- f
- mputation
- f
- mplete
- rder
- ratoire
- f
- llab
- ration
Intro dution A full NLLx example: Mueller-Navelet jets - - PowerPoint PPT Presentation
Intro dution A full NLLx example: Mueller-Navelet jets Pratial implementation of the omputation Results First alulation of Mueller Navelet jets at LHC at a omplete NLL BFKL o rder Samuel W allon Universit
h1(M 2
1 )
h2(M 2
2 )
s → t ↓ ←
va uum quantum numb erh′
1(M ′2 1 )
h′
2(M ′2 2 )
ha rd s ales: M 21 , M 2 2 ≫ Λ2 QCD
1 , M ′2 2 ≫ Λ2 QCD
QCD
where the t− hannel ex hanged state is the so- alled ha rd Pp → 0
and not b y itsm = 0 m = 0
θ → 0 = ⇒
sele t semi-ha rd p ro esses with s ≫ p2T i ≫ Λ2 QCD
where p2T i
a re t ypi al transverse s ale, allP
n(αS ln s)n
series (Balitski, F adin, Kuraev, Lipatov)A = + B @ + + · · · 1 C A + B @ + · · · 1 C A + · · · ∼ s ∼ s (αs ln s) ∼ s (αs ln s)2
this results in the ee tive BFKL ladder PSfrag repla ements gluon reggeon = "dressed gluon" ee tive vertex= ⇒ σh1 h2→anything
tot
= 1 s ImA ∼ sαP(0)−1
with αP(0) − 1 = C αs(C > 0)
Leading Log P omeron Balitsky , F adin, Kuraev, Lipatov 5 / 361 = p2 2 = 0, 2p1 · p2 = s)
writed4ki = s
2 dαi dβi d2k⊥i
(k = Eu l. ↔ k⊥ = Mink.)t− hannel
gluons have non-sense pNS
= 2
s p2/1
PSfrag repla ements⇒
set α1 = 0 andR dβ1 ⇒ Φγ∗→γ∗(k1, r − k1)
impa t fa to r⇒
set βn = 0 andR dαn ⇒ Φγ∗→γ∗(−kn, −r + kn)
β ր α ց γ∗ γ∗ r − k1 k1 k2 kn α1 α2
M = is (2π)2 Z d2k k2 Φup(k, r − k) Z d2k′ k′2 Φdown(−k′, −r + k′) ×
δ+i∞
Z
δ−i∞
dω 2πi „ s s0 «ω Gω(k, k′, r)
αn−1
← −
multi-Regge kinemati sβ2 βn αq, ¯
qβq, ¯
q 6 / 36αS P
n(αS ln s)n
resummation impa t fa to rs a re kno wn in some ases at NLLγ∗ → γ∗
at t = 0 (Ba rtels, Colferai, Giesek e, Kyrieleis, Qiao) fo rw a rd jet p ro du tion (Ba rtels, Colferai, V a a)γ∗
L → ρL
in the fo rw a rd limit (Ivanov, KT → ρT
t wist 3 transition at LL has b een p erfo rmedp(p1) p(p2)
jet1
(k⊥1, φ1)jet2
(k⊥2, φ2)φ1 φ2 − π
la rge + rapidit y la rge⊥
plane Beam axis 8 / 36| {z }
LL BFKL Green fun tion− →
strong de o rrelation b et w een the relative azimutal angle jets, in ompatible with p¯p
T evatron| {z }
NLL BFKL Green fun tionP(αs ln s)n
NLL w as in ludedP omeron
state, and not inside the jet verti es Sabio V era, S hw ennsen Ma rquet, Ro yk, k′ =
Eu lidian t wk k
NLL jet vertex: PSfrag repla ementsk k′ k − k′ k′
11 / 36kt
algo rithms (IR safe but time|pi| = transverse
energy depΩ = (yi, φi)
in y − φ plane dene transverse energyΩc 8 > > < > > :
yJ = |p1| y1 + |p2| y2 pJ φJ = |p1| φ1 + |p2| φ2 pJ
PSfrag repla ements pa rton1 (Ω1, |p1|) pa rton2 (Ω2, |p2|)Ω = (yi, φi)
in y − φ plane If distan es |Ωi − Ωc|2 ≡ (yi − yc)2 + (φi − φc)2 < R2 (i = 1 and i = 2 )= ⇒
pa rtons 1 and 2 a re in the same|p1| + |p2| max(|p1|, |p2|)R
13 / 36k, k′ =
Eu lidian t w0, x k k, x S(2)
J (k⊥; x) = δ
“ 1 − xJ x ” |k| δ(2)(k − kJ)
14 / 36k, k′ =
Eu lidian t wS(3,cone)
J
(k′, k − k′, xz; x) =
PSfrag repla ements0, x k k, x
S(2)
J (k, x) Θ
„h
|k−k′|+|k′| max(|k−k′|,|k′|)Rcone
i2 − ˆ ∆y2 + ∆φ2˜«
PSfrag repla ements0, x k k′ k − k′, x z k, x(1 − z)
+ S(2)
J (k − k′, xz) Θ
„ˆ ∆y2 + ∆φ2˜ − h
|k−k′|+|k′| max(|k−k′|,|k′|)Rcone
i2«
PSfrag repla ements0, x k k′ k − k′, x z k, x(1 − z)
+ S(2)
J (k′, x(1 − z)) Θ
„ˆ ∆y2 + ∆φ2˜ − h
|k−k′|+|k′| max(|k−k′|,|k′|)Rcone
i2« ,
15 / 36→ αS(µR)
IR se to r: PDF have IRkT
x1 x2 ↓ k1, φ1 ↓ k2, φ2 kJ,1, φJ,1, xJ,1 kJ,2, φJ,2, xJ,2
dσ d|kJ,1| d|kJ,2| dyJ,1 dyJ,2 = Z
dφJ,1 dφJ,2
Z
d2k1 d2k2
× Φ(kJ,1, xJ,1, −k1) × G(k1, k2, ˆ s) × Φ(kJ,2, xJ,2, k2)
with Φ(kJ,2, xJ,2, k2) =R dx2 f(x2)V (k2, x2) f ≡
PDFxJ = |kJ |
√s eyJ
17 / 36Cm ≡ Z dφJ,1 dφJ,2 cos ` m(φJ,1 − φJ,2 − π) ´ × Z d2k1 d2k2 Φ(kJ,1, xJ,1, −k1) G(k1, k2, ˆ s) Φ(kJ,2, xJ,2, k2). m = 0 = ⇒
ross-se tiondσ d|kJ,1| d|kJ,2| dyJ,1 dyJ,2 = C0 m > 0 = ⇒
azimutal de o rrelationcos(mϕ) ≡ cos ` m(φJ,1 − φJ,2 − π) ´ = Cm C0
18 / 361 π √ 2
` k2
1
´iν− 1
2 einφ1
de ompω(n, ν) = ¯ αsχ0 ` |n|, 1
2 + iν
´
with χ0(n, γ) = 2Ψ(1) − Ψ` γ + n
2
´ − Ψ ` 1 − γ + n
2
´
(Ψ(x) = Γ′(x)/Γ(x), ¯αs = Ncαs/π
)= ⇒
master fo rmula:Cm = (4 − 3 δm,0) Z dν Cm,ν(|kJ,1|, xJ,1) C∗
m,ν(|kJ,2|, xJ,2)
„ ˆ s s0 «ω(m,ν)
withCm,ν(|kJ|, xJ) = Z dφJ d2k dx f(x)V (k, x)Em,ν(k) cos(mφJ)
at NLL, same master fo rmula: just hange ω(m, ν) and V 19 / 36∂ ∂ν
it a tsω(n, ν) = ¯ αsχ0 „ |n|, 1 2 + iν « + ¯ α2
s
" χ1 „ |n|, 1 2 + iν « − πb0 2Nc χ0 „ |n|, 1 2 + iν « −2 ln µ2
R − i ∂
∂ν ln Cn,ν(|kJ,1|, xJ,1) Cn,ν(|kJ,2|, xJ,2) ff # , | {z } 2 ln |kJ,1| · |kJ,2| µ2
R
20 / 36s/s0)n + αs P(αs ln ˆ s/s0)n
(ˆs = x1 x2 s )
at LL s0 is a rbitra ry natural hoi e: s0 = √s0,1 s0,2 s0,i fo r ea hˆ s
is not an external s ale (x1,2 a re integrateds0,1 = (|kJ,1| + |kJ,1 − k1|)2 → s′
0,1 =
x2
1
x2
J,1
k2
J,1
s0,2 = (|kJ,2| + |kJ,2 − k2|)2 → s′
0,2 =
x2
2
x2
J,2
k2
J,2
9 > > > > > = > > > > > ; ˆ s s0 → ˆ s s′ = xJ,1 xJ2 s |kJ,1| |kJ,2| = eyJ,1−yJ,2 ≡ eY
s0 → s′
ae ts the BFKL NLL Green fun tion the impa t fa to rs:ΦNLL(ki; s′
0,i) = ΦNLL(ki; s0,i) +
Z d2k′ ΦLL(k′
i) KLL(k′ i, ki)1
2 ln s′
0,i
s0,i
(1) numeri al stabilities (non azimuthal averaging0,i
after numeri al integration) (1) an b e used to test s0 → λ s0 dep enden e 21 / 36αsχ(1)(γ, ω)
with shifted p¯ αsχ0(γ, 0) + ¯ α2
sχ1(γ, 0)
ω(0, ν)
isω(0, ν) = ¯ αsχ(1)(γ, ω(0, ν))
fo r ω(n, ν) numeri ally . there is no need fo r any jet vertex imp rovement b e auseγ
and 1 − γ pµR = µF
(this is impR)
W e use a ν grid (with a dense sampling a round 0) all numeri al al ulations a re done in Mathemati a w e use Cuba integration routines (in p ra ti e V egas): p re ision 10−2 fo r 500.000 max p= ⇒
14 minimal stable basi blo ks to b e evaluated numeri ally 23 / 366 7 8 9 10 0.01 0.1 1
PSfrag repla ementsC0 ˆ
nb GeV2
˜ = σ Y
Dierential ross se tion in dep enden e7 8 9 10 0.5 0.5 1.0 1.5 7 8 9 10 0.2 0.1 0.1 0.2
PSfrag repla ementsδC0 ˆ
nb GeV2
˜ δC0 ˆ
nb GeV2
˜ Y Y
Relative ee t7 8 9 10 0.2 0.1 0.1 0.2 0.3 7 8 9 10 0.06 0.04 0.02 0.02 0.04 0.06
PSfrag repla ementsδC0 ˆ
nb GeV2
˜ δC0 ˆ
nb GeV2
˜ Y Y
Relative ee t7 8 9 10 0.2 0.4 0.6 0.8 1.0 1.2
PSfrag repla ementsC1 C0 = cos ϕ
Y
erro r bands = erro rs due to the Monte Ca rlo integration dots = results6 7 8 9 10 0.2 0.4 0.6 0.8 1.0 1.2 6 7 8 9 10 0.2 0.4 0.6 0.8 1.0 1.2
PSfrag repla ementsC1 C0 = cos ϕ C1 C0 = cos ϕ
Y Y
Ee t6 7 8 9 10 0.2 0.4 0.6 0.8 1.0 1.2 6 7 8 9 10 0.2 0.4 0.6 0.8 1.0 1.2
PSfrag repla ementsC1 C0 = cos ϕ C1 C0 = cos ϕ
Y Y
Ee tcos ϕ
is still rather µR = µF and s0 dep endentpJ,1 pJ,2 p3 q
they a repJ,1 pJ,2
29 / 36|p1| ∼ |p2|
) PSfrag repla ementspJ,1 pJ,2
this ma y ho w ever not mean that the region |p1| ∼ |p2| is p erfe tly trustable even in a BFKL t yp e7 8 9 10 1.000 0.500 0.100 0.050 0.010 0.005
PSfrag repla ementsC0 ˆ
nb GeV2
˜ = σ Y
bands = erro rs due to the Monte Ca rlo integration dots = based7 8 9 10 0.2 0.4 0.6 0.8 1.0 1.2
PSfrag repla ementsC1 C0 = cos ϕ
Y
bands = erro rs due to the Monte Ca rlo integration dots = based7 8 9 10 0.2 0.4 0.6 0.8
PSfrag repla ementsC2 C0 = cos 2ϕ
Y
bands = erro rs due to the Monte Ca rlo integration dots = based6 7 8 9 10 0.2 0.4 0.6 0.8 1.0 1.2 6 7 8 9 10 0.2 0.4 0.6 0.8 1.0 1.2
PSfrag repla ementsC1 C0 = cos ϕ C1 C0 = cos ϕ
Y Y
Ee t6 7 8 9 10 0.2 0.4 0.6 0.8 1.0 1.2 6 7 8 9 10 0.2 0.4 0.6 0.8 1.0 1.2
PSfrag repla ementsC1 C0 = cos ϕ C1 C0 = cos ϕ
Y Y
Ee tcos ϕ
is still rather µR = µF and s0 dep endent7 8 9 10 0.2 0.4 0.6 0.8
PSfrag repla ementsC2 C1 = cos 2ϕ/cos ϕ
Y
bands = erro rs due to the Monte Ca rlo integration dots = based