Post on 27-Oct-2019
MEPS@NLO Results Conclusions
NLO merging in tt+jets
Marek Schonherr
Institute for Particle Physics Phenomenology
MPI ATLAS-PhenoGroup Meeting – 15/01/2013
Phys.Rev.D88(2013)014040∗
∗in collaboration with S. Hoche, J. Huang, G. Luisoni, J. Winter
Marek Schonherr IPPP Durham
NLO merging in tt+jets 1
MEPS@NLO Results Conclusions
Contents
1 MEPS@NLO
Introduction of the method to merge NLOPS
for pp→ tt, pp→ tt+ 1j, pp→ tt+ 2j,etc.used in SHERPA.
2 Results
Results for top pair production at the Tevatron,with emphasis on the forward-backward asymmetry.
3 Conclusions
Marek Schonherr IPPP Durham
NLO merging in tt+jets 2
MEPS@NLO Results Conclusions
The SHERPA event generator framework
�������������������������
�������������������������
������������������������������������������������������������
������������������������������������������������������������
������������������������������������
������������������������������������
������������������������������������
������������������������������������
���������������������������������������������������������
���������������������������������������������������������
������������
������������
������������������
������������������
��������������������
��������������������
������������������
������������������
��������������������
��������������������
������������������
������������������
��������������������
��������������������
��������������������
��������������������
������������������������
������������������������
������������������������
������������������������
��������������������
��������������������
������������������
������������������������
������������
������������������
���������������
���������������
������������������
������������������
�������������������������
�������������������������
���������������
���������������
������������������
������������������
������������������������
������������������
������������������
������������
������������
������������������
������������������
����������
����������
����������
����������
������������������
������������������ ���
������������
���������������
������������������
������������������
���������������
���������������
����������
����������
������������������
������������������
������������
������������
���������������
���������������
��������������������
��������������������
������������
������������
������������������
������������������
������������������������������������
��������������������������������������
����������
������������
���������������
���������������
���������������
���������������
• Two multi-purpose Matrix Element (ME) generatorsAMEGIC++ JHEP02(2002)044, EPJC53(2008)501
COMIX JHEP12(2008)039, PRL109(2012)042001
• A Parton Shower (PS) generatorCSSHOWER++ JHEP03(2008)038
• A multiple interaction simulationa la Pythia AMISIC++ hep-ph/0601012
• A cluster fragmentation moduleAHADIC++ EPJC36(2004)381
• A hadron and τ decay package HADRONS++
• A higher order QED generator usingYFS-resummation PHOTONS++ JHEP12(2008)018
• A minimum bias simulation SHRiMPS to appear
Sherpa’s traditional strength is the perturbative part of the eventMEPS (CKKW), S-MC@NLO, MENLOPS, MEPS@NLO
→ full analytic control mandatory for consistency/accuracyMarek Schonherr IPPP Durham
NLO merging in tt+jets 3
JHEP02(2009)007
MEPS@NLO Results Conclusions
MethodPS LO
LO
LO
Parton showers
resummation of (soft-)collinear limit→ intrajet evolution
• matrix elements (ME) and parton showers (PS) are approximations indifferent regions of phase space
• MEPS combines multiple LOPS – keeping either accuracy• NLOPS elevate LOPS to NLO accuracy• MENLOPS supplements core NLOPS with higher multiplicities LOPS
•Marek Schonherr IPPP Durham
NLO merging in tt+jets 4
MEPS@NLO Results Conclusions
Method
ME
LO LO LO LO
Matrix elements
fixed-order in αs→ hard wide-angle emissions→ interference terms
• matrix elements (ME) and parton showers (PS) are approximations indifferent regions of phase space
• MEPS combines multiple LOPS – keeping either accuracy• NLOPS elevate LOPS to NLO accuracy• MENLOPS supplements core NLOPS with higher multiplicities LOPS
•Marek Schonherr IPPP Durham
NLO merging in tt+jets 5
MEPS@NLO Results Conclusions
Method
ME
PS LO
LO
LO
LO
LO
LO LO
MEPS (CKKW,MLM)
Catani, Krauss, Kuhn, Webber JHEP11(2001)063
Lonnblad JHEP05(2002)046
Hoche, Krauss, Schumann, Siegert JHEP05(2009)053
Lonnblad, Prestel JHEP02(2013)094
• matrix elements (ME) and parton showers (PS) are approximations indifferent regions of phase space
• MEPS combines multiple LOPS – keeping either accuracy• NLOPS elevate LOPS to NLO accuracy• MENLOPS supplements core NLOPS with higher multiplicities LOPS
•Marek Schonherr IPPP Durham
NLO merging in tt+jets 6
MEPS@NLO Results Conclusions
MethodPS NLO
NLO
NLO
NLOPS (MC@NLO,POWHEG,S-MC@NLO)
Frixione, Webber JHEP06(2002)029
Nason JHEP11(2004)040, Frixione et.al. JHEP11(2007)070
Hoche, Krauss, MS, Siegert JHEP09(2012)049
• matrix elements (ME) and parton showers (PS) are approximations indifferent regions of phase space
• MEPS combines multiple LOPS – keeping either accuracy• NLOPS elevate LOPS to NLO accuracy• MENLOPS supplements core NLOPS with higher multiplicities LOPS
•Marek Schonherr IPPP Durham
NLO merging in tt+jets 7
MEPS@NLO Results Conclusions
Method
ME
PS NLO
NLO
NLO
LO
LO
LO LO
MENLOPS
Hamilton, Nason JHEP06(2010)039
Hoche, Krauss, MS, Siegert JHEP08(2011)123
Gehrmann, Hoche, Krauss, MS, Siegert JHEP01(2013)144
• matrix elements (ME) and parton showers (PS) are approximations indifferent regions of phase space
• MEPS combines multiple LOPS – keeping either accuracy• NLOPS elevate LOPS to NLO accuracy• MENLOPS supplements core NLOPS with higher multiplicities LOPS
•Marek Schonherr IPPP Durham
NLO merging in tt+jets 8
MEPS@NLO Results Conclusions
Method
ME
PS NLO
NLO
NLO
NLO
NLO
NLO NLO
MEPS@NLO
Lavesson, Lonnblad JHEP12(2008)070
Hoche, Krauss, MS, Siegert JHEP04(2013)027
Gehrmann, Hoche, Krauss, MS, Siegert JHEP01(2013)144
Lonnblad, Prestel JHEP03(2013)166
• matrix elements (ME) and parton showers (PS) are approximations indifferent regions of phase space
• MEPS combines multiple LOPS – keeping either accuracy• NLOPS elevate LOPS to NLO accuracy• MENLOPS supplements core NLOPS with higher multiplicities LOPS
• MEPS@NLO combines multiple NLOPS – keeping either accuracyMarek Schonherr IPPP Durham
NLO merging in tt+jets 9
MEPS@NLO Results Conclusions
Method
ME
PS NLO
NLO
NLO
NLO
NLO
NLO NLO
MEPS@NLO
Lavesson, Lonnblad JHEP12(2008)070
Hoche, Krauss, MS, Siegert JHEP04(2013)027
Gehrmann, Hoche, Krauss, MS, Siegert JHEP01(2013)144
Lonnblad, Prestel JHEP03(2013)166
• matrix elements (ME) and parton showers (PS) are approximations indifferent regions of phase space
• MEPS combines multiple LOPS – keeping either accuracy• NLOPS elevate LOPS to NLO accuracy• MENLOPS supplements core NLOPS with higher multiplicities LOPS
• MEPS@NLO combines multiple NLOPS – keeping either accuracyMarek Schonherr IPPP Durham
NLO merging in tt+jets 10
MEPS@NLO Results Conclusions
Contents
1 MEPS@NLO
Introduction of the method to merge NLOPS
for pp→ tt, pp→ tt+ 1j, pp→ tt+ 2j,etc.used in SHERPA.
2 Results
Results for top pair production at the Tevatron,with emphasis on the forward-backward asymmetry.
3 Conclusions
Marek Schonherr IPPP Durham
NLO merging in tt+jets 11
MEPS@NLO Results Conclusions
Recent results
Fixed-mulitplicity NLOPS (S-MC@NLO)
• pp→W + 0, 1, 2, 3jets – SHERPA+BLACKHAT
Hoche, Krauss, MS, Siegert Phys.Rev.Lett.110(2013)052001
• pp→ jets – SHERPA+BLACKHAT Hoche, MS Phys.Rev.D86(2012)094042
• pp→ ttbb – SHERPA+OPENLOOPS
Cascioli, Maierhofer, Moretti, Pozzorini, Siegert arXiv:1309.0500
Multijet merging at NLO accuracy (MEPS@NLO)
• pp→W + jets – SHERPA+BLACKHAT Hoche, Krauss, MS, Siegert JHEP04(2013)027
• e+e− → jets – SHERPA+BLACKHAT
Gehrmann, Hoche, Krauss, MS, Siegert JHEP01(2013)144
• pp→ h+ jets – SHERPA+GOSAM Hoche, Krauss, MS, Siegert, in YR3 arXiv:1307.1347
• pp→ tt+ jets – SHERPA+GOSAM
Hoche, Huang, Luisoni, MS, Winter Phys.Rev.D88(2013)014040
• pp→ 4`+ jets – SHERPA+OPENLOOPS
Cascioli, Hoche, Krauss, Maierhofer, Pozzorini, Siegert arXiv:1309.5912
Marek Schonherr IPPP Durham
NLO merging in tt+jets 12
MEPS@NLO Results Conclusions
Results – pp→ tt+ jets – AFB
Skands, Webber, Winter JHEP07(2012)151
Parton showers and the tt-asymmetryat the Tevatron
• if colour coherence is respected,PS creates an asymmetrybecause of asymmetric colourflow
• HERWIG respects colourcorrelations through angularordering
(b)_
__
_qq
(a)
q
t
t t
qqqq q
ttt t
q q
• CSSHOWER++ (CS dipoles, 1→2 splittings, recoil to large-Nc partner)→ respects colour correlations by choice of radiating dipoles/recoil partners
⇒ it is important to respect colour-correlations
Marek Schonherr IPPP Durham
NLO merging in tt+jets 13
MEPS@NLO Results Conclusions
Results – pp→ tt+ jets – AFB
Hoche, Huang, Luisoni, MS, Winter arXiv:1306.2703
Sherpa+GoSam
S-MC@NLO, H(A) = 0µ =
√1/2 . . .
√2 kT
PS, B → Bµ =
√1/2 . . .
√2 kT
10−2
10−1
1
10 1Transverse momentum of top-quark pair
dσ
/d
log(
p T,t
t/G
eV)
[pb]
0 0.5 1 1.5 2 2.5
0.9
1.0
1.1
log(pT,tt/GeV)
PS/
MC
@N
LO
Importance ofNc = 3 colour coherence
(SHERPA’s S-MC@NLO)vs.Nc →∞ colour coherence
(SHERPA’s CSSHOWER++)
• small effect on standard (rapidityblind) observables, e.g. p⊥,tt→ some destructive interferenceat large p⊥,tt
• large effect on AFB(p⊥,tt)→ subleading colour termsincrease asym. radiation pattern
Marek Schonherr IPPP Durham
NLO merging in tt+jets 14
MEPS@NLO Results Conclusions
Results – pp→ tt+ jets – AFB
Hoche, Huang, Luisoni, MS, Winter arXiv:1306.2703
Sherpa+GoSam
S-MC@NLO, H(A) = 0µ =
√1/2 . . .
√2 kT
PS, B → Bµ =
√1/2 . . .
√2 kT
0 10 20 30 40 50 60 70-0.3
-0.2
-0.1
0
0.1
0.2
0.3Transverse momentum dependent forward-backward asymmetry
pT, tt [GeV]
AFB(p
T,t
t)
Importance ofNc = 3 colour coherence
(SHERPA’s S-MC@NLO)vs.Nc →∞ colour coherence
(SHERPA’s CSSHOWER++)
• small effect on standard (rapidityblind) observables, e.g. p⊥,tt→ some destructive interferenceat large p⊥,tt
• large effect on AFB(p⊥,tt)→ subleading colour termsincrease asym. radiation pattern
Marek Schonherr IPPP Durham
NLO merging in tt+jets 15
MEPS@NLO Results Conclusions
Results – pp→ tt+ jets – AFB
• Definition of forward-backward asymmetry of an observable O
AFB(O) =
dσttdO
∣∣∣∣∆y>0
− dσttdO
∣∣∣∣∆y<0
dσttdO
∣∣∣∣∆y>0
+dσttdO
∣∣∣∣∆y<0
• AFB is ratio of expectation values→ conventional scale variations by factor 2 will largely cancel for→ uncertainty on AFB
⇒ use different functional forms of the scale defintion that behave differentlyin ∆y > 0 and ∆y < 0 for a realistic estimate of uncertainty
Marek Schonherr IPPP Durham
NLO merging in tt+jets 16
MEPS@NLO Results Conclusions
Results – pp→ tt+ jets – AFB
Hoche, Huang, Luisoni, MS, Winter arXiv:1306.2703
Setup: pp→ tt+ jets• purely perturbative calculation
(no hadronisation, MPI, etc.)
• 0,1 jets @ NLOQcut = 7 GeV
• virtual MEs from GOSAM
• perturbative scale variationsµR/F ∈ [ 1
2 , 2]µdef
µQ ∈ [ 1√2,√
2]µcore
• variation of merging parameterQcut ∈ {5, 7, 10} GeV
Meps@Nlo µcore = mtt
Meps@Lo µcore = mtt
0 50 100 150 200 250 300
0.01
0.02
0.03
0.04
0.05
0.06Transverse momentum of the top quark
p⊥,t [GeV]d
σ/dp⊥,t[pb/GeV
]
• scale choices: αk+ns (µeff) = αk
s (µ)αs(t1) · · ·αs(tn)1) µcore = mtt
2) µcore = µQCD = 2 |pipj |i, j . . . Nc →∞ colour partners, chooses between s, t, u
⇒ different behaviour for forward/backward configurationsMarek Schonherr IPPP Durham
NLO merging in tt+jets 17
MEPS@NLO Results Conclusions
Results – pp→ tt+ jets – AFB
Hoche, Huang, Luisoni, MS, Winter arXiv:1306.2703
Setup: pp→ tt+ jets• purely perturbative calculation
(no hadronisation, MPI, etc.)
• 0,1 jets @ NLOQcut = 7 GeV
• virtual MEs from GOSAM
• perturbative scale variationsµR/F ∈ [ 1
2 , 2]µdef
µQ ∈ [ 1√2,√
2]µcore
• variation of merging parameterQcut ∈ {5, 7, 10} GeV
Meps@Nlo µcore = µQCD
Meps@Lo µcore = µQCD
0 50 100 150 200 250 300
0.01
0.02
0.03
0.04
0.05
0.06Transverse momentum of the top quark
p⊥,t [GeV]d
σ/dp⊥,t[pb/GeV
]
• scale choices: αk+ns (µeff) = αk
s (µ)αs(t1) · · ·αs(tn)1) µcore = mtt
2) µcore = µQCD = 2 |pipj |i, j . . . Nc →∞ colour partners, chooses between s, t, u
⇒ different behaviour for forward/backward configurationsMarek Schonherr IPPP Durham
NLO merging in tt+jets 18
MEPS@NLO Results Conclusions
Results – pp→ tt+ jets – AFB
Hoche, Huang, Luisoni, MS, Winter arXiv:1306.2703
Setup: pp→ tt+ jets• purely perturbative calculation
(no hadronisation, MPI, etc.)
• 0,1 jets @ NLOQcut = 7 GeV
• virtual MEs from GOSAM
• perturbative scale variationsµR/F ∈ [ 1
2 , 2]µdef
µQ ∈ [ 1√2,√
2]µcore
• variation of merging parameterQcut ∈ {5, 7, 10} GeV
Meps@Nlo µcore = µQCD
Meps@Nlo µcore = mtt
Meps@Lo µcore = µQCD
Meps@Lo µcore = mtt
0 50 100 150 200 250 300
0.01
0.02
0.03
0.04
0.05
0.06Transverse momentum of the top quark
p⊥,t [GeV]d
σ/dp⊥,t[pb/GeV
]
• scale choices: αk+ns (µeff) = αk
s (µ)αs(t1) · · ·αs(tn)1) µcore = mtt
2) µcore = µQCD = 2 |pipj |i, j . . . Nc →∞ colour partners, chooses between s, t, u
⇒ different behaviour for forward/backward configurationsMarek Schonherr IPPP Durham
NLO merging in tt+jets 19
MEPS@NLO Results Conclusions
Results – pp→ tt+ jets – AFB
Hoche, Huang, Luisoni, MS, Winter arXiv:1306.2703
CDF data Phys.Rev.D87(2013)092002
bb
bb
b
b
b
Reconstructed top
Sherpa+GoSam
b CDF dataPhys. Rev. D87 (2013) 092002
Meps@Nlo µcore = µQCDperturbative uncertaintyMeps@Nlo µcore = mttperturbative uncertainty
0 10 20 30 40 50 60 70-0.3
-0.2
-0.1
0
0.1
0.2
0.3Transverse momentum dependent forward-backward asymmetry
pT, tt [GeV]
AFB(p
T,t
t)
pp→ tt+jets (0,1 @ NLO)
• AFB(p⊥,tt) NLO accurate in allbut the first bin
• tops reconstructed from decayproducts (jets, lepton, MET)
• no EW corrections
Marek Schonherr IPPP Durham
NLO merging in tt+jets 20
MEPS@NLO Results Conclusions
Results – pp→ tt+ jets – AFB
Hoche, Huang, Luisoni, MS, Winter arXiv:1306.2703
CDF data Phys.Rev.D87(2013)092002
b
b
b
b
Parton level
Sherpa+GoSam
b CDF dataPhys. Rev. D87 (2013) 092002Meps@Nlo µcore = µQCDperturbative uncertaintyMeps@Nlo µcore = mttperturbative uncertainty
0 0.5 1 1.5 20
0.1
0.2
0.3
0.4
0.5
0.6
0.7Rapidity dependent forward-backward asymmetry
∆y, tt
AFB(∆
y,tt)
b
b
b
b
Parton level
Sherpa+GoSam
b CDF dataPhys. Rev. D87 (2013) 092002Meps@Nlo µcore = µQCDperturbative uncertaintyMeps@Nlo µcore = mttperturbative uncertainty
350 400 450 500 550 600 650 700 7500
0.1
0.2
0.3
0.4
0.5
0.6
0.7Mass dependent forward-backward asymmetry
mtt [GeV]
AFB(m
tt)
• parton level (exact top quarks)• no EW corrections (≈ 20%) effected• right qualitative bahviour, but consistently below data
Marek Schonherr IPPP Durham
NLO merging in tt+jets 21
MEPS@NLO Results Conclusions
Conclusions
• MEPS merging methods have evolved to NLO: MEPS@NLO
• taking colour correlations properly into account already produces anasymmetry in parton showers
• asymmetry increases when parton shower includes subleading colour terms• decreasing uncertainties LOPS/MEPS/NLOPS →MEPS@NLO
• uncertainties on asymmetry observables should be evaluated by choosingdifferent functional forms with different properties towards the asymmetry
• can be improved by adding higher order calculations• (N)NLL resummation• NNLO corrections Barnreuther, Czakon, Mitov Phys.Rev.Lett.109(2012)132001
Czakon, Mitov JHEP01(2013)080
Czakon, Fiedler, Mitov Phys.Rev.Lett.110(2013)252004
current release SHERPA-2.0.β2, when fully tuned SHERPA-2.0.0http://sherpa.hepforge.org
Marek Schonherr IPPP Durham
NLO merging in tt+jets 22
MEPS@NLO Results Conclusions
Thank you for your attention!
Marek Schonherr IPPP Durham
NLO merging in tt+jets 23
MEPS@NLO Results Conclusions
Results – pp→ tt+ jets
Hoche, Huang, Luisoni, MS, Winter arXiv:1306.2703
Sherpa+GoSam
Qcut = 5 GeVQcut = 7 GeVQcut = 10 GeV
10−2
10−1
1
10 1Transverse momentum of the top-quark pair
dσ
/d
log(
p T,t
t/G
eV)
[pb]
0 0.5 1 1.5 2 2.5
0.9
1.0
1.1
log(pT,tt/GeV)
Qcu
t/
7G
eV
Sherpa+GoSam
NLO µcore = µQCDNLO µcore = mttLO µcore = µQCDLO µcore = mtt
0 10 20 30 40 50 60 70-0.3
-0.2
-0.1
0
0.1
0.2
0.3Transverse momentum dependent forward-backward asymmetry
pT, tt [GeV]
AFB(p
T,t
t)
• very small Qcut dependence
• scale variation shrinks going LO to NLO (both factor and functional form)
Marek Schonherr IPPP Durham
NLO merging in tt+jets 24