SLIDE 1 Perturbative spectra in gauge theories with gravity duals
Christoph Sieg
Niels Bohr International Academy Niels Bohr Institute
22.02.10, Nordic String Meeting, Hannover C.S., A. Torrielli: 0505071
- F. Fiamberti, A. Santambrogio, C.S., D. Zanon:
0712.3522 0806.2095 0806.2103 0811.4594
- F. Fiamberti, A. Santambrogio, C.S.:
0908.0234
- J. Minahan, O. Ohlsson Sax, C.S.:
0908.2463 0912.3460
SLIDE 2
Outline
Introduction and overview Perturbative calculations Conclusions and outlook
SLIDE 3 AdS/CFT correspondence
energy E (semicl.) strings
- anom. dim. γ
- comp. operators
OL ∝ tr φi1 . . . φiL integrable systems (asymptot.) Bethe ansätze
[Beisert, Dippel, Staudacher] [Arutyunov, Frolov, Staudacher]
string gauge
dressing phase
[Beisert, Hernandez, Lopez] [Beisert, Eden, Staudacher] [Minahan, Zarembo] [Bena, Polchinski, Roiban] [Kazakov, Marshakov, Minahan, Zarembo]
λ ≫ 1 λ ≪ 1
SLIDE 4 AdS/CFT correspondence
energy E (semicl.) strings
- anom. dim. γ
- comp. operators
OS ∝ tr φ DS φ integrable systems (asymptot.) Bethe ansätze
[Beisert, Dippel, Staudacher] [Arutyunov, Frolov, Staudacher]
string gauge
dressing phase
[Beisert, Hernandez, Lopez] [Beisert, Eden, Staudacher]
dilatation operator
[Beisert, Kristjansen, Staudacher]
D OL = γ OL Feynman graph computations in the flavour SU(2) subsector: 1-loop:
[Berenstein, Maldacena, Nastase]
2-loops: [Gross, Mikhailov, Roiban] checks at higher loops:
[Gross, Mikhailov, Roiban] [Beisert, McLoughlin, Roiban] [Fiamberti, Santambrogio, CS, Zanon] [Fiamberti, Santambrogio, CS] [Minahan, Zarembo] [Bena, Polchinski, Roiban] [Kazakov, Marshakov, Minahan, Zarembo]
λ ≫ 1 λ ≪ 1 [D,Qi] = 0
SLIDE 5 AdS/CFT correspondence
energy E (semicl.) strings
- anom. dim. γ
- comp. operators
OS ∝ tr φ DS φ integrable systems (asymptot.) Bethe ansätze
[Beisert, Dippel, Staudacher] [Arutyunov, Frolov, Staudacher]
string gauge
dressing phase
[Beisert, Hernandez, Lopez] [Beisert, Eden, Staudacher]
integral eq. for f(λ)
[Eden, Staudacher]
√ λ π − 3 π ln 2 λ 2π2 − ζ(2) λ2 16π4
[Benna, Benvenuti, Klebanov, Sardicchio] [Casteill, Kristjansen] [Alday, Arutyunov, Benna, Eden, Klebanov] [Basso, Korchemsky, Kotanski] [Minahan, Zarembo] [Bena, Polchinski, Roiban] [Kazakov, Marshakov, Minahan, Zarembo]
λ ≫ 1 λ ≪ 1 S ≫ 1 λ ≫ 1 λ ≪ 1
[Gubser, Klebanov, Polyakov] [Frolov, Tseytlin] [Kruczenski] [Kotikov, Lipatov, Velizhanin [Makeenko]
SLIDE 6 AdS/CFT correspondence
energy E (semicl.) strings
- anom. dim. γ
- comp. operators
OS ∝ tr φ DS φ integrable systems (asymptot.) Bethe ansätze
[Beisert, Dippel, Staudacher] [Arutyunov, Frolov, Staudacher]
string gauge
dressing phase
[Beisert, Hernandez, Lopez] [Beisert, Eden, Staudacher]
Konishi multiplet tr φiφi tr φ D2 φ tr φ [φ,Z] Z
γasy =
λkγk
[Minahan, Zarembo] [Bena, Polchinski, Roiban] [Kazakov, Marshakov, Minahan, Zarembo]
λ ≫ 1 λ ≪ 1 S = 2
SLIDE 7 AdS/CFT correspondence
energy E (semicl.) strings
- anom. dim. γ
- comp. operators
OS ∝ tr φ DS φ integrable systems (asymptot.) Bethe ansätze
[Beisert, Dippel, Staudacher] [Arutyunov, Frolov, Staudacher]
string gauge
dressing phase
[Beisert, Hernandez, Lopez] [Beisert, Eden, Staudacher]
Konishi multiplet tr φiφi tr φ D2 φ tr φ [φ,Z] Z
γasy =
λkγk finite size effects wrapping interactions
[Ambjørn, Janik, Kristjansen] [Schafer-Nameki] [Janik, Lukowski] [C.S.,Torrielli]
γ = γasy + λ4γ4 + . . .
[Arutyunov, Frolov]
Y-system
[Gromov, Kazakov, Vieira]
(gen.) Lüscher formula
[Bajnok,Janik]
L = 4 loop Feynman graphs
[Fiamberti,Santambrogio,C.S.,Zanon] [Minahan, Zarembo] [Bena, Polchinski, Roiban] [Kazakov, Marshakov, Minahan, Zarembo]
λ ≫ 1 λ ≪ 1 S = 2
SLIDE 8 AdS4/CFT3 (ABJM) correspondence
[Aharony, Bergman, Jafferis, Maldacena]
Type II A ST AdS4 × CP3 ⇔ 3-dim. N = 6 CS theory
energy E (semicl.) strings
- anom. dim. γ
- comp. operators
OJ ∝ tr φn1 . . . φnJ integrable systems (asymptot.) Bethe ansätze
[Gromov, Vieira] [Ahn, Nepomechie]
magnon dispersion relation E =
2 − Q
[Beisert, Dippel, Staudacher] [Beisert] [Arutyunov, Frolov, Zamaklar]
h(λ)2 = ? h(λ)2 =
λ 4π2
unexpectedly simple in AdS5/CFT4
[Minahan, Zarembo] [Bak, Rey] [Arutyunov, Frolov] [Stefanski] [Gromov, Vieira]
λ ≫ 1 λ ≪ 1
SLIDE 9 AdS4/CFT3 (ABJM) correspondence
[Aharony, Bergman, Jafferis, Maldacena]
Type II A ST AdS4 × CP3 ⇔ 3-dim. N = 6 CS theory
energy E (semicl.) strings
- anom. dim. γ
- comp. operators
OJ ∝ tr φn1 . . . φnJ integrable systems (asymptot.) Bethe ansätze
[Gromov, Vieira] [Ahn, Nepomechie]
magnon dispersion relation E =
2 − Q
[Beisert, Dippel, Staudacher] [Beisert] [Arutyunov, Frolov, Zamaklar]
BMN limit giant magnons
[Nishioka, Takayanagi] [Gaiotto, Giombi, Yin] [Grignani, Harmark, Orselli]
quantum corr.
[McLoughlin, Roiban, Tseytlin]
two loops
[Nishioka, Takayanagi] [Minahan, Zarembo] [Bak, Rey]
λ 2 −
2 ln 2 π
λ2 h(λ)2 = ? h(λ)2 =
λ 4π2
unexpectedly simple in AdS5/CFT4
[Minahan, Zarembo] [Bak, Rey] [Arutyunov, Frolov] [Stefanski] [Gromov, Vieira]
λ ≫ 1 λ ≪ 1
SLIDE 10 AdS4/CFT3 (ABJM) correspondence
[Aharony, Bergman, Jafferis, Maldacena]
Type II A ST AdS4 × CP3 ⇔ 3-dim. N = 6 CS theory
energy E (semicl.) strings
- anom. dim. γ
- comp. operators
OJ ∝ tr φn1 . . . φnJ integrable systems (asymptot.) Bethe ansätze
[Gromov, Vieira] [Ahn, Nepomechie]
magnon dispersion relation E =
2 − Q
[Beisert, Dippel, Staudacher] [Beisert] [Arutyunov, Frolov, Zamaklar]
BMN limit giant magnons
[Nishioka, Takayanagi] [Gaiotto, Giombi, Yin] [Grignani, Harmark, Orselli]
quantum corr.
[McLoughlin, Roiban, Tseytlin]
two loops
[Nishioka, Takayanagi] [Minahan, Zarembo] [Bak, Rey]
λ 2 −
2 ln 2 π
λ2 + λ4(−16 + 4ζ(2)) h(λ)2 = ? h(λ)2 =
λ 4π2
unexpectedly simple in AdS5/CFT4 four loops
[Minahan, Ohlsson Sax, C.S.] [Minahan, Zarembo] [Bak, Rey] [Arutyunov, Frolov] [Stefanski] [Gromov, Vieira]
λ ≫ 1 λ ≪ 1
SLIDE 11 Renormalization of composite operators
composite operator OL = L
- f length L (with L scalar fields)
two-point functions of composite operators: tree level
L (x), 1 , OB L (y)
x y = δAB (x − y)2∆ , ∆ = L
SLIDE 12 Renormalization of composite operators
composite operator OL = L
- f length L (with L scalar fields)
two-point functions of composite operators: with loop corrections
L (x), V2L, OB L (y)
x
V
y = δAB (x − y)2∆ , ∆ = L + γ + . . .
SLIDE 13 Renormalization of composite operators
composite operator OL = L
- f length L (with L scalar fields)
two-point functions of composite operators: with loop corrections
L (x), V2L, OB L (y)
x
V
y = δAB (x − y)2∆ , ∆ = L + γ + . . . renormalization of composite operators in a CFT in D = 4 − 2ε dimensions Oa
L,ren = Za bOb L,bare ,
D = µ d dµ ln Z(λµ2ε)
SLIDE 14 Renormalization of composite operators
composite operator OL = L
- f length L (with L scalar fields)
two-point functions of composite operators: with loop corrections
L (x), V2L, OB L (y)
x
V
y = δAB (x − y)2∆ , ∆ = L + γ + . . . renormalization of composite operators in a CFT in D = 4 − 2ε dimensions Oa
L,ren = Za bOb L,bare ,
D = µ d dµ ln Z(λµ2ε) anomalous dimensions: eigenvalues of the dilatation operator D =
λkDk D OL = γ OL
SLIDE 15
Bethe ansatz in the flavour SU(2) subsector
complex fields: φ =
1 √ 2(φ1 + iφ2), ψ = 1 √ 2(φ3 + iφ4), Z = 1 √ 2(φ5 + iφ6)
ψ only as internal flavour in Feynman diagrams
SLIDE 16 Bethe ansatz in the flavour SU(2) subsector
complex fields: φ =
1 √ 2(φ1 + iφ2), ψ = 1 √ 2(φ3 + iφ4), Z = 1 √ 2(φ5 + iφ6)
ψ only as internal flavour in Feynman diagrams map to integrable spin chain of length L OL = tr(φ . . . φ
M
ZZZ . . . Z
) ↔ BPS operator tr(Z . . . Z) ↔ ferromagnetic vaccum impurities φ ↔ spin excitations (magnons) dilatation operator D ↔ Hamiltonian H anomalous dimensions γ ↔ energies E
SLIDE 17 Bethe ansatz in the flavour SU(2) subsector
complex fields: φ =
1 √ 2(φ1 + iφ2), ψ = 1 √ 2(φ3 + iφ4), Z = 1 √ 2(φ5 + iφ6)
ψ only as internal flavour in Feynman diagrams map to integrable spin chain of length L OL = tr(φ . . . φ
M
ZZZ . . . Z
) ↔ BPS operator tr(Z . . . Z) ↔ ferromagnetic vaccum impurities φ ↔ spin excitations (magnons) dilatation operator D ↔ Hamiltonian H = HXXX1/2 + . . . anomalous dimensions γ ↔ energies E
- perator mixing problem solved by the asymptotic Bethe ansatz
M
pj = 0 , eipjL =
M
ˆ S(uj, uk)e2iθ(uj, uk) , E =
M
π2 sin2 pj 2 − 1
conservation matrix part dressing phase single magnon dispersion relation two-particle S-matrix
SLIDE 18 One loop
i tr(ψ [Z,φ]) = i
−i tr( ¯ ψ [¯ φ,¯ Z]) = −i
−
b
=
b
−
b
= =
SLIDE 19 One loop
i tr(ψ [Z,φ]) = i
−i tr( ¯ ψ [¯ φ,¯ Z]) = −i
−
b
= −
λ (4π)2ε
b
−
b
= finite = finite D1 = 2
λ (4π)2
−
L
Pi i+1
SLIDE 20 Chiral functions
b
→ − + = − χ(1) + {} {1}
b b
→ − − + = χ(1, 2) {} − {1} − {1} + {1, 2}
b b b
{a1, . . . , an} =
L
Pi+a1 i+a1+1 . . . Pi+an i+an+1 = χ(1, 2, 3) − {} + 3{1} − 2{1, 2} − {1, 3} + {1, 2, 3}
SLIDE 21
Two loops
◮ all diagrams (apart from reflections, one-loop wave function ren.) ◮ finiteness Fiamberti, Santambrogio, CS, Zanon ◮ generalized finiteness ⇒ absence of χ() to all orders CS, to appear
R = 1 R = 2 R = 3 χ() χ(1) − χ(1, 2) − −
SLIDE 22
Two loops
◮ all diagrams (apart from reflections, one-loop wave function ren.) ◮ finiteness Fiamberti, Santambrogio, CS, Zanon ◮ generalized finiteness ⇒ absence of χ() to all orders CS, to appear
R = 1 R = 2 R = 3 χ() χ(1) − χ(1, 2) − −
SLIDE 23
Two loops
◮ all diagrams (apart from reflections, one-loop wave function ren.) ◮ finiteness Fiamberti, Santambrogio, CS, Zanon ◮ generalized finiteness ⇒ absence of χ() to all orders CS, to appear
R = 1 R = 2 R = 3 χ() χ(1) − χ(1, 2) − − D2 = 4χ(1) − 2[χ(1, 2) + χ(2, 1)]
SLIDE 24 Checks at higher loops
Feynman diagrams with χ(a1, . . . , an) at k loops are simplest if: purely chiral → n = k
- f maximum range → maxa1,...,an − mina1,...,an = k − 1
D4 = + 200χ(1) − 150[χ(1, 2) + χ(2, 1)] + 8(10 + ǫ3a)χ(1, 3) − 4χ(1, 4) + 60[χ(1, 2, 3) + χ(3, 2, 1)] + (8 + 2β + 4ǫ3a − 4iǫ3b + 2iǫ3c − 2iǫ3d)χ(1, 3, 2) + (8 + 2β + 4ǫ3a + 4iǫ3b − 2iǫ3c + 2iǫ3d)χ(2, 1, 3)] − (4 + 4iǫ3b + 2iǫ3c)[χ(1, 2, 4) + χ(1, 4, 3)] − (4 − 4iǫ3b − 2iǫ3c)[χ(1, 3, 4) + χ(2, 1, 4)] − (12 + 2β + 4ǫ3a)χ(2, 1, 3, 2) + (18 + 4ǫ3a)[χ(1, 3, 2, 4) + χ(2, 1, 4, 3)] − (8 + 2ǫ3a + 2iǫ3b)[χ(1, 2, 4, 3) + χ(1, 4, 3, 2)] − (8 + 2ǫ3a − 2iǫ3b)[χ(2, 1, 3, 4) + χ(3, 2, 1, 4)] − 10[χ(1, 2, 3, 4) + χ(4, 3, 2, 1)]
SLIDE 25 Checks at higher loops
Feynman diagrams with χ(a1, . . . , an) at k loops are simplest if: purely chiral → n = k
- f maximum range → maxa1,...,an − mina1,...,an = k − 1
D4 = + 200χ(1) − 150[χ(1, 2) + χ(2, 1)] + 8(10 + ǫ3a)χ(1, 3) − 4χ(1, 4) + 60[χ(1, 2, 3) + χ(3, 2, 1)] + (8 + 2β + 4ǫ3a − 4iǫ3b + 2iǫ3c − 2iǫ3d)χ(1, 3, 2) + (8 + 2β + 4ǫ3a + 4iǫ3b − 2iǫ3c + 2iǫ3d)χ(2, 1, 3)] − (4 + 4iǫ3b + 2iǫ3c)[χ(1, 2, 4) + χ(1, 4, 3)] − (4 − 4iǫ3b − 2iǫ3c)[χ(1, 3, 4) + χ(2, 1, 4)] − (12 + 2β + 4ǫ3a)χ(2, 1, 3, 2) + (18 + 4ǫ3a)[χ(1, 3, 2, 4) + χ(2, 1, 4, 3)] − (8 + 2ǫ3a + 2iǫ3b)[χ(1, 2, 4, 3) + χ(1, 4, 3, 2)] − (8 + 2ǫ3a − 2iǫ3b)[χ(2, 1, 3, 4) + χ(3, 2, 1, 4)] − 10[χ(1, 2, 3, 4) + χ(4, 3, 2, 1)]
β = 4ζ(3) is leading coeff. of the dressing phase
Beisert, McLoughlin, Roiban
SLIDE 26 Checks at higher loops
Feynman diagrams with χ(a1, . . . , an) at k loops are simplest if: purely chiral → n = k
- f maximum range → maxa1,...,an − mina1,...,an = k − 1
D4 = + 200χ(1) − 150[χ(1, 2) + χ(2, 1)] + 8(10 + ǫ3a)χ(1, 3) − 4χ(1, 4) + 60[χ(1, 2, 3) + χ(3, 2, 1)] + (8 + 2β + 4ǫ3a − 4iǫ3b + 2iǫ3c − 2iǫ3d)χ(1, 3, 2) + (8 + 2β + 4ǫ3a + 4iǫ3b − 2iǫ3c + 2iǫ3d)χ(2, 1, 3)] − (4 + 4iǫ3b + 2iǫ3c)[χ(1, 2, 4) + χ(1, 4, 3)] − (4 − 4iǫ3b − 2iǫ3c)[χ(1, 3, 4) + χ(2, 1, 4)] − (12 + 2β + 4ǫ3a)χ(2, 1, 3, 2) + (18 + 4ǫ3a)[χ(1, 3, 2, 4) + χ(2, 1, 4, 3)] − (8 + 2ǫ3a + 2iǫ3b)[χ(1, 2, 4, 3) + χ(1, 4, 3, 2)] − (8 + 2ǫ3a − 2iǫ3b)[χ(2, 1, 3, 4) + χ(3, 2, 1, 4)] − 10[χ(1, 2, 3, 4) + χ(4, 3, 2, 1)]
important for leading wrapping correction
Fiamberti, Santambrogio, CS, Zanon
SLIDE 27 Finite size effects / wrapping interactions
dilatation operator Dk = k + 1
- Dk
- k + 1 at order k, i.e. ∼ λk
composite operator OL = L
action of the dilatation operator: k < L: DkOL =
Dk
+
Dk
+ . . . +
Dk
+ . . . k ≥ L: DkOL =
Dk
+
Dk
universal part finite size effects: wrapping interactions
Dk from integrability: only in the asymptotic limit k < L
SLIDE 28 Calculation of wrapping effects
Fiamberti, Santambrogio, C.S., Zanon: 0712.3522, 0806.2095, 0806.2103, 0811.4594 Fiamberti, Santambrogio, C.S.: 0908.0234
- 1. analyze the properties of the wrapping interactions
C.S., Torrielli: 0505071
- 2. use efficient formalism → N = 1 superfields
- 3. compute all k-loop diagrams? No!
use known asymptotic dilatation operator Dk
◮ correct it for the application to OL with k = L ◮ need appropriate basis for flavour permutations
→ chiral functions
◮ only have to compute subtraction and wrapping
- 4. compute the divergences of the loop integrals analytically
◮ we improved the Gegenbauer polynomial x-space technique
(correct treatment of traceless products in numerators)
◮ we introduced recursion chains for the radial integrations,
→ at k = 11 loops: 225 975 instead of 39 916 800 terms
SLIDE 29
Our results as tests of AdS/CFT
Fiamberti, Santambrogio, C.S., Zanon: 0712.3522, 0806.2095 ◮ four-loop result of O4 = tr(φ [φ,Z] Z)
γ4 = −2496 + 576ζ(3) − 1440ζ(5)
◮ matches result from the string integrable model [Bajnok, Janik]
now available also at five-loops
[Bajnok, Hegedus, Janik, Lukowski]
SLIDE 30
Our results as tests of AdS/CFT
Fiamberti, Santambrogio, C.S., Zanon: 0712.3522, 0806.2095 ◮ four-loop result of O4 = tr(φ [φ,Z] Z)
γ4 = −2496 + 576ζ(3) − 1440ζ(5)
◮ matches result from the string integrable model [Bajnok, Janik]
now available also at five-loops
[Bajnok, Hegedus, Janik, Lukowski] Fiamberti, Santambrogio, C.S.: 0908.0234 ◮ five-loop result of O5 = tr(φ [φ,Z] ZZ)
γ5 = 6664 + 1152ζ(3) + 3840ζ(5) − 2240ζ(7)
◮ matches result from the string integrable model [Beccaria, Forini, Lukowski, Zieme]
SLIDE 31
Our results as tests of AdS/CFT
Fiamberti, Santambrogio, C.S., Zanon: 0712.3522, 0806.2095 ◮ four-loop result of O4 = tr(φ [φ,Z] Z)
γ4 = −2496 + 576ζ(3) − 1440ζ(5)
◮ matches result from the string integrable model [Bajnok, Janik]
now available also at five-loops
[Bajnok, Hegedus, Janik, Lukowski] Fiamberti, Santambrogio, C.S.: 0908.0234 ◮ five-loop result of O5 = tr(φ [φ,Z] ZZ)
γ5 = 6664 + 1152ζ(3) + 3840ζ(5) − 2240ζ(7)
◮ matches result from the string integrable model [Beccaria, Forini, Lukowski, Zieme] Fiamberti, Santambrogio, C.S., Zanon: 0806.2103, 0811.4594 ◮ L ≤ 11 loop results of OL = tr(φZ . . . Z) in β-deformed N = 4 SYM ◮ match results from the string integrable model [Beccaria, De Angelis]
SLIDE 32 Conclusions and outlook
◮ perturbative computations important:
first computation of the 4-loop anomalous dimension of the Konishi
◮ N = 1 supergraphs is an efficient tool (finiteness theorems) ◮ refined tests at four and five loops ◮ wrapping increases transcendentality (degree of harmonic sums) ◮ all results confirm the duality and are in accord with the Y-system ◮ still non-trivial cancellations and simplifications
→ more efficient formalism: required for calculations at higher orders and beyond the critical wrapping orders Thank you!