Off-shell effects in t + /Z production at the LHC · arXiv:1907.09359 [hep-ph] arXiv:1912.09999 [hep-ph] Introduction In the absence of convincing evidence for new resonances effects,precise

Off-shell effects in tt + γ/Zproduction at the LHC

Giuseppe Bevilacqua

MTA-DE Particle Physics Research Group, Debrecen

LHCP2020

May 25, 2020

In collaboration with H. B. Hartanto, M. Kraus, T. Weber and M. Worek

Based on: arXiv:1803.09916 [hep-ph] arXiv:1809.08562 [hep-ph]

arXiv:1907.09359 [hep-ph] arXiv:1912.09999 [hep-ph]

https://arxiv.org/abs/1803.09916




Introduction

In the absence of convincing evidence for new resonances effects, precisemeasurements of SM observables are key to look for effects of New Physics(NP) at the LHC

This is especially true for the top quark sector, where NP effects are expectedto be more prominent due to the large mass scale

https://twiki.cern.ch/twiki/bin/view/CMSPublic/PhysicsResultsCombined

tt, ttj

ttγ

ttZ

As luminosity increases, LHCallows to probe associated ttchannels with more precision

We’ll focus on tt+ γ andtt+ Z(Z → νν) productionat Run II of the LHC

G. Bevilacqua LHCP2020 2/22

https://twiki.cern.ch/twiki/bin/view/CMSPublic/PhysicsResultsCombined

Motivationsttγ

- Direct probe of top quark electric charge Melnikov et al. ’11 ...

↪→ σttγ ∼ Q2t , while indirect through other processes

- Probe of the effective ttγ interaction Baur et al. ’05 Aguilar Saavedra ’09

↪→ - dimension-six SM EFT Schulze et al. ’16 Maltoni et al. ’16 ...

- top-quark anomalous couplings ...

- Irreducible background to direct BSM searches

ttZ

- Direct probe of ttZ coupling Rontsch and Schulze ’15 ...

↪→ search for deviations from SM expectations

- Dark Matter searches at LHC (tt + EmissT ) Arina et al. ’16 Haisch et al. ’16

↪→ in several models, DM couples preferentially to top quarks

- Background to searches for rare processes↪→ e.g. search of same-sign leptons + jets + EmissT


Predictions for ttγ/Z at LHC: state of the artttγ

- NLO QCD & EW: on-shell, inclusive Duan et al. ’09,’11,’16 Maltoni et al. ’16

↪→ top quarks undecayed: general idea about size of NLO corrections

- NLO QCD: on-shell, with radiative decays Melnikov, Scharf and Schulze ’11

↪→ QCD corrections to production and decays, with spin correlations and γ radiation from decays

- NLO QCD: on-shell, with PS effects Kardos and Trocsanyi ’14

↪→ top decays in PS, no spin correlations, only direct-photon contribution

- NLO QCD: complete off-shell G.B, Hartanto, Kraus, Weber and Worek ’18

↪→ resonant and non-resonant diagrams, interferences and finite-width effects (dilepton channel)

ttZ

- NLO QCD: on-shell, inclusive Lazopoulos et al. ’08

- NLO QCD: on-shell, with NLO decays Rontsch and Schulze ’14

- NLO QCD & EW: on-shell, with PS effects Kardos et al. ’12 Frixione et al. ’15

- NLO QCD & EW + NNLL Kulesza et al. ’18, ’20 Broggio et al. ’17, ’19

- NLO QCD: complete off-shell G.B, Hartanto, Kraus, Weber and Worek ’19


















Narrow Width Approximation

Based on the narrow-width limit:

• cross section factorizes into on-shell tt (+X) production ⊗ decays↪→ only double-resonant diagrams are retained

• non-factorizable contributions are suppressed by powers of Γt/mt = O(1%)for sufficiently inclusive observables

• huge simplification, but several contributions have to be put together to accountfor all resonant structures...

e.g. ttγ


Beyond NWARepresentative Feynman diagrams for the case of ttZ(Z → νν):

double-resonant (DR) single-resonant (SR) non-resonant (NR)

• all resonant and non-resonant contributions are taken into account

• most complete calculation of ME’s at fixed perturbative order

• computationally demanding: O(10) increase in complexity w.r.t. NWA

”Off-shell effects” = finite-width effects for tops and W ’s

+ SR + NR

+ interferences among DR, SR, NR


The HELAC-NLO framework

→ C++/ROOT analysis framework for event Ntuples- reweighting for multiple scale/PDFs

- kinematical selection of events

- histogramming tools


Predictions for ttγ at√s = 13 TeV


Predictions for off-shell ttγ

10−2

10−1

dσ/dpT,γ

[fb/GeV

] LO

NLO

0 50 100 150 2000.60.81

1.21.4

pT,γ [GeV]

NLO

/LO

pp→ tt(γ)→ bbe+νeµ−νµγ @ 13 TeV-dilepton channel-

µ = mt/2

HELAC-NLO

G.B., Hartanto, Kraus, Weber and Worek, arXiv:1803.09916 [hep-ph]

10−2

10−1

dσ/dpT,γ

[fb/GeV

] LO

NLO

0 50 100 150 2000.60.81

1.21.4

pT,γ [GeV]

NLO

/LO

µ = HT /4

HELAC-NLO

Dynamical scale µR = µF = HT /4 improves perturbative stabilityHT =

∑i={b,b,`±,γ} pTi + 6pT

Comparative analysis of different approaches of ttγ modeling in progressATLAS-CONF-2019-042



https://cds.cern.ch/record/2690350

ttγ: off-shell vs on-shell

10−5

10−4

10−3

10−2

10−1

100

dσ/p

T(b

avg)

[fb/G

eV]

off-shell NWA (HT /4)

NWA LOdec (HT /4) NWA (mt/2)

NWA LOdec (mt/2)

0.6

0.8

1

Rat

ioto

off-s

hel

l

0 100 200 300 400 500 600

0.6

0.8

1

pT (bavg)

Rati

oto

off-s

hel

lImpact of off-shell effects

and radiative decays


G.B., Kraus, Hartanto, Weber and Worek, arXiv:1912.09999 [hep-ph]

HELAC-NLO

10−4

10−3

10−2

10−1

100

dσ/M

(bl+

) min

[fb/G

eV]

off-shell NWA (HT /4)

NWA LOdec (HT /4) NWA (mt/2)

NWA LOdec (mt/2)

0.4

0.6

0.8

1

1.2

Rat

ioto

off

-shel

l

0 20 40 60 80 100 120 140 160 180 200

0.5

1

M(bl+)min

Rat

ioto

off-s

hel

l

HELAC-NLO

σNLO, NWA = 7.33 fb ↔ σNLO, NWA (LO decays) = 4.63 fb



Predictions for off-shell ttγ

10−6

10−5

10−4

10−3

10−2

10−1

100

101

dσ/pT(b

avg)[fb/G

eV]

full NWADR SRNR

0 100 200 300 400 500 6000

0.25

0.50

0.75

1

pT (bavg)

Ratio

tofull

Analysis of Double-, Single- andNon-Resonant phase space regionsKauer and Zappenfeld ’01 ...


G.B., Kraus, Hartanto, Weber and Worek, arXiv:1912.09999 [hep-ph]

HELAC-NLO

10−7

10−5

10−3

10−1

101

dσ/M

(bl+) m

in[fb/G

eV]

full NWADR SRNR

0 20 40 60 80 100 120 140 160 180 2000

0.25

0.50

0.75

1

M(bl+)min

Ratio

tofull

HELAC-NLO

DR → |M(t)−mt| < nΓt ∧ |M(t)−mt| < nΓt (n = 15)

NR → |M(t)−mt| > nΓt ∧ |M(t)−mt| > nΓt

SR → full− (DR + NR)



Predictions for ttZ(Z → νν) at√s = 13 TeV


tt+ EmissT

While leading to the same visible final state, off-shell tt and ttZ(Z → νν)have pretty different kinematics

0 100 200 300 400 500

10−5

10−4

10−3

10−2

10−1

pT,miss [GeV]

1/σ·dσ/dpT,m

iss

[1/G

eV]

tt ttZHELAC-NLO

200 400 600 800 1000 1200 1400

10−6

10−5

10−4

10−3

10−2

HT [GeV]

1/σ·dσ/dH

T[1

/GeV

]

tt ttZHELAC-NLO

G.B, Hartanto, Kraus, Weber and Worek, arXiv:1907.09359 [hep-ph]

[distributions normalized to one]

ttZ populates high pmissT regions which are important for BSM studies.However tt dominates over ttZ(Z → νν) by several orders of magnitude (atinclusive level).

↪→ Is it important to strive for higher accuracy in ttZ? Yes!



Determining the CP nature of spin-0 mediators in tt + DM productionHaisch, Pani and Polesello, arXiv:1611.09841 [hep-ph]

(tt, tW , ttW ) (WW,WZ,ZZ...)

(ttZ)−→

- Distribution of events in the(EmissT ,mT2) plane for thedifferent backgrounds andfor one example of DM signal[Mφ = 100 GeV , Mχ = 1 GeV ]

- The area in the upper rightcorner above the black lineis the region selected in theanalysis

- With 300 fb−1, assuming 20% systematics for SM backgrounds, it should be possibleto resolve the (pseudo-)scalar nature of the mediator up to mass ≈ 200 GeV

- Discovery reach depends on syst. uncertainty of SM backgrounds, dominated by ttZ



Off-shell ttZ(Z → νν): differential cross sections

10−1.5

10−1

dσ/d

cosθ l

l[fb]

mt +mZ/2 HT /3 ET /3

E′T /3 E′′T /3

1

1.1

1.2

1.3

NLO

/LO

0 0.2 0.4 0.6 0.8 1

1

1.5

cos θll

scaledep

enden

ce LO mt +mZ/2 HT /3

ET /3 E′T /3 E′′T /3

HELAC-NLO


cos θll = tanh(∆yll/2)

- Sensitive to the nature of DM mediatorHaisch et al. ’16

- Differential K-factors far from constant!

- µ = mt +mZ/2 : +4%↔ 27%

- µ = HT /3 : −5%↔ 10%

- µ = E′′T /3 : −1%↔ 14%



Summary

• Achieving the highest precision for tt+ V requires to include non-resonant,finite-width and interference effects.

• Systematic comparisons between predictions at different levels of accuracy(on-shell vs off-shell) are important to assess the reliability of variousapproximations

• We have computed predictions for a variety of off-shell tt+X processes(dilepton channel) with the package HELAC-NLO:

- pp→ tt+ γ - pp→ tt+ Z

- pp→ tt+ j - pp→ tt+W± NEW! arXiv:2005.09427 [hep-ph]

• Our predictions are available in the form of ROOT Ntuples. We can provide themalong with a user-friendly software to extract physical results for your analysis.If interested, contact us!

A Zoom meeting room is available for further discussions.Please find the details on next slide →



Details of Zoom Meeting room (May 25, 16:30-17:00 CERN time)

Topic: Giuseppe Bevilacqua’s Zoom Meeting

Date, Time: May 25, 2020 04:30 PM Paris

Click here to join

(same password of this parallel session)

In case of problems, just email me( giuseppe.bevilacqua AT science.unideb.hu )


https://zoom.us/j/95869099250?pwd=TThkaEc4TXVEemhmdkd4ODZKWG12UT09

Backup slides


Off-shell ttγ: fixed vs dynamical scalesA collection of distributions based on the scale choice µR = µF = mt/2

0

2

4

6

dσ/d∆Rb1,γ

[fb

]

LO

NLO

0 1 2 3 4 5

1

2

3

∆Rb1,γ

NL

O/

LO


0

2

4

6

dσ/d∆Rl 1,γ

[fb

]

LO

NLO

0 1 2 3 4 5

1

2

3

∆Rl1,γ

NL

O/

LO

0

2

4

6

dσ/d∆Rl 2,γ

[fb

]

LO

NLO

0 1 2 3 4 5

123

∆Rl2,γ

NL

O/

LO

10−3

10−2

10−1

dσ/dpT,b

1[fb/GeV

] LO

NLO

0 100 200 300 400

0.5

1

1.5

pT,b1 [GeV]

NLO

/LO

10−3

10−2

10−1

dσ/dpT,b

2[fb/GeV

] LO

NLO

0 50 100 150 200

0.5

1

pT,b2 [GeV]

NLO

/LO

10−3

10−2

10−1

dσ/dpT,l1[fb/GeV

] LO

NLO

0 100 200 300

0.5

1

pT,l1 [GeV]

NLO

/LO



Off-shell ttγ: fixed vs dynamical scalesA collection of distributions based on the scale choice µR = µF = HT /4

0

2

4

6

dσ/d∆Rb1,γ

[fb

]

LO

NLO

0 1 2 3 4 5

1

2

3

∆Rb1,γ

NL

O/

LO


0

2

4

6

dσ/d∆Rl 1,γ

[fb

]

LO

NLO

0 1 2 3 4 5

1

2

3

∆Rl1,γ

NL

O/

LO

0

2

4

6

dσ/d∆Rl 2,γ

[fb

]

LO

NLO

0 1 2 3 4 5

123

∆Rl2,γ

NL

O/

LO

10−3

10−2

10−1

dσ/dpT,b

1[fb/GeV

] LO

NLO

0 100 200 300 400

0.5

1

1.5

pT,b1 [GeV]

NLO

/LO

10−3

10−2

10−1

dσ/dpT,b

2[fb/GeV

] LO

NLO

0 50 100 150 200

0.5

1

pT,b2 [GeV]

NLO

/LO

10−3

10−2

10−1

dσ/dpT,l1[fb/GeV

] LO

NLO

0 100 200 300

0.5

1

pT,l1 [GeV]

NLO

/LO



Off-shell ttγ: comparisons to Run II data”Theory NLO” = off-shell ttγ

(HELAC-NLO)”Madgraph5” = double-resonant ttγ (LO)

ATLAS-CONF-2019-042

σexp = 44.2± 0.9 (stat)+2.6−2.4 (syst) fb = 44.2± 2.6 fb

σTheory NLO = 39.50+0.56−2.18 (scale)+1.04

−1.18 (PDF) fb


https://cds.cern.ch/record/2690350

Off-shell ttZ(Z → νν): differential cross sections

Checks on dimensionful observables


10−4

10−3

dσ/dm

ll[fb/G

eV]


E′T /3 E′′T /3

1

1.2

NLO

/LO

0 100 200 300 400 500

1

1.5

mll [GeV]

scaledep

endence LO mt +mZ/2 HT /3

ET /3 E′T /3 E′′T /3

HELAC-NLO

dilepton inv. mass

10−5

10−4

10−3

10−2

dσ/dpT,l[fb/G

eV]


E′T /3 E′′T /3

0.8

1

1.2

NLO

/LO

0 50 100 150 200 250 300 350 4000.5

1

1.5

2

pT,l [GeV]

scaledep


ET /3 E′T /3 E′′T /3

HELAC-NLO

charged lepton’s pT

10−6

10−5

10−4

10−3

dσ/dH

T[fb/G

eV]


E′T /3 E′′T /3

0.8

1

1.2

NLO

/LO

200 400 600 800 1000 1200 1400

0.5

1

1.5

2

HT [GeV]

scaledep


ET /3 E′T /3 E′′T /3

HELAC-NLO

transverse energy (HT )

- µ = mt + mZ/2 → NLO gets outside LO uncertainties

- µ = HT/3, ET/3, ... → improves perturbative convergence




Documents

Off-shell effects in t + /Z production at the LHC · arXiv:1907.09359 [hep-ph] arXiv:1912.09999 [hep-ph] Introduction In the absence of convincing evidence for new resonances effects,precise