[PPT] - Search for heavy right-handed gauge bosons decaying into boosted PowerPoint Presentation

SLIDE 1

Debarati Roy

n behalf of the ATLAS Collaboration

ILHC-ICTP2019

1

Search for heavy right-handed gauge bosons decaying into boosted heavy neutrinos with the ATLAS detector at √s = 13 TeV

SLIDE 2

2

Standard Model (SM) and beyond

SM :

=> Extremely successful theory. => Guided through new particle discoveries (Higgs boson glorifies its success in 2012!)

Couple of experimental observations

SM cannot explain direct towards new Physics ✨✨👼 ✨✨ => Neutrino oscillation concludes neutrino has a very small mass. => Several searches performed in LHC to explain origin of a very small neutrino mass!

¯ q q WR NR ` W ∗

R

` ¯ q q

Left Right Symmetric Model

(LRSM) : => Restores parity by introducing right-handed gauge bosons (WR) & right-handed neutrinos (NR). => Small neutrino mass can be explained via its coupling to NR via mass mixing matrix. Final state => 2 jets + 2 leptons (resolved topology)

mWR ~ TeV

mNR ~ GeV

SLIDE 3

3

Extending phase space with boosted topology

Several searches performed in ATLAS (& in CMS) with resolved topology.

[TeV]

R

W

m [TeV]

R

N

m

ATLAS

1

=13 TeV, 36.1 fb s channel ee ,

R

N Majorana

L

g =

R

g

Obs. 95% CL limit
Exp. 95% CL limit

σ 1 ±

Exp. limit

σ 2 ±

Exp. limit

1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 0.5 1 1.5 2 2.5 3 3.5 4

Latest ATLAS result in resolved scenario

mNR << mWR :

Less explored phase space

with limited discovery potential estimation => Sensitivity drops with resolved topology.

More efficient to consider

boosted scenario!

mNR << mWR

arXiv:1904.12679

First time we looked at

possibility for boosted heavy neutrinos in ATLAS with 80 fb-1

f data at 13 TeV.

SLIDE 4

4

Boosted heavy neutrino search : Introduction

Final state consists of a large radius jet &

two leptons.

Electron (e) & muon (μ) final states looked

at separately with no flavour mixing.

A balancing topology between hardest e (e1)
r μ (μ1) & highest mass large radius jet (j)

along with 2nd hardest e (e2) or μ (μ2) inside that large radius jet gives well shaped detector level variables.

Different NR mass computation performed

between e & μ final states due to nature of jet reconstruction :

e channel : Mass of large radius jet (e energy

part of j energy, a distinguishing feature of this search) μ channel : Mass of large radius jet & μ2.

[GeV]

j

m

200 400 600 800 1000

Events / 40 GeV

2 −

10

1 −

10 1 10

2

10

3

10

300, 3000 400, 4000 500, 5000 600, 6000

SR (ee)

R

W

, M

R

N

Signal: M

[GeV], [GeV]

1

= 13 TeV, L = 80 fb s ATLAS Simulation

[GeV]

2 µ j,

m

200 400 600 800 1000

Events / 40 GeV

2 −

10

1 −

10 1 10

2

10

3

10

300, 3000 400, 4000 500, 5000 600, 6000

) µ µ SR (

R

W

, M

R

N

Signal: M

[GeV], [GeV]

1

= 13 TeV, L = 80 fb s ATLAS Simulation

arXiv:1904.12679

SLIDE 5

5

Boosted heavy neutrino search : Analysis Selection

Object Selection :

Exactly 2 leptons & at least 1 large radius trimmed jet.
Isolated e1/μ1 & non-isolated e2/μ2 (2nd hardest leptons allowed to

be close to large radius jet).

Highest mass large radius (R = 1.0) jet (j) used with pT > 200 GeV, |η|

< 2.0 (mj > 50 GeV in e final state).

pT,e1/e2 > 26 GeV, |η| < 2.47 excluding crack region. pT,μ1/μ2 > 28 GeV,

|η| < 2.5. Topological Cuts :

Azimuthal separation (dΦ)

between e1/μ1 & j > 2.0.

∆R between e2/μ2 & j < 1.0.

Further Background Reduction Cuts :

Dilepton invariant mass (mll) >

200 GeV.

dΦ between e1(μ1) & e2(μ2) > 1.5.

0.34 0.21 0.12 0.08 0.48 0.44 0.31 0.20 0.46 0.37 0.24 0.38 0.26 0.27

2000 3000 4000 5000 6000

[GeV]

R

W

m

100 200 300 400 500 600 700 800

[GeV]

R

N

m

0.40 0.34 0.23 0.14 0.44 0.41 0.30 0.19 0.39 0.31 0.20 0.32 0.20 0.20

Signal selection efficiency ATLAS Simulation

Muon channel Electron channel

arXiv:1904.12679

More boosted More boosted

SLIDE 6

6

mWR : The discriminating variable for region definition

mWR Computation in e final state :

Invariant mass of j + e1.

mWR Computation in μ final state :

Invariant mass of j + μ1 + μ2.

Control Region (CR : mWR < 2 TeV)

shows reasonable data-mc agreement including statistical uncertainty.

Signal Region (SR : mWR > 2 TeV).
A Validation Region (VR) studied

with a hard e inside j balanced by a μ to conclude that data can be well predicted by mc (when a hard e inside j).

arXiv:1904.12679

Events / 100 GeV

1 10

2

10

3

10

4

10

Data t t Z+jet(s) Single-t W+jet(s) Diboson MC stat. unc.

CR (ee)

1

= 13 TeV, 80 fb s ATLAS

[GeV]

j,e1

m

400 600 800 1000 1200 1400 1600 1800 2000

Data / Pred.

0.5 1 1.5 2

Events / 100 GeV

1 10

2

10

3

10

Data t t Single-t Diboson MC stat. unc.

e) µ VR (

1

= 13 TeV, 80 fb s ATLAS

[GeV]

1 µ j,

m

500 1000 1500 2000 2500 3000

Data / Pred.

0.5 1 1.5 2

↑ ↑↑ ↑

SLIDE 7

7

Performance of large radius jet with a hard e inside

Large radius jet reconstruction in ATLAS based on energy clusters

calibrated at hadronic scale.

Effect of a non-negligible fraction of

EM clusters in j investigated in terms

f jet energy scale (JES) & jet mass

scale (JMS) as a function of ratio of energy of e to the energy of j.

A weak dependence

(within scale expected uncertainty range) concludes no additional correction factor needs to be implemented.

arXiv:1904.12679

j

/ E

e2

E

0.2 0.4 0.6 0.8 1

JMS

0.99 1 1.01 1.02 1.03 1.04 1.05

(150 GeV)

R

N

(3 TeV), m

R

W

m (300 GeV)

R

N

(3 TeV), m

R

W

m (400 GeV)

R

N

(4 TeV), m

R

W

m (500 GeV)

R

N

(5 TeV), m

R

W

m

= 13 TeV s ATLAS Simulation

SLIDE 8

8

Performance of large radius jet with a hard e inside

Large radius jet reconstruction in ATLAS based on energy clusters

calibrated at hadronic scale.

Effect of a non-negligible fraction of

EM clusters in j investigated in terms

f jet energy scale (JES) & jet mass

scale (JMS) as a function of ratio of energy of e to the energy of j.

A weak dependence

(within scale expected uncertainty range) concludes no additional correction factor needs to be implemented.

arXiv:1904.12679

j

/ E

e2

E

0.2 0.4 0.6 0.8 1

JES

0.99 1 1.01 1.02 1.03 1.04 1.05

(150 GeV)

R

N

(3 TeV), m

R

W

m (300 GeV)

R

N

(3 TeV), m

R

W

m (400 GeV)

R

N

(4 TeV), m

R

W

m (500 GeV)

R

N

(5 TeV), m

R

W

m

= 13 TeV s ATLAS Simulation

SLIDE 9

9

Performance of large radius jet with a hard e inside

JMS as a function of generator level large radius jet mass shows

reasonable behaviour :

Effect of a non-negligible fraction of

EM clusters in j investigated in terms

f jet energy & jet mass scales as a

function of ratio of energy of e to the energy of j.

A weak dependence

(within scale expected uncertainty range) concludes no additional correction factor needs to be implemented.

arXiv:1904.12679

j

/ E

e2

E

0.2 0.4 0.6 0.8 1

JES

0.99 1 1.01 1.02 1.03 1.04 1.05

(150 GeV)

R

N

(3 TeV), m

R

W

m (300 GeV)

R

N

(3 TeV), m

R

W

m (400 GeV)

R

N

(4 TeV), m

R

W

m (500 GeV)

R

N

(5 TeV), m

R

W

m

= 13 TeV s ATLAS Simulation

[GeV]

gen j

m

100 200 300 400 500 600

3

10 ×

JMS-1

1 − 0.8 − 0.6 − 0.4 − 0.2 − 0.2 0.4 0.6 0.8 1

Events normalised to bin entries

0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4

(300 GeV)

R

N

(3 TeV), m

R

W

m

1

= 13 TeV, 80 fb s ATLAS Simulation

Events mostly concentrated at the JMS expected value equal to unity.

SLIDE 10

10

Overlap Removal (OR) Strategy for e close to hadronic activity

In signal topology e2 always close to a real jet

=> Standard OR in ATLAS removes jet or e if within ∆R < 0.4 : Thus signal efficiency drops off ! => A modified OR approach followed for e2 : Within ∆R ~ 0.04 of e & jet, events dominated with a true e mis-reconstructed as a jet. Thus events with ∆R > 0.04 selected. => Further standard e efficiency correction factor cannot be used. Thus in VR additional criterion applied : a b-tagged jet & data-mc comparison done within 0.04 < ∆R <0.4.

arXiv:1904.12679

Events / 98 GeV

1 10

2

10

3

10

4

10

Data t t Single-t MC stat. unc.

e) µ VR ( < 0.4

,e

small

j y

R ∆ 0.04 <

1

= 13 TeV, 80 fb s ATLAS

[GeV]

e T

p

100 200 300 400 500 600 700 800

Data / Pred.

0.5 1 1.5 2

↓

Residual disagreement in addition to statistical, theory & b-tagging uncertainties quantified as an additional efficiency correction factor uncertainty.

SLIDE 11

11

Background Estimation : A fit extrapolation from CR to SR

500 1000 1500 2000 2500 3000 3500 4000 4500

3 −

10

2 −

10

1 −

10 1 10

2

10

3

10

4

10

5

10

Events/200 GeV Electron channel

1

= 13 TeV, 80 fb s ATLAS

R

W

, M

R

N

Signal: M

[GeV], [GeV]

300, 3000 400, 4000 Data Z+jets fit BG fit BG uncertainties

500 1000 1500 2000 2500 3000 3500 4000 4500

[GeV]

j,e1

m

2 − 2

Significance

500 1000 1500 2000 2500 3000 3500 4000 4500

4 −

10

3 −

10

2 −

10

1 −

10 1 10

2

10

3

10

4

10

5

10

Events/200 GeV Muon channel

1

= 13 TeV, 80 fb s ATLAS

R

W

, M

R

N

Signal: M

[GeV], [GeV]

300, 3000 400, 4000 Data Z+jets fit BG fit BG uncertainties 500 1000 1500 2000 2500 3000 3500 4000 4500

[GeV]

2 µ 1, µ j,

m

4 − 2 − 2

Significance

A data-driven CR fit (range 600-1800GeV) performed & extrapolated

to SR.

Different steeply falling functional forms tested in CR, best taken with

respect to mc closure (in CR & VR) & GOF.

As Zjets dominates in higher mass range, a Zjets mc fit (range

400-4000GeV) parameters used in resultant fit to data.

Fitted uncertainty includes extreme variations in SR yield using

different fit ranges in CR as well as modelling uncertainty in Zjets mc & statistical uncertainty of fit.

arXiv:1904.12679

SLIDE 12

12

Estimation of limit with a single bin Poissonian counting expt.

2000 3000 4000 5000 [GeV]

R

W

m 500 1000 1500 [GeV]

R

N

m ATLAS

1

= 13 TeV, 80 fb s

Muon channel

Obs. 95% CL
Exp. 95% CL
Obs. resolved 95% CL

Band σ

Exp. 1

Band σ

Exp. 2

Not covered Excluded

2000 3000 4000 5000 [GeV]

R

W

m 500 1000 1500 [GeV]

R

N

m ATLAS

1

= 13 TeV, 80 fb s

Electron channel

Obs. 95% CL
Exp. 95% CL
Obs. resolved 95% CL

Band σ

Exp. 1

Band σ

Exp. 2

Not covered Excluded

SR yields from signal, background

& systematic uncertainties as a set

f nuisance parameters used for

likelihood fit to data as a single bin.

Lower limits on masses of NR &

WR determined by profiled likelihood test statistic with CLs method.

Excluded region extends upto mWR

~ 4.8 TeV in e & 5 TeV in μ ( mNR ~ 0.4-0.5 TeV).

arXiv:1904.12679

Assumptions : gWR = gWL NeR, NμR, N𝜐R at same mass

SLIDE 13

13

Summary & Outlook

As LHC energy & luminosity increase extended phase space

becoming more suitable to explore massive resonances & thus more crucial to study boosted topologies => boosted heavy neutrinos looked into for the first time!

Till now as no new physics can be reached in LHC with standard

topologies & reconstructed standard objects we need to focus more on unusual topologies & objects which present a challenge to standard reconstruction techniques => a large radius jet with a hard electron inside an example (a common final state for many BSM physics to explore in boosted scenario) that also results into small background.

Further tuning for this search in order to gain more signal efficiency