high-energy amplitudes in n=4 sym - uni-regensburg.demaa29312/conference/talks/balitsky.pdf · i....

High-energy amplitudes in N=4 SYM

I. Balitsky

JLAB & ODU

Regensburg 15 July 2014

I. Balitsky (JLAB & ODU) High-energy amplitudes in N=4 SYM Regensburg 15 July 2014 1 / 50

Outline

Regge limit in a conformal theory.High-energy scattering and Wilson lines.Evolution equation for color dipoles.Leading order: BK equation.NLO BK kernel in N = 4.NLO amplitude in N = 4 SYMConclusions 1

Light-ray operators.“BFKL” representation of 4-point correlation function in N = 4SYM.“DGLAP” representation of 4-point correlation function.Anomalous dimensions from DGAP vs BFKL representations.Conclusions 2

Outline

Regge limit in a conformal theory.High-energy scattering and Wilson lines.Evolution equation for color dipoles.Leading order: BK equation.NLO BK kernel in N = 4.NLO amplitude in N = 4 SYMConclusions 1Light-ray operators.“BFKL” representation of 4-point correlation function in N = 4SYM.“DGLAP” representation of 4-point correlation function.Anomalous dimensions from DGAP vs BFKL representations.Conclusions 2

Conformal four-point amplitude

A(x, y, x′, y′) = (x− y)4(x′ − y′)4N2c 〈O(x)O†(y)O(x′)O†(y′)〉

O = TrZ2 (Z = 1√2(φ1 + iφ2)) - chiral primary operator

In a conformal theory the amplitude is a function of two conformal ratios

A = F(R,R′)

R =(x− y)2(x′ − y′)2

(x− x′)2(y− y′)2 , R′ =(x− y)2(x′ − y′)2

(x− y′)2(x′ − y)2

At large Nc

A(x, y, x′, y′) = A(λ; x, y, x′, y′) λ = g2Nc − ‘t Hooft coupling

Our goal is perturbative expansion and resummation of (λ ln s)n at large energiesin the next-to-leading approximation

(λ ln s)n(cLOn + cNLO

Conformal four-point amplitude

A(x, y, x′, y′) = (x− y)4(x′ − y′)4N2c 〈O(x)O†(y)O(x′)O†(y′)〉

O = TrZ2 (Z = 1√2(φ1 + iφ2)) - chiral primary operator

In a conformal theory the amplitude is a function of two conformal ratios

A = F(R,R′)

R =(x− y)2(x′ − y′)2

(x− x′)2(y− y′)2 , R′ =(x− y)2(x′ − y′)2

(x− y′)2(x′ − y)2

At large Nc

A(x, y, x′, y′) = A(λ; x, y, x′, y′) λ = g2Nc − ‘t Hooft coupling

Our goal is perturbative expansion and resummation of (λ ln s)n at large energiesin the next-to-leading approximation

(λ ln s)n(cLOn + cNLO

Regge limit in the coordinate space

Regge limit: x+ → ρx+, x′+ → ρx′+, y− → ρ′y−, y′− → ρ′y−− ρ, ρ′ →∞

,y’_ y’_

x+ , x_ , y_y+

x’− , x’_

Full 4-dim conformal group: A = F(R, r)

R =(x− y)2(x′ − y′)2

(x− x′)2(y− y′)2 →ρ2ρ′2x+x′+y−y′−

(x− x′)2⊥(y− y′)2

⊥→ ∞

r =[(x− y)2(x′ − y′)2 − (x′ − y)2(x− y′)2]2

(x− x′)2(y− y′)2(x− y)2(x′ − y′)2

→[(x′ − y′)2

⊥x+y− + x′+y′−(x− y)2⊥ + x+y′−(x′ − y)2

⊥ + x′+y−(x− y′)2⊥]2

(x− x′)2⊥(y− y′)2

⊥x+x′+y−y′−

4-dim conformal group versus SL(2,C)

Regge limit: x+ → ρx+, x′+ → ρx′+, y− → ρ′y−, y′− → ρ′y−−ρ, ρ′ →∞

,y’_ y’_

x+ , x_ , y_y+

x’− , x’_

Regge limit symmetry: 2-dim conformal group SL(2,C) formed fromP1,P2,M12,D,K1 and K2 which leave the plane (0, 0, z⊥) invariant.

Pomeron in a conformal theory

A(x, y; x′, y′) s→∞=

∫dν f+(ω(λ, ν))F(λ, ν)Ω(r, ν)Rℵ(λ,ν)/2

L. Cornalba (2007)f+(ω) = eiπω−1

sinπω - signature factor

Ω(r, ν) =sin νρsinh ρ

, cosh ρ =

- solution of the eqn (H3 + ν2 + 1)Ω(r, ν) = 0

The dynamics is described by:ω(λ, ν) - pomeron intercept,andF(λ, ν) - “pomeron residue”.

A(x, y; x′, y′) s→∞=

∫dν f+(ω(λ, ν))F(λ, ν)Ω(r, ν)Rω(λ,ν)/2

Pomeron intercept ω(ν, λ) is known in two limits:

1. λ→ 0 : ω(ν, λ) =λ

πχ(ν) + λ2ω1(ν) + ...

χ(ν) = 2ψ(1)− ψ(12 + iν)− ψ(1

2 − iν) - BFKL intercept,

ω1(ν) - NLO BFKL intercept Lipatov, Kotikov (2000)

2. λ→∞ : AdS/CFT ⇒ ω(ν, λ) = 2− ν2 + 42√λ

2 = gravition spin , next term - Brower, Polchinski, Strassler, Tan (2006)Cornalba, Costa, Penedones (2007)

Now: two more terms (∼ λ−1 and ∼ λ−3/2)

A(x, y; x′, y′) s→∞=

The function F(ν, λ) in two limits:

1. λ→ 0 : F(ν, λ) = λ2F0(ν) + λ3F1(ν) + ...

F0(ν) = π sinhπν4ν cosh3 πν

Cornalba, Costa, Penedones (2007)

F1(ν) = see below G. Chirilli and I.B. (2009)

2. λ→∞ : AdS/CFT ⇒ ω(ν, λ) = π3ν2 1 + ν2

sinh2 πν+ ...

L.Cornalba (2007)

We calculate F1(ν) (and confirm ω1(ν)) using the expansion of high-energyamplitudes in Wilson lines (color dipoles)

A(x, y; x′, y′) s→∞=

The function F(ν, λ) in two limits:

1. λ→ 0 : F(ν, λ) = λ2F0(ν) + λ3F1(ν) + ...

F0(ν) = π sinhπν4ν cosh3 πν

Cornalba, Costa, Penedones (2007)

F1(ν) = see below G. Chirilli and I.B. (2009)

2. λ→∞ : AdS/CFT ⇒ ω(ν, λ) = π3ν2 1 + ν2

sinh2 πν+ ...

L.Cornalba (2007)

We calculate F1(ν) (and confirm ω1(ν)) using the expansion of high-energyamplitudes in Wilson lines (color dipoles)

Light-cone expansion and DGLAP evolution in the NLO

µk > 2 2

µ2 - factorization scale (normalization point)

k2⊥ > µ2 - coefficient functions

k2⊥ < µ2 - matrix elements of light-ray operators (normalized at µ2)

OPE in light-ray operators (x− y)2 → 0

Tjµ(x)jν(y) =xξ

2π2x4

π(ln x2µ2 + C)

]ψ(x)γµγ

ξγν [x, y]ψ(y) + O(1x2 )

[x, y] ≡ Peig∫ 1

0 du (x−y)µAµ(ux+(1−u)y) - gauge link

µk > 2 2

OPE in light-ray operators (x− y)2 → 0

Tjµ(x)jν(y) =xξ

2π2x4

π(ln x2µ2 + C)

]ψ(x)γµγ

ξγν [x, y]ψ(y) + O(1x2 )

[x, y] ≡ Peig∫ 1

0 du (x−y)µAµ(ux+(1−u)y) - gauge link

µk > 2 2

Renorm-group equation for light-ray operators⇒ DGLAP evolution ofparton densities (x− y)2 = 0

µ2 ddµ2 ψ(x)[x, y]ψ(y) = KLOψ(x)[x, y]ψ(y) + αsKNLOψ(x)[x, y]ψ(y)

Four steps of an OPE

To factorize an amplitude into a product of coefficient functionsand matrix elements of relevant operators.To find the evolution equations of the operators with respect tofactorization scale.To solve these evolution equations.To convolute the solution with the initial conditions for the evolutionand get the amplitude

DIS at high energy: relevant operators

At high energies, particles move along straight lines⇒the amplitude of γ∗A→ γ∗A scattering reduces to the matrix element ofa two-Wilson-line operator (color dipole):

A(s) =

∫d2k⊥4π2 IA(k⊥)〈B|TrU(k⊥)U†(−k⊥)|B〉

U(x⊥) = Pexp[ig∫ ∞−∞

du nµAµ(un + x⊥)]

Wilson line

Formally, means the operator expansion in Wilson lines

DIS at high energy: relevant operators

At high energies, particles move along straight lines⇒the amplitude of γ∗A→ γ∗A scattering reduces to the matrix element ofa two-Wilson-line operator (color dipole):

A(s) =

∫d2k⊥4π2 IA(k⊥)〈B|TrU(k⊥)U†(−k⊥)|B〉

U(x⊥) = Pexp[ig∫ ∞−∞

du nµAµ(un + x⊥)]

Wilson line

Formally, means the operator expansion in Wilson linesI. Balitsky (JLAB & ODU) High-energy amplitudes in N=4 SYM Regensburg 15 July 2014 12 / 50

Rapidity factorization

+ +...

η - rapidity factorization scale

Rapidity Y > η - coefficient function (“impact factor”)Rapidity Y < η - matrix elements of (light-like) Wilson lines with rapiditydivergence cut by η

Uηx = Pexp

[ig∫ ∞−∞

dx+Aη+(x+, x⊥)]

Aηµ(x) =

∫d4k

(2π)4 θ(eη − |αk|)e−ik·xAµ(k)

Spectator frame: propagation in the shock-wave background.

Boosted Field

Each path is weighted with the gauge factor Peig∫

dxµAµ

. Quarks and gluonsdo not have time to deviate in the transverse space⇒ we can replace thegauge factor along the actual path with the one along the straight-line path.

z z’

Wilson Line

[ x→ z: free propagation]×[Uab(z⊥) - instantaneous interaction with the η < η2 shock wave]×[ z→ y: free propagation ]

High-energy expansion in color dipoles

+ +...

The high-energy operator expansion is

T jµ(x)jν(y) =

∫d2z1d2z2 ILO

µν (z1, z2, x, y)TrUηz1

U†ηz2

+ NLO contribution

High-energy expansion in color dipoles

+ +...

η - rapidity factorization scale

Evolution equation for color dipoles

trUηx U†ηy =

∫d2z

(x− y)2

(x− z)2(y− z)2 [trUηx U†ηy trUη

x U†ηy

− NctrUηx U†ηy ] + αsKNLOtrUη

x U†ηy + O(α2s )

(Linear part of KNLO = KNLO BFKL)I. Balitsky (JLAB & ODU) High-energy amplitudes in N=4 SYM Regensburg 15 July 2014 15 / 50

Evolution equation for color dipoles

To get the evolution equation, consider the dipole with the rapidies upto η1 and integrate over the gluons with rapidities η1 > η > η2. Thisintegral gives the kernel of the evolution equation (multiplied by thedipole(s) with rapidities up to η2).

αs(η1 − η2)Kevol ⊗

Evolution equation in the leading order

TrUxU†y = KLOTrUxU†y+ ... ⇒

ddη〈TrUxU†y〉shockwave = 〈KLOTrUxU†y〉shockwave

y(a) (b) (c) (d)

x xx* xx

Uabz = TrtaUztbU†z ⇒ (UxU†y)η1 → (UxU†y)η1 +αs(η1−η2)(UxU†z UzU†y)η2

⇒ Evolution equation is non-linear

Non linear evolution equation

U(x, y) ≡ 1− 1Nc

TrU(x⊥)U†(y⊥)

BK equation

ddηU(x, y) =

∫d2z (x− y)2

(x− z)2(y− z)2

U(x, z) + U(z, y)− U(x, y)− U(x, z)U(z, y)

I. B. (1996), Yu. Kovchegov (1999)Alternative approach: JIMWLK (1997-2000)

Non-linear evolution equation

U(x, y) ≡ 1− 1Nc

TrU(x⊥)U†(y⊥)

BK equation

ddηU(x, y) =

∫d2z (x− y)2

(x− z)2(y− z)2

LLA for DIS in pQCD⇒ BFKL (LLA: αs 1, αsη ∼ 1)

Non-linear evolution equation

U(x, y) ≡ 1− 1Nc

TrU(x⊥)U†(y⊥)

BK equation

ddηU(x, y) =

∫d2z (x− y)2

(x− z)2(y− z)2

LLA for DIS in pQCD⇒ BFKL (LLA: αs 1, αsη ∼ 1)

LLA for DIS in sQCD⇒ BK eqn (LLA: αs 1, αsη ∼ 1, αsA1/3 ∼ 1)

(s for semiclassical)

Conformal invariance of the BK equation

Formally, a light-like Wilson line

[∞p1 + x⊥,−∞p1 + x⊥] = Pexp

ig∫ ∞−∞

dx+ A+(x+, x⊥)

is invariant under inversion (with respect to the point with x− = 0).

Indeed,(x+, x⊥)2 = −x2

⊥ ⇒ after the inversion x⊥ → x⊥/x2⊥ and x+ → x+/x2

⊥ ⇒

[∞p1+x⊥,−∞p1+x⊥] → Pexp

ig∫ ∞−∞

x2⊥,

x⊥x2⊥

= [∞p1+x⊥x2⊥,−∞p1+

x⊥x2⊥

⇒The dipole kernel is invariant under the inversion V(x⊥) = U(x⊥/x2⊥)

TrVxV†y =αs

∫d2z6z4

(x− y)2 6z4

(x− z)2(z− y)2 [TrVxV†z TrVzV†y − NcTrVxV†y]

[∞p1 + x⊥,−∞p1 + x⊥] = Pexp

ig∫ ∞−∞

dx+ A+(x+, x⊥)

[∞p1+x⊥,−∞p1+x⊥] → Pexp

ig∫ ∞−∞

x2⊥,

x⊥x2⊥

= [∞p1+x⊥x2⊥,−∞p1+

x⊥x2⊥

TrVxV†y =αs

∫d2z6z4

(x− y)2 6z4

[∞p1 + x⊥,−∞p1 + x⊥] = Pexp

ig∫ ∞−∞

dx+ A+(x+, x⊥)

⊥ ⇒

[∞p1+x⊥,−∞p1+x⊥] → Pexp

ig∫ ∞−∞

x2⊥,

x⊥x2⊥

= [∞p1+x⊥x2⊥,−∞p1+

x⊥x2⊥

TrVxV†y =αs

∫d2z6z4

(x− y)2 6z4

[∞p1 + x⊥,−∞p1 + x⊥] = Pexp

ig∫ ∞−∞

dx+ A+(x+, x⊥)

⊥ ⇒

[∞p1+x⊥,−∞p1+x⊥] → Pexp

ig∫ ∞−∞

x2⊥,

x⊥x2⊥

= [∞p1+x⊥x2⊥,−∞p1+

x⊥x2⊥

TrVxV†y =αs

∫d2z6z4

(x− y)2 6z4

SL(2,C) for Wilson lines

S− ≡i2

(K1 + iK2), S0 ≡i2

(D + iM12), S+ ≡i2

(P1 − iP2)

[S0, S±] = ±S±,12

[S+, S−] = S0,

[S−, U(z, z)] = z2∂zU(z, z), [S0, U(z, z)] = z∂zU(z, z), [S+, U(z, z)] = −∂zU(z, z)

z ≡ z1 + iz2, z ≡ z1 + iz2, U(z⊥) = U(z, z)

Conformal invariance of the evolution kernel

[S−,TrUxU†y] =αsNc

∫dz K(x, y, z)[S−,TrUxU†zTrUzU†y]

⇒[x2 ∂

∂x+ y2 ∂

∂y+ z2 ∂

]K(x, y, z) = 0

In the leading order - OK. In the NLO - ?

S− ≡i2

(K1 + iK2), S0 ≡i2

(D + iM12), S+ ≡i2

(P1 − iP2)

[S0, S±] = ±S±,12

[S+, S−] = S0,

z ≡ z1 + iz2, z ≡ z1 + iz2, U(z⊥) = U(z, z)

⇒[x2 ∂

∂x+ y2 ∂

∂y+ z2 ∂

]K(x, y, z) = 0

In the leading order - OK. In the NLO - ?

S− ≡i2

(K1 + iK2), S0 ≡i2

(D + iM12), S+ ≡i2

(P1 − iP2)

[S0, S±] = ±S±,12

[S+, S−] = S0,

z ≡ z1 + iz2, z ≡ z1 + iz2, U(z⊥) = U(z, z)

⇒[x2 ∂

∂x+ y2 ∂

∂y+ z2 ∂

]K(x, y, z) = 0

In the leading order - OK. In the NLO - ?I. Balitsky (JLAB & ODU) High-energy amplitudes in N=4 SYM Regensburg 15 July 2014 22 / 50

Expansion of the amplitude in color dipoles in the NLO

+ +...

The high-energy operator expansion is

TO(x)O(y) =

∫d2z1d2z2 ILO(z1, z2)TrUη

z1U†ηz2

∫d2z1d2z2d2z3 INLO(z1, z2, z3)[

TrTnUηz1

U†ηz3TnUη

z3U†ηz2 − TrUη

z1U†ηz2]

In the leading order - conf. invariant impact factor

ILO =x−2+ y−2

π2Z21Z2

2, Zi ≡

(x− zi)2⊥

x+− (y− zi)

y+CCP, 2007

NLO impact factor

zz’z1

(a) (b)

INLO(x, y; z1, z2, z3; η) = − ILO × λ

z212z2

[lnσs4Z3 −

The NLO impact factor is not Möbius invariant⇒ the color dipole with thecutoff η is not invariantHowever, if we define a composite operator (a - analog of µ−2 for usual OPE)

[TrUηz1

U†ηz2]conf

= TrUηz1

U†ηz2

∫d2z3

z213z2

23[TrTnUη

z1U†ηz3

TnUηz3

U†ηz2 − NcTrUη

z1U†ηz2] ln

z213z2

23+ O(λ2)

the impact factor becomes conformal in the NLO.I. Balitsky (JLAB & ODU) High-energy amplitudes in N=4 SYM Regensburg 15 July 2014 24 / 50

Operator expansion in conformal dipoles

TO(x)O(y) =

∫d2z1d2z2 ILO(z1, z2)TrUη

z1U†ηz2conf

∫d2z1d2z2d2z3 INLO(z1, z2, z3)[

TrTnUηz1

U†ηz3TnUη

z3U†ηz2 − TrUη

z1U†ηz2]

INLO = − ILO λ

∫dz3

z213z2

z212e2ηas2

z213z2

3 − iπ + 2C]

The new NLO impact factor is conformally invariant⇒ TrUη

z1U†ηz2 conf is Möbius invariant

We think that one can construct the composite conformal dipole operator orderby order in perturbation theory.

Analogy: when the UV cutoff does not respect the symmetry of a local operator,the composite local renormalized operator in must be corrected by finitecounterterms order by order in perturbaton theory.

Definition of the NLO kernel

In general

TrUxU†y = αsKLOTrUxU†y+ α2s KNLOTrUxU†y+ O(α3

α2s KNLOTrUxU†y =

TrUxU†y − αsKLOTrUxU†y+ O(α3s )

We calculate the “matrix element” of the r.h.s. in the shock-wave background

〈α2s KNLOTrUxU†y〉 =

ddη〈TrUxU†y〉 − 〈αsKLOTrUxU†y〉+ O(α3

Subtraction of the (LO) contribution (with the rigid rapidity cutoff)⇒

prescription in the integrals over Feynman parameter v

Typical integral∫ 1

1(k − p)2

⊥v + p2⊥(1− v)

ln(k − p)2

⊥p2⊥

In general

1(k − p)2

⊥v + p2⊥(1− v)

ln(k − p)2

⊥p2⊥

In general

1(k − p)2

⊥v + p2⊥(1− v)

ln(k − p)2

⊥p2⊥

In general

1(k − p)2

⊥v + p2⊥(1− v)

ln(k − p)2

⊥p2⊥

Gluon part of the NLO BK kernel: diagrams

(II) (III) (IV) (V)

(VI) (VII) (VIII) (IX) (X)

(XIV)(XI) (XIII)(XII) (XV)

Diagrams for 1→3 dipoles transition

(XXVI) (XXVII)

(XVI) (XVII) (XVIII) (XIX) (XX)

(XXI) (XXIV) (XV)(XXII) (XXIII)

(XVIII) (XXIX) (XXX)

Diagrams for 1→3 dipoles transition

(XXXI) (XXXIII) (XXXIV)(XXXII)

"Running coupling" diagrams

(I) (II) (III) (IV) (V)

(VI) (VII) (VIII) (IX) (X)

1→ 2 dipole transition diagrams

k’ kk’

(c)(b)(a) (d) (e)

(f) (g) (h) (i) (j)

x x x x*

x x x x

a aaaa

Gluino and scalar loops

Evolution equation for color dipole in N = 4 (I.B. and G. Chirilli)

TrUηz1

U†ηz2

∫d2z3

z213z2

1− αsNc

3+2 ln

]× [TrTaUη

z1U†ηz3

TaUηz3

z1U†ηz2]

− α2s

∫d2z3d2z4

z212z2

z213z2

z212z2

z213z2

24 − z223z2

z213z2

z214z2

× Tr[Ta,Tb]Uηz1

Ta′Tb′U†ηz2+ TbTaUη

z1[Tb′ ,Ta′ ]U†ηz2

(Uηz3

)aa′(Uηz4− Uη

z3)bb′

NLO kernel = Non-conformal term + Conformal term.

Non-conformal term is due to the non-invariant cutoff α < σ = e2η in the rapidityof Wilson lines.

For the conformal composite dipole the result is Möbius invariant

Evolution equation for color dipole in N = 4 (I.B. and G. Chirilli)

TrUηz1

U†ηz2

∫d2z3

z213z2

1− αsNc

3+2 ln

]× [TrTaUη

z1U†ηz3

TaUηz3

z1U†ηz2]

− α2s

∫d2z3d2z4

z212z2

z213z2

z212z2

z213z2

24 − z223z2

z213z2

z214z2

× Tr[Ta,Tb]Uηz1

Ta′Tb′U†ηz2+ TbTaUη

z1[Tb′ ,Ta′ ]U†ηz2

(Uηz3

)aa′(Uηz4− Uη

z3)bb′

NLO kernel = Non-conformal term + Conformal term.

Non-conformal term is due to the non-invariant cutoff α < σ = e2η in the rapidityof Wilson lines.

For the conformal composite dipole the result is Möbius invariant

Evolution equation for composite conformal dipoles in N = 4

[TrUη

z1U†ηz2]conf

∫d2z3

z213z2

[1− αsNc

4ππ2

][TrTaUη

z1U†ηz3

TaUz3U†ηz2 − NcTrUη

z1U†ηz2]conf

− α2s

∫d2z3d2z4

z213z2

24z234

z212z2

z214z2

23+[1 +

z212z2

z213z2

24 − z214z2

z213z2

z214z2

× Tr[Ta,Tb]Uη

z1Ta′Tb′U†ηz2

+ TbTaUηz1

[Tb′ ,Ta′ ]U†ηz2[(Uη

z3)aa′(Uη

z4)bb′ − (z4 → z3)]

Now Möbius invariant!

NLO Amplitude in N=4 SYM theory: factorization in rapidity

<Y<YB YA

y’x’

+ +...

x y x y

y’y’x’

(x− y)4(x′ − y′)4〈TO(x)O†(y)O(x′)O†(y′)〉

∫d2z1⊥d2z2⊥d2z′1⊥d2z′2⊥IFa0(x, y; z1, z2)[DD]a0,b0(z1, z2; z′1, z

′2)IFb0(x′, y′; z′1, z

a0 = x+y+(x−y)2 , b0 =

x′−y′−(x′−y′)2 ⇔ impact factors do not scale with energy

⇒ all energy dependence is contained in [DD]a0,b0 (a0b0 = R)

<Y<YB YA

y’x’

+ +...

x y x y

y’y’x’

(x− y)4(x′ − y′)4〈TO(x)O†(y)O(x′)O†(y′)〉

′2)IFb0(x′, y′; z′1, z

Impact factor (R =(x−y)2z2

12x+y+Z1Z2

- conf. ratio)

IFa0 =

∫dν∫

dz0 R12 +iν

4χ(ν)− 81 + 4ν2 −

cosh2 πν

])Ua0(z0, ν)

Ua(z0, ν) =1π2

∫d2z1d2z2

( z212

z210z2

) 12−iνU(z1, z2), U ≡ 1− 1

NcTrUz1 U†z2

<Y<YB YA

y’x’

+ +...

x y x y

y’y’x’

(x− y)4(x′ − y′)4〈TO(x)O†(y)O(x′)O†(y′)〉

′2)IFb0(x′, y′; z′1, z

Dipole-dipole scattering χ(γ) ≡ 2C − ψ(γ)− ψ(1− γ), γ ≡ 12 + iν

[DD] =

∫dν∫

dz0( z2

z210z2

) 12 +iν( z2

z210z2

) 12−iν

+ iν;λ)Rω(ν)/2

D(γ;λ) =Γ(−γ)Γ(γ − 1)

Γ(1 + γ)Γ(2− γ)

[ 4χ(ν)

1 + 4ν2 −π2

3+ iπ

N2c − 42N2

]+ O(g4)

<Y<YB YA

y’x’

+ +...

x y x y

y’y’x’

(x− y)4(x′ − y′)4〈TO(x)O†(y)O(x′)O†(y′)〉

′2)IFb0(x′, y′; z′1, z

Result : (G.A. Chirilli and I.B.)

F(ν) =N2

N2c − 1

4π4α2s

cosh2 πν

2− 2π2

cosh2 πν− 8

1 + 4ν2 + iπN2

c − 4N2

]I. Balitsky (JLAB & ODU) High-energy amplitudes in N=4 SYM Regensburg 15 July 2014 35 / 50

Conclusions

High-energy operator expansion in color dipoles works at the NLOlevel.

The NLO BK kernel in for the evolution of conformal compositedipoles in N = 4 SYM is Möbius invariant in the transverse plane.The NLO BK kernel agrees with NLO BFKL equation.The correlation function of four Z2 operators is calculated at theNLO BFKL order.

Conclusions

High-energy operator expansion in color dipoles works at the NLOlevel.The NLO BK kernel in for the evolution of conformal compositedipoles in N = 4 SYM is Möbius invariant in the transverse plane.The NLO BK kernel agrees with NLO BFKL equation.The correlation function of four Z2 operators is calculated at theNLO BFKL order.

NLO BFKL and anomalous dimensions of light-ray operators

Gluon operators of leading twist

Ogn ≡ Fa

−i∇n−2− Fai

Anomalous dimension (in gluodynamics)

γn =2παsNc

[− 1

n(n− 1)− 1

(n + 1)(n + 2)+ ψ(n + 1) + γE −

+ O(α2s )

BFKL gives γ(n, αs) at the non-physical point n→ 1

γn =[ALO BFKL

n + ωBNLO BFKLn + ...

](αsNc

)nω ≡ n− 1

LO: Jaroszewicz (1980), NLO: Lipatov, Fadin, Camici, Ciafaloni (1998)

ω = n− 1→ 0 corresponds to Fa−i∇

ω−1− Fai

− - some non-local operator.

Q: which one? A: gluon light-ray (LR) operator

Ogn ≡ Fa

−i∇n−2− Fai

γn =2παsNc

[− 1

n(n− 1)− 1

(n + 1)(n + 2)+ ψ(n + 1) + γE −

+ O(α2s )

γn =[ALO BFKL

](αsNc

)nω ≡ n− 1

ω−1− Fai

Ogn ≡ Fa

−i∇n−2− Fai

γn =2παsNc

[− 1

n(n− 1)− 1

(n + 1)(n + 2)+ ψ(n + 1) + γE −

+ O(α2s )

γn =[ALO BFKL

](αsNc

)nω ≡ n− 1

ω−1− Fai

Ogn ≡ Fa

−i∇n−2− Fai

γn =2παsNc

[− 1

n(n− 1)− 1

(n + 1)(n + 2)+ ψ(n + 1) + γE −

+ O(α2s )

γn =[ALO BFKL

](αsNc

)nω ≡ n− 1

ω−1− Fai

Light-ray operators and parton densities

Gluon light-ray (LR) operator of twist 2

Fa−i(x′+ + x⊥)[x′+, x+]abFb i

− (x+ + x⊥)

Forward matrix element - gluon parton density

zµzν〈p|Faµξ(z)[z, 0]abFbξ

ν (0)|p〉µ z2=0= 2(pz)2

0dxB xBDg(xB, µ) cos(pz)xB

Evolution equation (in gluodynamics)

µ2 ddµ2 Fa

−i(x′+ + x⊥)[x′+, x+]abFb i− (x+ + x⊥)

∫ x′+

x+dz′+

∫ z′+

x+dz+ K(x′+, x+; z′+, z+;αs)Fa

−i(z′+ + x⊥)[z′+, z+]abFb i− (z+ + x⊥)

“Forward” LR operator

F(L+, x⊥) =

∫dx+ Fa

−i(L+ + x+ + x⊥)[L+ + x+, x+]abFbi−(x+ + x⊥)

Light-ray operators and parton densities

Expansion in (“forward”) local operators

F(L+, x⊥) =

∞∑n=2

Ln−2+

(n− 2)!Og

n(x⊥), Ogn ≡

∫dx+Fa

−i∇n−2− Fai

−(x+, x⊥)

Evolution equation for F(L+, x⊥)

dµF(L+, x⊥) =

0du Kgg(u, αs)F(uL+, x⊥)

⇒ γn(αs) = −∫ 1

0du un−2Kgg(u, αs) µ

ddµOg

n = −γn(αs)Ogn

u−1Kgg - DGLAP kernel

u−1Kgg(u) =2αsNc

]+− 2 +

1112δ(u)

)+ higher orders in αs

Light-ray operators

Conformal LR operator (j = 12 + iν)

Fµ(L+, x⊥) =

∫ ∞−∞

dν2π

(L+)−32 +iνFµ1

2 +iν(x⊥)

Fµj (x⊥) =

∫ ∞0

dL+ L1−j+ Fµ(L+, x⊥)

Evolution equation for “forward” conformal light-ray operators

⇒ µ2 ddµ2Fj(z⊥) =

0du Kgg(u, αs)uj−2Fj(z⊥)

⇒ γj(αs) is an analytical continuation of γn(αs)

Supermultiplet of LR operators

Gluino and scalar LR operators

Λ(L+, x⊥) =i2

∫dx′+ [λa(L+ + x+ + x⊥)[x′+ + x+, x+]abσ−∇−λb(x+ + x⊥) + c.c]

Φ(L+, x⊥) =

∫dx′+ φa,I(L+ + x+ + x⊥)[x′+ + x+, x+]ab∇2

−φb,I(x+ + x⊥)

Λj(x⊥) =

∫ ∞0

dL+ L−j+1+ Λ(L+, x⊥), Φj(x⊥) =

∫ ∞0

dL+ L−j+1+ Φ(L+, x⊥)

SU4 singlet LR operators.

S1j(x⊥) = Fj(x⊥) +j− 1

8Λj(x⊥)− j(j− 1)

8Φj(x⊥)

S2j(x⊥) = Fj(x⊥)− 18

Λj(x⊥) +j(j + 1)

24Φj(x⊥)

S3j(x⊥) = Fj(x⊥)− j + 24

Λj(x⊥)− (j + 1)(j + 2)

8Φj(x⊥)

All operators have the same anomalous dimension

γS1j (αs) ≡ γj(αs) =

πNc[ψ(j− 1) + C] + O(α2

s ), γS2j = γS1

j+1, γS3j = γS1

Three-point CF of LR operator and two locals

Three-point correlation function of local operators (O = φaIφ

aI - Konishi

operator)

(µ2z223)−γK 〈S1n(z1)O(z2)O(z3)〉 =

∑ cn[1 + (−1)n]

z212z2

13z223

(z12−

z212−

z13−

)n(µ−2z223

z212z2

) 12γ(n,αs)

x223⊥(µ2x2

23⊥)γK

∫dx1+〈S1n(x1+ , x1⊥)O(x2− , x2⊥)O(x3− , x3⊥)〉

= −2πicn[1 + (−1)n]

x423⊥

Γ(1 + 2n + γn)

Γ2(1 + n + 12γn)

( x223⊥

x2−x3−

)1+ γn2

(x212⊥

x2−+

x213⊥

)−1−n−γn

CF of light-ray operator and two local operators

x223⊥(µ2x2

23⊥)γK 〈S1j(x1⊥)O(x2− , x2⊥)O(x3− , x3⊥)〉

= −2πicj[1 + eiπj]

x423⊥

Γ(1 + 2j + γj)

Γ2(1 + j + 12γj)

( x223⊥

x2−x3−

)1+γj2

(x212⊥

x2−+

x213⊥

)−1−j−γj

Light-cone OPE

CF of light-ray operator and two local operators “in forward kinematics”

x223⊥(µ2x2

23⊥)γK

∫dx3−〈S

µj (x1⊥)O(L− + x3− , x2⊥)O(x3− , x3⊥)〉

c(j, αs)[1 + eiπj]Lj−(uu)j[

x212⊥u + x2

13⊥ u]1+j

( uuµ−2x223⊥

[x212⊥u + x2

13⊥ u]2

) 12γ(j,αs)

Limit x23⊥ ≡ ∆→ 0

∆2(µ2∆2)γK

∫dx3−〈S+j (x1⊥)O(L− + x3− , x2⊥)O(x3− , x3⊥)〉

=Γ2(1 + j + γ

)Γ(2 + 2j + γ)

c(j, αs)[1 + eiπj]Lj−

(x212⊥)1+j(µ2x2

12⊥)−γ(j,αs)(µ∆)γ(j,αs)

Light-cone OPE

CF of light-ray operator and two local operators “in forward kinematics”

x223⊥(µ2x2

23⊥)γK

∫dx3−〈S

µj (x1⊥)O(L− + x3− , x2⊥)O(x3− , x3⊥)〉

c(j, αs)[1 + eiπj]Lj−(uu)j[

x212⊥u + x2

13⊥ u]1+j

( uuµ−2x223⊥

[x212⊥u + x2

13⊥ u]2

) 12γ(j,αs)

Limit x23⊥ ≡ ∆→ 0

∆2(µ2∆2)γK

∫dx3−〈S+j (x1⊥)O(L− + x3− , x2⊥)O(x3− , x3⊥)〉

=Γ2(1 + j + γ

)Γ(2 + 2j + γ)

c(j, αs)[1 + eiπj]Lj−

(x212⊥)1+j(µ2x2

12⊥)−γ(j,αs)(µ∆)γ(j,αs)

Light-cone OPE

Light-ray operators in (+) and (-) directions:

Φ+(L+, x⊥) =

∫dx′+ ∇−φa,I(L+ + x+ + x⊥)[x′+ + x+, x+]ab∇−φb,I(x+ + x⊥)

Φ+j (x⊥) =

∫ ∞0

dL+ L−j+1+ Φ(L+, x⊥)

Φ−(L−, x⊥) =

∫dx′− ∇+φ

a,I(L− + x− + x⊥)[x′− + x−, x−]ab∇+φb,I(x− + x⊥)

Φ−j (x⊥) =

∫ ∞0

dL− L−j+1+ Φ(L−, x⊥)

and similarly for Λj’s and Gj’s

CF of two LRs : 〈S+j= 3

2 +iν(x1⊥)S−

j′= 32 +iν′

(x3⊥)〉 =δ(ν − ν′)a(j, αs)

(x213⊥)j+1(x2

13⊥µ2)γ(j,αs)

“DGLAP” represenation of 4-point CFs

⇒ Light-cone OPE

x212⊥(µ2x2

12⊥)γK

∫dx1+ O(L+ + x2+ , x1⊥)O(x2+ , x2⊥)

∫ 12 +i∞

12−i∞

dj c(j, αs)[1 + eiπj]Lj+(µ2x2

12⊥)γ(j,αs)S+j (x1⊥)

⇒ “DGLAP” result (at x212 → 0)

µ−4(µ4x212x2

34)2+γk

∫dx2+dx3− 〈O(L+ + x2+ , x1⊥)O(x2+ , x2⊥)O(L− + x4− , x3⊥)O(x4− , x4⊥)〉

∫ 12 +i∞

12−i∞

dj C(j, αs)[1 + eiπj]Lj+(µ2x2

12⊥)12γ(j,αs)〈Sµj (x2⊥)O(L− + x4− , x3⊥)O(x4− , x4⊥)〉

∫ 12 +i∞

12−i∞

dj C(j, αs)

[1 + eiπj](L+L−)j[x2

12⊥v + x213⊥ v]1+j

( x212x2

[x213⊥v + x2

14⊥ v]2

) 12γ(j,αs)

(vv)j+ 12γ(j,αs)

⇒ Light-cone OPE

x212⊥(µ2x2

12⊥)γK

∫dx1+ O(L+ + x2+ , x1⊥)O(x2+ , x2⊥)

∫ 12 +i∞

12−i∞

12⊥)γ(j,αs)S+j (x1⊥)

µ−4(µ4x212x2

34)2+γk

∫ 12 +i∞

12−i∞

∫ 12 +i∞

12−i∞

dj C(j, αs)

12⊥v + x213⊥ v]1+j

( x212x2

[x213⊥v + x2

14⊥ v]2

) 12γ(j,αs)

(vv)j+ 12γ(j,αs)

⇒ Light-cone OPE

x212⊥(µ2x2

12⊥)γK

∫dx1+ O(L+ + x2+ , x1⊥)O(x2+ , x2⊥)

∫ 12 +i∞

12−i∞

12⊥)γ(j,αs)S+j (x1⊥)

µ−4(µ4x212x2

34)2+γk

∫ 12 +i∞

12−i∞

∫ 12 +i∞

12−i∞

dj C(j, αs)

12⊥v + x213⊥ v]1+j

( x212x2

[x213⊥v + x2

14⊥ v]2

) 12γ(j,αs)

(vv)j+ 12γ(j,αs)

⇒ Light-cone OPE

x212⊥(µ2x2

12⊥)γK

∫dx1+ O(L+ + x2+ , x1⊥)O(x2+ , x2⊥)

∫ 12 +i∞

12−i∞

12⊥)γ(j,αs)S+j (x1⊥)

µ−4(µ4x212x2

34)2+γk

∫ 12 +i∞

12−i∞

∫ 12 +i∞

12−i∞

dj C(j, αs)

12⊥v + x213⊥ v]1+j

( x212x2

[x213⊥v + x2

14⊥ v]2

) 12γ(j,αs)

(vv)j+ 12γ(j,αs)

Light-cone limit of Cornalba’s formula

Cornalba’s formula as x212⊥ → 0

[x212x2

34]2+γK

8π2L+L−

∫dν

tanhπνν cosh2 πν

( x212⊥x2

34⊥uuvv[x2

13⊥v + x214v]2

) 12 +iν(L2

+L2−uuvv

x212⊥x2

)ℵ(ν)/2f+

8π2∫ 1

0dv∫ 1

2 +i∞

12−i∞

dξ f(ℵ(ξ)

)ξ ≡ 1

2+ iν

× (vv)1−ξ+ℵ(ξ)2 cosπξ(ξ − 1

2 ) sin3 πξ

B(2− ξ + ℵ(ξ)

13⊥v + x214v]2+ℵ(ξ)

( x212⊥x2

34⊥[x2

13⊥v + x214v]2

)−ξ−ℵ(ξ)2(L+L−)1+ℵ(ξ)

= “BFKL” representation of the amplitude in the (Regge) + (x212 → 0) limit

DGLAP vs BFKL for 4-point CFs

“DGLAP” result (leading twist)

µ−4(µ4x212x2

34)2+γk

∫ 12 +i∞

12−i∞

dj C(j, αs)

12⊥v + x213⊥ v]1+j

( x212x2

[x213⊥v + x2

14⊥ v]2

) 12γ(j,αs)

(vv)j+ 12γ(j,αs)

BFKL result (ℵ(ξ, αs) = αsNcπ χ(ξ) + ... - pomeron intercept)

[x212x2

34]2+γK

8π2∫ 1

0dv∫ 1

2 +i∞

12−i∞

dξ f(ℵ(ξ)

)ξ ≡ 1

2+ iν

2 ) sin3 πξ

B(2− ξ + ℵ(ξ)

13⊥v + x214v]2+ℵ(ξ)

( x212⊥x2

34⊥[x2

13⊥v + x214v]2

)−ξ−ℵ(ξ)2(L+L−)1+ℵ(ξ)

We compare these results around ξ ∼ j− 1 ∼ 0

⇒ 1 + ℵ(ξ, αs) = j and γ(j, αs) = −2ξ − ℵ(ξ)

ℵ(ξ)−2ℵ(ξ)ℵ′(ξ) ' αsNc

πξ+ζ(3)α2

ξ+... → γj = −2

π(j− 1)+ [0 + ζ(3)(j− 1)]

( αsNc

π(j− 1)))3

µ−4(µ4x212x2

34)2+γk

∫ 12 +i∞

12−i∞

dj C(j, αs)

12⊥v + x213⊥ v]1+j

( x212x2

[x213⊥v + x2

14⊥ v]2

) 12γ(j,αs)

(vv)j+ 12γ(j,αs)

[x212x2

34]2+γK

8π2∫ 1

0dv∫ 1

2 +i∞

12−i∞

dξ f(ℵ(ξ)

)ξ ≡ 1

2+ iν

2 ) sin3 πξ

B(2− ξ + ℵ(ξ)

13⊥v + x214v]2+ℵ(ξ)

( x212⊥x2

34⊥[x2

13⊥v + x214v]2

)−ξ−ℵ(ξ)2(L+L−)1+ℵ(ξ)

πξ+ζ(3)α2

ξ+... → γj = −2

π(j− 1)+ [0 + ζ(3)(j− 1)]

( αsNc

π(j− 1)))3

µ−4(µ4x212x2

34)2+γk

∫ 12 +i∞

12−i∞

dj C(j, αs)

12⊥v + x213⊥ v]1+j

( x212x2

[x213⊥v + x2

14⊥ v]2

) 12γ(j,αs)

(vv)j+ 12γ(j,αs)

[x212x2

34]2+γK

8π2∫ 1

0dv∫ 1

2 +i∞

12−i∞

dξ f(ℵ(ξ)

)ξ ≡ 1

2+ iν

2 ) sin3 πξ

B(2− ξ + ℵ(ξ)

13⊥v + x214v]2+ℵ(ξ)

( x212⊥x2

34⊥[x2

13⊥v + x214v]2

)−ξ−ℵ(ξ)2(L+L−)1+ℵ(ξ)

πξ+ζ(3)α2

ξ+... → γj = −2

π(j− 1)+ [0 + ζ(3)(j− 1)]

( αsNc

π(j− 1)))3

Result

γj = −2αsNc

π(j− 1)+ [0 + ζ(3)(j− 1)]

( αsNc

π(j− 1)))3

is an anomalous dimension of light-ray operator F∇j−2F(x⊥)

Conclusion and Outlook

1 Conclusion:NLO BFKL gives anomalous dimensions of light-ray operatorsF+i∇ω−1F i

+ as ω → 0 (in all orders in αs)

2 OutlookIn QCD3-point CF of 〈F∇ω1−1F(x1⊥) F∇ω2−1F(x2⊥) F∇ω3−1F(x3⊥)〉as ω1, ω2, ω3 → 0 (joint project with V. Kazakov and E. Sobko):

Conclusion and Outlook

1 Conclusion:NLO BFKL gives anomalous dimensions of light-ray operatorsF+i∇ω−1F i

+ as ω → 0 (in all orders in αs)2 Outlook

In QCD3-point CF of 〈F∇ω1−1F(x1⊥) F∇ω2−1F(x2⊥) F∇ω3−1F(x3⊥)〉as ω1, ω2, ω3 → 0 (joint project with V. Kazakov and E. Sobko):

high-energy amplitudes in n=4 sym - uni-regensburg.demaa29312/conference/talks/balitsky.pdf · i....

Documents

the dilatation operator of n=4 sym, amplitudes and yangian...

image amplitudes in reverse time migration/ · pdf fileimage...

all-multiplicity non-planar amplitudes in sym at 2-loops ·...

mellin amplitudes: the scattering amplitudes of...

a new identity for gauge theory amplitudes · a new...

henriette elvang, daniel z. freedman and michael kiermaier...

amplitudes térmicas no deserto

sym handleiding 400

parallel sym

allone-loopnmhvgluonamplitudesin n = 1 sym · 3.2 example:...

seismic amplitudes correction

geometric formulation of scattering amplitudes in n=4...

celestial amplitudes

high-energy qcd and wilson lines - los alamos … qcd and...

lectures on superstring amplitudes€¦ · lectures on...

springfield sym

sym sr150cc

generating tree amplitudes in n = 4 sym and n = 8 sg ·...

2009 - sym

amplitudes, for piano (draft)