Monte Carlo Developments Outline in the light of LHC data · Monte Carlo Developments in the light of LHC data Bryan Webber* University of Cambridge Outline Mostly newer developments

MC Developments & LHC Data POwLHC Workshop, KEK, Feb 2012

Monte Carlo Developmentsin the light of LHC data

Bryan Webber*

University of Cambridge

Outline

� Mostly newer developments.� New generators (Herwig++/Pythia8/Sherpa) (biased ;).� Emphasis on NLO and Underlying event.

Stefan Gieseke · Ringberg 2011 2/29

*with thanks to Stefan Gieseke


pp Event Generator

2

pp Event Generator



pp Event Generator

3

pp Event Generator



Hard Processes

4

Hard processes (tree level)

Processes at Born level (out of the box)� Hadron collider

QCD 2 → 2, tt, MinBias(γ,Z0)→ �+�−, W± → �±ν�, (Z0,W±)+ jetW+W−, W±Z0, Z0Z0, W±γ , Z0γh0, h0 +W±, h0 +Z0, h0 + jet, qqh0 (VBF), tth0

γ + jet, γγ� DIS

NC/CC/Photoproduction, γp → jets.� e+e−/γγ

e+e− → Z0, e+e− → qq, e+e− → �+�−, e+e− → W+W−,e+e− → Z0h0, e+e− → h0e+e−, e+e− → h0νeνe.γγ → W+W−, γγ → f f .

Also at NLO with POWHEG matching (more details later).



Hard Processes

5


Processes at Born level (out of the box)� Hadron collider

QCD 2 → 2, tt, MinBias(γ,Z0)→ �+�−, W± → �±ν�, (Z0,W±)+ jetW+W−, W±Z0, Z0Z0, W±γ , Z0γh0, h0 +W±, h0 +Z0, h0 + jet, qqh0 (VBF), tth0

γ + jet, γγ� DIS

NC/CC/Photoproduction, γp → jets.� e+e−/γγ

e+e− → Z0, e+e− → qq, e+e− → �+�−, e+e− → W+W−,e+e− → Z0h0, e+e− → h0e+e−, e+e− → h0νeνe.γγ → W+W−, γγ → f f .

Also at NLO with POWHEG matching (more details later).



Hard Processes

6


� BSM processes

� MSSM.

� Extra Dimensions.

� More under construction.

� Hard process and up to 3 body decays

created automatically from model file.

� Allows for simulation of full spin correlations.

� Anything else via LesHouchesFileReader.



Hard Processes

7


Example event, only MSSM hard process.

Full cascade decay chain w/ spin correlations

MSSM, UED, RS included in Herwig++ (since 2.1).

[Martyn Gigg, Peter Richardson, EPJC 51 (2007) 989]

Finite width effects and 3 body decays (since 2.3)

[M.A. Gigg, P. Richardson, arXiv:0805.3037]

All automatically built.

Inclusive or exclusive process specification.



Hard Processes

8


Sherpa:All tree level processes via AMEGIC++, COMIX, built–in MEgenerators. New models via FeynRules.Pythia:Many processes built–in. Pythia 8.1 can link back to Pythia 6.4processes. Rest via LHEF.



Parton Showers

9

Parton showers

Herwig++

� New parton shower variables introduced for Herwig++

[SG, P. Stephens and B. Webber, JHEP 0312 (2003) 045 [hep-ph/0310083]]

� More under development → dipole shower, based upon

Catani–Seymour dipoles.

[SG, S. Platzer, 0909.5593]]

Sherpa

� CS-Shower default by now, always matched via CKKW

(see later).

Pythia

� pT ordered shower (simple matching).

� Interleaved with Multiple partonic interactions.


[S. Gieseke,

[S. Gieseke, P. Stephens, BW, JHEP 0312 (2003) [hep-ph/0310083]]


Inclusive Jet Rates

10

[GeV]T

p100 200 300 400 500

dy

[pb/

GeV

]T

/dp

!2 d

1

10

210

310

410

510

610

ATLASHerwig++Pythia 6Pythia 8Sherpa

7000 GeV pp Jets

mcp

lots

.cer

n.ch

Herwig++ 2.5.2, Pythia 6.426, Pythia 8.157, Sherpa 1.3.1

ATLAS_2010_S8817804

Jet Transverse Momentum ()

100 200 300 400 5000.5

1

1.5Ratio to ATLAS

[GeV]T

p100 200 300 400

dy

[pb/

GeV

]T

/dp

!2 d1

10

210

310

410

510 ATLASHerwig++Pythia 6Pythia 8Sherpa

7000 GeV pp Jets

mcp

lots

.cer

n.ch


ATLAS_2010_S8817804


100 200 300 4000.5

1

1.5Ratio to ATLAS

[GeV]T

p100 200 300 400 500

dy

[pb/

GeV

]T

/dp

!2 d

1

10

210

310

410

510

610


7000 GeV pp Jets

mcp

lots

.cer

n.ch


ATLAS_2010_S8817804


100 200 300 400 5000.5

1

1.5Ratio to ATLAS

[GeV]T

p100 200 300 400 500

dy

[pb/

GeV

]T

/dp

!2 d

-110

1

10

210

310

410

510

610


7000 GeV pp Jets

mcp

lots

.cer

n.ch


ATLAS_2010_S8817804


100 200 300 400 5000.5

1

1.5Ratio to ATLAS

[GeV]T

p100 200 300 400

dy

[pb/

GeV

]T

/dp

!2 d

-110

1

10

210

310

410

510ATLASHerwig++Pythia 6Pythia 8Sherpa

7000 GeV pp Jets

mcp

lots

.cer

n.ch


ATLAS_2010_S8817804


100 200 300 4000.5

1

1.5Ratio to ATLAS

|y| < 0.3 0.3 < |y| < 0.8 0.8 < |y| < 1.2

1.2 < |y| < 2.1 2.1 < |y| < 2.8

Outline



Outline



anti-kt, R=0.4mcplots.cern.ch

Outline



ATLAS data


Transverse Thrust

11

Transverse thrust

)C

ln(1-T-10 -5

) C1/

N

dN

/d ln

(1-T

-210

-110

CMSHerwig++Perugia 2010Pythia 8

7000 GeV pp Jets

mcp

lots

.cer

n.ch

Herwig++ 2.5.0, Pythia 6.424, Pythia 8.145

CMS_2011_S8957746

Central Transverse Thrust (cms1-pt090)

-10 -50.6

0.8

1

1.2

1.4

Ratio to CMS

)C

ln(1-T-10 -5

) C1/

N

dN

/d ln

(1-T

-210

-110

CMSHerwig++Perugia 2010Pythia 8

7000 GeV pp Jets

mcp

lots

.cer

n.ch

Herwig++ 2.5.0, Pythia 6.424, Pythia 8.145

CMS_2011_S8957746

Central Transverse Thrust (cms1-pt125)

-10 -50.6

0.8

1

1.2

1.4

Ratio to CMS



Integral Jet Shapes

12

Integral jet shapes

not too hard, central (30 < pT/GeV < 40;0 < |y|< 0.3)



Integral Jet Shapes

13

Integral jet shapes

harder, more forward (80 < pT/GeV < 110;1.2 < |y|< 2.1)



Integral Jet Shapes

14

Integral jet shapes

Jet substructure in CMS. t and W tagging from jet substructure.

(GeV/c)T

Jet p300 400 500 600 700 800 900 1000 1100 1200

Top

Mis

tag

Rat

e

0

0.02

0.04

0.06

0.08

0.1

0.12 DataPythia 6 Tune Z2, Stat ErrorPythia 6 Tune D6T, Stat ErrorPythia 8 Tune 1, Stat ErrorHerwig++ Tune 23, Stat Error

Top Tagging AlgorithmCMS Preliminary

= 7 TeVs at -135.97 pb

(GeV/c)jetT

p200 300 400 500 600 700 800 900 1000

W M

ista

g R

ate

0

0.05

0.1

0.15

0.2

0.25

0.3Data

Pythia 6 Tune Z2, Stat Error

Pythia 6 Tune D6T, Stat Error

Herwig++ Tune 23, Stat Error

= 7TeVs at -134.7 pbCMS PreliminaryJet Pruning Algorithm

[CMS PAS JME-10-013]



Forward + Central Jets

15

12 8 Summary

(GeV/c)Tcentral jet p40 60 80 100 120 140

pb/(G

eV/c)

c !d T/d

p"2 d

1

10

210

310

410

510DataPYTHIA 6 (D6T)PYTHIA 6 (Z2)PYTHIA 8 (Tune 1)POWHEG (+PYTHIA 6)CASCADE

| < 2.8!|

-1 = 3.14 pbint=7 TeV, Ls+ X, cent

+ jetfwd

jet#CMS, pp

(a)

(GeV/c)Tforward jet p40 60 80 100 120 140

pb/(G

eV/c)

f !d T/d

p"2 d

1

10

210

310

410

510DataPYTHIA 6 (D6T)PYTHIA 6 (Z2)PYTHIA 8 (Tune 1)POWHEG (+PYTHIA 6)CASCADE

| < 4.7! 3.2 < |


+ jetfwd

jet#CMS, pp

(b)

(GeV/c)Tcentral jet p40 60 80 100 120 140

pb/(G

eV/c)

c !d T/d

p"2 d

1

10

210

310

410

510DataHERWIG 6 (+JIMMY) HERWIG++POWHEG (+HERWIG)HEJ

| < 2.8!|


+ jetfwd

jet#CMS, pp

(c)

(GeV/c)Tforward jet p40 60 80 100 120 140

pb/(G

eV/c)

f !d T/d

p"2 d

1

10

210

310

410

510DataHERWIG 6 (+JIMMY) HERWIG++POWHEG (+HERWIG)HEJ

| < 4.7! 3.2 < |


+ jetfwd

jet#CMS, pp

(d)

Figure 7: Differential cross sections as a function of jet pT for dijet events with at least onecentral jet ((a) and (c)) and one forward jet ((b) and (d)), compared to predictions from severalmodels. The error bars on all data points (which, in (a) and (c), are smaller than the size ofthe markers) reflect just statistical uncertainties, with systematic uncertainties plotted as greybands.

Outline



CMS data

anti-kt, R=0.5

Outline



[CMS, 1202.0704]

Outline



HEJ: Andersen & Smillie, JHEP 01 (2010) 039; 06 (2011) 010

CASCADE: Jung & Salam, EPJC 19

(2001) 351

Outline




Matrix Element Corrections

16

Matrix element correctionsHard ME correction (e.g. in DY)

� Light: collinear/soft

regions.

� Dark: Dead region, filled

with extra hard emissions

— not accessable by parton

shower.

� To be complemented by

soft matrix element

corrections.

Also for V∗ → qq, t–decay (2.0)

gg → h0(2.2),

-2

-1.5

-1

-0.5

0

1 1.5 2 2.5 3t/M

2

s/M2

Simplest matching.



Merging ME with Parton Shower

17

Matching tree level ME and parton showers

� Problem: have multiple tree level MEs for X+0,1, . . .n jets.

� Jets well separated and inclusive.

� Merge this into one exclusive multijet sample.

� Idea: use Sudakov form factors to

disallow “+ anything softer”

(which is normally inside an inclusive ME).

� That’s done in the CKKW(-L) approach. Catani, Krauss, Kuhn, Webber,

JHEP 0111:063,2001, Krauss JHEP 0208:015,2002, L. Lonnblad, JHEP 0205:046,2002, Gleisberg, Hoche,

Winter, Schalicke, Schumann.

� Alternative: MLM matching. M.L. Mangano

� Systematic study and comparison of implementations.

J. Alwall, S. Hoche, F. Krauss, N. Lavesson, L. L”onnblad, F. Maltoni, M.L. Mangano, M. Moretti,

C.G. Papadopoulos, F. Piccinini, S. Schumann, M. Treccani, J. Winter, M. Worek, EPJC53:473-500,2008.


Catani, Krauss, Kuhn, BW,



18


� Separates ME and parton shower at intermediate scale Qini.� Parton shower fills region below Qini.� All emissions resolvable above Q0.

Merges ME and parton shower at scale Qini.




19

Matching tree level ME and PS — trouble?

Hard emission, to be complemented by parton shower.

p⊥ ordered shower. Angular

ordering from additional ve-

tos.

Angular ordered shower.

Some softer emissions

before hardest one.

Potential holes in phase space −→ truncated showers.

S. Hoche, F. Krauss, S. Schumann, F. Siegert, JHEP 0905:053,2009.

K. Hamilton, P. Richardson, J. Tully, JHEP 0911:038,2009.




20



Parton level merging for illustration.

Instabilities at Qini removed.




21



Hadron level with matching uncertainty band vs OPAL.Stefan Gieseke · Ringberg 2011 12/29



22


Sherpa CS shower, matched with Z0 +Njet jets vs CDF data.

Nmax = 0Nmax = 1Nmax = 2Nmax = 3data

10 1

10 2

10 3

10 4

σ(N

jet)

(sca

led

tofirs

tbin

)

1 2 3

0.6

0.8

1

1.2

1.4

Njet

MC

/dat

a

S. Hoche, F. Krauss, S. Schumann, F. Siegert, JHEP 0905:053,2009.

Reached remarkable stability wrt Qini variation.



Matching NLO with PS

23

Matching MC and NLO

�O�MC@NLO =O(0)

�B+ V+

�1

0

dxP(x)−A(x)

x

�

+�

dxO(x)R(x)−P(x)

x.

Observations/remarks:

� Events with n and n+1 legs are seperately finite. No

cancellation of large weights.

� NLO result can be recovered strictly upon expansion in

powers of α (with parton shower emission).

� Interface to MC program very well defined.

� Dropping µ → 0 is only a power correction.




24

Matching MC and NLO

�O�MC@NLO =O(0)

�B+ V+

�1

0

dxP(x)−A(x)

x

�

+�

dxO(x)R(x)−P(x)

x.

Three types of matching

1. MC@NLO (classic, Frixione and Webber).

2. Simpler: parton shower with P(x) = A(x).

3. Or, also simpler, P(x) = R(x).




25

Matching MC and NLO

�O�MC@NLO =O(0)�

B+ V+� 1

0dx

P(x)−A(x)

x

�

+�

dxO(x)R(x)−P(x)

x.

1. Classic MC@NLO (Frixione and Webber)� A(x) = FKS subtraction terms� P(x) and phase space specific for HERWIG.� Generic, calculate once and for all.

(Usually, A(x) and P(x) factorize off B.)� New for every process.




26

Matching MC and NLO

�O�MC@NLO =O(0)

�B+ V+

�1

0

dxP(x)−A(x)

x

�

+�

dxO(x)R(x)−P(x)

x.

2. ‘Custom’ parton shower

e.g. with Catani–Seymour subtraction kernels

� CS subtraction already used in many NLO calculations.

� P(x) = A(x), so terms vanish.

� R(x)−A(x) already in NLO parton level program.

⇒ (almost) no need to modify NLO calculation!




27

Matching MC and NLO

�O�MC@NLO =O(0)

�B+ V+

�1

0

dxP(x)−A(x)

x

�

+�

dxO(x)R(x)−P(x)

x.

3. Simpler in a different way, P(x) = R(x)

� R(x)−A(x) now only needed as integral

available in NLO parton level program.

� No n+1 body events.

� ≥ 1 PS emission from R(x) as splitting kernel →POWHEG.

� Positive weights (terms �= 0 are σ incl

NLO).

� Further emissions from (truncated) standard PS.



MC@NLO

28

MC@NLO

� Introduced 2002 Frixione, Webber, JHEP 0206:029,2002 [hep-ph/0204244].

� Extended to heavy quarksFrixione, Nason, Webber, JHEP 0308:007,2003 [hep-ph/0305252].

� further extensions to many processes (single top etc.)� MC@NLO customised to use with HERWIG.� Some processes in Herwig++ as well

e+e− → jets, DY, W�, h0 decayLatunde–Dada 0708.4390, 0903.4135, Latunde–Dada, Papaefstatiou, 0901.3685.

� MC@NLO package adopted to Herwig++ as well.S. Frixione, F. Stoeckli P. Torrielli and B.R. Webber, 1010.0568.


[Frixione, BW, JHEP 06 (2002) 029]

[Frixione, Nason, BW, JHEP 08 (2003) 007]

Frixione, Stoekli, Torrielli, BW, JHEP 1101:053,2011 [1010.0568]


MC@NLO with Herwig

29

MC@NLO

Examples with Herwig++ (solid) Herwig6 (dash)

h0 → WW → lν lν , tt, pt(b) rel to t (right).

(no spin corr dotted)

S. Frixione, F. Stoeckli P. Torrielli and B.R. Webber, 1010.0568.


Frixione, Stoekli, Torrielli, BW, JHEP 1101:053,2011 [1010.0568]


MC@NLO with Herwig

30

) [nb]-l+ l!*" BR(Z/# Zfid$

0.4 0.45 0.5 0.55

) [nb

]% l

!±

BR(

W#

±Wfid$

4

4.5

5

5.5

= 7 TeV)sData 2010 (MSTW08HERAPDF1.5ABKM09JR09

total uncertainty sys&sta

uncertainty

68.3% CL ellipse area

-1 L dt = 33-36 pb'

ATLAS

) [nb]-l+ l!*" BR(Z/# Zfid$

0.4 0.45 0.5 0.55

) [nb

]% l

!±

BR(

W#

±Wfid$

4

4.5

5

5.5

Figure 6: Inclusive W and Z boson production: Cross sections extrapolated to full phasespace and compared to NNLO theory predictions [76] (left); Measured and predicted fidu-cial cross sections times leptonic branching ratios [77], compared to predictions obtainedusing a variety of pdf sets (right).

( ) (

0 0.5 1 1.5 2 2.5 3 3.5 4

Lept

on c

harg

e as

ymm

etry

-0.3

-0.2

-0.1

0

0.1

0.2

0.3

-1) 35 pb% l!ATLAS (extrapolated data, W-1) 36 pb%µ !CMS (W-1) 36 pb%µ !LHCb (W

MSTW08 prediction (MC@NLO, 90% C.L.)CTEQ66 prediction (MC@NLO, 90% C.L.)HERA1.0 prediction (MC@NLO, 90% C.L.)

ATLAS+CMS+LHCbPreliminary

=7 TeVs

> 20 GeVlTp

Figure 7: A comparison of ATLAS, CMS and LHCb results on the lepton charge asymmetrywith theoretical predictions based on several different pdf sets [75].

12

Outline



Lepton charge asymmetry in W → µν

ATLAS-CONF-2011-129


POWHEG

31

POWHEG

� Alternative proposed by P. Nason.

� Modified Sudakov FF for first emission.

� Angular ordered Parton Shower tricky (see below).

� Truncated Shower adds in missing radiation afterwards.

� Finally evolution with ‘ordinary’ Parton Shower.

[Nason, hep-ph/0409146; Nason, Ridolfi hep-ph/0606275]

Recently systematically extended.

� POWHEG formulation independent of the event generator

implementation.

� Worked out for different subtraction schemes.

[Frixione, Nason, Ridolfi, 0707.3081, 0707.3088; Frixione, Nason, Oleari, 0709.2092]



POWHEG

32

POWHEG

Angular ordered showers and POWHEG

p⊥ ordered shower. Angularordering from additional ve-tos.

Angular ordered shower.Some softer emissionsbefore hardest one.

Need truncated showers.



POWHEG in Herwig++

33

POWHEG in Herwig++

� First implementation of method for e+e− annihilation[O. Latunde–Dada, SG, B. Webber, hep-ph/0612281]

� Many more processes now available with release:DY (γ∗/Z0/W±),h0,h0Z0,h0W±,W+W−,W±Z0,Z0Z0

[K. Hamilton, P. Richardson and J. Tully, 0806.0290, 0903.4345, Hamilton, JHEP 1101:009]

� Finished, out with next release: VBF, γγ[D’Errico, Richardson, 1106.2983, 1106.3939]

� and with contributed code:e+e− → jets, tt, t−decay,W�,h0 −decay

[ O. Latunde-Dada, 0812.3297, Eur. Phys. J. C 58, 543 (2008)]

[A. Papaefstathiou and O. Latunde-Dada, JHEP 0907, 044]

� includes full truncated showers.� Interface to PowhegBox straightforward.� More processes underway (SUSY pair prod. . . ).


[O. Latunde-Dada, S. Gieseke, BW, JHEP 0702, 051]


POWHEG in Herwig++

34

POWHEG in Herwig++

POWHEG in Herwig++ with full truncated shower.

[K. Hamilton, JHEP 1101:009]

VV production. Phase space of radiated gluon properly filled.



POWHEG in Herwig++

35

POWHEG in Herwig++

Higgs production in VBF. (POWHEG, MEC, LO+PS)

[L. D’Errico, P. Richardson 1106.2983]



POWHEG in Herwig++

36

POWHEG in Herwig++18 6 Results

2.5GeV2<Q2<5GeV2 5GeV2<Q2<10GeV2 10GeV2<Q2<20GeV2 20GeV2<Q2<50GeV2 50GeV2<Q2<100GeV2

5×10−5<x<10−4

10−4 <x<2×10−4

2×10−4 <x<3.5×10−4

3.5×10−4 <x<10−3

10−4 <x<2×10−4

2×10−4 <x<3.5×10−4

3.5×10−4 <x<7×10−4

7×10−4<x<2×10−3

2×10−4 <x<5×10−4

5×10−4 <x<8×10−4

8×10−4<x<1.5×10−3

1.5×10−3 <x<4×10−2

5×10−4 <x<1.4×10−3

1.4×10−3 <x<3×10−3

3×10−3 <x<×10−2

8×10−4<x<3×10−3

3×10−3 <x<2 ×10−2

1 NdE

∗ T/dη∗/G

eV

η∗0 5 0 5

0 5

0 5

0 5

0

2

0

2

0

2

0

2

0.2

−0.2

0.2

−0.2

0.2

−0.2

0.2

−0.2

0

2

0.2

−0.2

Hw

Hw++

Hw++, ME correction

Hw++, POWHEG

Fig. 6: The inclusive transverse energy flow 1N dE∗

T/dη∗ at different values of x and

Q2 for the low Q2 sample from [102]. The lower frame shows (Data −Theory)/Data and the yellow band gives the one sigma variation.

DIS transverse Energy

vs H1 Data

[H1, EPJC 12 (2000) 595]




POWHEG in Herwig++

37

POWHEG in Herwig++6.1 Deep Inelastic Scattering 19

100GeV2<Q2<400GeV2 400GeV2<Q2<1100GeV2 1100GeV2<Q2<100000GeV2

2.51×10−3<x<6.31×10−3

6.31×10−3<x<1.58×10−2

1.58×10−2<x<3.98×10−2

6.31×10−3<x<1.58×10−2

1.58×10−2<x<3.98×10−2

3.98×10−2<x

1 NdE

∗ T/dη∗/G

eV

η∗

0 5

0 5

0 5

0

2

4

0.2

−0.2

0

2

4

0.2

−0.2

0

2

4

0.2

−0.2

0

2

4

0.2

−0.2

Hw

Hw++

Hw++, ME correction

Hw++, POWHEG

Fig. 7: The inclusive transverse energy flow 1N dE∗

T /dη∗ at different values of x and

Q2 for the high Q2 sample from [102]. The lower frame shows (Data −Theory)/Data and the yellow band gives the one sigma variation.

DIS transverse Energy

vs H1 Data

[H1, EPJC 12 (2000) 595]




POWHEG in Sherpa

38

NLO in Sherpa

� Automated POWHEG matching approach.� Only virtuals needed → Binoth Les Houches Accord.

POWHEGME+PS (1-jet) × 2.1LO+PS × 2.1

10−5

10−4

10−3

10−2

Higgs p⊥

dσ/dph ⊥

1 10 1 10 2 10 3

0.6

0.8

1

1.2

1.4

ph⊥[GeV]

Ratio

POWHEGME+PS (1-jet) × 1.17

LO+PS × 1.34

10−10

10−9

10−8

10−7

10−6

10−5

10−4

10−3

10−2

Transverse momentum of leading jet

dσ/dp⊥(jet

1)[pb/GeV

]

10 2 10 3

0.6

0.8

1

1.2

1.4

p⊥(jet 1) [GeV]

Ratio

[Hoeche, Krauss, Schonherr, Siegert, JHEP 1104:024]

gg → h0 (left) WW+jets (right)



POWHEG + Pythia

39

NLO in Pythia

� pT-ordered shower → no truncated showers for POWHEG.� Variants in phase space coverage more versatile.

Interpolation between 1/p2t and 1/p4

t .

10-4

10-3

10-2

10-1

0 20 40 60 80 100

dP

/ d

p⊥ [G

eV

-1 ]

p⊥ [GeV]

(a)

POWHEGPythia Default (Power)

Pythia Damp, k = 2Pythia Damp, k = 1

Pythia Wimpy

10-7

10-6

10-5

10-4

10-3

10-2

100 200 300 400 500 600 700 800 900 1000

dP

/ d

p⊥ [G

eV

-1 ]

p⊥ [GeV]

(b)

POWHEGPythia Default (Power)

Pythia Damp, k = 2Pythia Damp, k = 1

Pythia Wimpy

First emission off tt pair. [R. Corke, T. Sjostrand, EPJC69:1 (2010)]



Inclusive b-jets

40

Measurements of pairs of such b-tagged jets have been made [38] in the kinematic regionpT > 40 GeV and |y| < 2.1, with in addition the b-jets well separated in the azimuthalplane (Fig. 3 left). In this case the gluon splitting g → bb is expected to make a rather smallcontribution. Note that these measurements have defined a b-jet as a jet which is matchedin angle to one or more b-hadrons. NLO QCD calculations describe the data well overthe measured range. Furthermore, measurements of the inclusive b-jet cross section havebeen made over the range 20 < pT < 400 GeV and |y| < 2.1 and show some discrepanciesat higher rapidity y, as well as some divergence between different NLO+PS calculations(POWHEG/MC@NLO), by up to 30%.

An innovative study of the angular correlations between b-quarks [39] shows that a rangeof perturbative calculations fail to describe the angular distribution of b-hadron pairs injet events (Fig. 3 right). Specifically, when normalized to the rate at wide angles, up to50% divergence between data and theory, and between different approximations of QCD,is seen at small angles.

The measurements to date suggest that a better understanding of the g → bb vertexmay well be required in order to accurately and correctly describe b-jet production overthe kinematic range accessible at the LHC.

Measurements of jets containing charm have also been made [40], using D∗-mesondecays as a tag, but these are much more sensitive to soft physics due to the lower c mass.Significant discrepancies are seen between data and MC simulations at low z, where z isthe fraction of the jet momentum carried by the D∗.

Figure 3: Left: Inclusive double-differential b-jet cross-section as a function of pT fordifferent rapidity ranges, compared to the predictions of several Monte Carlo models [38];Right: Differential BB production cross section as a function of the angular separation∆R = [(yB − yB)

2 + (φB − φB)2]

1

2 (the Monte Carlo prediction is normalized to the region∆R > 2.4 (shaded)) [39].

7

ATLAS, EPJC 71 (2011) 1846 [1109.6833]


ME+NLO+PS

41

MENLOPS

ME+PS merging with lowest multiplicity at NLO.

Test generic method with Pythia. ynm in tt+jets[Hamilton, Nason, JHEP 1006:039]



ME+NLO+PS

42

MENLOPS

ME+PS merging with lowest multiplicity at NLO.

MENLOPS (3-jet)ME+PS (3-jet) × 1.04POWHEG

10−7

10−6

10−5

10−4

10−3

HT

dσ/dH

T[pb/GeV

]

10 2 10 3

0.6

0.8

1

1.2

1.4

HT [GeV]

Ratio

MENLOPS (3-jet)ME+PS (3-jet) × 1.04POWHEG

0

0.02

0.04

0.06

0.08

0.1

Azimuthal decorrelation of leading and second leading jet

dσ/d

∆φ(jet

1,jet

2)[pb]

0 0.5 1 1.5 2 2.5 3

0.6

0.8

1

1.2

1.4

∆φ(jet 1, jet 2)

Ratio

WW+jets implementation in Sherpa.[Hoeche, Krauss, Schonherr, Siegert, 1009.1127]



Hadron Decays

43

Decays

� Herwig++ decays quite sophisticated

� Specialized decayers for majority of channels.

Mesons and baryons.

� Up to 5-body decays.

� Spin correlations.

� Running widths.

� Photon radiation off charged hadrons.

∼ 500 particles and ∼ 6500 decay modes included.



Hadron Decays

44

Decays

� Herwig++ decays quite sophisticated� Specialized decayers for majority of channels.

Mesons and baryons.� Up to 5-body decays.� Spin correlations.� Running widths.� Photon radiation off charged hadrons.

� Sherpa� Similar level of sophistication.� - less decay modes.� + Mixing in B decays with interference.

� Pythia� No spin correlations.� Relies on EvtGen.� Photons from PHOTOS.� τ decays from TAUOLA.



Min-bias/Underlying Event

45

Min Bias/Underlying Event

Herwig++

MPI model with independent hard processes, showers and

colour reconnection. Min bias without integrated diffraction.

Pythia

MPI interleaved with showering. Many tune families.

Sherpa

MPI model with independent hard processes. New model with

integrated diffraction under development.



Colour Reconnection

46

Colour reconnection at hadron colliders

� Colour preconfinement� Shorten colour string/lower mass clusters.



Colour reconnection at hadron colliders

� Colour preconfinement� Shorten colour string/lower mass clusters.


Colour Reconnection


CR in Herwig++

48

Colour reconnection (CR) in Herwig++

i

j

k

l

Extend cluster hadronization:� QCD parton showers provide

pre-confinement ⇒colour-anticolour pairs

� → clusters� CR in the cluster hadronization

model: allow reformation ofclusters, e.g. (il)+(jk)

� Allow CR if the cluster mass decreases,

Mil +Mkj < Mij +Mkl,

� Accept alternative clustering with probability preco (modelparameter) ⇒ this allows to switch on CR smoothly

Stefan Gieseke · Ringberg 2011 25/29 S Gieseke, C Rohr, A Siodmok, arXiv:1110.2675


CR in Herwig++

49




� → clusters

� CR in the cluster hadronizationmodel: allow reformation ofclusters, e.g. (il)+(jk)






CR in Herwig++

50











CR in Herwig++

51











MB/UE in Herwig++

52

Min Bias/Underlying Event in Herwig++

New colour reconnection model vital.

Read off from ATLASHerwig++ 2.4Herwig++ 2.5

1.6

1.8

2

2.2

2.4

2.6

2.8

3

3.2Charged particle multiplicity as function of η (0.9 TeV,Nch ≥ 6)

1/NevdNch/d

η

-2 -1 0 1 2

0.6

0.8

1

1.2

1.4

η

MC/data

Read off from ATLASHerwig++ 2.4Herwig++ 2.5

10−4

10−3

10−2

10−1

Charged particle density (0.9 TeV,Nch ≥ 6)

1/NevdNev/dNch

10 20 30 40 50 600.60.8

11.21.41.61.8

2

Nch

MC/data



MB/UE in Herwig++

53


Tunes with only one energy dependent parameter possible√s dependence usually tricky.

LHC 900 GeV

ATLAS dataHw++ 2.5.0MU900-1Hw++ 2.5.0 UE-EE-1

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7Transverse ∑ p⊥ density vs. ptrk1⊥ ,

√s = 900 GeV

�d2

∑p⊥/d

ηd

φ�[GeV]

1 2 3 4 5 6 7 8 9 10

0.6

0.8

1

1.2

1.4

p⊥ (leading track) [GeV]

MC/data


0

0.2

0.4

0.6

0.8

1

Transverse �p⊥� vs. Nchg,√s = 900 GeV

�p⊥�[GeV]

2 4 6 8 10 12 14 16 18 20

0.6

0.8

1

1.2

1.4

Nchg

MC/data


0

0.2

0.4

0.6

0.8

1

1.2

1.4

p⊥ density vs. ∆φ, ptrk1⊥ > 2.5 GeV,√s = 900 GeV

�d2p⊥/d

ηd

φ�

0 0.5 1 1.5 2 2.5 3

0.6

0.8

1

1.2

1.4

|φ| (w.r.t. leading track) [rad]

MC/data



MB/UE in Herwig++

54


Tunes with only one energy dependent parameter possible√s dependence usually tricky.

LHC 7 TeV

ATLAS dataHw++ 2.5.0 UE7-1Hw++ 2.5.0 UE-EE-1

0

0.2

0.4

0.6

0.8

1

1.2

1.4Transverse ∑ p⊥ density vs. ptrk1⊥ ,

√s = 7 TeV

�d2

∑p⊥/d

ηd

φ�[GeV]

2 4 6 8 10 12 14 16 18 20

0.6

0.8

1

1.2

1.4

p⊥ (leading track) [GeV]

MC/data


0

0.2

0.4

0.6

0.8

1

1.2

Transverse �p⊥� vs. Nchg,√s = 7 TeV

�p⊥�[GeV]

5 10 15 20 25 30

0.6

0.8

1

1.2

1.4

Nchg

MC/data


0

0.2

0.4

0.6

0.8

1

1.2

Nchg density vs. ∆φ, ptrk1⊥ > 3.0 GeV,√s = 7 TeV

�d2Nchg/d

ηd

φ�

0 0.5 1 1.5 2 2.5 3

0.6

0.8

1

1.2

1.4

|φ| (w.r.t. leading track) [rad]

MC/data



MB/UE in Pythia

55

Fig. 5 (left) shows the corrected stable particle scalar�

pT density at√s = 7 TeV in the

transverse region as a function of the pT of the leading particle (pleadT ) from the cluster-based

measurement. The summed particle pT in the plateau characterizes the mean contribution of

the Underlying Event to jet energies. The higher number density implies a higher pT density as

well. Most of the MC tunes considered show 10-15% lower�

pT than the data in the plateau

part of the transverse region.

Figure 5: (left) Corrected scalar�

pT density of stable particles (pT > 500 MeV, |η| < 2.5)at

√s = 7 TeV in the transverse region as a function of the pT of the leading particle (pleadT )

from the cluster-based Underlying Event measurement by ATLAS [12]. (right) Corrected mean

pT of charged particles at√s = 7 TeV for pleadT > 1 GeV as a function of the charged-particle

multiplicity in the transverse region from the track-based Underlying Event measurement by

ATLAS [11]. For both plots, the error bars show the statistical uncertainty while the shaded

areas show the combined statistical and systematic uncertainty.

Fig. 5 (right) shows the corrected mean pT of charged particles �pT � at√s = 7 TeV versus

the charged-particle multiplicity nch in the transverse region from the track-based UE measure-

ment. The correlation between �pT � and nch in each region is sensitive to the amount of hard

versus soft processes contributing to the UE. Although not shown here, the profile in the away

region is very similar to that of the transverse region, showing a monotonic increase of �pT �with nch. The models tend to overestimate �pT � in both the transverse and toward regions.

7 Conclusions

Data from the LHC provide a new energy scale for studying soft QCD. Charged-particle multi-

plicities have been measured by ATLAS in various regions of phase space, helping to disentangle

the contribution coming from diffractive processes. The results of these measurements indicate

a deficit of activity in models that were previously tuned to data from the Tevatron. Activ-

ity coming from the Underlying Event has been measured by ATLAS using track-based and

cluster-based methods, providing statistically independent results. The activity measured in

data is generally above the predictions from current model tunes.

8 MPI@LHC 2011

Pythia 6 tunes have difficulty fitting UE

Outline




MB/UE in Pythia

56

Min Bias/Underlying Event in PythiaNew development: x dependent matter distribution, i.e.

Gaussian with x dependent width a(x) = a0(1+ a1 log1/x).→ harder scattering more central. [Corke, Sjostrand, JHEP 1105:009]

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

0 0.5 1 1.5 2 2.5

(1 /

N)

dN

/ d

bM

PI

no

rm

bMPInorm

(a)

SGDYZ0

Z’



MB/UE in Pythia

57

Min Bias/Underlying Event in Pythia

New development: x dependent matter distribution, i.e.


0

0.02

0.04

0.06

0.08

0.1

0 5 10 15 20 25

(1 /

N)

dN

/ d

NM

PI

NMPI

(a) DY

SGLog

0

0.02

0.04

0.06

0.08

0.1

0 5 10 15 20 25

(1 /

N)

dN

/ d

NM

PI

NMPI

(b) Z0

SGLog

0

0.02

0.04

0.06

0.08

0.1

0.12

0 5 10 15 20 25

(1 /

N)

dN

/ d

NM

PI

NMPI

(c) Z’

SGLog

Results in �NMPI� dependent on hardness.



MB/UE in Pythia

58

1.0

1.5

2.0

2.5

3.0

3.5

4.0

-2 -1 0 1 2

<d2 !

p T /

d"d#

> [

GeV

]

pTlead [GeV]

ATLAS 7 TeVLog 7 TeV

Overlap 7 TeVATLAS 900 GeV

Log 900 GeVOverlap 900 GeV

10-6

10-4

10-2

100

102

104

0 10 20 30 40 50 60 70 80 90 100

1/N e

v dN e

v / d

N ch

Nch

ATLAS 7 TeV (x 100)Log 7 TeV (x 100)

Overlap 7 TeV (x 100)ATLAS 900 GeV


0.0

0.5

1.0

1.5

2.0

2.5

0 2 4 6 8 10 12 14 16 18 20

<d2 N c

h / d"

d#>

pTlead [GeV]

Towardregion




0.0

1.0

2.0

3.0

4.0

5.0

0 2 4 6 8 10 12 14 16 18 20

<d2 !

p T /

d"d#

> [

GeV

]

pTlead [GeV]

Towardregion




0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

0 2 4 6 8 10 12 14 16 18 20

<d2 N c

h / d"

d#>

pTlead [GeV]

Transverseregion




0.0

0.5

1.0

1.5

2.0

0 2 4 6 8 10 12 14 16 18 20

<d2 !

p T /

d"d#

> [

GeV

]

pTlead [GeV]

Transverseregion




Figure 11: Tune 4C, using the log profile, and with a raised p⊥0 in the MPI framework,compared against an overlap profile with p = 1.6, also with a raised p⊥0, and LHC data

affect the results shown here. Just this change leads to a rise in the tail of the chargedmultiplicity distributions, with an increase in activity in all regions of the underlying event,as expected from the considerations of the previous sections. This behaviour is most closelymatched by an overlap function with p = 1.6, against which we can compare the results.The simplest way to remove this excess activity is a retuning of the p⊥0 parameter of theMPI framework, in this case achieved by raising pref⊥0 = 2.085 → 2.15GeV. This rise doesnot greatly affect the relative slope of a0, as constrained in Sec. 3.1. The results are shownin Fig. 11 for the same distributions as Fig. 10.

After this retuning, the log profile shows some promise. For the charged multiplicity

19

Log = Pythia 8 with x-dependent profile

Outline




MB/UE in Pythia

57

Min Bias/Underlying Event in PythiaNew development: x dependent matter distribution, i.e.


0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

0 0.5 1 1.5 2 2.5

(1 /

N)

dN

/ d

bM

PI

norm

bMPInorm

(a)

SGDYZ0

Z’


Outline



Better fits to LHC and CDF UE data

η


Conclusions

59

• Many new developments in shower MCs

• Parton showers well established

• NLO available for very many processes

• MENLOPS matching with many (LO) legs becoming available

• LHC data have led to developments in MB/UE simulation. Good tunes available now

• First round of LHC data are well described!

Documents

Monte Carlo Developments Outline in the light of LHC data · Monte Carlo Developments in the light of LHC data Bryan Webber* University of Cambridge Outline Mostly newer developments