Measurements of Higgs boson production and properties in ...€¦ · Introduction V.R.Tavolaro 04.08.2016 2 • Clean ﬁnal state with two highly energetic photons • Final state

Date and place

Measurements of Higgs boson production and properties in the di-photon decay channel using

the CMS detector

Vittorio Raoul Tavolaro on behalf of the CMS collaboration

38th International Conference on High Energy Physics, 4 August 2016, Chicago IL

Introduction

V.R.Tavolaro 204.08.2016

• Clean final state with two highly energetic photons

• Final state fully reconstructed with high resolution

• Very small branching fraction (~0.2%)

• Large backgrounds (ɣɣ, ɣj, jj)

• Exclusive categories targeting:gluon-gluon fusion (ggH),vector boson fusion (VBF)and ttH production modes

• Search for a narrow peak on a falling background in mass distribution

• 2016 dataset analysed,12.9/fb collected at 13 TeV (HIG-16-020) [GeV] HM

120 122 124 126 128 130

H+X

) [pb

]

→(p

p σ

1−10

1

10

210 = 13 TeVs

LHC

HIG

GS

XS W

G 2

016

H (NNLO+NNLL QCD + NLO EW)→pp

qqH (NNLO QCD + NLO EW)→pp

WH (NNLO QCD + NLO EW)→pp ZH (NNLO QCD + NLO EW)→pp

ttH (NLO QCD + NLO EW)→pp

bbH (NNLO QCD in 5FS, NLO QCD in 4FS)→pp

tH (NLO QCD)→pp

γ γTransformed BDT0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Even

ts/0

.02

1

10

210

310

410

510

610

710 Datajet jet jetγγ γ

MC stat. uncert.

ggHVBFVHttH

Preliminary CMS (13TeV)-112.9 fb

Simulation background =125 GeVH

, mγγ→SM H

Analysis strategy


• Photons energy reconstructed and corrected, rejection of fake photons • Events categorised into classes (S/B, mass resolution, additional particles) to improve

the analysis sensitivity • Extraction of signal through fit of di-photon invariant mass spectrum in each category

Di-photoncategorisation

Even

ts /

GeV

0

1000

2000

3000

4000

5000

6000DataS+B fitB component

σ1 ±σ2 ±

Untagged 3 =0.90µ=125.09 GeV, Hm --

Preliminary CMS TeV) (13-1 12.9 fbγγ→H

(GeV)γγm100 110 120 130 140 150 160 170 180150−

100−50−0

50100150200 B component subtracted

Even

ts /

GeV

0

1000

2000

3000

4000

5000


σ1 ±σ2 ±



(GeV)γγm100 110 120 130 140 150 160 170 180150−

100−50−0

50100150200 B component subtractedEv

ents

/ G

eV

0

200

400

600

800

1000

1200

1400

1600

1800

2000

DataS+B fitB component

σ1 ±σ2 ±



(GeV)γγm100 110 120 130 140 150 160 170 180

100−

50−

0

50


Even

ts /

GeV

0

200

400

600

800

1000

1200

1400

1600

1800

2000


σ1 ±σ2 ±



(GeV)γγm100 110 120 130 140 150 160 170 180

100−

50−

0

50


Even

ts /

GeV

0

100

200

300

400

500


σ1 ±σ2 ±



(GeV)γγm100 110 120 130 140 150 160 170 180

60−40−20−0


Even

ts /

GeV

0

100

200

300

400

500


σ1 ±σ2 ±



(GeV)γγm100 110 120 130 140 150 160 170 180

60−40−20−0


Even

ts /

GeV

02468

10121416182022


σ1 ±σ2 ±



(GeV)γγm100 110 120 130 140 150 160 170 180

5−

0

5


Even

ts /

GeV

02468

10121416182022


σ1 ±σ2 ±



(GeV)γγm100 110 120 130 140 150 160 170 180

5−

0

5


(GeV)- e+em80 82 84 86 88 90 92 94 96 98 100

Even

ts/(0

.5 G

eV)

0

5000

10000

15000

20000

25000

30000

35000Data

(MC)-e+ e→Z

MC stat. uncert.


) > 0.9429, R19min(R) < 1.442η, 1ηmax(

(GeV)γγm105 110 115 120 125 130 135 140

Even

ts /

( 0.5

GeV

)

0

10

20

30

40

50

60

70

80

FWHM = 3.58 GeV

All Categoriesγγ→H

Simulation

modelParametric

= 1.88 GeVeffσ

13 TeVCMS Simulation Preliminary

Photon Energy


• Electro-magnetic calorimeter (ECAL) response:

• corrected for change in time

• inter-calibrated to be uniform in η/φ

• adjustment of absolute scale

• Energy and its uncertainty corrected for local and global shower containment:

• regression targeting Etrue/Ereco

• Scale vs time and resolution calibration: Z→ee peak used as reference

• Corrected energies and resolutions used in the analysis

σeff = 1.18GeV in best cat.

(GeV)T

Z boson p0 20 40 60 80 100 120 140 160 180 200

data

/sim

ulat

ion

0.80.9

11.11.2

pt| <

10

mm

µµ

- z

sele

cted

fract

ion

of |z

0

0.2

0.4

0.6

0.8

1

CMS Preliminary (13 TeV)-112.6 fb

, 25 nsµµ →Z simulationdata

(GeV)γγ

Tp

0 50 100 150 200 250

| < 1

0 m

mtru

e -

zse

lect

edFr

actio

n of

|z

0.5

0.6

0.7

0.8

0.9

1

1.1

True vertex efficiency

Average vertex probability

CMS 13 TeV Simulation Preliminary

)-1Data PU scenario (12.9 fb = 125 GeV)

H (mγγ →H

Vertex identification


• Vertex assignment correct within 1 cm → negligible impact on mass resolution

• No ionization in the tracker for photons

• Multi-variate approach for vertex identification

• exploit kinematic correlations and track distribution imbalance

• direction of conversion tracks, when present

• Second MVA estimates probability of correct vertex choice, used for di-photon classification

• Method validated on Z→μμ (ɣ+j for converted ɣ) events, where vertex found after removing muon tracks

εvtx = 80%

Even

ts/0

.02

0

100

200

300

400

500

600

Dataee MC→Z

MC syst.

310× Preliminary CMS (13 TeV)-112.9 fb

Photon ID BDT score0.8− 0.6− 0.4− 0.2− 0 0.2 0.4 0.6 0.8 1

Dat

a/M

C

0.51

1.5

Photon selection


• MVA based photon ID classifier:

• goal: discriminate between prompt and fake photons

• inputs are shower shape variables, particle-flow isolations, kinematic of photon, median energy density (ρ)

• Double-photon trigger selection based on transverse energy, mɣɣ, isolation and electromagnetic shower shapes variables

• Minimal pre-selection, similar to but tighter than trigger selection

ttH leptonic tagttH tags

• (sub)lead pT/mɣɣ > 1/2(1/4) • at least one lepton (ℓ=e,μ), away

from Z peak • ≥ 2 jets • ≥ 1 b-jet

• (sub)lead pT/mɣɣ > 1/2(1/4) • 0 leptons • ≥ 5 jets • ≥ 1 b-jet

ttH tags

VBF tags

Untagged

ɣɣTagged events


Feynman Diagrams

Andrea Carlo Marini

November 25, 2015

1 Higgs

ggŸH

g

g

p

p

H

qq̄ŸH

q̄

q

p

p

HZ/W

qq̄ŸVH

q

qZ/W

p

pH

Z/W

ppŸtt̄H

t̄

t

p

p

Hg t̄g

g

t

p

p

Hq t̄q

g

t

p

p

H

1

ttH hadronic tag

ttH leptonic tagttH tags

• (sub)lead pT/mɣɣ > 1/2(1/4) • at least one lepton (ℓ=e,μ), away

from Z peak • ≥ 2 jets • ≥ 1 b-jet

• (sub)lead pT/mɣɣ > 1/2(1/4) • 0 leptons • ≥ 5 jets • ≥ 1 b-jet

VBF tags

ttH tags

VBF tags

Untagged

ɣɣTagged events


• Identify events with 2 jets through a MVA • inputs: pT/mɣɣ of both photons, pT of both jets, mjj, Δηjj,

Zeppenfeld variable, Δφɣɣjj

• 2 jets with pT1 > 30GeV, pT2 >20 GeV, |η| < 4.7, mjj >250 GeV

• VBF Classification BDT combines di-jet and diphoton BDT: 2 data categories (VBF tag 0-1)

Feynman Diagrams

Andrea Carlo Marini

November 25, 2015

1 Higgs

ggŸH

g

g

p

p

H

qq̄ŸH

q̄

q

p

p

HZ/W

qq̄ŸVH

q

qZ/W

p

pH

Z/W

ppŸtt̄H

t̄

t

p

p

Hg t̄g

g

t

p

p

Hq t̄q

g

t

p

p

H

1

Feynman Diagrams

Andrea Carlo Marini

November 25, 2015

1 Higgs

ggŸH

g

g

p

p

H

qq̄ŸH

q̄

q

p

p

HZ/W

qq̄ŸVH

q

qZ/W

p

pH

Z/W

ppŸtt̄H

t̄

t

p

p

Hg t̄g

g

t

p

p

Hq t̄q

g

t

p

p

H

1

ttH hadronic tag


Even

ts/0

.02

1

10

210

310

410

510

610

710 Datajet jet jetγγ γ

MC stat. uncert.

ggHVBFVHttH


Simulation background =125 GeVH

, mγγ→SM H

“Untagged” events


• Multivariate discriminator is used to separate diphoton pairs with signal-like kinematics, high photon ID scores and good mass resolution from background

• Input variables: pT of the two photons rescaled by pair mass, η, cos(Δφ), photon ID scores, relative mass resolutions under correct/incorrect vertex selection hypotheses, vertex probability estimate

• 4 event categories (“Untagged” events)

• Categorise events into classes (S/B, resolution) to improve the analysis sensitivity ttH tags

VBF tags

Untagged

ɣɣ

Unta

gged

3

Unta

gged

2

Unta

gged

1

Unta

gged

0

Even

ts /

GeV

0

1000

2000

3000

4000

5000


σ1 ±σ2 ±



(GeV)γγm100 110 120 130 140 150 160 170 180150−

100−50−0


(GeV)γγm105 110 115 120 125 130 135 140

Even

ts /

( 0.5

GeV

)

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

2.2

FWHM = 2.42 GeV

Untagged 0γγ→H

Simulation

modelParametric

= 1.18 GeVeffσ

13 TeVCMS Simulation Preliminary

Signal and background models


• Fully parametric signal model from simulation

• physical nuisances allowed to float

• corrections and data/MC efficiency scale factors applied

• Background model data driven:

• background functional form treated as discrete nuisance parameter

• for each category, use different functional forms(sums of exponentials, sums of power law terms, Laurent series and polynomials)

Even

ts /

GeV

0

2

4

6

8

10


σ1 ±σ2 ±

TTH Hadronic Tag =0.95µ=126.0 GeV, Hm --


(GeV)γγm100 110 120 130 140 150 160 170 1804−2−02468


Even

ts /

GeV

0

1

2

3

4

5

6

7

8

9


σ1 ±σ2 ±

TTH Leptonic Tag =0.95µ=126.0 GeV, Hm --


(GeV)γγm100 110 120 130 140 150 160 170 1802−

0

2

4


Even

ts /

GeV

024

68

1012

14

1618

2022


σ1 ±σ2 ±

VBF Tag 0 =0.95µ=126.0 GeV, Hm --


(GeV)γγm100 110 120 130 140 150 160 170 180

5−

0

5


Even

ts /

GeV

0

20

40

60

80

100


σ1 ±σ2 ±

VBF Tag 1 =0.95µ=126.0 GeV, Hm --


(GeV)γγm100 110 120 130 140 150 160 170 180

10−

0

10

20


Mass spectra, tagged events


ttH leptonic

VBF tag 0VBF tag 1

ttH hadronic

Even

ts /

GeV

02468

10121416182022


σ1 ±σ2 ±



(GeV)γγm100 110 120 130 140 150 160 170 180

5−

0

5


Even

ts /

GeV

0

100

200

300

400

500


σ1 ±σ2 ±



(GeV)γγm100 110 120 130 140 150 160 170 180

60−40−20−0


Even

ts /

GeV

0

200

400

600

800

1000

1200

1400

1600

1800

2000


σ1 ±σ2 ±



(GeV)γγm100 110 120 130 140 150 160 170 180

100−

50−

0

50


Even

ts /

GeV

0

1000

2000

3000

4000

5000


σ1 ±σ2 ±



(GeV)γγm100 110 120 130 140 150 160 170 180150−

100−50−0


Mass spectra, untagged events


Untagged 0Untagged 1

Untagged 2Untagged 3

Even

ts /

GeV

0

1000

2000

3000

4000

5000

6000

7000

8000

9000


σ1 ±σ2 ±

All categories=0.95µ=126.0 GeV, Hm --


(GeV)γγm100 110 120 130 140 150 160 170 180

100−

0

100


S/(S

+B) W

eigh

ted

Even

ts /

GeV

0

1000

2000

3000

4000

5000


σ1 ±σ2 ±

S/(S+B) weightedAll categories=0.95µ=126.0 GeV, Hm

--


(GeV)γγm100 110 120 130 140 150 160 170 180

100−

0

100


Mass spectra, all categories


(GeV)Hm116 118 120 122 124 126 128 130 132 134

Loca

l p-v

alue

10−10

9−10

8−10

7−10

6−10

5−10

4−10

3−10

2−10

1−10

1σ1 σ2

σ3

σ4

σ5

σ6

Observed

Expected

= 125.09 GeVHmExpected

Preliminary CMSγγ→H

(13TeV)-112.9 fb

Results


• Significance at 125.09 GeV: 5.6σ observed (6.2σ expected)

• Maximum observed significance is 6.1σ at 126.0 GeV

• Best-fit signal strength

µ2− 0 2 4 6 8

-1.2+1.5 1.91

ttHµ

-0.8+0.9 1.61

VBFµ

-0.23+0.25 0.77

ggHµ

0.18−0.21+ = 0.95

combinedµ

ProfiledHm

σ 1±Combined

σ 1±Per category

SMµ=µ

Preliminary CMS

γγ→H

TeV) (13-1 12.9 fb

= 1VHµ

ggH,ttHµ

0.5− 0 0.5 1 1.5 2 2.5 3

VBF,

VHµ

0

0.5

1

1.5

2

2.5

3

3.5

4Best Fit

SM

σ1

σ2

Preliminary CMSγγ→H

TeV) (13-112.9 fb

ProfiledHm

• Signal strength measured is measured in bosonic and fermionic components

Results


• Production mechanism signal strengths measurements compatible with SM

• More details on mass: see Ulascan Sarica’s talk“Combined results of the 125 GeV Higgs boson on the mass, tensor structure, and couplings measured by the CMS detector”, today@1:10pm

decorr|γγ/mmσ

0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 0.045 0.05

Even

ts/0

.000

5

0

20

40

60

80

100

120

140

160

310×

Data

(MC)-e+ e→Z

MC syst. uncert.


) < 1.442η, 1ηmax(

(TeV)s7 8 9 10 11 12 13 14

(fb)

fid.

σ

20

30

40

50

60

70

80

90

100Preliminary CMS

(13 TeV)-1 (8 TeV) + 12.9 fb-119.7 fb

γγ→H)Hbest-fit mData (

syst. uncertainty)=125.09 GeVHmSM (

- norm. LHC Higgs XSWG YR4MC@NLOA- acc.

Fiducial cross section


• Fiducial cross section measured profiling mH:

• Theoretical prediction for mH=125.09 GeV

• Different event categorisation: 3 mass resolution categories

Summary


• CMS H→ɣɣ results using 12.9/fb of 13 TeV collision data collected in 2016 have been presented

• Observation (6.1σ peak significance) of the Higgs boson in the di-photon channel

• Best fit signal strength is

• Bosonic and fermionic components of signal strength are observed

• Fiducial cross-section is measured to be


BACKUP

Event yields per category


Event yields per category


Vκ0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

fκ

1−

0.5−

0

0.5

1

1.5

2

2.5

3 ) fκ, Vκq(

0

2

4

6

8

10

Best Fitσ1σ2

SM

Preliminary CMS TeV) (13-112.9 fb

γγ→H ProfiledHm

γκ0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

gκ

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2 ) gκ, γκq(

0

2

4

6

8

10

Best Fitσ1σ2

SM

Preliminary CMS TeV) (13-112.9 fb

γγ→H ProfiledHm

• Measurement of coupling modifiers to vector bosons and fermions (kV, kf) and to photons and gluons (kɣ,kg)

Analysis flowchart


Energy regression

Vtx ID BDT

Vtx prob BDT

Pho ID BDT

DiPho BDT

DiJet BDT

VBF BDT

Untagged cats.

VBF cats.

Scale and res. corr.

E σE/E

ttH cats.

ttH req.

• Vertex ID BDT:

• if conversions are present

• Vertex probability BDT

Vertex BDTs inputs


Preselection details


• pT>30 (20) GeV, pT/mɣɣ > 1/3 (1/4) for (sub)leading-pT photon

• |η|<2.5, removing 1.44<|η|<1.57, electron veto

• either R9>0.8, or charged hadron isolation < 20 GeV, or charged hadron isolation relative to pT <0.3

BDTɣ ID inputs


BDTɣ ID


• BDTɣ ID of the lowest-scoring photon for data and simulation, after preselection

BDT score of the photon ID1− 0.8− 0.6− 0.4− 0.2− 0 0.2 0.4 0.6 0.8 1

Even

ts

210

310

410

510

610

710

810DataSimulation:

4 = 125 GeV)x10H

(mγγ→Htotal background+stat.uncert.γ-γ

-jetγjet-jet

DataSimulation:

4 = 125 GeV)x10H


-jetγjet-jet

DataSimulation:

4 = 125 GeV)x10H


-jetγjet-jet

DataSimulation:

4 = 125 GeV)x10H


-jetγjet-jet

DataSimulation:

4 = 125 GeV)x10H


-jetγjet-jet

DataSimulation:

4 = 125 GeV)x10H


-jetγjet-jet

(13 TeV)-112.9 fbCMS Preliminary

BDTɣɣ inputs


BDTɣɣ systematic uncertainties



Even

ts/0

.04

0

20

40

60

80

100

120

310×

Data

(MC)-e+ e→Z

MC syst. uncert.


• Impact on BDTɣɣ of systematic uncertainties associated to:

• relative energy resolution (±5% relative shift)

• photon ID BDT (± 0.03 shift plus linearly increasing term)

• Z→ee events with electrons reconstructed as photons

ttH leptonic tag details


ttH hadronic tag details


Systematic uncertainties


• Theory uncertainties (PDFs, αS, QCD scale, underlying event and parton shower, H→ɣɣ branching fraction)

• ggH contamination in VBF and ttH tagged categories

• Trigger efficiency, integrated luminosity, vertex efficiency, preselection

• Non-unformity of light collection, non-linearity, detector simulation, modeling of the material budget, shower shape corrections

• Photon energy scale and resolution

• BDTɣ ID and per-photon energy resolution

• Jet energy scale and smearings

• b-tagging efficiency, gluon-splitting fraction, parton shower, ID efficiency for e and μ

µ2− 0 2 4 6 8-1.4+2.0TTH Leptonic Tag 1.15

-1.2+1.6TTH Hadronic Tag 2.10

-0.8+1.2VBF Tag 1 0.04

-0.6+0.8VBF Tag 0 1.58

-0.6+0.9Untagged 3 1.33

-0.31+0.32Untagged 2 0.44

-0.29+0.33Untagged 1 1.04

-0.5+0.6Untagged 0 0.95

0.18−0.21+ = 0.95

combinedµ

ProfiledHm

σ 1±Combined

σ 1±Per category

SMµ=µ

Preliminary CMS

γγ→H

TeV) (13-1 12.9 fb

Per category signal strength


Fiducial cross section: mass spectra


Even

ts /

GeV

0

100

200

300

400

500

600

700

800


σ1 ±σ2 ±

category 0 decorr

/M |Mσ=126.0 GeVHm --


(GeV)γγm100 110 120 130 140 150 160 170 180

60−

40−

20−


Even

ts /

GeV

0

200

400

600

800

1000

1200

1400

1600


σ1 ±σ2 ±

category 1 decorr

/M |Mσ=126.0 GeVHm --


(GeV)γγm100 110 120 130 140 150 160 170 180

80−60−40−20−0


Even

ts /

GeV

0

500

1000

1500

2000

2500

3000

3500

4000

4500


σ1 ±σ2 ±

category 2 decorr

/M |Mσ=126.0 GeVHm --


(GeV)γγm100 110 120 130 140 150 160 170 180

150−100−

50−0


Documents

Measurements of Higgs boson production and properties in ...€¦ · Introduction V.R.Tavolaro 04.08.2016 2 • Clean ﬁnal state with two highly energetic photons • Final state