Status and perspectives - Durham University

Loop-induced as sole production mechanism Loop-induced as additional production mechanism Tools Conclusion

Status and perspectiveson

loop-induced processes

Marek Schonherr

Universitat Zurich

ATLAS–CMS Monte Carlo Generators Workshop

Marek Schonherr 1/20


Disclaimer

I am an author of the SHERPA event generator, but I will try to keep thatbias out of the following summary.

This review cannot be exhaustive, but I will focus on most importantprocesses/process classes and recent developments/tools.

I will restrict myself to published results, with the implicit understandingthat modern automated one-loop programs can calculate these processeswithout a dedicated publication.



Introduction

• Common signature gg → EW bosonsexceptions e.g. same sign top production etc.

• Distinguish two cases

1) Loop-induced as sole production mode, e.g. gg → h→ few cases with EFT approximation

2) Tree-level production in qq dominant→ gauge invariant subset of NNLO calculation→ no guarantee that dominant part of NNLO correction

• Important production mechanism even if tree-level production exists→ scales with gg/qq luminosities

• Implementations available in many parton level gens (MCFM, etc.)→ will focus on particle level implementations

• when qq exists, then part of NNLO codes (for VV , VH etc.)

• Generally known only at LO, at most estimates of NLO corrections



Higgs production in gluon fusion

• for inclusive ggh coupling HEFT working well, correction toloop-induced flat and incorporated in every implementation

• for pp → h + n jets HEFT good enough to be reweighted→ O(1) corrections throughout phase space

⇒ can be incorporated in all the multijet merged at NLO machineryavailable for HEFT

• available in MADGRAPH, OPENLOOPS, GOSAM

Hirschi, Mattelaer JHEP10(2015)146



Higgs production in gluon fusionBuschmann, Goncalves, Kuttimalai, MS, Krauss, Plehn JHEP02(2015)038

pp → H + jets production (ggF)

• correction factor/weight

r(n)t =

|M(n)(mt)|2|M(n)(mt →∞)|2

• loops from OPENLOOPS

• construct MEPS@NLO fromreweighted S-MC@NLO

• factorised approach forunknown top massdependence in Vn, otherwiseexact NLO mass dependence

0 100 200 300 400 500 600

-410

-310

-210

-110

1 tmH+0j NLO

tmH+1j NLO

tmH+2j LO

tmH+jets NLO

∞ →tmH+jets NLO

GeVfb

T,Hdp

σd

[GeV]T,H

p0 100 200 300 400 500 600

0.2

0.4

0.6

0.8

1

1.2

∞→tm(N)LO

tm(N)LO

SHERPA+OPENLOOPS

dσn = dΦn r(n)t

[Bn + Vn +

∫dΦ1 Dn

]PSn + dΦn+1

[r

(n+1)t Rn − r

(n)t Dn

]Marek Schonherr 5/20


Interference in Higgs production

Buschmann, Goncalves, Kuttimalai, MS, Kraus, Plehn JHEP02(2015)038

100 200 300 400 500 600 700 800 900 1000

-410

-310

-210

-110qq

gg

=(0,0)t,gκ

=(1,0)t,gκ

=(0,1)t,gκ

GeVfb

4ldm

σd ZZ→pp

SMHΓ=HΓ

[GeV]4lm100 200 300 400 500 600 700 800 900 10000

1

2

3

=(0,1)t,gκLO=(1,0)t,gκLO

100 200 300 400 500 600 700 800 900 1000

-410

-310

-210

-110qq

gg

=(0,0)t,gκ

=(1,0)t,gκ

=(0,1)t,gκ

GeVfb

4ldm

σd ZZ→pp

SMHΓ=25HΓ

[GeV]4lm100 200 300 400 500 600 700 800 900 10000

1

2

3

=(0,1)t,gκLO=(1,0)t,gκLO

investigate new physics in pp → 4`, qq @ NLO, gg @ LOparametrised in κ framework, SM: (κt , κg ) = (1, 0)




Coradeschi, de Florian, Dixon, Fidanza, Hoche, Ita, Li, Mazzitelli PRD92(2015)013004

-20

-10

0

10

∆m

H@MeVD

M jj>400GeV

VBF

GF

Sum

0 1 2 3 4 5 6 7 80.001

0.01

0.1

1

10

ÈDΗjj min

Signal@fbD

mass shift in mγγ in VBF/GF–background interferencequickly increases with ΓH




Coradeschi, de Florian, Dixon, Fidanza, Hoche, Ita, Li, Mazzitelli PRD92(2015)013004

0 5 10 15 20-350

-300

-250

-200

-150

-100

-50

0

G� GSM

Dm

HΓΓ@MeVD

mass shift in mγγ in VBF/GF–background interferencequickly increases with ΓH



Double/Triple Higgs production in gluon fusion

Maltoni, Vryonidou, Zaro JHEP11(2014)079

• reweight in dΦB with rBt ,in dΦR with rRt

• factorised approach forvirtual corrections

100755033251413

103

102

101

pp→HH

LOFT

pp→HH

NLOFTap

prox

Mad

Gra

ph5

aMC

@N

LO

MH=125 GeV, MSTW2008 (N)LO pdf (68%cl)

HH production at pp colliders

. .

-

√s[TeV]

σ(N

)LO[fb]

100755033251413

100

10−1

pp→HHH

LOFT

pp→HHH

NLOFT ap

prox

Mad

Gra

ph5

aMC

@N

LO

MH=125 GeV, MSTW2008 (N)LO pdf (68%cl)

HHH production at pp colliders

. .

-

√s[TeV]

σ(N

)LO[fb]



Double Higgs production in gluon fusionMaierhofer, Papaefstathiou JHEP03(2014)126

0j inc.ETclus = 50GeV

ETclus = 60GeV

ETclus = 70GeV

10−5

10−4

10−3

10−2

Separation between Higgs bosons

dσ/d

∆R(h,h)[pb]

1 2 3 4 5 6

0.5

1

1.5

2

∆R(h, h)

Ratio

0j inc.ETclus = 50GeV

ETclus = 60GeV

ETclus = 70GeV

10−8

10−7

10−6

10−5

10−4

hh p⊥

dσ/dphh ⊥[pb/GeV

]100 200 300 400 500 600

0.6

0.8

1

1.2

1.4

phh⊥ [GeV]Ratio

• LO merging for pp → hh + 0, 1 jets in MLM scheme

• use exact loop-induced, no reweighting

• using HERWIG+++OPENLOOPS



Associated Higgs production

• only gg → HZ , gg → HZg , gq → HZq and gg → Hγg

• relative size of contribution strongly dependent on observable




Associated Higgs productionHespel, Maltoni, Vryonidou JHEP06(2015)065

g

gg

H

Z

g

g

g

H

Z

g

g

g

H

Z

ggg

g

H

g

Zq

H

Z

q q

H

Z

q

q

q

H

Z

gZ

H

gq

q

MLM merge loop-inducedZH productiongg → ZHgg → ZHg and gq → ZHq

powerwimpy

scale uncertaintiesZH + ZHj, merged

MadGraph5aMC@NLO

PHT [GeV]

5004003002001000

partonpowerwimpy

scale uncertaintiesZH

LHC14TeVMSTW2008LO

-

.

PHT [GeV]

dσ/d

PH T

[fb/b

in]

5004003002001000

1

0.1

0.01

0.001




Goncalves, Krauss, Kuttimalai, Maierhofer PRD92(2015)7,073006

pp → ``H + jets production

• MEPS@NLO for qqMEPS@LOOP2 for gg

• care for qg → ZHq:

→ part of NLO ZHj→ in loop-induced as gauge→ inv. subset of NNLO ZHj

10−3

10−2

10−1

dσ

/d(|E

mis

sT−

pll T|/

pll T)

[pb]

Sherpa+OpenLoops

MEPS@NLO [scale variations]MEPS@NLO 0-jet exclusive

MEPS@Loop2 [scale variations]

MEPS@Loop2 [k-factor variations]

MEPS@Loop2 0-jet exclusive

0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40|Emiss

T − pllT|/pll

T

0.0

0.2

0.4

0.6

0.8

Rat

ioto

inc.

sam

ple

pp → Z [→`` ]H[→ inv] + jets




Goncalves, Krauss, Kuttimalai, Maierhofer PRD92(2015)7,073006

pp → ``H + jets production


• care for qg → ZHq:

→ part of NLO ZHj→ in loop-induced as gauge→ inv. subset of NNLO ZHj

[GeV]BDRSHm

80 100 120 140 160

GeVfb

BDRS

Hdm

σd

-410

-310

-210

MEPS@NLO HZ+0,1j HZ+0,1j2MEPS@Loop

Sherpa+OpenLoops

[GeV]BDRSHm

80 100 120 140 1600.20.30.40.5 /MEPS@NLO2MEPS@Loop

F/CA0.18 C

[GeV]BDRSHm

80 100 120 140 1600.20.40.60.8 jet-veto/incl

pp → Z [→``]H[→bb] + jets



Vector boson pair production

• relative size of contribution strongly dependent on observable,but typically < 10%




Vector boson pair productionCascioli, Hoche, Krauss, Maierhofer, Pozzorini, Siegert arXiv:1309.5912

pp → 4`+ 0, 1 jetsand pp → 2`2ν + 0, 1 jets


• includes loop-inducedgg → 4`,gg → 4`+ g , gq → 4`+ q,qq → 4`+ g

Sherpa+OpenLoops

MEPS@NLO 4ℓ+ 0, 1jµF,R/2 · · · 2µF,RµQ/

√2 · · ·

√2µQ

Unc. quad. sum

MC@NLO 4ℓNLO 4ℓ

0

10

20

30

40

50

60

70

80

90

Azimuthal lepton distance (ATLAS, Njet = 0)

dσ/d

∆φℓℓ[fb]

0.80.91.01.11.2

dσ/d

σMEPS@NLO

0 0.5 1 1.5 2 2.5 3

0.02

0.04

0.06

0.08 MEPS@LOOP2 4ℓ+ 0, 1j

∆φℓℓ

dσ/d

σMEPS@NLO



Vector boson pair productionCascioli, Hoche, Krauss, Maierhofer, Pozzorini, Siegert arXiv:1309.5912

pp → 4`+ 0, 1 jetsand pp → 2`2ν + 0, 1 jets


• includes loop-inducedgg → 4`,gg → 4`+ g , gq → 4`+ q,qq → 4`+ g

Sherpa+OpenLoops

MEPS@NLO 4ℓ+ 0, 1jµF,R/2 · · · 2µF,R

µQ/√2 · · ·

√2µQ

Unc. quad. sum

MC@NLO 4ℓNLO 4ℓ+ 1j

0

0.1

0.2

0.3

0.4

0.5

Lepton pair mass (CMS, Njet = 1)

dσ/dm

ℓℓ[fb/GeV]

0.80.91.01.11.2

dσ/d

σMEPS@NLO

50 100 150 200 250 300

0.02

0.04

0.06

0.08 MEPS@LOOP2 4ℓ+ 0, 1j

mℓℓ [GeV]

dσ/d

σMEPS@NLO



Multiboson production

• many process, generally very small cross section




Tools

Process specific tools:

• gg2VV+HERWIG/PYTHIA Kauer JHEP12(2013)082

- gg → 2`2ν, 4`

...

General tools:

• MADGRAPH+HERWIG/PYTHIA Hirschi, Mattelaer JHEP10(2015)146

- all on-shell 2→ 2 and 2→ 3 processes, some 2→ 4 in SM- MLM merging possible with PYTHIA6

• SHERPA+OPENLOOPS several publications

- all off-shell 2→ 2 and 2→ 3, e.g. gg → 4`g- MEPS@LOOP

2 merging possible



Timings – MADGRAPH

Timing for 10k unweighted events (number of phase space points needed)




Timings – OPENLOOPS

Timing per phase space point using COLLIER

Process ms/point Process ms/point Process ms/point

gg → h 0.11 gg → hh 1.3gg → hg 1.6 gg → hhg 47 gg → γg 2.7gg → hgg 50 gg → γgg 117gg → he+e− 2.0 gg → e+e− 0.28 gg → e+e−γ 4.3gg → he+e−g 61 gg → e+e−g 4.9 gg → e+e−γg 128

Process ms/point Process ms/point

gg → e+e−µ+µ− 8.9 gg → e+νeµ−νµ 2.0gg → e+e−µ+µ−g 238 gg → e+νeµ−νµg 72gg → e+e−e+e− 25 gg → e+νee−νe 6.2gg → e+e−e+e−g 638 gg → e+νee−νeg 202



Conclusion

• all processes in SM available by several toolsprocess specific, MADGRAPH+HERWIG/PYTHIA,SHERPA+OPENLOOPS, SHERPA+GOSAM

• improve EFT used otherwiseinclude important subleading process where qq channel exists,gg has different propertiescalculate signatures inaccessible otherwise

• NLO corrections are two-loop, usually with many mass scales→ generally not available

• NLO K -factors are estimated to be large ≈ 2

• selected BSM processes appeared in the literature

• timings/efficiencies vary between tools



Thank you for your attention!


Documents

Status and perspectives - Durham University