window around peak at 125 GeV Higgs

slide-10
SLIDE 10

10 EPS-HEP 2013, Stockholm

Main systematics

Z→ττ

  • 8% tau-Id efficiency
  • Category efficiency 0-8%
  • Tau energy scale 3% (shape)

W+jets

  • Normalization 10-20%
  • Includes
  • Statistical unc. In

control region

  • Unc. of extrapolation

factor

  • Unc. of MC bias of

topological variable

QCD

  • Normalization 10%
  • Shape (bin-by-bin)

uncertainty in low stat categories

Z→ll

  • Normalization

20%/30% for ee/μμ

slide-11
SLIDE 11

11 EPS-HEP 2013, Stockholm

VH analysis

WH channel ZH channel

W(l)H(τhτh)

  • Main when one or two

fake τh (reducible)

  • Z→ll+2(jet→τh) fakes

reduced by vetoing events with 2nd lepton All VH channels: Reducible backgrounds with jet→l/τh fakes measured from data using fake-rate method Z(ll)H(LL)

  • Four final states
  • eτh, μτh, τhτh, eμ
  • Require Z→ll candidate
  • Irreducible background

from ZZ, reducible from Z→ll+2jets W(l)H(lτh)

  • Main background

irreducible WZ

  • Same-sign of two

leptons to suppress Z→ll+(jet→τh fake)

slide-12
SLIDE 12

12 EPS-HEP 2013, Stockholm

VH results

WH channel ZH channel

Combined VH limit

  • Sensitivity of ~3xSM
  • Small excess

consistent with both SM Higgs at 125 GeV and background

H(125) signal injected

slide-13
SLIDE 13

13 EPS-HEP 2013, Stockholm

Limits

◎ Combining all channels (VH included) ◎ Observed exclusion limit of 1.81 x SM at 125 GeV, while

0.76 x SM expected

◎ Result consistent with SM Higgs expectation

H(125) signal injected

slide-14
SLIDE 14

14 EPS-HEP 2013, Stockholm

Signal cross section

◎ Signal strength (σ/σSM) by channel (left) and by category (right)

at 125 GeV

◎ Combined signal strength σ/σSM = 1.1 ± 0.4 ◎ Consistent between channels and categories

EPS-HEP 2013, Stockholm

slide-15
SLIDE 15

15 EPS-HEP 2013, Stockholm

Significance & mass

◎ Observe broad excess over whole

probed mass range

◎ Consistent with SM Higgs expectation ◎ Max. significance of 2.93σ at 120 GeV ◎ 2.85σ obs. (2.62σ exp.) at 125 GeV EPS-HEP 2013, Stockholm ◎ Most favoured mass of the observed

excess

○ From likelihood fit

◎ mH = 120+9

  • 7(stat+sys) GeV

○ Consistent with

mH(4l) = 125.8 ± 0.5 GeV

slide-16
SLIDE 16

16 EPS-HEP 2013, Stockholm

Conclusions

◎ Search for H→ττ with the full 2011+2012

dataset collected by CMS was shown

◎ A wide excess compatible with SM Higgs is

  • bserved

○ Obs. limit of 1.81xSM at 125 GeV, while

0.76xSM exp.

○ Signal strength of 1.1 ± 0.4 ○ Significance of 2.85σ obs. (2.62 exp.)

at 125 GeV

◎ Strong indication that the new particle

decays to taus!

slide-17
SLIDE 17

17 EPS-HEP 2013, Stockholm

VBF H→μτh event candidate

Thank you!

slide-18
SLIDE 18

18 EPS-HEP 2013, Stockholm

Additional material

slide-19
SLIDE 19

19 EPS-HEP 2013, Stockholm

Luminosity

After high quality selection

  • 4.9/fb at 7 TeV (2011)
  • 19.4/fb at 8 TeV (2012)
slide-20
SLIDE 20

20 EPS-HEP 2013, Stockholm

VBF H→μτh event candidate

slide-21
SLIDE 21

21 EPS-HEP 2013, Stockholm

Systematic uncertainties (exp)

slide-22
SLIDE 22

22 EPS-HEP 2013, Stockholm

Systematic uncertainties (th)

slide-23
SLIDE 23

23 EPS-HEP 2013, Stockholm

Sensitivity break-down

by channel by category

  • Expected limits for full 2011+2012 dataset
  • Sensitivity of 0.76 x SM at 125 GeV
  • Sensitivity lead by the μτh channel
slide-24
SLIDE 24

24 EPS-HEP 2013, Stockholm

Combined mass distribution

  • Combine channels and categories weighting each by S/B
  • S – expected signal
  • B – fitted background
  • Both S and B in window around peak at 125 GeV Higgs
slide-25
SLIDE 25

25 EPS-HEP 2013, Stockholm

Consistency with SM H(125)

H(125) as background H(125) signal injected

  • Result consistent with background + SM Higgs mH = 125 GeV
slide-26
SLIDE 26

26 EPS-HEP 2013, Stockholm

Event yields and efficiency

μτh eτh τhτh eμ

slide-27
SLIDE 27

27 EPS-HEP 2013, Stockholm

Mass distributions (1-jet)

  • 1-jet, high-pT: second

most sensitive category

  • ggH dominated
  • Jet and high-pT

requirements improve resolution and S/B

slide-28
SLIDE 28

28 EPS-HEP 2013, Stockholm

Mass distributions (0-jet)

  • 0-jet:
  • Low S/B
  • Background fit only
  • To constrain

uncertainties in signal sensitive categories

slide-29
SLIDE 29

29 EPS-HEP 2013, Stockholm

Extension to low mass

  • 8 TeV (2012) only:
  • Limit extended down to 90 GeV
  • Combining the eτh, μτh, eμ and τhτh channels. The μμ and VH channels

not included.

H(125) as background

slide-30
SLIDE 30

30 EPS-HEP 2013, Stockholm

Jets

◎ Used for event categorisation

○ Specially important as tags of VBF H

◎ “Fake” jets from pileup

○ High-ET jets from overlapping pileup jets

◎ Discriminate “fake” jets with MVA using

○ Track-vertex association ○ Jet shape

◎ Reduces background in VBF category

by a factor of ~2

slide-31
SLIDE 31

31 EPS-HEP 2013, Stockholm

Missing Transverse Energy

◎ Used to reject background and for full

di-tau mass reconstruction

○ Specially important as tags of VBF H

◎ Resolution of MET degrades with pileup

○ <19> interactions in 2012

◎ Use MVA (BDT regression) using 5

different MET estimators as input

◎ Test: compare components of the recoil

to the hard probe (e.g. in Z→μμ)

○ Standard MET (particle-flow) vs MVA MET

◎ Significant improvement in resolution,

pileup dependence suppressed