ttH¯ Theory Overview

Stefano Pozzorini

Zurich University in collaboration with Stefan Guindon, Laura Reina, Chris Neu

8th Workshop of LHC Higgs Section Working Group CERN, Geneva, 22 January 2015 Outline

1 ttH¯ Signal

2 Backgrounds for ¯b →

3 Backgrounds for H γγ →

4 Backgrounds for H W, ZZ, ττ multi-leptons → →

5 Scale uncertainties in NLO matching & merging Theory Challenges in ttH¯ searches

Large multi-particle backgrounds (tt¯+ jets, ttV¯ + jets, ttγγ¯ , VV + jets) Key priority is precision for backgrounds NLO automation powerful but 2 → 4 CPU intensive

NLO matching & merging crucial various new methods (FxFx, MEPS@NLO, MINLO, UNLOPS,. . . ) various automated tools( MG5 aMC@NLO, Sherpa, Minlo–PowHeg,. . . )

Theory uncertainty estimates nontrivial still limited experience in NLO matching+merging framework sophisticated analyses (data-driven extrapolations, reweighting, . . . )

S. Pozzorini (Zurich University) ttH¯ Theory 8th HXSWG Workshop 1 / 17 Strategy of ttH¯ subgroup

New priorities dominant sources of theory systematics ↔ emphasis on backgrounds and NLO Monte Carlo methods close intereaction with ATLAS/CMS to identify theory priorities

Goals and guiding principles recommend state-of-the-art methodology and encourage use of all relevant tools support/coordinate NLO simulations in ATLAS/CMS understanding and assessment of theory uncertainties (choices and variations of renormalisation+factorisation+resummation+merging scales in multi-scale processes)

S. Pozzorini (Zurich University) ttH¯ Theory 8th HXSWG Workshop 2 / 17 Outline

1 ttH¯ Signal

2 Backgrounds for H b¯b →

3 Backgrounds for H γγ →

4 Backgrounds for H W W, ZZ, ττ multi-leptons → →

5 Scale uncertainties in NLO matching & merging NLO tools for ttH¯ I

¯ NLO ttH cross section 1400 _ σ(pp → ttH + ) [fb] [Beenakker et al ’01; Reina, Dawson ’01] 1200 √s = 14 TeV

MH = 120 GeV 1000 µ0 = mt + MH/2 moderate uncertainty (9% scale, 8% PDF+αS ) LO 800 NLO

sufficient for mid-term future 600

400

200

0 0.2 0.5 1 2 5

µ/µ0 Publicly available NLO+ tools 0.001 0.0009 POWHEG+PYTHIA 6 POWHEG+PYTHIA 8 MG5 aMC@NLO (2011) 0.0008 POWHEG+HERWIG NLO 0.0007 Powhel samples (2011) 0.0006 0.0005 (H) [pb/GeV] T 0.0004

[Jaeger et al, 1501.04498] /dp Powheg Box 0.0003 0.0002 Sherpa+OpenLoops or GoSam 0.0001 0 0 50 100 150 200 250 300 350 ∼ 10% uncertainty for inclusive observables, 1.2

more if sensitive to jet radiation 1

0.8 0 50 100 150 200 250 300 350

pT(H) [GeV]

S. Pozzorini (Zurich University) ttH¯ Theory 8th HXSWG Workshop 3 / 17 NLO tools for ttH¯ II

Spin-correlated top decays tth¯ (γγ) : corr 0.65 Frame 1 tth¯ (γγ) : uncor in Powheg, MG5 aMC@NLO, Sherpa ttγγ¯ : corr ttγγ¯ : uncor 0.6 (NLO production)×(LO decays) ℓℓ

θ 0.55 dN cos

Breit–Wigner smearing d 1 0.5 spin correlations via ttH¯ (+) tree amplitudes 0.45 ⇒ mandatory for signal and all backgrounds! [Biswas et al ’14] 0.4 -1 -0.6 -0.2 0.2 0.6 1

cos θℓℓ

tt-H production at the 13 TeV LHC Towards NLO EW corrections 10-1

LO weak corrections to ttH¯ in MG5 aMC@NLO LO+NLO QCD LO+NLO QCD+Weak 10-2 −10% at high-pT per bin [pb] σ 2 → 2, 3, 4 NLO EW automation in OpenLoops 10-3 [1412.5157] and Recola [1411.0916] [Frixione et al ’14] MadGraph5_aMC@NLO 10-4 ⇒ relevant for signal and backgrounds at high ratio over LO pT 1.4 1.2 1 relative contributions 0.4

0.2 QCD Weak HBR 0 -0.2 0 200 400 600

pT(H) [GeV] S. Pozzorini (Zurich University) ttH¯ Theory 8th HXSWG Workshop 4 / 17 NLO tools for ttH¯ III

Towards NLO merging for ttH¯ + 0, 1 jets parton-level ttHj¯ Sherpa+GoSam [Deurzen et al, ’13] Sherpa and MG5 aMC@NLO support NLO merging ⇒ more reliable prediction and uncertainty for extra jet activity (impact on ttH¯ analysis?)

Off-shell+EW effects in pp `ν + 2j + 4b at LO [Denner at al, 1412.5290] →

tiny ttH¯ signal–background interference νℓ νℓ W+ W+ ℓ+ ℓ+ g ¯ t t ttb¯ b QCD bckg receives +16% EW cont b t b t and −8% QCD–EW interference! b b¯ + g W b¯ b t d d extra +12% from off-shell top decays t W− W− ¯u ¯u g g ¯t ¯t ⇒ calls for detailed off-shell studies at NLO b¯ b¯

S. Pozzorini (Zurich University) ttH¯ Theory 8th HXSWG Workshop 5 / 17 Outline

1 ttH¯ Signal

2 Backgrounds for H b¯b →

3 Backgrounds for H γγ →

4 Backgrounds for H W W, ZZ, ττ multi-leptons → →

5 Scale uncertainties in NLO matching & merging Irreducible ttb¯ ¯b background

NLO ttb¯ ¯b [Bredenstein et al ’09/’10; Bevilacqua et al ’09]; ttb¯ ¯b dominates ttH¯ (b¯b) systematics NLO reduces uncertainty from 80% to 20–30%

NLO+PS ttb¯ ¯b 5F scheme (mb = 0) with Powhel [Garzelli et al ’13/’14] ttb¯ ¯b MEs cannot describe collinear g → b¯b splittings ⇒ inclusive tt¯+b-jets simulation requires NLO merging tt¯+ 0, 1, 2 jets

NLO+PS ttb¯ ¯b 4F scheme (mb > 0) with Sherpa+OpenLoops [Cascioli et al ’13] ttb¯ ¯b MEs cover full b-quark phase space ⇒ inclusive tt¯+b-jets simulation possible

S. Pozzorini (Zurich University) ttH¯ Theory 8th HXSWG Workshop 6 / 17 S–MC@NLO ttb¯ ¯b 4F scheme [Cascioli et al ’13] Good perturbative stability but unexpected MC@NLO enhancement ttb ttbb ttbb (mbb > 100) +71%+14% +66%+15% +63%+17% σLO[fb] 2644−38%−11% 463.3−36%−12% 123.4−35%−13% +34%+5.6% +29%+5.4% +26%+6.5% σNLO[fb] 3296−25%−4.2% 560−24%−4.8% 141.8−22%−4.6% σNLO/σLO 1.251 .211 .15 +32%+3.9% +24%+2.0% +20%+8.1% σMC@NLO[fb] 3313−25%−2.9% 600−22%−2.1% 181−20%−6.0% σMC@NLO/σNLO 1.011 .071 .28

Large enhancement ( 30%) in Higgs region from double g b¯b splittings ∼ →

⇒ understand PS and matching systematics!!

S. Pozzorini (Zurich University) ttH¯ Theory 8th HXSWG Workshop 7 / 17 tt¯+ jets background and merging at NLO

NLO tt¯+ 2 jets [Bevilacqua, Czakon, Papadopoulos, Worek ’10/’11] reduces uncertainty from 80% to 15%

NLO merging (FxFx, MEPS@NLO, UNLOPS,. . . ) NLO and log accuracy for 0, 1, . . . n jets 0-jet NLO+PS tt¯ 1-jet NLO+PS tt¯+ 1 j separated via kT-algo at merging scale Qcut ...... smooth PS–MEs transition ↔ MEs with PS-like ≥ n jets NLO+PS tt¯+ n j scale and Sudakov FFs

NLO merging for tt¯+ 0, 1 jets FxFx with Madgraph5/aMC@NLO [Frederix, Frixione ’12] MEPS@NLO with Sherpa+GoSam [H¨ocheet al ’13]

S. Pozzorini (Zurich University) ttH¯ Theory 8th HXSWG Workshop 8 / 17 MEPS@NLO for tt¯+ 0, 1, 2 jets (Sherpa+OpenLoops)

[H¨oche,Krauss, Maierh¨ofer,S. P. , Sch¨onherr,Siegert ’14]

■ ☛❝☞✉✌✐✈✍ ☞✐❣✎✏ ✑✍✏ ♠✉☞✏✐✒ ☞✐❝✐✏② ✶

❪ Consistency with LO merging and NLO+PS

❜

✛

❬

✮

✚

✙

✘

✗

✖ ✞ ✡

✶ ☎ decent (10–20%) mutual agreement

✔

✕

✭

✓

❧ ✲❥✁

❃ ✹ ✵ ●✂❱

❄

✞ ✷ ✶ ☎

Reduction of µR, µF , µQ variations

✸

✞

✶ ☎

❧ ✲❥✁

❃ ✵ ●✂❱

✻ Nlight−jet ≥ 0 1 2

❄

✄ ☎ ✳ ✶ ✠

✞ LO 48% 65% 80%

✶ ☎

❙ ▲❖ ❅◆

▼ ❊ NLO 17% 20% 20–30%

❧ ✲❥✁

✆ ✺ ❙ ▲❖

✶✳ ✄ ▼ ❊ ❅

❃ ✵ ●✂❱

✽

❄

✞ ✟

❙ ▲❖

✝▼ ❈ ❅◆

✄ ☎✳ ☎ ✶ ✶ ☎

❙ ❤❡♣❛✰❖♣ ❡ ♥▲♦♦♣ Merging scale choice and dependence

❖

❧✲❥ ❡

✷

♣ ❃ ✹✵ ●❱

▲

♦ ❄

✁ ± GeV and  dependence

◆ Qcut = 30 10 10%

♦

❅

✳✺

✶ ✐

✁

❙

❛

❘

❊ ✶

▼ ⇒ NLO precision for tt¯+ 0, 1, 2 jets

❖

❧✲❥ ❡

✷

♣ ❃ ✻✵ ●❱

▲

♦

❄

✁

◆

♦

❅

✳✺

✶ ✐

✁

❙

❛ Next steps

❘

❊ ✶

▼ improve CPU performance (ongoing)

❖

❧✲❥ ❡

✷

♣ ❃ ✽✵ ●❱

▲

♦

❄ ✞

✟ consolidate understanding of theory uncertainty

♦

❅

✳✺

✶ ✐

✞

❙

❛

❘

❊ ✶

▼ combination with ttb¯ ¯b

✶ ✷ ✸

✁ ¯ ◆

S. Pozzorini (Zurich University) ✂✄☎ ✆✝ ttH Theory 8th HXSWG Workshop 9 / 17 Outline

1 ttH¯ Signal

2 Backgrounds for H b¯b →

3 Backgrounds for H γγ →

4 Backgrounds for H W W, ZZ, ττ multi-leptons → →

5 Scale uncertainties in NLO matching & merging Relevant backgrounds and existing predictions

Relevant backgrounds tt¯+ 1, 2γ, single-top+1, 2γ, jets+1, 2γ H+jets with H → γγ and HF jets

NLO+PS for ttγ¯ and ttγγ¯ with Powhel [Kardos, Trocsanyi ’14] (available also in MG5 MC@NLO and Sherpa+OpenLoops)

2 10 10− ' and 14% NLO scale 9 PowHel LHE 1.25 PowHel-NLO 5 8 LO PowHel NLO 7 uncertainty (µ = H /2) 6 13 TeV, µ = Hˆ /2 T 0 [fb/GeV] ⊥ 2 [fb]

5 2 σ 4 , γ δ0 = 0.4 1 3 3 γ 10− p , γ > 30 GeV mild MC effects for inclusive 2 ⊥ ∆R(γ1, γ2), yγ < 2.5 d 5 8 TeV, µ = mt 1 | |

σ/ δ0 = 0.4

observables d 1.6 2 1.3 1.0 1.1 0.7 1.0 ratio K-factor 0.4 0.9

1 5 10− 2 5 1 2 5 10 0 50 100 150 200 250 300 350 400 450 500 ξ mγ1, γ2 [GeV]

2

Isolation of hard photons (Powhel) 2 SMC Rγ, = 0.1 10− SMC δ0 = 0.1 5 [fb/GeV] 2

realistic cone isolation at NLO+PS level , γ 1 2 γ m

d 10 3 − ˆ loose Frixione isolation at generation-cut level σ/ 13 TeV, µ = H /2 d 5 ⊥ 1.1 ⇒ avoid fragmentation component 1.0 ratio 0.9 0 50 100 150 200 250 300 350 400 450 500

mγ1, γ2 [GeV]

S. Pozzorini (Zurich University) ttH¯ Theory 8th HXSWG Workshop 10 / 17 Some open issues in tt¯+photons Production of tt¯+multiple photons

ATLAS/CMS requires sample with Nγ =1,2 photons from matrix elements and/or shower ⇒ requires merging of tt¯+ 0, 1, 2 γ

Radiative top decays [hep-ph/0604120] ttγ¯ at NLO including radiative top decays (t → b`+ν + γ) [Melnikov, Schulze ’11]

1 1 γ in production γ in production 0.8 γ in decay 0.8 γ in decay d ) σ d γ NLO X R ( NLO T 0.6 σ

p 0.6 , d d d , σ ) d NLO γ R

( 0.4 X

NLO 0.4 T σ p d d 0.2 0.2

0 0 0 20 40 60 80 100 120 140 160 180 200 1 2 3 pT(γ) [GeV] R(γ, bjet) < photons from top decays dominate up to pT(γ) ∼ 60 GeV (smooth isolation, Rγi > 0.4) ⇒ requires detailed studies for tt¯+ 2γ

S. Pozzorini (Zurich University) ttH¯ Theory 8th HXSWG Workshop 11 / 17 Outline

1 ttH¯ Signal

2 Backgrounds for H b¯b →

3 Backgrounds for H γγ →

4 Backgrounds for H W W, ZZ, ττ multi-leptons → →

5 Scale uncertainties in NLO matching & merging Need of accurate Monte Carlo predictions

ttV¯ +jets with = W, /γ (dominant MC systematics) ∗ enters signal through extra jet emissions requires precise MC for ttW¯ + 0, 1, 2 jets (same signature as ttH¯ (H → WW → `νjj)! + − tt¯+ Z/γ(→ ` ` ) requires off-shell effects down to small m``

VV +jets (subdominant MC systematics) requires inclusive sample with NLO precision up to 2 jets HF jets crucial (genuine V V b¯b component and light-jet mistags) mainly WZ → 3` but also ZZ → 4`

S. Pozzorini (Zurich University) ttH¯ Theory 8th HXSWG Workshop 12 / 17 Existing predictions and ongoing studies

2 1 (a) PowHel-NLO 5 µ0/2 µ 2µ0 PowHel-LHE NLO+PS ttV¯ with Powhel [Garzelli et al ’12] ≤ ≤ 2

± GeV] -1 tt¯+ W /Z with NWA uncorrelated decays / 10 5 √s = 7TeV [fb t

, mt = 172.9GeV

σ 2 ⊥ d p -2 mZ = 91.1876GeV

K = 1.1–1.35 d 10 µ0 = mt + mZ/2 5 CTEQ6.6M mild NLO uncertainties (10% scale, 8% PDFs) 2 1.1 1.2 1.05 LHE/NLO 1.1 1.0 1.0 -fact 0.95 K-fact 0.9 0.9 0.8 K LHE/NLO 0 50 100 150 200 250 300 350 400 450 500

p ,t [GeV] ⊥ Ongoing Sherpa+OpenLoops and MG5 aMC@NLO studies

pT of W-top-antitop system ttV¯ + 0, 1 jets NLO merging Sherpa+OpenLoops

[pb/GeV] 3

T 10− + − + − p

¯ /d tt` ` with off-shell Z/γ → ` ` possible σ d

4 jets possible (but HF-jets tricky) 10− MEPSNLO30 R1F1Q1 WZ + 0, 1, 2 MEPSNLO30 R0.5F1Q1 MEPSNLO30 R2F1Q1 MEPSNLO30 R1F0.5Q0.5

5 MEPSNLO30 R1F2Q2 10− 1.4 1.2 1

MC/Data 0.8 0.6 0 50 100 150 200 250 300 350 400 pT [GeV]

S. Pozzorini (Zurich University) ttH¯ Theory 8th HXSWG Workshop 13 / 17 Outline

1 ttH¯ Signal

2 Backgrounds for H b¯b →

3 Backgrounds for H γγ →

4 Backgrounds for H W W, ZZ, ττ multi-leptons → →

5 Scale uncertainties in NLO matching & merging Scale dependencies in NLO matching and merging

Prescriptions for scale choices and variations basis for definition of intrinsic precision of NLO Monte Carlo

matching (µQ) and merging (Qcut) involve two technical scales; no widely accepted prescription for their choice and variation (Qcut debate)!

S. Pozzorini (Zurich University) ttH¯ Theory 8th HXSWG Workshop 14 / 17 Benefits of NLO matching and merging (top-pair pT in tt¯+ 0, 1 jets)

pT of top-antitop pair pT of top-antitop pair

Sherpa+OpenLoops Sherpa+OpenLoops µR = ξ mt

[pb/GeV] 1 [pb/GeV] 1 T T p p

/d /d renorm. scale σ σ d d

1 1 10− 10− SMCNLO R1F1Q1 MEPSNLO30 R1F1Q1 SMCNLO R0.5F1Q1 MEPSNLO30 R0.5F1Q1 SMCNLO R2F1Q1 MEPSNLO30 R2F1Q1 2 SMCNLO R1F0.5Q0.5 MEPSNLO30 R1F0.5Q0.5 10− 2 µ = µ = ξ m SMCNLO R1F2Q2 10− MEPSNLO30 R1F2Q2 Q t 1.4 1.4 NLO 1.2 1.2 factorisation scale 1 1

MC/Data 0.8 MC/Data 0.8 0.6 0.6 resummation scale 0 50 100 150 200 250 300 350 400 0 50 100 150 200 250 300 350 400 pT [GeV] pT [GeV]

pT of top-antitop pair pT of top-antitop pair −→ 1 Sherpa+OpenLoops Sherpa+OpenLoops NLO merging 1 [pb/GeV] 1 [pb/GeV] 10− T T p p

/d 2 /d 10− σ σ

d d 10% accuracy over 3 1 10− 10− LOPS R1F1Q1 MEPSLO30 R1F1Q1 4 whole spectrum! 10− LOPS R0.5F1Q1 MEPSLO30 R0.5F1Q1

LO LOPS R2F1Q1 MEPSLO30 R2F1Q1 5 2 10− LOPS R1F0.5Q0.5 10− MEPSLO30 R1F0.5Q0.5 LOPS R1F2Q2 MEPSLO30 R1F2Q2 error estimate more 6 10− 1.4 1.4 1.2 1.2 realistic! 1 1

MC/Data 0.8 MC/Data 0.8 0.6 0.6 0 50 100 150 200 250 300 350 400 0 50 100 150 200 250 300 350 400 pT [GeV] pT [GeV] matching merging −→

S. Pozzorini (Zurich University) ttH¯ Theory 8th HXSWG Workshop 15 / 17 Choice and variation of merging scale Qcut?!

Virtue of small Qcut

NLO accuracy over whole pT spectrum of experimentally resolved jets

Formal problem at (very) small Qcut [Hamilton, Nason, Oleari, Zanderighi ’12] 2 when Qcut approaches Sudakov peak, i.. αS log (Qcut/Qhard) ∼ 1 1.5 ⇒ fake subleading logs generates αS uncertainty

⇒ Optimal Qcut choice? Uncertainty? k log10( jet resolution 0 1 [GeV]) ⊥ → 2 [pb] 10 Sherpa+OpenLoops ) 10 1 (quantitative recipe) /GeV Proposal 01 d

( 1 10 1 consider jet-kT dist ⇒ max Qcut sensitivity 10− /d log

σ 2 MEPSNLO30 R1F1Q1 d 10 − MEPSNLO5 R1F1Q1 push down to Sudakov region MEPSNLO10 R1F1Q1 Qcut 10 3 − MEPSNLO20 R1F1Q1 4 MEPSNLO40 R1F1Q1 10− take local Qcut dependence as uncertainty estimate 1.4 1.2 stop before you exceed NLO scale dependence 1

MC/Data 0.8 0.6 0.5 1 1.5 2 2.5 3 3.5 Example: MEPS@LO for ¯ jets d tt + 0, 1 log10( 01/GeV)

Qcut = 40 ... 5 GeV ⇒ less than 10% uncertainty for Qcut ≥ 10 GeV

S. Pozzorini (Zurich University) ttH¯ Theory 8th HXSWG Workshop 16 / 17 Summary and Outlook

New ttH¯ priorities and activities emphasis on backgrounds and NLO matching+merging lot of exchange between theory ↔ ATLAS/CMS

Next meetings and activities start meetings (soon) focus ttH¯ activities on “top priorities” and new meetings to survey TH and ATLAS/CMS progress theory requiremets from new Run2 analyses (fat jets, MEM, . . . )?

Guarantee coherence of MC simulations tool comparisons based on standard setup (coming soon) theory agreement on uncertainty estimates detailed recommendations for ATLAS/CMS

S. Pozzorini (Zurich University) ttH¯ Theory 8th HXSWG Workshop 17 / 17 Recommendations and Priorities Theory priorities and recommendations (loose selection)I ttH¯ signal NLO merging of ttH + 0, 1 jets NLO decays for top and Higgs

H b¯b backgrounds → use 4F NLO+PS for tt¯+ b-jets; study systematics of g → b¯b shower splittings use NLO merging for tt¯+ 0, 1(2) jets combination of 4F ttb¯ ¯b & tt¯+ jets sound prescription for resummation/merging scale uncertainties NLO+PS for tt¯+ -jets EW contributions and EW–QCD interferences

H γγ backgrounds → use NLO+PS tools for tt¯+photons merging of tt¯+ 0, 1, 2 γ and radiative top decays

S. Pozzorini (Zurich University) ttH¯ Theory 8th HXSWG Workshop 18 / 17 Theory priorities and recommendations (loose selection)II

H W W, ZZ, ττ multi-lepton backgrounds → → NLO merging for ttW¯ + 0, 1(2) jets NLO merging for ttZ/γ¯ + 0, 1 jets with off-shell Z/γ → `+`− (also from top decays) NLO predictions for WZ + b/c-jets

see more details at: https://twiki.cern.ch/twiki/bin/view/LHCPhysics/LHCHXSWGTthMeetingsSummary

S. Pozzorini (Zurich University) ttH¯ Theory 8th HXSWG Workshop 19 / 17