[PPT] - Coherent Showers in Decays of Coloured Resonances Helen Brooks & PowerPoint Presentation

SLIDE 1

VINCIA VINCIA

Coherent Showers in Decays of Coloured Resonances

Helen Brooks & Peter Skands (Monash University)

HEP Seminar Lund, June 2019

45 90 135 180 225 270 315 1 2 3 4 5 6

log10(aRF

g/qqsAK) as a function of θjk in A COM frame

log(E/GeV) = 0.0 log(E/GeV) = 0.2 log(E/GeV) = 0.4 log(E/GeV) = 0.6 log(E/GeV) = 0.8 log(E/GeV) = 1.0 log(E/GeV) = 1.2 log(E/GeV) = 1.4 log(E/GeV) = 1.6 log(E/GeV) = 1.8

PRODUCTION DECAY(S) RF ANTENNA PATTERN

A new shower model based on “Resonance- Final” antennae (with mass- and helicity-dependence)

IF antenna IF antenna II antenna

⊗

R F a n t e n n a

⊗

SLIDE 2

mt < Qevol < Qcut

Coherence in Resonance Decays

HE LEN B RO O K S & P E TE R SK A N D S

2

MO NA S H U.

VINCIA

Note: interference between production and decay will

ccur at scales < Γ; not the topic of this talk

PRODUCTION DECAY(S)

IF antenna IF antenna II antenna

⊗

R F a n t e n n a

⊗

๏In narrow width approximation,

Factorise production and decay of resonances;
These stages are showered independently.

Goal is to shower the resonance-final antenna in decay coherently, without modifying the invariant mass of the resonance, needed for resonance-aware matching.

√s < Qevol < Qcut

SLIDE 3

Prime Motivation: Top Quark Mass

HE LEN B RO O K S & P E TE R SK A N D S

3

MO NA S H U.

VINCIA

arXiv:1801.03944 “... the very minimal message that can be drawn from our work is that, in order to assess a meaningful theoretical error in top-mass measurements, the use of different shower models, associated with different NLO+PS generators, is mandatory.”

SLIDE 4

Uncertainties

HE LEN B RO O K S & P E TE R SK A N D S

4

MO NA S H U.

๏Fixed-order accuracy (μR) + PDFs (μF) + matching/merging (e.g. hdamp) ๏Parton shower ambiguities from logarithmic accuracy

→ Estimate by comparing different shower architectures

๏

+ systematic parametric variations

→ To reduce, need systematic improvements:

๏

At LL / LC: coherence & “optimised” choices (for muR, evolution scale, recoil strategies, …)

๏

Beyond LL / LC: genuine subleading colour (beyond optimised LC) and higher-order corrections to shower kernels (beyond optimised LL)

๏+ Mass Effects, Finite-Width Effects, Polarisation Effects

VINCIA

๏+ Non-perturbative: Renormalon pole mass ambiguity ≲ ΛQCD ,

colour-reconnections,︎ MPI, beam remnant treatment, hadronisation, hadron rescattering, hadron and τ decays, …

( )

SLIDE 5

Dipoles vs Antennae (in resonance decays)

HE LEN B RO O K S & P E TE R SK A N D S

5

MO NA S H U.

๏Dipole showers

Each branching has a well-defined

“radiator” and a “recoiler”, with distinct kinematics maps.

Neglect contribution from

resonance as radiator (partition can even become negative).

In principle free to choose recoiler,

e.g. W in t → W b

๏Antenna Showers

Agnostic as to who is the radiator;

smooth transition in kinematics

๏

Interpolates between collinear limits

Coherence built in; cannot neglect

resonance’s contribution

Recoil strategy relates to antenna

factorisation

VINCIA

kinematics

t b ?

X b −t

t → b W : Top sits at rest (does not radiate) Bottom quark radiates; recoils against the

nly other final-state parton, W.

More branchings: ambiguous what recoiler to use for parton colour-connected to top t → b W: Antenna between bottom and crossed top. Kinematics map with X = W ⟹ W acquires recoil More branchings: unambiguous. Parton colour- connected to top participates in the RF antenna; rest = X collectively acquire the recoil.

*

*Note: the original dipole shower, ARIADNE, is of the type I here call “antenna shower”

SLIDE 6

Current Status of Resonance Decay Showers

HE LEN B RO O K S & P E TE R SK A N D S

6

MO NA S H U.

VINCIA

Shower Type Decay shower? Coherence? Pythia 8 [hep-ph/0010012]

[hep-ph/0408302]

Dipole 3 7 Sherpa [1412.6478] Catani-Seymour 7

(production only)

(3) Herwig 7 (˜ q)

[1810.06493]

Angular-ordered 3 3 Herwig 7 (dip)

[1810.06493]

Catani-Seymour (3)

(on-shell only)

(3) Vincia - NEW! Antenna 3 3

Slide from H. Brooks

Dire? Dipole ✔? *

*: not completely sure about status

( )

*Via ME corrections

*

(no RF dipole) (no RF dipole)

✔?

SLIDE 7

RF Showers 1: Antenna Functions

HE LEN B RO O K S & P E TE R SK A N D S

7

MO NA S H U.

VINCIA

b∗ t b t∗ t b

+

Ant = |

|

2/ Born

pa

<latexit sha1_base64="Klc4ZT8wGVueyowW9OUqStxjSNY=">AB6nicbVA9SwNBEJ2LXzF+RS1tFoNgFe6ioIVFwMYyovmA5Ah7m7lkyd7esbsnhCM/wcZCEVt/kZ3/xk1yhSY+GHi8N8PMvCARXBvX/XYKa+sbm1vF7dLO7t7+QfnwqKXjVDFsljEqhNQjYJLbBpuBHYShTQKBLaD8e3Mbz+h0jyWj2aSoB/RoeQhZ9RY6SHp03654lbdOcgq8XJSgRyNfvmrN4hZGqE0TFCtu56bGD+jynAmcFrqpRoTysZ0iF1LJY1Q+9n81Ck5s8qAhLGyJQ2Zq78nMhpPYkC2xlRM9L3kz8z+umJrz2My6T1KBki0VhKoiJyexvMuAKmRETSyhT3N5K2IgqyoxNp2RD8JZfXiWtWtW7qNbuLyv1mzyOIpzAKZyDB1dQhztoQBMYDOEZXuHNEc6L8+58LFoLTj5zDH/gfP4AR8CNxg=</latexit>

pX

<latexit sha1_base64="ys/KDT7FtvSrefrabnbx1KBsoh0=">AB6nicbVA9SwNBEJ2LXzF+RS1tFoNgFe6ioIVFwMYyovmA5Ah7m7lkyd7esbsnhCM/wcZCEVt/kZ3/xk1yhSY+GHi8N8PMvCARXBvX/XYKa+sbm1vF7dLO7t7+QfnwqKXjVDFsljEqhNQjYJLbBpuBHYShTQKBLaD8e3Mbz+h0jyWj2aSoB/RoeQhZ9RY6SHpd/rlilt15yCrxMtJBXI0+uWv3iBmaYTSMEG17npuYvyMKsOZwGmpl2pMKBvTIXYtlTRC7WfzU6fkzCoDEsbKljRkrv6eyGik9SQKbGdEzUgvezPxP6+bmvDaz7hMUoOSLRaFqSAmJrO/yYArZEZMLKFMcXsrYSOqKDM2nZINwVt+eZW0alXvolq7v6zUb/I4inACp3AOHlxBHe6gAU1gMIRneIU3RzgvzrvzsWgtOPnMfyB8/kDOhyNvQ=</latexit>

pj

<latexit sha1_base64="z0CLNADCfwBGsz16o/McPHsu8g=">AB6nicbVA9SwNBEJ3zM8avqKXNYhCswl0UtLAI2FhGNB+QHGFvs5es2ds7dueEcOQn2FgoYusvsvPfuEmu0MQHA4/3ZpiZFyRSGHTdb2dldW19Y7OwVdze2d3bLx0cNk2casYbLJaxbgfUcCkUb6BAyduJ5jQKJG8Fo5up3ri2ohYPeA4X5EB0qEglG0n3Se+yVym7FnYEsEy8nZchR75W+uv2YpRFXyCQ1puO5CfoZ1SiY5JNiNzU8oWxEB7xjqaIRN342O3VCTq3SJ2GsbSkM/X3REYjY8ZRYDsjikOz6E3F/7xOiuGVnwmVpMgVmy8KU0kwJtO/SV9ozlCOLaFMC3srYUOqKUObTtG4C2+vEya1Yp3XqneXZRr13kcBTiGEzgDy6hBrdQhwYwGMAzvMKbI50X5935mLeuOPnMEfyB8/kDVWSNzw=</latexit>

pk

<latexit sha1_base64="w0EQAkNDfaLwaF7qtdIYNvQk4DI=">AB6nicbVA9SwNBEJ2LXzF+RS1tFoNgFe6ioIVFwMYyovmA5Ah7m7lkyd7esbsnhCM/wcZCEVt/kZ3/xk1yhSY+GHi8N8PMvCARXBvX/XYKa+sbm1vF7dLO7t7+QfnwqKXjVDFsljEqhNQjYJLbBpuBHYShTQKBLaD8e3Mbz+h0jyWj2aSoB/RoeQhZ9RY6SHpj/vlilt15yCrxMtJBXI0+uWv3iBmaYTSMEG17npuYvyMKsOZwGmpl2pMKBvTIXYtlTRC7WfzU6fkzCoDEsbKljRkrv6eyGik9SQKbGdEzUgvezPxP6+bmvDaz7hMUoOSLRaFqSAmJrO/yYArZEZMLKFMcXsrYSOqKDM2nZINwVt+eZW0alXvolq7v6zUb/I4inACp3AOHlxBHe6gAU1gMIRneIU3RzgvzrvzsWgtOPnMfyB8/kDVuiN0A=</latexit>

pk

<latexit sha1_base64="w0EQAkNDfaLwaF7qtdIYNvQk4DI=">AB6nicbVA9SwNBEJ2LXzF+RS1tFoNgFe6ioIVFwMYyovmA5Ah7m7lkyd7esbsnhCM/wcZCEVt/kZ3/xk1yhSY+GHi8N8PMvCARXBvX/XYKa+sbm1vF7dLO7t7+QfnwqKXjVDFsljEqhNQjYJLbBpuBHYShTQKBLaD8e3Mbz+h0jyWj2aSoB/RoeQhZ9RY6SHpj/vlilt15yCrxMtJBXI0+uWv3iBmaYTSMEG17npuYvyMKsOZwGmpl2pMKBvTIXYtlTRC7WfzU6fkzCoDEsbKljRkrv6eyGik9SQKbGdEzUgvezPxP6+bmvDaz7hMUoOSLRaFqSAmJrO/yYArZEZMLKFMcXsrYSOqKDM2nZINwVt+eZW0alXvolq7v6zUb/I4inACp3AOHlxBHe6gAU1gMIRneIU3RzgvzrvzsWgtOPnMfyB8/kDVuiN0A=</latexit>

pX

<latexit sha1_base64="ys/KDT7FtvSrefrabnbx1KBsoh0=">AB6nicbVA9SwNBEJ2LXzF+RS1tFoNgFe6ioIVFwMYyovmA5Ah7m7lkyd7esbsnhCM/wcZCEVt/kZ3/xk1yhSY+GHi8N8PMvCARXBvX/XYKa+sbm1vF7dLO7t7+QfnwqKXjVDFsljEqhNQjYJLbBpuBHYShTQKBLaD8e3Mbz+h0jyWj2aSoB/RoeQhZ9RY6SHpd/rlilt15yCrxMtJBXI0+uWv3iBmaYTSMEG17npuYvyMKsOZwGmpl2pMKBvTIXYtlTRC7WfzU6fkzCoDEsbKljRkrv6eyGik9SQKbGdEzUgvezPxP6+bmvDaz7hMUoOSLRaFqSAmJrO/yYArZEZMLKFMcXsrYSOqKDM2nZINwVt+eZW0alXvolq7v6zUb/I4inACp3AOHlxBHe6gAU1gMIRneIU3RzgvzrvzsWgtOPnMfyB8/kDOhyNvQ=</latexit>

pa

<latexit sha1_base64="Klc4ZT8wGVueyowW9OUqStxjSNY=">AB6nicbVA9SwNBEJ2LXzF+RS1tFoNgFe6ioIVFwMYyovmA5Ah7m7lkyd7esbsnhCM/wcZCEVt/kZ3/xk1yhSY+GHi8N8PMvCARXBvX/XYKa+sbm1vF7dLO7t7+QfnwqKXjVDFsljEqhNQjYJLbBpuBHYShTQKBLaD8e3Mbz+h0jyWj2aSoB/RoeQhZ9RY6SHp03654lbdOcgq8XJSgRyNfvmrN4hZGqE0TFCtu56bGD+jynAmcFrqpRoTysZ0iF1LJY1Q+9n81Ck5s8qAhLGyJQ2Zq78nMhpPYkC2xlRM9L3kz8z+umJrz2My6T1KBki0VhKoiJyexvMuAKmRETSyhT3N5K2IgqyoxNp2RD8JZfXiWtWtW7qNbuLyv1mzyOIpzAKZyDB1dQhztoQBMYDOEZXuHNEc6L8+58LFoLTj5zDH/gfP4AR8CNxg=</latexit>

pj

<latexit sha1_base64="z0CLNADCfwBGsz16o/McPHsu8g=">AB6nicbVA9SwNBEJ3zM8avqKXNYhCswl0UtLAI2FhGNB+QHGFvs5es2ds7dueEcOQn2FgoYusvsvPfuEmu0MQHA4/3ZpiZFyRSGHTdb2dldW19Y7OwVdze2d3bLx0cNk2casYbLJaxbgfUcCkUb6BAyduJ5jQKJG8Fo5up3ri2ohYPeA4X5EB0qEglG0n3Se+yVym7FnYEsEy8nZchR75W+uv2YpRFXyCQ1puO5CfoZ1SiY5JNiNzU8oWxEB7xjqaIRN342O3VCTq3SJ2GsbSkM/X3REYjY8ZRYDsjikOz6E3F/7xOiuGVnwmVpMgVmy8KU0kwJtO/SV9ozlCOLaFMC3srYUOqKUObTtG4C2+vEya1Yp3XqneXZRr13kcBTiGEzgDy6hBrdQhwYwGMAzvMKbI50X5935mLeuOPnMEfyB8/kDVWSNzw=</latexit>

yaj ≡ saj/(sAK + sjk), µ

…

, µ2

a ≡ m2 a/(sAK + sjk),

collinear

= ⇒ zk ∼ yak = 1 − yaj (+2µ2

j) ,

za ∼ yAK = 1 − yjk

<latexit sha1_base64="ASWFlizDJe0CbU7GuWOR0ZD2qMc=">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</latexit>

Define dimensionless invariants:

Labeling: AI KF | {z }

pre-branching

→ aI jF kF | {z }

post-branching

<latexit sha1_base64="LW8v+2NvWNsOBoZepGIEU+knU=">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</latexit>

<latexit sha1_base64="R4jQO8Hn2S9B4gXzYJx529W8RSI=">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</latexit>

Polynomial(s) chosen such that all helicity components remain positive-definite → same forms as FF, IF, II :*

*: difference is 1/(sAK + sjk) normalisation and phase-space map

N.B.: sαβ ≡ 2pα · pβ throughout!

SLIDE 8

Example: Collinear Limits

HE LEN B RO O K S & P E TE R SK A N D S

8

MO NA S H U.

VINCIA

Example: qq antenna limits

Can rewrite antenna as: aRF

g/qq =

1 sAK 2 6 6 6 4 2yak yajyjk − 2µ2

a

y2

aj

− 2µ2

k

y2

jk

| {z }

soft

+ yaj yjk + yjk yaj | {z }

collinear

+ n.s. 3 7 7 7 5 Define Q2 ≡ sjk; y ≡

Q2 sAK ;

z ≡

sak sAK

⇒

saj sAK = 1 + y − z

aRF

g/qq = 1

Q2 2z(1 + y) 1 + y − z + (1 + y − z) − 2m2

k

Q2 + O(y)

+ n.s.

In collinear limit, y → 0 lim

y→0 aRF g/qq = 1

Q2 1 + z2 1 − z − 2m2

k

Q2

= 1

Q2 Pq→gq(z, ˜ µ) N.B. Need to sum over neighbouring antennae for gg collinear limit.

Labeling: AI KF | {z }

pre-branching

→ aI jF kF | {z }

post-branching

<latexit sha1_base64="LW8v+2NvWNsOBoZepGIEU+knU=">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</latexit>

N.B.: sαβ ≡ 2pα · pβ throughout!

yαβ = sαβ sAK + sjk

<latexit sha1_base64="NVSQ1PniNbRy7oBg8rLboV8JbM=">ACInicbZDLSsNAFIYn9VbrerSzWARBKEkVAXQtWN4KaCvUATysl0o6dXJiZCXkWdz4Km5cKOpK8GctlnY1h8Gfr5zDmfO70acSWa30ZuYXFpeSW/Wlhb39jcKm7vNGQYC0LrJOShaLkgKWcBrSumOG1FgoLvctp0B9ejevORCsnC4F4NI+r40AuYxwgojTrF82EnsYFHfbBdqiDF9j2BJBETvN0BC5vU3yEtXkYpGmnWDL5lh43liZKaFMtU7x0+6GJPZpoAgHKduWGSknAaEY4TQt2LGkEZAB9Ghb2wB8Kp1kfGKDzTpYi8U+gUKj+nfiQR8KYe+qzt9UH05WxvB/2rtWHlnTsKCKFY0IJNFXsyxCvEoL9xlghLFh9oAEUz/FZM+6ISUTrWgQ7BmT543jUrZOi5X7k5K1asjzaQ/voEFnoFXRDaqhOiLoCb2gN/RuPBuvxofxNWnNGdnMLpqS8fMLrxKk/A=</latexit>

SLIDE 9

Helicity Structure for Gluon Splittings

HE LEN B RO O K S & P E TE R SK A N D S

9

MO NA S H U.

VINCIA

Example: XGsplitIF

Labeling: AI KF | {z }

pre-branching

→ aI jF kF | {z }

post-branching

<latexit sha1_base64="LW8v+2NvWNsOBoZepGIEU+knU=">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</latexit>

N.B.: sαβ ≡ 2pα · pβ throughout!

Note sum of ++ antennae have same singularities as sum of +- ones => same singular terms

btained when summing over helicity of emitted gluon irrespective of parent helicities

Identical to RF modulo nonsingular terms

yαβ = sαβ sAK + sjk

Helicity conservation; go to zero when “-“ daughter gets x→1

“Helicity Flip” proportional to mass squared

SLIDE 10

Helicity Structure for Gluon Emissions

HE LEN B RO O K S & P E TE R SK A N D S

10

MO NA S H U.

VINCIA

a(++ → + + +) = 1 sAK  1 yajyjk + (1 − α)1 − 2yaj yjk − µ2

a

y2

aj

,

(C.28) a(++ → + − +) = 1 sAK (1 − yaj)3 + (1 − yjk)2 − 1 yajyjk − µ2

a(1 − yjk − yaj)2(1 − yaj)

y2

aj

+ 3 − y2

aj

(C.29)

a(++ → − − +) = 1 sAK µ2

ay2 jk

y2

aj

(C.30)

a(+− → + + −) = 1 sAK (1 − yaj)3 yajyjk − µ2

a(1 − yaj)2

y2

aj

,

(C.31) a(+− → + − −) = 1 sAK (1 − yjk)2 yajyjk + (1 − α)1 − 2yaj yjk − µ2

a(1 − yjk)2

y2

aj

+ 2yaj − yjk

a(+− → − − −)

= 1 sAK µ2

ay2 jk

y2

aj

(C.32)

Example: QGemitIF

MHV

Labeling: AI KF | {z }

pre-branching

→ aI jF kF | {z }

post-branching

<latexit sha1_base64="LW8v+2NvWNsOBoZepGIEU+knU=">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</latexit>

N.B.: sαβ ≡ 2pα · pβ throughout!

Note sum of ++ antennae have same singularities as sum of +- ones => same singular terms

btained when summing over helicity of emitted gluon irrespective of parent helicities

“Helicity Flip” proportional to mass squared “Helicity Flip” proportional to mass squared NMHV

Identical to RF modulo nonsingular terms

Helicity conservation => Suppressed when xk or xa -> 0

yαβ = sαβ sAK + sjk

<latexit sha1_base64="NVSQ1PniNbRy7oBg8rLboV8JbM=">ACInicbZDLSsNAFIYn9VbrerSzWARBKEkVAXQtWN4KaCvUATysl0o6dXJiZCXkWdz4Km5cKOpK8GctlnY1h8Gfr5zDmfO70acSWa30ZuYXFpeSW/Wlhb39jcKm7vNGQYC0LrJOShaLkgKWcBrSumOG1FgoLvctp0B9ejevORCsnC4F4NI+r40AuYxwgojTrF82EnsYFHfbBdqiDF9j2BJBETvN0BC5vU3yEtXkYpGmnWDL5lh43liZKaFMtU7x0+6GJPZpoAgHKduWGSknAaEY4TQt2LGkEZAB9Ghb2wB8Kp1kfGKDzTpYi8U+gUKj+nfiQR8KYe+qzt9UH05WxvB/2rtWHlnTsKCKFY0IJNFXsyxCvEoL9xlghLFh9oAEUz/FZM+6ISUTrWgQ7BmT543jUrZOi5X7k5K1asjzaQ/voEFnoFXRDaqhOiLoCb2gN/RuPBuvxofxNWnNGdnMLpqS8fMLrxKk/A=</latexit>

“Gluon collinear partitioning” interpolates between GP (α=1) and GGG (α=0)

ARIADNE VINCIA DEF

SLIDE 11

RF Showers 2: Evolution Variables

HE LEN B RO O K S & P E TE R SK A N D S

11

MO NA S H U.

VINCIA

Emissions: Q2

evol =

sajsjk sjk + sAK ζ = sjk + sAK sAK Splittings: Q2

evol = (sjk + 2m2 q)(saj − m2 q)

sAK + sjk + 2m2

q

ζ = sak sAK

Labeling: AI KF | {z }

pre-branching

→ aI jF kF | {z }

post-branching

<latexit sha1_base64="LW8v+2NvWNsOBoZepGIEU+knU=">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</latexit>

Same since mj = 0 for emission

Generalisation

f ARIADNE

pT

Resonance Mass “Dead Cone” Resonance Mass “Dead Cone” Final-state quark “Dead Cone”

N.B.: sαβ ≡ 2pα · pβ throughout!

SLIDE 12 ๏2→3 phase-space factorisation:

RF Showers 3: Phase-Space Factorisation

HE LEN B RO O K S & P E TE R SK A N D S

12

MO NA S H U.

VINCIA

dΦn+1 = dΦant × dΦn

Labeling: AI KF | {z }

pre-branching

→ aI jF kF | {z }

post-branching

<latexit sha1_base64="LW8v+2NvWNsOBoZepGIEU+knU=">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</latexit>

N.B.: sαβ ≡ 2pα · pβ throughout!

≡ ·

I Factorisation is exact, not just in soft, collinear limits I Preserves invariant mass of resonance: pA = pa I Preserves invariant mass of system of recoilers:

pA = pK + pX = ⇒ m2

X = (pA − pK)2 ≡ (pa − pj − pk)2 = m2 X0

<latexit sha1_base64="/0Af+l2Fr+v1290LbeRAu4kt41g=">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</latexit>

dΦa!jk+{X} = 1 (4π)5 dsajdsjkdφ m2

A

dΩK

dΦant = 1 16π2 dsajdsjk λ1/2(m2

A, m2 AK, m2 K)

dφ 2π

dΦA!K+{X} = 1 8(2π)2 λ1/2(m2

A, m2 AK, m2 K)

m2

A

dΩK

Same form as the final-final antenna phase space … with mX = mAK as one of the Born parameters

SLIDE 13

RF Showers 4: Kinematics Map (Recoil)

HE LEN B RO O K S & P E TE R SK A N D S

13

MO NA S H U.

VINCIA

I Construct in A rest frame, and rotate such that K is along z. I Specify system X only recoils longitudinally. I Rotate about z by φ (flatly sampled). I Boost back to lab frame. I For each recoiler i, boost pi by pX0 − pX

Note!

If we fix to just one recoiler i.e. A → RKX, a → rjkX then CANNOT simulatenously preserve m2

A, m2 R and m2 AK.

Replace A → A − X everywhere.

I Antenna mass is modified! I Phase space normalisation is modified! I Mass used everywhere is (pA − pX)2 - not same as propagator!

Labeling: AI KF | {z }

pre-branching

→ aI jF kF | {z }

post-branching

<latexit sha1_base64="LW8v+2NvWNsOBoZepGIEU+knU=">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</latexit>

N.B.: sαβ ≡ 2pα · pβ throughout!

*Note the prescription defined here is similar to one recently implemented in Herwig7 by Cormier et al., arXiv:1810.06493

SLIDE 14

Effect of Kinematics Map

HE LEN B RO O K S & P E TE R SK A N D S

14

MO NA S H U.

VINCIA

Consider average recoil |∆~ pW |, after first and second emission(s). Recoil after first:

100 101 pT evol [GeV] 20 40 60 80 |∆~ pW | [GeV] PYTHIA 8 (W recoil map) VINCIA (W recoil map) VINCIA (default map)

Recoil after second:

100 101 pT evol [GeV] 20 40 60 |∆~ pW | [GeV] PYTHIA 8 (W recoil map) VINCIA (W recoil map) VINCIA (default map)

Second branching: Collective RF map → less recoil to W First branching: there is only the W

SLIDE 15

(Coherence In Production)

HE LEN B RO O K S & P E TE R SK A N D S

15

MO NA S H U.

VINCIA

50 100 150 200 250 pT (t¯ t) [GeV] 1 2 Ratio to HERWIG 7 ang −1.00 −0.75 −0.50 −0.25 0.00 0.25 0.50 0.75 1.00 AF B(pT (t¯ t)) p¯ p → t¯ t √s = 1.96 TeV PS only (no MPI) PYTHIA 8 HERWIG 7 angular HERWIG 7 dipole VINCIA

Forward-backwards asymmetry: AFB(O) =

dσ dO

∆y>0 − dσ

dO

∆y<0

dσ dO

∆y>0 + dσ

dO

∆y<0

Coherent showers include part of the real emission correction that generates a FB asymmetry that becomes negative for large pT (t¯ t). [1205.1466]

(b) _ _ _ _ q q (a) q t t t q q q q q t t q q

Well-studied effect in p-pbar collisions Top quark FB asymmetry

PS, Webber, Winter JHEP 1207 (2012) 151

Coherent showers produce a pTdependent asymmetry Herwig7 dipole shower exhibits exactly same behaviour as VINCIA

SLIDE 16

Coherence in Decay

HE LEN B RO O K S & P E TE R SK A N D S

16

MO NA S H U.

VINCIA

Plot antenna function in top centre of mass frame (b along z):

45 90 135 180 225 270 315 1 2 3 4 5 6

log10(aRF

g/qqsAK) as a function of θjk in A COM frame

log(E/GeV) = 0.0 log(E/GeV) = 0.2 log(E/GeV) = 0.4 log(E/GeV) = 0.6 log(E/GeV) = 0.8 log(E/GeV) = 1.0 log(E/GeV) = 1.2 log(E/GeV) = 1.4 log(E/GeV) = 1.6 log(E/GeV) = 1.8 45 90 135 180 225 270 315 0.2 0.4 0.6 0.8 1.0

aRF

g/qq

Pgq(z)/Q2 as a function of θjk in A COM frame

log(E/GeV) = 0.0 log(E/GeV) = 0.2 log(E/GeV) = 0.4 log(E/GeV) = 0.6 log(E/GeV) = 0.8 log(E/GeV) = 1.0 log(E/GeV) = 1.2 log(E/GeV) = 1.4 log(E/GeV) = 1.6 log(E/GeV) = 1.8

Antenna function is consistent with Altarelli-Parisi splitting function in (quasi-)collinear direction, coherence results in a suppression in the backwards direction.

Ratio to AP kernel Log of antenna function

SLIDE 17 ๏VINCIA gives narrower b-jets than Pythia 8

Effect survives MPI + hadronisation

B-Jet Profiles

HE LEN B RO O K S & P E TE R SK A N D S

17

MO NA S H U.

VINCIA

0.0 0.1 0.2 0.3 0.4 0.5 0.6 r 0.5 1.0 1.5 Ratio to PYTHIA 8 10−2 10−1 100 101 (r) pp → t¯ t → b¯ b`+`−⌫¯ ⌫, √s = 13 TeV PS only (no MPI) pT bj ∈ [30, 50] GeV Qcut ∈ [0.5, 1.0] GeV PYTHIA 8 VINCIA 0.0 0.1 0.2 0.3 0.4 0.5 0.6 r 0.6 0.8 1.0 1.2 1.4 Ratio to PYTHIA 8 10−1 100 101 (r) pp → t¯ t → b¯ b`+`−⌫¯ ⌫, √s = 13 TeV PS + MPI + had pT bj ∈ [30, 50] GeV Qcut ∈ [0.5, 1.0] GeV PYTHIA 8 VINCIA

Shower only Shower + MPI + Hadr

Tentative conclusion: more coherence ~ more wide-angle suppression?

*Also agrees with intuition from dipole language where “top dipole” can be negative

Differential jet shape ρ(r)

SLIDE 18

Matching with POWHEG

HE LEN B RO O K S & P E TE R SK A N D S

18

MO NA S H U.

VINCIA

I Use POWHEG v2 (t¯

tdec)1 (no need for exact finite width effects)

I Very similar setup to matching

with PYTHIA in 2.

I Veto hardest emission in

production with

Vincia:QmaxMatch = 1

I Veto hardest emission in decay

with UserHooks interface ATLAS dileptonic t¯ t @ 8 TeV [1709.09407]

100 150 200 250 300 350 pT e + pT µ[GeV ] 0.8 1.0 1.2 Ratio to ATLAS data 10−4 10−3 10−2

1

d

d(pT e+pT µ)[GeV −1]

pp → t¯ t → b¯ b`+`−⌫¯ ⌫, √s = 8 TeV PS only (no MPI) ATLAS data PY8+POWHEG VINCIA+POWHEG H7(ang)+POWHEG PYTHIA 8 @LO

1[1412.1828],[1509.0907] 2[1801.03944] 3Thanks to S. Ferrario Ravasio for providing an interface to H7

24

SLIDE 19

168 170 172 174 176 178 mbj+[GeV ] 0.8 1.0 1.2 Ratio to PYTHIA 8 0.00 0.05 0.10 0.15 0.20 0.25 0.30

d dmbj + [pb/GeV]

pp → t¯ t → b¯ be+µ−νe¯ νµ, √s = 8 TeV PS +POWHEG v2 (no MPI) PYTHIA 8 VINCIA HERWIG 7 (ang) PYTHIA 8 (W recoil) VINCIA (W recoil)

Top Mass Profile @ 8 TeV : Parton Level

HE LEN B RO O K S & P E TE R SK A N D S

19

MO NA S H U.

VINCIA

p¯ p → t¯ t @ 8 TeV: mbj`⌫

Monte-Carlo “truth” (parton-level) analysis:

I Assumes we can reconstruct pν and match correct `, bj pair.

(looking under the hood / “cheating”)

Plot from H. Brooks

SLIDE 20

Top Mass Profile @ 8 TeV

HE LEN B RO O K S & P E TE R SK A N D S

20

MO NA S H U.

VINCIA

20 40 60 80 100 120 140 160 180 200 mbjµ[GeV ] 0.8 0.9 1.0 1.1 1.2 Ratio to PYTHIA 8 10−4 10−3 10−2

dσ dmbj µ [pb/GeV]

pp → t¯ t → b¯ be+µ−νe¯ νµ, √s = 8 TeV PS+ MPI+had+POWHEG v2 PYTHIA 8 VINCIA HERWIG 7 (ang)

p¯ p → t¯ t @ 8 TeV: mbjµ

Full hadron-level analysis: choose pairing for `, bj that minimise average mass. Again, note endpoint. Note Endpoint (example of a realistic observable)

Plot from H. Brooks

SLIDE 21

Summary

HE LEN B RO O K S & P E TE R SK A N D S

21

MO NA S H U.

๏VINCIA can now do

production and decay

f top quarks
With full mass and

helicity dependence

๏Based on new

“resonance-final” antennae

Coherent top+b (&

top+g) radiation patterns

Collective recoil

kinematics

VINCIA

Coming soon... PYTHIA 8.3 → Watch this space!

SLIDE 22

Outlook

HE LEN B RO O K S & P E TE R SK A N D S

22

MO NA S H U.

๏Finite-width effects

Baseline naive model:
+ some alternatives (with Rob Verheyen)

VINCIA

IF antenna IF antenna II antenna

⊗

R F a n t e n n a R F a n t e n n a

⊗ Q > Γ Q > Γ Q < Γ

I F a n t e n n a

Note: we do not expect these effects to be large for top decays, cf e.g., Khoze & Sjöstrand Phys.Lett. B328 (1994) 466-476

SLIDE 23

Shower Architectures

HE LEN B RO O K S & P E TE R SK A N D S

23

MO NA S H U.

๏Sum over all dipoles / antennae should reproduce the

leading log

VINCIA

Type Singularities Coherence? No dead Examples soft collinear zones? DGLAP part. full 7 7 Angular full+veto full+veto 3 7 H7 ˜ q Dipole part. part. 7 3 Pythia 8 C-S part. part. 3 3 Sherpa, H7 dip Antenna (global) full part. 3 3 Vincia Antenna (sector) full full+veto 3 3 Vincia

Table from H. Brooks