Soft QCD (and PYTHIA)

(1)

Soft QCD (and PYTHIA)

Torbj¨ orn Sj¨ ostrand

Department of Astronomy and Theoretical Physics Lund University

S¨olvegatan 14A, SE-223 62 Lund, Sweden

ATLAS SM Workshop, Annecy, 4 February 2015

(2)

The structure of an event

An event consists of many di↵erent physics steps:

An event generator like PYTHIA attempts to describe all of it . . .

. . . and fails!

Torbj¨orn Sj¨ostrand Soft QCD (and PYTHIA) slide 2/29

(3)

The structure of an event

An event consists of many di↵erent physics steps:

An event generator like PYTHIA attempts to describe all of it . . . . . . and fails!

(4)

PYTHIA 8.2 status

PYTHIA is general-purpose, but with special ties to soft QCD Origin in hadronization studies — Lund string model (⇠ 1980) First realistic model for MPI (1987) and continued interest Gradual evolution

PYTHIA 6 (Fortran) definitely frozen September 2013 PYTHIA 8.100 released October 2007

PYTHIA 8.200 released October 2014 PYTHIA 8.205 released 23 January 2015 Smooth transition 8.1! 8.2

but rearranged directories and new build procedure.

New releases 3 – 4 times a year (not all subversions).

More refinements, fewer completely new features.

More interfaces: LHAPDF6, EvtGen, ProMC, . . .

(5)

Hard QCD & SM & BSM, status and news

Built-in library of common processes, but no own ME generator (unlike Sherpa).

Relies on MadGraph, PowHeg, . . . , for higher-order processes.

LHEFv3 ! MadGraph5 aMC@NLO, . . . Plethora of matching and merging techniques:

CKKW-L, MLM (AlpGen & MadGraph), NL³, UMEPS, PowHeg, aMC@NLO, UNLOPS, FxFx.

Minor upgrades to ISR and FSR dipole-style showers;

emission of W and Z gauge bosons.

NNPDF2.3 QCD+QED PDF sets (2 LO, 1 NLO, 1 NNLO).

Improved handling of colour sextets.

Long-lived R-hadrons (flexible ) also top).

Polarized ⌧^± decays updated & extended.

More NRQCD charmonium & bottomonium processes.

(6)

Soft QCD, status and news

New default handling of g! qq, notably g ! bb:

Add missing mass-correction term in g! QQ threshold region.

Suppress radiation for mqq! m^dipole based on ME for H! gg ! gqq.

Net result: g! bb rate ⇠ unchanged (⇠ LEP), m_bb shifted to lower masses.

String fragmentation extended to handle multiple junctions, but basic machinery unchanged.

New beam remnant model.

Many new colour reconnection models:

most general, some top-specific.

New tunes: Monash 2013 (MB+UE, now default), A14 (UE).

(7)

Multiparton Interactions (MPIs)

Seek unified description of hard jets, UE and MB.

Perturbative origin ) p? d.o.f. essential

Screening) dp²_?/p⁴_?! dp_?²/(p_?² + p_?0² )² ) finite Screening p_?0 energy-dependent, p_?0⇡ 2 3 GeV.

p_?-ordered generation, interleaved with ISR and FSR dP

dp_?=

✓dPMPI

dp_? +XdPISR

dp_? +XdPFSR

dp_?

◆

⇥ exp

✓ Z p_?max p_?

✓dPMPI

dp_?⁰ +XdPISR

dp_?⁰ +XdPFSR

dp_?⁰

◆ dp⁰_?

◆

Hardest MPI standard PDFs, softer modified, including flavour and momentum correlations.

Tuneable impact-parameter picture.

Colour reconnection needed.

Colours hook up with nontrivial beam remnants.

(8)

The new QCD-based beam remnant model (1)

Jesper Christiansen & Peter Skands (in preparation):

new frameworks for beam remnants and colour reconnection;

sharing philosophy, not quite separately available.

New beam remnant model

The beam remnant model comes after the perturbative machinery Overall idea of the model:

A game of conservation laws Add the minimal required amount of extra particles

MPI 1 MPI 2

Beam Remnant 2

...

Beam Remnant 1

- Example of two scattered gluons from a proton:

Flavour conservation

Add two up and one down quark

Baryon number conservation Turn two quarks into a diquark

Energy/momentum conservation

Choose x according to modified PDFs and rescale to match overall momentum conservation

Jesper Roy Christiansen (Lund) Non pertubative colours November 3, MPI@LHC 4 / 15

? Flavour conservation

? Baryon number conservation

? Energy/momentum conservation: modifed PDFs

(9)

The new QCD-based beam remnant model (2)

New beam remnant model - colour conservation

Possible colour states for the two gluons:

8 ⌦ 8 = 27 10 10 8 8 1

27

2 C & 2 AC + 1 gluon

10

0 C & 3 AC + 0 gluon (junction)

10

3 C & 0 AC + 1 gluon (junction)

8

1 C & 1 AC + 0 gluon

1

0 C & 0 AC + 0 gluon (not allowed)

Examples of the 27 and the 8 configurations:

MPI 1 MPI 2

Beam Remnant 2

...

Beam Remnant 1

MPI 1 MPI 2

Beam Remnant 2

...

Beam Remnant 1

Old model: 8⌦ 8 = 8, i.e. minimal nonvanishing colour, no junctions in BR (but possible in the event as a whole)

(10)

The new QCD-based beam remnant model (3)

Random walk in colour space unrealistic: partons are correlated.

Included as simple suppression exp( M/k),

where M is multiplet size and k is a free parameter

Comparisons to data

Relative large x and small p_?) forward physics Comparison to forward TOTEM measurements.

10 % di↵erence between no and maximal saturation The old model is similar to maximal saturation

Data Max saturation No saturation Monash 0

0.5 1 1.5 2 2.5 3 3.5 4 4.5

Charged particle| |at 7 TeV, track p >40 MeV, for N_ch 1

dN/d

5.4 5.6 5.8 6 6.2 6.4

0.6 0.8 1 1.2 1.4

| |

MC/Data

(arXiv:1205.4105)

(11)

Colour Reconnection (CR)

CR needed to explain e.g.hp?i(nch), 1987 and now:

N ch

0 50 100

[GeV]〉 T p〈

0.6 0.8 1 1.2 1.4 1.6

ATLAS Pythia 8 Pythia 8 (no CR)

7000 GeV pp Soft QCD (mb,diff,fwd)

mcplots.cern.ch 200k events≥Rivet 1.8.2,

Pythia 8.175 ATLAS_2010_S8918562

> 0.5 GeV/c) > 1, pT (Nch vs Nch Average pT

0 50 100

0.5 1

1.5 Ratio to ATLAS

CR reduces total string length ) reduces hadronic multiplicity

(12)

The new QCD-based CR model (1)

New model relies on two main principles

? SU(3) colour rules give allowed reconnections

Possible reconnections

Ordinary string reconnection

(qq: 1/9, gg: 1/8, model: 1/9)

Triple junction reconnection

(qq: 1/27, gg: 5/256, model: 2/81)

Double junction reconnection

(qq: 1/3, gg: 10/64, model: 2/9)

Zipping reconnection

(Depends on number of gluons)

? minimal measure gives preferred reconnections

⇡P

dipolesln(1 + m²_ij/m²₀) measure of string length,/ nhadronic

(13)

The new QCD-based CR model (2)

Comparison with LHC data:

Comparison to LHC data

Data New model Monash 0

0.1 0.2 0.3 0.4 0.5 0.6 0.7

/K⁰_Sversus rapidity at s = 7 TeV

N()/N(K

0 S)

0 0.5 1 1.5 2

0.6 0.8 1 1.2 1.4

|y|

MC/Data

(arXiv:1102.4282)

Data New model Monash

0 0.2 0.4 0.6 0.8 1 1.2 1.4

/K⁰_Sversus transverse momentum at s = 7 TeV

N()/N(K

0 S)

0 2 4 6 8 10

0.6 0.8 1 1.2 1.4 1.6

pT[GeV/c]

MC/Data

(arXiv:1102.4282)

Can describe /K_s ratios (tuned)

many baryons from junctions, but few from regular strings (big-mass string systems cut up!)

Distinguish new model from old model

Observables to distinguish junction baryons from diquark baryons

Best observable found so far can be seen on the right (again hadron decays are turned o↵)

Still looking for more observables

The di↵erence between Monash and the diquark curve can be understood by looking at the masses of the strings

Multiplicity

0 50 100 150 200 250 300

>Baryons<N

0 5 10 15 20 25 30 35 40 45

All (new model) Junctions (new model) diquark (new model) All (Monash)

Distinguish new model from old model

Observables to distinguish junction baryons from diquark baryons

Best observable found so far can be seen on the right (again hadron decays are turned o↵)

Still looking for more observables

The di↵erence between Monash and the diquark curve can be understood by looking at the masses of the strings

[GeV]) string Log (M

-1 -0.5 0 0.5 1 1.5 2 2.5 3 3.5 4

dLog(M)dN evtN1

0 0.5 1 1.5 2 2.5

3 All strings

Junctions Ordinary strings Monash

(14)

A top mass puzzle

t ⇡ 1.5 GeV

W ⇡ 2 GeV

Z⇡ 2.5 GeV 9=

;) c⌧ ⇡ 0.1 fm : p “pancakes” have passed,

MPI/ISR/FSR for p_? 2 GeV, inside hadronization colour fields.

t

t W b

2

Experiment mtop [GeV] Error due to CR Reference World comb. 173.34±0.76 310 MeV (40%) arXiv:1403.4427

CMS 172.22±0.73 150 MeV (20%) CMS-PAS-TOP-14-001 D0 174.98±0.76 100 MeV (13%) arXiv:1405.1756

1. Great job in reducing the errors

2. CR is one of the dominant systematics

3. Why is the CR uncertainty going down when there are

-no advances on the theoretical understanding

-no measurements to constrain it

A puzzle about mtop

(S. Argyropoulos) 1. Great job in reducing the errors.

2. CR is one of the dominant systematics.

3. Why is the CR uncertainty going down when there are

• no advances in theoretical understanding, and

• no measurements to constrain it?

(15)

New Pythia 8.2 CR models

PYTHIA 8.1: one CR model, joining some MPIs, with reduced . Late/early resonance decays: after/before CR.

S. Argyropoulos & TS: JHEP 11 (2014) 043

Basic idea: produce range of models to study how big mtop

could be without contradicting data.

Top CR as afterburner:

toy / stealth models

• forced random

• forced nearest

• forced farthest

• forced smallest

• smallest only for top

Top CR on equal footing:

more sophisticated / fragile

• swap

• move

• swap + flip

• move + flip so as to reduce also for MB/UE

(16)

E↵ects on top mass before tuning

CR off default forced random

100 120 140 160 180 200 220 240

0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 0.045

Reconstructed top mass, m_W [75, 85]GeV, p_T(jets) >40 GeV

mtop[GeV]

1/NdN/dmtop[GeV1]

mtop relative to no CR:

model mtop mtop

[GeV] rescaled default (late) 0.415 +0.209 default early +0.381 +0.285 forced random 6.970 6.508

.

Asymmetric spread:

mtop < 0 easy, mtop > 0 difficult.

Parton showers already prefer minimal . Main e↵ect from jet broadening, some from jet–jet angles.

(17)

E↵ects on top mass after tuning

No publicly available measurements of UE in top events.

• Afterburner models tuned to ATLAS jet shapes in tt events ) high CR strengths disfavoured.

• Early-decay models tuned to ATLAS minimum bias data

) maximal CR strengths required to (almost) match hp?i(nch).

model mtop

rescaled default (late) +0.239 forced random 0.524

swap +0.273

mtop relative to no CR

Excluding most extreme (unrealistic) models

m^max_top m^min_top ⇡ 0.50 GeV

(in line with Sandho↵, Skands & Wicke) New: mtop ⇡ 0 in QCD-based model .

Studies of top events could help constrain models:

• jet profiles and jet pull (skewness)

• underlying event

(18)

Rescattering Rescattering

Often assume that MPI =

. . . but should also include

Same order in ↵s, same propagators, but

• one PDF weight less smaller

• one jet less QCD radiation background 2 3 larger than 2 4 will be tough to find direct evidence.

Rescattering grows with number of “previous” scatterings:

Tevatron LHC

Min Bias QCD Jets Min Bias QCD Jets

Normal scattering 2.81 5.09 5.19 12.19

Single rescatterings 0.41 1.32 1.03 4.10 Double rescatterings 0.01 0.04 0.03 0.15 R. Corke & TS, JHEP 01 (2010) 035

(19)

An x-dependent proton size

Reasonable to assume that low-x partons are more spread out:

⇢(r , x)/ 1 a³(x) exp

✓ r² a²(x)

◆

with a(x) = a0

✓

1 + a1ln1 x

◆

a1 ⇡ 0.15 tuned to rise of ND

a0 tuned to value of ND, given PDF, p_?0, . . .

0.7 0.8 0.9 1 1.1 1.2

10² 10³ 10⁴ 10⁵

b2 eik [fm]

E_CM [GeV]

(b) a1 = 0.00 a1 = 0.15 a1 = 1.00

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 / dbMPInorm

bMPInorm (a)

SG DY Z⁰ Z’

Consequence: collisions at large x will have to happen at small b, and hence further large-to-medium-x MPIs are enhanced, but . . . a1 > 0 not favoured by tunes so far!

(20)

ATLAS questions: new CR & BR

Q: Do you expect we will be able to obtain better agreement with the early 13 TeV data using the new colour reconnection & beam remnant models in Pythia 8.2?

New model is appealing, but big junction rate is problematical — too much of a good thing? Further work/ideas needed.

Not now possible to guarantee better overall agreement.

Q: And which of their parameters should we be paying most attention to?

A: m0 and minimumGainJun. Current defaults are before tunes.

Better values:

ColourReconnection:m0 = 2.8

ColourReconnection:minimumGainJun = -0.65 MultipartonInteractions:pT0Ref = 2.15 ColourReconnection:allowDoubleJunRem = off More options coming: suppress CR for high relative velocities.

(21)

ATLAS questions: MB & UE unification

Q: Could these new features potentially help in the unification of MPI tune parameters between MB and UE datasets?

A: Glass half full or half empty? Monash MB+UE not too bad!

0 50 100

)ChProb(n

10-7

10-6

10-5

10-4

10-3

10-2

10-1

1 >0.5, |η|<2.5)

1, pT Ch≥ Chg. Mult. (n

Pythia 8.185 Data from New J.Phys. 13 (2011) 053033

ATLAS PY8 (Monash 13) PY8 (4C) PY8 (2C)

/Nbins 2 χ5%

0.0

± 2.7

0.0

± 2.7

±0.0 9.6

V I N C I A R O O T

pp _{7000 GeV}

nCh

0 50 100

Theory/Data

0.6 0.8 1 1.2 1.4

0 5 10 15 20

T/dpCh dnT)/pπ/(2Ch1/n

10-8

10-7

10-6

10-5

10-4

10-3

10-2

10-1

1 10

|<2.5) η

>0.5, | 1, pT Ch≥ T (n p

/Nbins 2 χ5%

0.0

± 1.5

0.0

± 0.8

±0.1 5.8

V I N C I A R O O T

pp _{7000 GeV}

[GeV]

pT

0 5 10 15 20

Theory/Data

0.6 0.8 1 1.2 1.4

0 50 100 150 200

)ChProb(n

10-6

10-5

10-4

10-3

10-2

10-1

1

|<2.5) η

>0.1, | 2, pT Ch≥ Soft Chg. Mult. (n

/Nbins 2 χ5%

0.0

± 4.5

0.0

± 4.9

0.1

± 19.7

V I N C I A R O O T

pp 7000 GeV

nCh

0 50 100 150 200

Theory/Data

0.6 0.8 1 1.2 1.4

0 5 10 15 20

T/dpCh dnT)/pπ/(2Ch1/n

10-9

10-7

10-5

10-3

10-1

10

|<2.5) η

>0.1, | 2, pT Ch≥ T (n p

/Nbins 2 χ5%

±0.0 4.3

±0.1 7.3

0.2

± 15.5

V I N C I A R O O T

pp 7000 GeV

[GeV]

pT

0 5 10 15 20

Theory/Data

0.6 0.8 1 1.2 1.4

Figure 18: Min-bias pp collisions at 7 TeV. Charged-multiplicity and p?distributions, with standard (top row) and soft (bottom row) fiducial cuts, compared to ATLAS data [91].

28

0 5 10 15 20

)φ∆η∆>/(Tp∑<

0 0.5 1 1.5 2 2.5

>0.1)

|<2.5, pT (|η )> vs pT1 TRNS <Sum(pT

Pythia 8.185 Data from Phys.Rev. D83 (2011) 112001 ATLAS

PY8 (Monash 13) PY8 (4C) PY8 (2C)

/Nbins 2 χ5%

0.0

± 0.3

0.1

± 0.9 0.2

± 11.2

V I N C I A R O O T

pp 7000 GeV

(hardest track) [GeV]

pT1

0 5 10 15 20

Theory/Data

0.6 0.8 1 1.2 1.4

0 5 10 15 20

)φ∆η∆>/(Ch<n

0 0.5 1 1.5 2 2.5

>0.5)

|<2.5, pT (|η

> vs pT1 TRNS <nCh

Pythia 8.185 Data from Phys.Rev. D83 (2011) 112001 ATLAS

PY8 (Monash 13) PY8 (4C) PY8 (2C)

/Nbins 2 χ5%

0.0

± 0.2

0.0

± 0.4

0.2

± 10.4

V I N C I A R O O T

pp 7000 GeV

(hardest track) [GeV] pT1

0 5 10 15 20

Theory/Data

0.6 0.8 1 1.2 1.4

Figure 22: pp collisions at 7 TeV. UE (“Transverse region”) transverse-momentum sum density (left) and charged-track density (right), compared with ATLAS data [98].

|⌘| < 2.5 [98]. As is now well known the Tevatron extrapolations (represented here by Tune 2C) predicted a UE level which was 10% – 20% below the LHC data. Both the current default tune 4C (which included LHC data) and the Monash 2013 tune exhibit significantly better agreement with the LHC measurements, with the Monash one giving a slight additional improvement in the ²_5%values.

We conclude that the Monash 2013 tune parameters are appropriate for both min-bias and UE studies.

3.5 Identified Particles at LHC

While the description of inclusive charged particles, discussed in the previous section, is acceptable, larger discrepancies emerge when we consider the spectra of identified particles. We here focus on strange particles, in particular K_S⁰mesons and ⁰hyperons in figs.23and24, respectively. The experimental measurements come from CMS [99]. Additional comparisons to strange-particle spectra (K , , and ) are collected in appendixB.2.

In the K_S⁰rapidity distribution, shown in the left-hand pane of fig.23, we observe that tune 4C exhibits a mild underproduction, of about 10%. Though it might be tempting to speculate whether this could indicate some small reduction of strangeness suppression in pp collisions, however, we already noted in section2.1that the strangeness production in ee collisions also needed to be increased by about 10%. After this adjustment, we see that the overall K_S⁰yield in the Monash 2013 tune is fully consistent with the CMS measurement. Nonetheless, we note that the momentum distribution is still not satisfactorily described, as shown in the right-hand pane of fig.23. Our current best guess is therefore that the overall rate of strange quarks is consistent, at least in the average min-bias collision (dedicated comparisons in high-multiplicity samples would still be interesting), but that the phase- space distribution of strange hadrons needs more work. Similarly to the case in ee collisions, cf. fig.6, the model predicts too many very soft kaons, though we do not currently know whether there is a dynamic link between the ee and pp observations.

For strange baryons, we note that the increase in the ⁰fraction in ee collisions (cf. fig.5) does 32

Show-stopper di↵erence? Or fewer data) better tune?

Some pieces of the puzzle likely still missing.

(22)

ATLAS questions: tuning

Q: ATLAS has recently demonstrated, for example with the A14 tuning, the eigentune method of performing excursions in the covariance matrix about the tune minimum to fixed ² with the aim of discovering a minimal set of systematic parameter bounds.

What are your thoughts on more formal methods of quantifying the stability of tune parameters?

A: A14 is pretty formal to me — and very impressive!

Only complaint: MB omitted. How bad? What price to improve?

Root of evil: ² meaningful only if some parameter set is perfect (perfect model, perfect data).

Cf. PDFs: ² = 1 does not work) pragmatism.

In total too many parameters: A14 benefits from by-hand Monash!

Some well-known correlations: (a, b), (↵s, p_?0), . . .

(23)

ATLAS questions: sphericity misbehaviour (1)

Q: We observe that minimum bias data prefer a higher sphericity topology than MC predictions at large value ofP

|p?| where we expect to be transitioning to ’jetty’ events. How could this e↵ect be compensated in the current models?

(24)

ATLAS questions: sphericity misbehaviour (2)

A: Competition between moreP

p_? by more particles or by jets.

ISR experiments: no clear jet structure ifP

p_? trigger!

Sphericity momentum tensor : S^↵ = P

ip^↵_i p_i P

i|p(?)i|² , ↵, = x, y (, z) collinear unsafe — infamous since PETRA days (⇠ 1980).

Sensitive to jet fragmentation function, resonance decays, . . . Recommend linearized form : L^↵ =

P

i p_i^↵p_i

|p(?)i|

P

i|p(?)i|, ↵, = x, y (, z) But disagreement also for Thrust (linear) so not (only) problem.

0) Current tunes: Monash, A14?

1) Multijets? CKKW-L studies to check!

2) Impact parameter profile? (Double Gaussian, exponential.) 3) Minijet jet profiles: PYTHIA jets too narrow? NLO ↵s? 4) Exclude particles above/below p_? cuts to narrow down.

(25)

ATLAS questions: helical gluon packing

Q: Models of helical gluon packing can better describe the azimuthal ordering of soft hadron emissions. What are your thoughts on such extensions?

low-p_? depleted:

7

Hadronisation effects should become more evident when measurements are made in regions of the phase space domi- nated by the production of low-pTparticles. Figure 4 shows the power spectra measured in the low-pTenhanced sample (nch> 10, pT>100 MeV and maximal pT< 1 GeV). A significant amount of correlations in observed in the data in both SEand Sηdistributions compared to the PHOJET and PYTHIA based models. HERWIG++ gives a seemingly bet- ter description for the SEdistribution yet it seems its predic- tion is more of an artefact of an enhanced single jet structure given the fact the model overestimates the measurements in the inclusive event selection (Fig. 3). The interpretation of this measurement in terms of the azimuthal ordering of hadrons related to the properties of the gluon field is discussed in Section VIII.1.

Figure 5 shows the power spectra SEand Sηfor the cor- rected data and MC predictions in the low-pTdepleted region (nch> 10, pT>500 MeV and maximal pT< 10 GeV). In principle, this should be the best understood part of the phase space, with a suppressed diffractive component, lower sensitivity to hadronisation effects and best available model tunes.

-1 ] [GeV ω

0 0.5 1 1.5 2 2.5 3

) -1ω(ES

0 0.2

0.4 ATLAS

= 7 TeV s

)<10 GeV

>10, max(pT nch

|<2.5 η

>500 MeV, | pT

Data 2010 PHOJET PYTHIA8 4C PYTHIA6 AMBT2b HERWIG++ UE7-2

ξ

0 1 2 3 4 5 6

) -1ξ(ηS

0 0.2 0.4 0.6 0.8

ATLAS

= 7 TeV s

)<10 GeV

>10, max(pT nch

|<2.5

>500 MeV, | η pT

FIG. 5. Corrected data from the low-pTdepleted sample compared to particle-level predictions from various MC models using conventional hadronisation algorithms. The top and bottom plots are for the SEand Sηpower spectra, respectively. The error bars correspond to the combined statistical and systematic uncertainties.

However, we find that all models significantly overestimate the size of the principal peak structure in both SEand Sη.

Comparison of Figs 3, 4 and 5 show that the azimuthal correlations are qualitatively different in each subsample, and that the standard MC models fail to reproduce these accu- rately. Similar conclusion can be drawn for the measurement performed at s = 900 GeV (Appendix A).

ξ

0 2 4 6

) -1ξ(ηS

0 0.5 1 1.5

= 7 TeV s ATLAS

)<10 GeV

>10, max(pT nch

|<2.5 η

>500 MeV, | pT

PYTHIA6 MC09 (non-diffr.) without MPI [MSTP(81)=20]

without ISR [MSTP(61)=0]

without FSR [MSTP(71)=0]

Data 2010

ξ

0 2 4 6

) -1ξ(ηS

-0.2 -0.1 0 0.1 0.2

= 7 TeV s ATLAS

)<1 GeV

>10, max(pT nch

|<2.5

>100 MeV, |η pT

PYTHIA6 MC09 non diffractive + diffractive

without MPI [MSTP(81)=20]

without ISR [MSTP(61)=0]

without FSR [MSTP(71)=0]

Data 2010

FIG. 6. Corrected Sηdistributions compared with the particle level predictions of PYTHIA6 MC09 for various settings, using the non- diffractive pp scattering (full line) as the baseline. Top: low-pTde- pleted subsample. Bottom: low-pTenhanced subsample. The error bars correspond to the combined statistical and systematic uncertainties.

In the frame of the conventional QCD modelling, we have tried to identify the most likely source of the observed discrep- ancies. Figure 6 shows the sensitivity of the Sηdistribution to various components of the QCD modelling implemented in PYTHIA 6, taking as a baseline the non-diffractive pp scatter- ing scenario (indicated by the full line). In the low-pTde- pleted sample, the size of correlations varies strongly with the amount of multiple parton interactions (MPI), of initial state radiation (ISR) and with the amount of parton showering.

The data prefer modelling with enhanced radiation and/or enhanced MPI rate which can be achieved via careful adjustment of the relevant model parameters.

However, such an adjustment typically creates an even larger discrepancy in the low-pTenhanced region, where the

low-p_? enhanced:

6

• uncertainty of the unfolding technique: parametrised to cover the residual discrepancies in the scaling of 3 fold- ing iterations (Appendix B);

• uncertainty on the tracking efficiency estimate: domi- nated by the uncertainty on the inner detector material description, which translates into a variation of scaling factors by 5%;

• uncertainty due to the residual content of secondary tracks: set to 25% of the correction applied, with minimal value of 0.005 (based on MC studies) ;

• uncertainty due to the difference in the charged-particle multiplicity selection at the generator level and at the detector level: calculated in a model-independent way as a variation of the shape corresponding to the change of the averaged selected charged-particle multiplicity by one unit; and

• the uncertainty in the correction of the bias due the max(pT) cut: corresponds to a 5% variation of the track reconstruction efficiency.

-1 ] [GeV ω

0 0.5 1 1.5 2 2.5 3

) -1ω(ES

-0.1 0 0.1 0.2

0.3 ATLAS

= 7 TeV s

)<10 GeV

>10, max(pT nch

|<2.5 η

>100 MeV, | pT

ξ

0 1 2 3 4 5 6

) -1ξ(ηS

0 0.2 0.4 0.6

ATLAS = 7 TeV s

)<10 GeV

>10, max(pT nch

|<2.5

>100 MeV, | η pT

FIG. 3. Corrected data from the inclusive sample compared to particle-level predictions from various MC models using conventional hadronisation algorithms. The top and bottom plots are for the SEand Sηpower spectra, respectively. The error bars correspond to the combined statistical and systematic uncertainties.

All contributions to the systematic uncertainty are combined quadratically. The negative correlation between track reconstruction efficiency and secondary track content is neglected, making the uncertainty estimate more conservative.

VIII. RESULTS

The results of this analysis obtained for pp collisions at s

= 7 TeV are presented in this section. Results from this analysis repeated for s = 900 GeV are shown in Appendix A. The corrected data are compared with the predictions of several commonly used MC models: PYTHIA6, PHOJET, PYTHIA8 and HERWIG++.

Figure 3 shows the comparison for the inclusive event se- lection (nch> 10, pT>100 MeV and maximal pT< 10 GeV).

The principal peak structure observed in the power spectra for both SEand Sηis roughly reproduced by PYTHIA and PHO- JET models and overestimated by HERWIG++. The tail of the SEdistribution around 0.5 < ω < 1 rad/GeV is not reproduced by any of the models.

-1 ] [GeV ω

0 0.5 1 1.5 2 2.5 3

) -1ω(ES

-0.2 -0.1 0 0.1

ATLAS = 7 TeV s

)<1 GeV

>10, max(pT nch

|<2.5 η

>100 MeV, | pT

ξ

0 1 2 3 4 5 6

) -1ξ(ηS

-0.2 -0.1 0 0.1 0.2

ATLAS = 7 TeV s

)<1 GeV

>10, max(pT nch

|<2.5

>100 MeV, | η pT

FIG. 4. Corrected data from the low-pTenhanced sample compared to particle-level predictions from various MC models using conventional hadronisation algorithms. The top and bottom plots are for the SEand Sηpower spectra, respectively. The error bars correspond to the combined statistical and systematic uncertainties.

A: problem in modelling of low-p_? region, as above, but S⌘(⇠) =

⌧ 1 nch

Xexp(i(⇠⌘_j '_j))² <hn^chi

Tricky: in PYTHIA ' = 0(⇡) if px > 0(< 0) gives most of e↵ect ) sensitive to y structure and p? compensation, not helix?

(26)

ATLAS questions: MPI measurements (1)

Q: Can we think about a way to measure DPI/MPI contribution (specially at high p_?) without being so dependent on model(s)?

A1: Possibly. 4j ! + 3j helped, so why not 2 + 2j? W W ?

A2: Wishful thinking? Is everything describable by one number?

(27)

ATLAS questions: MPI measurements (1)

Q: Can we think about a way to measure DPI/MPI contribution (specially at high p_?) without being so dependent on model(s)?

A1: Possibly. 4j ! + 3j helped, so why not 2 + 2j? W W ? A2: Wishful thinking? Is everything describable by one number?

Diego Ciangottini 6 MPI@LHC2014, Cracow

Not enough…

Large systematics, mostly related to model dependence.!

!

It’s not possible yet to get any informations neither on energy

dependence nor on parton correlation

It’s still a long way to the final answers CMS results so far:!

4jets, W+2jets

CMS ongoing:!

3jets+gamma, ! same sign WW, !

double j/psi

(28)

ATLAS questions: MPI measurements (2)

Dependence on x scale, flavour (q vs. g), . . . ? PYTHIA: correlated PDFs preserve E and flavour;

flexible impact-parameter profile, x-dependent ditto, . . . Normally not considered by “true theorists”.

Simple extension 2! 3 ! 4 . . . MPIs?

Count minijet multiplicities

e.g. R = 0.3, 0.5, 0.7 and p_?min = 3, 5, 10 GeV.

Q: Any thoughts on testing the re-scattering framework in this context?

A: Rescattering and rp(x) ambitious failures (so far)!

In theory study odd jet multiplicities, but 2! 3 dominates.

(29)

ATLAS questions: UE measurements

Q: Do the traditional UE measurements make sense any more in view of extra jet contamination? Should we switch to more complicated analysis techniques — like looking at event shapes and measure UE only in certain topologies?

A1 (pileup): low-luminosity runs are essential!

A2 (more jets for higher Ecm):

alternative to ' regions (R. Field) is

modified“Swiss cheese” (J. Huston, V. Tano, CDF):

use cones around jets to define the p_? trigger and then the remainder is the UE.

Vary number of jets: 2, 3, 4 . . . Also: UE rises faster than jets!?

improvement of the jet studies. Another important question is whether the presence of a hard interaction in the event influences the spectator interactions.

The measurement of the energy in minimum bias events is important in its own right as it is used to estimate the effect of pile-up events on any signal at hadron colliders, where due to high instantaneous luminosity, several interactions may occur in the same bunch crossing. In this paper, we present a measurement of the energy deposited far from the jets in ¯pp interactions at s = 1800 and s = 630 GeV and compare our measurement with the energy observed in minimum bias events and with the predictions from two Monte Carlo models.

The study reported in this paper is complementary to our previous analysis [9], which examined the evolution of event structure in ¯pp interactions at s = 1800 GeV with E_T⁽¹⁾ over the 1–50 GeV range and found that the energy transverse to the leading jet rises rapidly in the 1-5 GeV range and is almost constant for E_T⁽¹⁾ 10 GeV.

FIG. 1: An example of a two jet event in the detector region under study. The cones used for the determination of the underlying event contribution are at ⌘ = ⌘⁽¹⁾and = ⁽¹⁾± 90 where (⌘⁽¹⁾, ⁽¹⁾) is the centroid of the highest ET jet in the event.

To study the underlying energy in jet events, we define two cones with radius R = ( )²+ ( )² = 0.7 centered at = ⁽¹⁾, and = ⁽¹⁾± 90⁰ where ( ⁽¹⁾, ⁽¹⁾) is the centroid of the highest energy jet in the event as shown in Figure 1. The sum of transverse momenta of all tracks in these two cones is labeled P_T^90,min and P_T^90,max where P_T^90,max is

3

(30)

Distributions to measure at 13 TeV

n_ch distribution, in wide and narrow ⌘ bins.

dnch/d⌘, dn_ch/dp_? inclusively and (1/nch)dnch/dp_? in multiplicity bins.

hp?i(nch),hnK_S⁰i(nch),hn⇤i(nch) (& other species) Forward-backward correlations, e.g. UA5-style;

use E_? distributions to extend y range as much as possible.

Profile of jet energy & particle flow, also for minijets.

njet distribution for varying (mini)jet width and p_? threshold:

R = 0.3, 0.5, 0.7, p_?min= 3, 5, 10 ?

UE characteristics as function of p_?trig, in ' sectors and by “Swiss cheese” with varying number of excluded jets.

(31)

Summary and Outlook

PYTHIA continues to evolve, even if slower.

Coming: hard di↵raction a la PomPyt + screening, physics, double onium production.

MPI key PYTHIA component since 30 years.

Original concepts still hold, but more sophisticated.

Colour reconnection remains one of big known unknowns.

Everything mixed up) experimental tests sometimes indecisive, e.g. rescattering and x-dependent proton size.

Key aspect is missing from understanding of low-p_? region:

better colour reconnection? hadron gas & collective flow?

(32)

Summary and Outlook

PYTHIA continues to evolve, even if slower.

Coming: hard di↵raction a la PomPyt + screening, physics, double onium production.

MPI key PYTHIA component since 30 years.

Original concepts still hold, but more sophisticated.

Colour reconnection remains one of big known unknowns.

Everything mixed up) experimental tests sometimes indecisive, e.g. rescattering and x-dependent proton size.

Key aspect is missing from understanding of low-p_? region:

better colour reconnection? hadron gas & collective flow?