Duality in 2D dilaton gravity
based uponhep-th/0609197with Roman Jackiw
Daniel Grumiller
CTP, LNS, MIT, Cambridge, Massachusetts
Supported by the European Commission, Project MC-OIF 021421
Brown University, December 2006
Outline
1 Gravity in 2D Models in 2D
Generic dilaton gravity action
2 Vienna School Approach Gravity as gauge theory All classical solutions
3 Duality
Casimir exchange Applications
Outline
1 Gravity in 2D Models in 2D
Generic dilaton gravity action
2 Vienna School Approach Gravity as gauge theory All classical solutions
3 Duality
Casimir exchange Applications
Geometry in 2D
As simple as possible but not simpler...
Riemann (Weyl+Ricci): N2(N122−1) components in N dimensions 4D: 20 (10 Weyl and 10 Ricci)
5D: 50 (35 Weyl and 15 Ricci) 10D: 825 (770 Weyl and 55 Ricci) 11D: 1210 (1144 Weyl and 66 Ricci) 3D: 6 (Ricci)
2D: 1 (Ricci scalar)→Lowest dimension with curvature 1D: 0
But: 2D Einstein-Hilbert: no equations of motion!
Number of graviton modes: N(N−3)2
Stuck already in the formulation of the model?
Have to go beyond Einstein-Hilbert in 2D!
Geometry in 2D
As simple as possible but not simpler...
Riemann (Weyl+Ricci): N2(N122−1) components in N dimensions 4D: 20 (10 Weyl and 10 Ricci)
5D: 50 (35 Weyl and 15 Ricci) 10D: 825 (770 Weyl and 55 Ricci) 11D: 1210 (1144 Weyl and 66 Ricci) 3D: 6 (Ricci)
2D: 1 (Ricci scalar)→Lowest dimension with curvature 1D: 0
But: 2D Einstein-Hilbert: no equations of motion!
Number of graviton modes: N(N−3)2
Stuck already in the formulation of the model?
Have to go beyond Einstein-Hilbert in 2D!
Geometry in 2D
As simple as possible but not simpler...
Riemann (Weyl+Ricci): N2(N122−1) components in N dimensions 4D: 20 (10 Weyl and 10 Ricci)
5D: 50 (35 Weyl and 15 Ricci) 10D: 825 (770 Weyl and 55 Ricci) 11D: 1210 (1144 Weyl and 66 Ricci) 3D: 6 (Ricci)
2D: 1 (Ricci scalar)→Lowest dimension with curvature 1D: 0
But: 2D Einstein-Hilbert: no equations of motion!
Number of graviton modes: N(N−3)2
Stuck already in the formulation of the model?
Have to go beyond Einstein-Hilbert in 2D!
Geometry in 2D
As simple as possible but not simpler...
Riemann (Weyl+Ricci): N2(N122−1) components in N dimensions 4D: 20 (10 Weyl and 10 Ricci)
5D: 50 (35 Weyl and 15 Ricci) 10D: 825 (770 Weyl and 55 Ricci) 11D: 1210 (1144 Weyl and 66 Ricci) 3D: 6 (Ricci)
2D: 1 (Ricci scalar)→Lowest dimension with curvature 1D: 0
But: 2D Einstein-Hilbert: no equations of motion!
Number of graviton modes: N(N−3)2
Stuck already in the formulation of the model?
Have to go beyond Einstein-Hilbert in 2D!
Spherical reduction
Line element adapted to spherical symmetry:
ds2= gµν(N)
|{z}
full metric
dxµdxν =gαβ(xγ)
| {z }
2Dmetric
dxαdxβ− φ2(xα)
| {z }
surface area
dΩ2S
N−2,
Insert into N-dimensional EH action IEH =κR
dNxp
−g(N)R(N): IEH =κ2π(N−1)/2
Γ(N−12 )
| {z }
N−2sphere
Z
d2x√
−gφN−2
| {z }
determinant
h
R+(N−2)(N−3) φ2
(∇φ)2−1
| {z }
Ricci scalar
i
Cosmetic redefinition X ∝√
λφN−2
:
IEH ∝ Z
d2x√
−gh
XR+ N−3
(N−2)X(∇X)2−λX(N−4)/(N−2)
| {z }
Scalar−tensor theory a.k.a.dilaton gravity
i
Spherical reduction
Line element adapted to spherical symmetry:
ds2= gµν(N)
|{z}
full metric
dxµdxν =gαβ(xγ)
| {z }
2Dmetric
dxαdxβ− φ2(xα)
| {z }
surface area
dΩ2S
N−2,
Insert into N-dimensional EH action IEH =κR
dNxp
−g(N)R(N): IEH =κ2π(N−1)/2
Γ(N−12 )
| {z }
N−2sphere
Z
d2x√
−gφN−2
| {z }
determinant
h
R+(N−2)(N−3) φ2
(∇φ)2−1
| {z }
Ricci scalar
i
Cosmetic redefinition X ∝√
λφN−2
:
IEH ∝ Z
d2x√
−gh
XR+ N−3
(N−2)X(∇X)2−λX(N−4)/(N−2)
| {z }
Scalar−tensor theory a.k.a.dilaton gravity
i
Spherical reduction
Line element adapted to spherical symmetry:
ds2= gµν(N)
|{z}
full metric
dxµdxν =gαβ(xγ)
| {z }
2Dmetric
dxαdxβ− φ2(xα)
| {z }
surface area
dΩ2S
N−2,
Insert into N-dimensional EH action IEH =κR
dNxp
−g(N)R(N): IEH =κ2π(N−1)/2
Γ(N−12 )
| {z }
N−2sphere
Z
d2x√
−gφN−2
| {z }
determinant
h
R+(N−2)(N−3) φ2
(∇φ)2−1
| {z }
Ricci scalar
i
Cosmetic redefinition X ∝√
λφN−2
:
IEH ∝ Z
d2x√
−gh
XR+ N−3
(N−2)X(∇X)2−λX(N−4)/(N−2)
| {z }
Scalar−tensor theory a.k.a.dilaton gravity
i
Outline
1 Gravity in 2D Models in 2D
Generic dilaton gravity action
2 Vienna School Approach Gravity as gauge theory All classical solutions
3 Duality
Casimir exchange Applications
Second order formulation
Similar action arises from string theory, from other kinds of dimensional reduction, from intrinsically 2D considerations, ...
Generic action:
I2DG =κ Z
d2x√
−gh
XR+U(X)(∇X)2−λV(X) i
(1)
Special case U=0,V =X2: EOM R=2λX I∝
Z d2x√
−gR2
Similarly f(R)Lagrangians related to (1) with U =0 String context: X =e−2φ, withφas string dilaton Conformal trafo to different model withU(X˜ ) =0:
V˜(X) = d
dXw(X) :=V(X)eQ(X)
| {z }
conformally invariant
, withQ(X) :=RX
dyU(y)
Second order formulation
Similar action arises from string theory, from other kinds of dimensional reduction, from intrinsically 2D considerations, ...
Generic action:
I2DG =κ Z
d2x√
−gh
XR+U(X)(∇X)2−λV(X) i
(1)
Special case U=0,V =X2: EOM R=2λX I∝
Z d2x√
−gR2
Similarly f(R)Lagrangians related to (1) with U =0 String context: X =e−2φ, withφas string dilaton Conformal trafo to different model withU(X˜ ) =0:
V˜(X) = d
dXw(X) :=V(X)eQ(X)
| {z }
conformally invariant
, withQ(X) :=RX
dyU(y)
Second order formulation
Similar action arises from string theory, from other kinds of dimensional reduction, from intrinsically 2D considerations, ...
Generic action:
I2DG =κ Z
d2x√
−gh
XR+U(X)(∇X)2−λV(X) i
(1)
Special case U=0,V =X2: EOM R=2λX I∝
Z d2x√
−gR2
Similarly f(R)Lagrangians related to (1) with U =0 String context: X =e−2φ, withφas string dilaton Conformal trafo to different model withU(X˜ ) =0:
V˜(X) = d
dXw(X) :=V(X)eQ(X)
| {z }
conformally invariant
, withQ(X) :=RX
dyU(y)
Selected list of models
Model U(X) λV(X) λw(X)
1. Schwarzschild (1916) −1
2X −λ −2λ√
X
2. Jackiw-Teitelboim (1984) 0 ΛX 12ΛX2
3. Witten BH (1991) −1
X −2b2X −2b2X
4. CGHS (1992) 0 −2b2 −2b2X
5.(A)dS2ground state (1994) −a
X BX a6=2: 2−aB X2−a
6. Rindler ground state (1996) −a
X BXa BX
7. BH attractor (2003) 0 BX−1 B ln X
8. SRG (N>3) −(N−2)XN−3 −λ2X(N−4)/(N−2) −λ2 N−2N−3X(N−3)/(N−2) 9. All above: ab-family (1997) −a
X BXa+b b6=−1: b+1B Xb+1
10.Liouvillegravity a beαX a6=−α: a+αb e(a+α)X
11. Reissner-Nordström (1916) −2X1 −λ2+Q2X −2λ2√
X−2Q2/√ X
12. Schwarzschild-(A)dS −2X1 −λ2−`X −2λ2√
X−23`X3/2
13. Katanaev-Volovich (1986) α βX2−Λ RXeαy(βy2−Λ)dy
14. Achucarro-Ortiz (1993) 0 Q2X − J
4X 3−ΛX Q2ln X+ J 8X 2−12ΛX2
15. Scattering trivial (2001) generic 0 const.
16. KK reduced CS (2003) 0 12X(c−X2) −1
8(c−X2)2
17. exact string BH (2005) lengthy −γ −(1+p
1+γ2) 18. Symmetric kink (2005) generic −XΠni=1(X2−Xi2) lengthy 19. KK red. conf. flat (2006) −12tanh(X/2) A sinh X 4A cosh(X/2)
20. 2D type 0A −1X −2b2X+b2 q28π −2b2X+b2 q28π ln X
Red: relevant for strings Blue: pioneer models
Outline
1 Gravity in 2D Models in 2D
Generic dilaton gravity action
2 Vienna School Approach Gravity as gauge theory All classical solutions
3 Duality
Casimir exchange Applications
First order formulation
Gravity as gauge theory (89: Isler, Trugenberger, Chamseddine, Wyler, 91: Verlinde, 92: Cangemi, Jackiw, Achucarro, 93: Ikeda, Izawa, 94: Schaller, Strobl)
Example: Jackiw-Teitelboim model (U=0, λV = ΛX ) [Pa,Pb] = ΛεabJ, [Pa,J] =εabPb, Non-abelian BF theory:
IBF = Z
XAFA= Z h
Xadea+Xaεabω∧eb+X dω+εabea∧ebΛXi field strength F =dA+ [A,A]/2 contains SO(1,2)connection A=eaPa+ωJ, coadjoint Lagrange multipliers XA
Generic first order action:
I2DG∝ Z h
Xa Ta
|{z}
torsion
+X R
|{z}
curvature
+
|{z}
volume
(XaXaU(X) +λV(X))i (2)
Ta=dea+εabω∧eb, R=dω,=εabea∧eb
First order formulation
Gravity as gauge theory (89: Isler, Trugenberger, Chamseddine, Wyler, 91: Verlinde, 92: Cangemi, Jackiw, Achucarro, 93: Ikeda, Izawa, 94: Schaller, Strobl)
Example: Jackiw-Teitelboim model (U=0, λV = ΛX ) [Pa,Pb] = ΛεabJ, [Pa,J] =εabPb, Non-abelian BF theory:
IBF = Z
XAFA= Z h
Xadea+Xaεabω∧eb+X dω+εabea∧ebΛXi field strength F =dA+ [A,A]/2 contains SO(1,2)connection A=eaPa+ωJ, coadjoint Lagrange multipliers XA
Generic first order action:
I2DG∝ Z h
Xa Ta
|{z}
torsion
+X R
|{z}
curvature
+
|{z}
volume
(XaXaU(X) +λV(X))i (2)
Ta=dea+εabω∧eb, R=dω,=εabea∧eb
Symmetries and equations of motion
Reinterpretation as Poisson-σmodel 94: Schaller, Strobl
IPSM = Z
dXI ∧AI +1
2PIJAJ∧AI
gauge field 1-forms: AI = (ω,ea), connection, Zweibeine target space coordinates: XI = (X,Xa), dilaton, aux. fields PIJ =−PJI, PXa=εabXb, Pab =εab(λV(X) +XaXaU(X)) Jacobi: PIL∂LPJK +perm(IJK) =0
Equations of motion (first order):
dXI+PIJAJ =0 dAI+ 1
2(∂IPJK)AK ∧AJ =0 Gauge symmetries (local Lorentz and diffeos):
δXI =PIJεJ
δAI =−dεI −
∂IPJK εKAJ
Symmetries and equations of motion
Reinterpretation as Poisson-σmodel 94: Schaller, Strobl
IPSM = Z
dXI ∧AI +1
2PIJAJ∧AI
gauge field 1-forms: AI = (ω,ea), connection, Zweibeine target space coordinates: XI = (X,Xa), dilaton, aux. fields PIJ =−PJI, PXa=εabXb, Pab =εab(λV(X) +XaXaU(X)) Jacobi: PIL∂LPJK +perm(IJK) =0
Equations of motion (first order):
dXI+PIJAJ =0 dAI+ 1
2(∂IPJK)AK ∧AJ =0 Gauge symmetries (local Lorentz and diffeos):
δXI =PIJεJ
δAI =−dεI −
∂IPJK εKAJ
Outline
1 Gravity in 2D Models in 2D
Generic dilaton gravity action
2 Vienna School Approach Gravity as gauge theory All classical solutions
3 Duality
Casimir exchange Applications
Constant dilaton vacua and generic solutions
Light-cone components and Eddington-Finkelstein gauge (87: Polyakov, 92: Kummer, Schwarz, 96: Klösch, Strobl)
Constant dilaton vacua:
X =const. , V(X) =0, R=λV0(X) Minkowski, Rindler or (A)dS only
isolated solutions (no constant of motion) Generic solutions in EF gaugeω0=e+0 =0, e−0 =1:
ds2=2eQ(X)du dX+eQ(X)(λw(X) +M)
| {z }
Killing norm
du2 (3)
Birkhoff theorem: at least one Killing vector∂u
one constant of motion: mass M one parameter in action: λ
dilaton is coordinate x0(residual gauge trafos!)
Constant dilaton vacua and generic solutions
Light-cone components and Eddington-Finkelstein gauge (87: Polyakov, 92: Kummer, Schwarz, 96: Klösch, Strobl)
Constant dilaton vacua:
X =const. , V(X) =0, R=λV0(X) Minkowski, Rindler or (A)dS only
isolated solutions (no constant of motion) Generic solutions in EF gaugeω0=e+0 =0, e−0 =1:
ds2=2eQ(X)du dX+eQ(X)(λw(X) +M)
| {z }
Killing norm
du2 (3)
Birkhoff theorem: at least one Killing vector∂u
one constant of motion: mass M one parameter in action: λ
dilaton is coordinate x0(residual gauge trafos!)
Outline
1 Gravity in 2D Models in 2D
Generic dilaton gravity action
2 Vienna School Approach Gravity as gauge theory All classical solutions
3 Duality
Casimir exchange Applications
Exchange spacetime mass with reference mass
Recall
ds2=eQ(X)h
2 du dX + (λw(X) +M)du2i Reformulate as
ds2=eQ(X)w(X)
| {z }
eQ( ˜˜X)
h
2 du dX w(X)
| {z }
dX˜
+(M 1
w(X)
| {z }
w( ˜˜ X)
+λ)du2i
Leads to dual potentials
U( ˜˜ X) =w(X)U(X)−eQ(X)V(X) V˜( ˜X) =− V(X)
w2(X) and to dual action˜I= ˜κR
d2x√
−gh
X R˜ + ˜U(∇X˜)2−MV˜i
Exchange spacetime mass with reference mass
Recall
ds2=eQ(X)h
2 du dX + (λw(X) +M)du2i Reformulate as
ds2=eQ(X)w(X)
| {z }
eQ( ˜˜X)
h
2 du dX w(X)
| {z }
dX˜
+(M 1
w(X)
| {z }
w( ˜˜ X)
+λ)du2i
Leads to dual potentials
U( ˜˜ X) =w(X)U(X)−eQ(X)V(X) V˜( ˜X) =− V(X)
w2(X) and to dual action˜I= ˜κR
d2x√
−gh
X R˜ + ˜U(∇X˜)2−MV˜i
Exchange spacetime mass with reference mass
Recall
ds2=eQ(X)h
2 du dX + (λw(X) +M)du2i Reformulate as
ds2=eQ(X)w(X)
| {z }
eQ( ˜˜X)
h
2 du dX w(X)
| {z }
dX˜
+(M 1
w(X)
| {z }
w( ˜˜ X)
+λ)du2i
Leads to dual potentials
U( ˜˜ X) =w(X)U(X)−eQ(X)V(X) V˜( ˜X) =− V(X)
w2(X) and to dual action˜I= ˜κR
d2x√
−gh
X R˜ + ˜U(∇X˜)2−MV˜i
PSM perspective
Trick: convert parameter in action to constant of motion Example (conformally transformed Witten BH, 92: Cangemi, Jackiw):
Z d2x√
−g[XR−λ]
integrate in abelian gauge field (F =∗dA) Z
d2x√
−g[XR+YF−Y] on-shell: dY =0, so Y =λ
apply trick to PSM: 4D Poisson manifold, 2 Casimirs (mass, charge)
duality: exchanges mass with charge
PSM perspective
Trick: convert parameter in action to constant of motion Example (conformally transformed Witten BH, 92: Cangemi, Jackiw):
Z d2x√
−g[XR−λ]
integrate in abelian gauge field (F =∗dA) Z
d2x√
−g[XR+YF−Y] on-shell: dY =0, so Y =λ
apply trick to PSM: 4D Poisson manifold, 2 Casimirs (mass, charge)
duality: exchanges mass with charge
Outline
1 Gravity in 2D Models in 2D
Generic dilaton gravity action
2 Vienna School Approach Gravity as gauge theory All classical solutions
3 Duality
Casimir exchange Applications
The ab-family
Schwarzschild, Jackiw-Teitelboim, ...
Useful 2-parameter family of models:
U=−a
X , V =Xa+b Duality:˜a=1−(a−1)/b andb˜=1/b Global structure:
b>0: reflection at origin b<0: reflection atρ-axis
hereξ=lnp
|b|andρ= (a−1)/p
|b|
The ab-family
Schwarzschild, Jackiw-Teitelboim, ...
Useful 2-parameter family of models:
U=−a
X , V =Xa+b Duality:˜a=1−(a−1)/b andb˜=1/b Global structure:
b>0: reflection at origin b<0: reflection atρ-axis
hereξ=lnp
|b|andρ= (a−1)/p
|b|
Liouville gravity
cf. e.g. 04: Nakayama
Specific case (“almost Weyl invariant”):
Z d2x√
−gh
XR+1
2(∇X)2−m2eXi Dual model:
Z d2x√
−gh
X R˜ −λi (conformally transformed) Witten BH!
Note: on-shell R=0 (in both formulations)
Liouville gravity
cf. e.g. 04: Nakayama
Specific case (“almost Weyl invariant”):
Z d2x√
−gh
XR+1
2(∇X)2−m2eXi Dual model:
Z d2x√
−gh
X R˜ −λi (conformally transformed) Witten BH!
Note: on-shell R=0 (in both formulations)
Limiting action for gravity in 2 + ε dimensions (ε → 0)
79: Weinberg, 93: Mann, Ross
Spherical reduction from 2+εto 2 dimensions:
U(X) =−1−ε
εX , V(X) =−ε(1−ε)X1−2/ε
Limitε→0 not well-defined! No suitable rescaling of fields and coupling constants possible! [recall: in action XR+U(X)(∇X)2] Solution: dualize, take limit in dual formulation, dualize back [why it works? MGS'model with U=0]
Result:
Z d2x√
−gh
XR+1
2(∇X)2−m2eXi .
This is the specific Liouville gravity model discussed previously!
Limiting action for gravity in 2 + ε dimensions (ε → 0)
79: Weinberg, 93: Mann, Ross
Spherical reduction from 2+εto 2 dimensions:
U(X) =−1−ε
εX , V(X) =−ε(1−ε)X1−2/ε
Limitε→0 not well-defined! No suitable rescaling of fields and coupling constants possible! [recall: in action XR+U(X)(∇X)2] Solution: dualize, take limit in dual formulation, dualize back [why it works? MGS'model with U=0]
Result:
Z d2x√
−gh
XR+1
2(∇X)2−m2eXi .
This is the specific Liouville gravity model discussed previously!
Limiting action for gravity in 2 + ε dimensions (ε → 0)
79: Weinberg, 93: Mann, Ross
Spherical reduction from 2+εto 2 dimensions:
U(X) =−1−ε
εX , V(X) =−ε(1−ε)X1−2/ε
Limitε→0 not well-defined! No suitable rescaling of fields and coupling constants possible! [recall: in action XR+U(X)(∇X)2] Solution: dualize, take limit in dual formulation, dualize back [why it works? MGS'model with U=0]
Result:
Z d2x√
−gh
XR+1
2(∇X)2−m2eXi .
This is the specific Liouville gravity model discussed previously!
Literature I
A. Jevicki, “Development in 2-d string theory,”
hep-th/9309115.
D. Grumiller, W. Kummer, and D. Vassilevich, “Dilaton gravity in two dimensions,” Phys. Rept. 369 (2002) 327–429,hep-th/0204253.
D. Grumiller and R. Jackiw, “Duality in 2-dimensional dilaton gravity,” Phys. Lett. B642 (2006) 530,hep-th/0609197.