H1 Collaboration Meeting Lund, 3 October 2002
Monte Carlos and NLO
Torbj ¨orn Sj ¨ostrand
Apology: not specifically ep, but photoproduction ∼ pp
Event Physics Overview PYTHIA Status
Improved Parton Showers Heavy Flavour Production Heavy Flavour Hadronization
Multiple Interactions Summary
Event physics overview
Structure of the basic generation process:
1) Hard subprocess:
dˆσ/dˆt, Breit-Wigners
2) Resonance decays:
includes correlations (where implemented)
3) Final-state parton showers:
(or matrix elements)
4) Initial-state parton showers:
(or matrix elements)
5) Multiple
parton–parton interactions
q
q Z0 Z0
h0
Z0
µ+ µ−
h0
W− W+
ντ
τ− s c
q → qg g → gg g → qq q → qγ
g q
Z0
6) Beam remnants:
colour-connected to rest of event
7) Hadronization (or fragmentation)
8) Normal decays:
hadronic, τ, charm, . . .
p p
b b
ud ud
u u
q g g q
hadrons
ρ+
π0
π+
γ γ
9) QCD interconnection effects:
e− e+
W− W+
q3 q4 q2 q1
π+
π+ BE
a) colour rearrangement (⇒ rapidity gaps?);
b) Bose-Einstein (within & between strings).
10) The forgotten/unexpected: a chain is never stronger than its weakest link!
PYTHIA Status
ll
PYTHIA
lh hh
JETSET 7.4 PYTHIA 5.7 SPYTHIA
4 March 1997 : PYTHIA 6.1
Currently PYTHIA 6.210 of 25 September 2002
∼ 58, 900 lines Fortran 77
Code, manual, sample main programs, more:
www.thep.lu.se/
∼torbjorn/Pythia.html
short writeup in T. Sj ¨ostrand, P. Ed ´en, C. Friberg, L. L ¨onnblad, G. Miu, S. Mrenna and E. Norrbin
Computer Phys. Commun. 135 (2001) 238 [hep-ph/0010017]
Lund Productions
Proudly Presents
!#"$%&%')(*')"$ +
,
1 2 354 687 9;:=<
>5?A@CBEDFGBIHKJML N HKJPOMHKQ
RGSUTVW=XSYTZ\[W]XS_^`T abZdcfe
hgfikjml&g
XS_Z8ZnV8opa_cfe
[n`
ikq8r8i
Z;stT
i
Z8Zuave
xwyi
` i
Tz[|{_a}Z8cv^
~k x ]dG_ =]}
|
__ ] ¡ ]_¢£]¤ ¥ §¦¨©ª
¢«¬]¬z®]¬x¯°v±8²f³Y´¶µY²uµY±
·¡¸ ]x
}¥z& }=|U¢b
m =_º¹» ¡
v
¼
¡½|¾& ]U© =
k¿n= ¡k
À
¢C¬k¨UYÁÂk }m}}Ã
®]Ä]Å}¦ Ä}_»¢b©
Æ}ÇÇÇÇÇÇ.Æ
Æ}ÈÈÈÉÉÈÈÈÈÈÈÈÈÈÈÈ.Æ
Æ}ÈÈÈÈÈÈÉÉÈÈÈÈÈÈÈÈÈÈÈÈÈÈ.Æ
Æ}ÈÈÈÈÈÈÈÈÉÉÈÈÈÈÈÈÈÈÈÈÈÈÈÈÈÈ.Æ
Æ}ÈÈÈÈÈÈÈÈÈÉÉÈÈÈÈÈÈÈÈÈÈÈÈÈÈÈÈÈ*Æ
Æ}ÈÈÈÈÈÈÈÈÈÉÉÈÈÈÈÈÈÈÈÈÈÈÈÈÈÈÈÈ*Æ
Æ}ÈÈÈÈÈÈÈÈÉÉÈÈÈÈÈÈÈÈÈÈÈÈÈÈÈÈ.ÆUÉ
Æ}ÈÈÈÈÈÈÉÉÈÈÈÈÈÈÈÈÈÈÈÈÈÈ.ÆÊÉÉ
ÉÉ&Æ}ÈÈÈÉÉÈÈÈÈÈÈÈÈÈÈÈ.Æ ÉÉ
ÉÉ ÉpƶËÌÍËÎÆ ÉÉ
ÉÉ ÉÉ ÉÉ
ÉÉ ÉÉ ÉÉ
ÉÉ ÉÉ
ÉÉ ÏÐ ÉÉ
ÉÉ ÉÉ
ÉÉ ÐÐ ÉÉ
ÉÉ ÏÏ ÉÉ
ÉÉ ÉÉ
ÉÉ
with 430 new or revised pages!
Subprocess summary
Processes Examples
QCD & related
Soft QCD low-p⊥; diffraction
Hard QCD qg → qg
Open heavy flavour qq → tt Closed heavy flavour gg → gJ/ψ
γγ physics γq → qg
DIS γ∗q → q
γ∗γ∗ physics γT∗ γL∗ → qq Electroweak SM
Single γ∗/Z0/W± qq → γ∗/Z0 (γ/γ∗/Z0/W±/f/g)2 qq → W+W− Light SM Higgs gg → h0
Heavy SM Higgs Z0LZ0L → WL+W−L SUSY BSM
h0/H0/A0/H± qq → h0A0 SUSY qq0 → ˜χ0i χ˜±j R
/ SUSY χ˜0i → bcs ⇒ junctions!
Other BSM
Technicolor qq0 → πtc0 πtc±
New gauge bosons qq → γ∗/Z0/Z00 Compositeness qq → e±e∗∓
Leptoquarks qg → `LQ
H±± (from LR-sym.) qq → H++H−−
Extra dimensions gg → G∗ → e+e− + New User-Defined Process Machinery
Photoproduction
(G.A. Schuler & TS, NPB407 (1993) 539, ZPC73 (1997) 677)
Direct VMD Anomalous
Direct: point-like Resolved: hadronic state
Spectrum of fluctuations γ ↔ qq ∝ dk⊥2 /k⊥2 alt. m ' 2k⊥; dm2/m2
? k⊥ < k0 ' 0.5 GeV: nonperturbative γ → qq hadronic physics ⇒ VMD
(Vector Meson Dominance) parameterized couplings to ρ0, ω, φ, J/ψ
σtotγ→ρ = P(γ → ρ) · σtotρ
PDF fiγ→ρ(x, µ2), σjetγ→ρ = . . .
beam remnants, multiple interactions, . . .
? k⊥ > k0: perturbative γ → qq
PDF calculable: anomalous part of γ
but σtotqq not ⇒ GVMD (Generalized VMD) geometric scaling ansatz σtotqq ∝ kV2 /k⊥2 ,
kV ' mρ/2 for light quarks
again hadronic character: beam remnants, . . .
Deeply Inelastic Scattering
(aim: extend photoproduction to Q2 ∼ m2ρ)
(Ch. Friberg & TS, EPJC13 (2000) 151, JHEP09 (2000) 10)
Virtual photon: γ∗q → q σtotγ∗p ' 4π2αem
Q2 F2(x, Q2) ' 4π2αem
Q2
X q,q
e2q xq(x, Q2)
but F2 → 0 for Q2 → 0 by gauge invariance, + limit doublecounting with photoproduction
σDISγ∗p ' Q2 Q2 + m2ρ
!2
4π2αem
Q2
X q,q
e2q xq(x, Q2)
where q(x, Q2) frozen for Q2 < Q20;
and prefactor ensures σDIS → 0 for Q2 → 0
O(αs) DIS =
( QCDC γ∗q → qg BGF γ∗g → qq
)
= dir
σLO DISγ∗p = σDISγ∗p − σdirγ∗p → σDISγ∗p exp
−σdirγ∗p σDISγ∗p
corresponds to Sudakov form factor
so 4 components: DIS + dir + VMD + GVMD
From Real to Virtual Photons
σtotγ∗p(W2, Q2) = σDISγ∗p exp
−σdirγ∗p σDISγ∗p
+ σdirγ∗p
+ W
2
Q2+W2
!3
σVMDγ∗p + σGVMDγ∗p
Direct photon: Q2 in ME expression
Resolved photon:
total cross section σtotγ→i dampened by dipole m2
m2 + Q2
!2
(fewer fluctuations, smaller size)
VMD: m = mρ, mω, mφ, mJ/ψ GVMD: m ' 2k⊥; in total
Z k02
dk⊥2 k⊥2
k2V k⊥2
4k⊥2 4k⊥2 + Q2
!2
fγ
T∗
i (x, µ2, Q2): SaS 1D (also dipole-based) fγ
L∗
i (x, µ2, Q2): simple multiplicative factor or Ch´yla (hep-ph/0006232) (1 − x)3 reduces doublecounting at large x
⊕ virtual photon flux (T & L)
Parton Shower approach
2 → n = (2 → 2) ⊕ ISR ⊕ FSR
q q
Q Q Q2
2 → 2 Q22
Q21
ISR
Q24 Q23
FSR
2 → 2 = hard scattering (on-shell) σ =
ZZZ
dx1 dx2 dˆt fi(x1, Q2) fj(x2, Q2) dˆσij dˆt FSR = Final-State Radiation; timelike shower Q2i = M2 > 0 decreasing + coherence
ISR = Initial-State Radiation; spacelike shower Q2i = −M2 > 0 increasing + ∼ coherence backwards evolution: start at hard scattering
Do not doublecount! Q2 > Q21, Q22, Q23, Q24 2 → 2 = most virtual = shortest distance
ME vs. PS
ME : Matrix Elements
+ systematic expansion in αs (‘exact’) + powerful for multiparton Born level + flexible phase space cuts
− loop calculations very tough
− negative cross section in collinear regions
⇒ unpredictive jet/event structure
− no easy match to hadronization
PS : Parton Showers
− approximate, to LL (or NLL)
− main topology not predetermined
⇒ inefficient for exclusive states
+ process-generic ⇒ simple multiparton + Sudakov form factors/resummation
⇒ sensible jet/event structure + easy to match to hadronization
Marriage desirable! But how?
Problems: • gaps in coverage?
• doublecounting of radiation?
• Sudakov?
• NLO consistency?
Merging
= smooth transition ME/PS, no sharp edge.
+ emissions can cover full phase space
− coherence not straightforward Want to reproduce
WME = 1 σ(LO)
dσ(LO + g) d(phasespace)
by shower generation + correction procedure
wanted z }| {
WME =
generated z }| {
WPS
correction z }| {
WME WPS Comments:
• Do not normalize WME to σ(NLO),
since extra work without clear gain (expect radiation also in events added by K-factor ≥ 1)
• Exponentiate ME correction by shower Sudakov form factor:
WactualPS (Q2) =
WME(Q2) exp −
Z Q2max
Q2 WME(Q02) dQ02
!
• Normally several shower histories
⇒ alternative approaches, largely equivalent
Final-state showers
Merging with γ∗/Z0 → qqg since long
(M. Bengtsson & TS, PLB185 (1987) 435, NPB289 (1987) 810)
. . . but problems with γ∗/Z0 → bbg noted:
Q2i = m2i gives wrong singularity structure, Q2i = m2i − m2i,onshell is relevant propagator!
WME = (. . .)
Q21Q22 − (. . .)
Q41 − (. . .) Q42
(also weight from splitting kernels in PS) Coloured decaying particle also radiates:
0 (t)
1 (b) 2 (W+) i
3 (g)
0 (t)
1 (b) 2 (W+)
i 3 (g)
ME 1
Q20Q21 matches PS b → bg
⇒ can merge PS with generic a → bcg ME
(E. Norrbin & TS, NPB603 (2001) 297)
Subsequent branchings q → qg: also matched to ME, with reduced energy of system
Calculate for 1 → 2 processes in SM + MSSM:
WME(x1, x2) = 1
σ(a → bc)
dσ(a → bcg) dx1 dx2 Depends on
• mass ratios r1 = mb/ma and r2 = mc/ma
• colour and spin structure
• vector vs. axial vector etc. (γ5)
colour spin γ5 example
1 → 3 + 3 — — (eikonal)
1 → 3 + 3 1 → 12 + 12 1, γ5, 1 ± γ5 Z0 → qq 3 → 3 + 1 12 → 12 + 1 1, γ5, 1 ± γ5 t → bW+ 1 → 3 + 3 0 → 12 + 12 1, γ5, 1 ± γ5 H0 → qq 3 → 3 + 1 12 → 12 + 0 1, γ5, 1 ± γ5 t → bH+ 1 → 3 + 3 1 → 0 + 0 1 Z0 → ˜q˜q 3 → 3 + 1 0 → 0 + 1 1 ˜q → ˜q0W+ 1 → 3 + 3 0 → 0 + 0 1 H0 → ˜q˜q 3 → 3 + 1 0 → 0 + 0 1 ˜q → ˜q0H+ 1 → 3 + 3 12 → 12 + 0 1, γ5, 1 ± γ5 χ → q˜q 3 → 3 + 1 0 → 12 + 12 1, γ5, 1 ± γ5 ˜q → qχ 3 → 3 + 1 12 → 0 + 12 1, γ5, 1 ± γ5 t → ˜tχ 8 → 3 + 3 12 → 12 + 0 1, γ5, 1 ± γ5 ˜g → q˜q 3 → 3 + 8 0 → 12 + 12 1, γ5, 1 ± γ5 ˜q → q˜g 3 → 3 + 8 12 → 0 + 12 1, γ5, 1 ± γ5 t → ˜t˜g
WME(x1, x2)
g emission rate for different colour, spin and parity structures r1 = r2 = 0.2, x3 = 0.3
θqg
3 → 3 + 8
1 → 3 + 3 8 → 3 + 3 3 → 3 + 1
0.9 0.91 0.92 0.93 0.94 0.95 0.96 0.97 0.98 0.99 1
0 0.02 0.04 0.06 0.08 0.1
R3bl
yc
Pythia 6.153 Pythia 6.152 Pythia 6.129 DELPHI ALEPH
Rbl3 (yc)
ECM = 91 GeV mb = 4.8GeV
ratio of 3-jets in b and uds (=l) events
0.96 0.98 1 1.02 1.04 1.06 1.08 1.1 1.12 1.14 1.16
0 0.02 0.04 0.06 0.08 0.1
R3bl
yc
Vector Axial vector Scalar Pseudoscalar
Rbl3 (yc)
ECM = mh/H/A
= 120 GeV mb = 4.8GeV reference light q from γ∗/Z∗
Initial-state showers
p⊥Z dσ/dp⊥Z
physical Z + 1 jet ‘exact’
LO
‘exact’
NLO virtual
resummation: physical p⊥Z spectrum
shower: ditto + accompanying jets (exclusive)
Merged with matrix elements for
qq → (γ∗/Z0/W±)g and qg → (γ∗/Z0/W±)q0:
(G. Miu & TS, PLB449 (1999) 313)
WME WPS
!
qq0→gW
= ˆt2 + ˆu2 + 2m2Wˆs
ˆs2 + m4W ≤ 1 WME
WPS
!
qg→q0W
= ˆs2 + ˆu2 + 2m2Wˆt
(ˆs − m2W)2 + m4W < 3
curve=ResBos
black=PYTHIA KPYTHIA = 1.4
Q2max = s
C. Bal ´azs, J. Huston and I. Puljak, PRD63 (2001) 014021
Problem: requires large primordial k⊥ ≈ 2 GeV
⇒ need BFKL/CCFM non-ordered evolution ? Modified algorithm also affects other processes
• prefer Q2max = s where no doublecounting
⇒ more radiation at large p⊥
• require ˆu = Q2 − ˆs(1 − z) < 0 in branchings
⇒ fewer but harder emissions
Similarly for Higgs production in mt → ∞ limit:
• gg → gh0 and qg → qh0 simple
• qq → gh0 nonsingular & small ⇒ add Challenges:
• gauge boson pairs (S. Burby)
• QCD 2 → 2 with ISR+FSR+interference
Production graphs
Examples of Q = c/b production diagrams, not ex- haustive:
g Q
g Q
Leading order
q
q Q
Q Leading order
g Q
g Q
g Pair creation
(with gluon emission)
g g
g
Q Q
Flavour excitation
g g
g
Q Q
Gluon splitting
g g
g
Q Q g
Gluon splitting (flavour excitation)
PS approach to heavy quarks
3 main sources (arbitrary names):
1) pair creation:
based on gg → QQ and qq → QQ with masses + additional showering
2) flavour excitation:
based on c and b content of standard PDF’s + Qg → Qg and Qq → Qq ME’s;
massive kinematics but massless ME’s;
with Q2 > m2Q (so PDF> 0) and Q2i < Q2; g → bb by backwards evolution (improved)
≈ t-channel graph of gg → QQ
3) gluon splitting:
ordinary 2 → 2 processes, e.g. gg → gg + g → QQ branching with threshold
q1 − 4m2Q/m2g (1 + 2m2Q/m2g)
≈ s-channel graphs of gg, qq → QQ
Avoid doublecounting:
for 2 → 2: Q2 = ˆp2⊥+ (m23 + m24)/2 (⇒ ˆs∼ 4Q> 2) for FSR: Q2max = m2max = 4Q2
for ISR: Q2max = Q2
Beam Remnant Physics
Strings normally ‘large’ mass, but at times small because of beam remnant structure or by g → qq in shower. Thus three hadronization mechanisms (regions):
1. Normal string fragmentation:
continuum of phase-space states.
2. Cluster decay:
low mass ⇒ exclusive two-body state.
3. Cluster collapse:
very low mass ⇒ only one hadron.
p+ π−
u u
c c
ud d
If collapse:
cd: D−, D∗−, . . .
cud: Λ+c , Σ+c , Σ∗+c , . . .
⇒ flavour asymmetries Can give D “drag” to larger xF than c quark.
PYTHIA predicted qualitative behaviour.
Quantitative one sensitive to details
⇒ develop model & tune
Improved description of when collapse occurs (mass spectrum ⇐ constituent quark masses)
example:
charm string in πp collision
2 2.5 3 3.5
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35
Msinglet
(1/N)dN/dM
(a)
1-body
2-body cluster
2-body string
3-body
0.01 0.1
100 1000 10000
Collapse rate
√s
(a) Default
Small quark masses
Even BRDF charm string
collapse rate in pp collisions
(variations)
and
1-body collapse: energy-momentum shuffling 2-body decay: smoother joining to string
picture (matched anisotropic decay)
But also normal string fragmentation:
c d z
p± = E ± pz
p−D = zp−c 0 < z < 1
⇒ p+D = m2⊥D
p−D = m2⊥D zp−c
normally
> m2⊥c
zp−c = p+c z i.e. again drag.
Technical components of modelling:
• Charm and bottom masses: c and b cross sec- tions (mc = 1.5, mb = 4.8)
• Light-quark masses: threshold for cluster mass spectrum, together with mc
(mu = md = 0.33, ms = 0.50)
• Beam remnant distribution function:
(p − g = ud0 + u in colour octet state) hadron asymmetries also without collapse
(uneven sharing, but not extremely so)
• Primordial k⊥: collapse rate at large p⊥ (Gaussian width 1 GeV)
• Threshold behaviour for non-collapse:
all at Dπ or gradually at Dπ, D∗π, Dρ, . . .
• Collapse energy–momentum conservation:
practical solution to mass δ function
(several models tried; not very sensitive)
Asymmetries and correlations
0.001 0.01 0.1 1 10
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 (1/N)dN/dxF
xF (a) Pair production
All channels WA82 WA92
0.001 0.01 0.1 1 10
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 (1/N)dN/dxF
xF (b) Pair production
All channels WA82 WA92
D+ D−
A(xF) =
#D−−#D+
#D−+#D+
in π−p
-0.2 0 0.2 0.4 0.6 0.8 1 1.2
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 A(xF)
xF Pair production (a)
All channels WA92, 350 GeV WA82, 340 GeV E791, 500 GeV E769, 250 GeV
1e-05 0.0001 0.001 0.01 0.1 1
0 2 4 6 8 10 12 14 16 18 (1/N)dN/dpT2
pT2 (c) Pair production
All channels WA92
1e-05 0.0001 0.001 0.01 0.1 1
0 2 4 6 8 10 12 14 16 18 (1/N)dN/dpT2
pT2 (d) Pair production
All channels WA92
D+ D−
A(p⊥) =
#D−−#D+
#D−+#D+
in π−p
-0.05 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4
0 2 4 6 8 10
A(pT2 )
pT2 (b)
Pair production All channels WA92, 350 GeV E791, 500 GeV
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
0 0.5 1 1.5 2 2.5 3 3.5
(1/N)dN/d∆φ
∆φ (a)
Pair production All channels WA92
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45
-4 -3 -2 -1 0 1 2 3 4
(1/N)dN/d∆y
∆y
(b) Pair production All channels WA92 data
φ correlations improved . . .
. . . but
y correlations worsened
Multiple Interactions
(TS & M. van Zijl, PRD36 (1987) 2019, J. Dischler & TS, EPJdir C2 (2001) 1)
Consequence of composite nature of hadrons:
Evidence:
• direct observation: AFS, UA1, CDF
• implied by width of multiplicity distribution + jet universality: UA5
• forward–backward correlations: UA5
• pedestal effect: UA1, H1
One new free parameter: p⊥min 1
2σjet =
Z s/4 p2⊥min
dσ
dp2⊥ dp2⊥
⇐
Z s/4 0
dσ dp2⊥
p4⊥
(p2⊥0 + p2⊥)2 dp2⊥
Measure of colour screening length d in hadron p⊥min hdi ≈ 1(= ¯h)
r r
d
resolved
r r
d
screened
λ ∼ 1/p⊥
hdi ∼ rp
qNpartons no correlations
∼ rp
Npartons with correlations?
Npartons ∼ Ng =
Z 1
∼4p2⊥min/s g(x, ∼ p2⊥min) dx Olden days:
xg(x, Q20) → const. for x → 0
⇒ Npartons ∼ ln s
4p2⊥min ∼ const.
Post-HERA:
xg(x, Q20) ∼ x− for x → 0, ∼ 0.08>
⇒ Npartons ∼ s 4p2⊥min
!
⇒ p⊥min ∼ 1
hdi ∼ Npartons ∼ s
Mean charged multiplicity in inelastic non-diffractive ‘minimum bias’:
‘New’ PYTHIA default:
p⊥min = (1.9 GeV)
s 1 TeV2
0.08
Importance:
• comparison of data at 630 GeV & 1.8 TeV
• extrapolations to LHC
Summary
• PYTHIA evolving – do not use old versions!
• Test photoproduction/DIS transition region.
• Many ongoing efforts to improve showers.
Objective not NLO but good description.
Merging: “NLO ME” ⇒ shower smoothly;
applicable to ISR and FSR alike.
• Heavy flavour production/hadronization understood in pp?
Perturbative by non-overlapping
LO + flavour excitation + gluon splitting.
Combined with string hadronization;
small string = cluster, with special treatment
• Multiple interactions getting to be orthodoxy!
CDF: min bias & underlying event agree.
Varying impact parameter ⇒ “hot spots”.