Search for the Standard Model Higgs Boson in the two photon decay - - PowerPoint PPT Presentation

Search for the Standard Model Higgs Boson in the two photon decay mode

Nicholas Wardle

QM IOP - April 4, 2012

1 / 21

◮ Introduction ◮ Analysis Strategy ◮ Results ◮ Conclusions

2 / 21

Introduction

Search performed at CMS on 4.8fb−1 pp dataset at √s = 7TeV Update from recent paper.

3 / 21

Event Selection

◮ Two high pT photons ◮ MVA ID quality cut to reject fakes.

◮ isolation variables to reject jets ◮ shower shape variables for π0

◮ Vertex from track information.

◮ Correct vertex found 85% of the time.

pT mγγ > 1 3, 1 4

4 / 21

Search Strategy

Sensitivity dominated by H → γγ peak Width of the peak is driven by the experimental resolution Search is performed in a ±2% window around the peak

5 / 21

Event Classification

◮ Photon Kinematics

◮ pγ1

T /mγγ, pγ2 T /mγγ

◮ ηγ1, ηγ2, cos(∆φ)

◮ Photon Discriminator

◮ IDMVAγ1, IDMVAγ2

◮ Event Resolution

◮ σ(correct)/mγγ, σ(wrong)/mγγ ◮ vtxprob

Combine with (∆M/MH) using a second BDT Inputs invariant with mH

6 / 21

Di-jet Category

Look for two high ET forward Jets. Jets back to back with di-photon system.

)

2

(GeV/c

γ γ

m

100 120 140 160 180

)

2

Events / ( 1 GeV/c

2 4 6 8 10 12 14 16 18

Data Bkg Model σ 1 ± σ 2 ± =120 GeV

H

1xSM m

• 1

= 7 TeV L = 4.76 fb s CMS preliminary

BDT >= 0.05 VBF Tag

7 / 21

Signal Model

Signal model derived from MC simulation. MC corrected for detector effects (data-driven measurements in Z → ee and Z → µµγ) Dominant systematics are energy scale and resolution )

2

(GeV/c

γ γ

m

100 110 120 130

)

2

Events / ( 0.35 GeV/c

1 2 3 4 5 6 7

Simulation Parametric Model

2

= 1.82 GeV/c

eff

σ

Simulation CMS preliminary

All Categories Combined

2

FWHM = 3.29 GeV/c 8 / 21

Background Model - Normalization

fbkg(mγγ) = N(m−a

γγ + bm−c γγ )

Total background normalization taken from fit to sidebands. Systematic included for remaining potential bias from wrong function choice. Single parameter in the Likelihood

9 / 21

Background Model - Shape

Define sidebands

• utside the signal

region Shape in BDT extracted from linear fits in mH

◮ Only assumption is variation as a

function of mass is linear

◮ Systematic included to account

for potential higher order terms

10 / 21

Results

BDT Bin 1 BDT Bin 2 BDT Bin 3 BDT Bin 4 BDT Bin 5 BDT Bin 6 BDT Bin 7 Di-jet

1 10

2

10

3

10

4

10

Data =124.0 GeV (x5)

H

Higgs, m Background σ 1 ± σ 2 ± = 7 TeV s CMS preliminary,

• 1

L = 4.8 fb

∫

Data =124.0 GeV (x5)

H

Higgs, m Background σ 1 ± σ 2 ±

Categories optimized so as to achieve the best sensitivity to signal. Likelihood constructed as a combination of counting experiments.

11 / 21

Exclusion Limits

)

2

(GeV/c

H

m 110 115 120 125 130 135 140 145 150

SM

) γ γ → (H σ /

95%CL

) γ γ → (H σ 0.5 1 1.5 2 2.5 3 3.5 4

• 1

= 7 TeV L = 4.8 fb s CMS preliminary

Observed CLs Limit Median Expected CLs Limit Expected CLs σ 1 ± Expected CLs σ 2 ±

◮ Exclusion limits extracted in the range 110 < mH < 150 ◮ Expected exclusion limit less than 1.5xSM in 110 < mH < 140 ◮ Several ranges excluded at 95% confidence level (CL).

12 / 21

Significance Scan

◮ Significance and best-fit

signal strength are extracted in the range 110 < mH < 150

◮ Largest local significance

around 3σ

◮ Look elsewhere effect

reduces this to 1.9σ

110 115 120 125 130 135 140 145 150

µ

1 2 3 (GeV)

H

m 110 115 120 125 130 135 140 145 150 p-value

• 4

10

• 3

10

• 2

10

• 1

10 1 σ 1 σ 2 σ 3

1xSM Higgs Median Expected (combined) Local P-value (no di-jet) Local P-value (di-jet only) Local P-value Observed p-value (Ensemble)

• 1

= 7 TeV L = 4.8 fb s CMS preliminary

13 / 21

Conclusions

◮ A search for the SM Higgs has been performed on 4.8fb−1 of pp collisions

at CMS

◮ Exclusions placed on the SM Higgs at several ranges of mass hypotheses. ◮ No significant excess is observed in the search range. ◮ Can expect to become sensitive to SM Higgs by summer with twice the

current dataset.

◮ The analysis is described in CMS public document HIG-12-001.

14 / 21

Backup

15 / 21

Analysis Workflow

16 / 21

VBF Selection

Jet Tag

ET > 30/20GeV mjj > 350GeV |ηj

1 − ηj 2 > 3.5|

|φjj − φγγ > 2.6| |0.5(ηj

1 − ηj 2) − ηγγ| < 2.5

17 / 21

Signal MC Corrections

MC additional smearing measured using Z → ee MC-data likelihood fit in categories.

(H) (GeV/c)

T

p

50 100 150 200 250

Fraction of events

0.2 0.4 0.6 0.8 1

=9.5 〉

PU

n 〈 Simulation CMS preliminary

< 10 mm

true

• z

reco

z

Vtx Probabilities measured in Z → µµγ as a function of boson pT, MC reweighted.

18 / 21

Signal Systematics

19 / 21

MVA Validation using Z → ee

MC simulation is validated using Z → ee MC-data comparisons.

20 / 21

CLs

Limit setting with the CLs procedure:

◮ Fit toy data, once conditionally with a

fixed µ and once with µ floating.

◮ Systematics are incorporated into

Likelihood, constraints randomized in the fits (nuisance parameters).

◮ Ratio of Likelihoods is the Profile

Likelihood, we use the test statistic - qµ = −2 ln L(µ, ˆ θ) L(ˆ µ, ˆ ˆ θ)

◮ Throw toys from model fit to data to

generate the distribution of under signal+background and background only hypotheses

◮ Exclude at 95% confidence when ratio of

p-values CLs = CLs+b/CLb is 0.05. =2) µ (

µ

Test Statistic q

5 10 15 20

Number of toys

1 10

2

10

3

10

4

10

5

10

6

10

) for signal+bkgd pseudo-data

µ

f(q ) for bkgd-only pseudo-data

µ

f(q

• bserved

µ

q

• 1

=0.2-0.9 fb

int

=7 TeV L s CMS Preliminary = 250 GeV

H

Higgs Combination at m 21 / 21