SLIDE 1
Stress Tensor Correlators of Various Black Hole Vacua in Two Dimensions
Hing Tong Cho Tamkang University, Taiwan 2012 Asia Pacific School/Workshop on Cosmology and Gravitation (March 4, 2012)
SLIDE 2 Outline
- I. Introduction
- II. Black hole vacua and mode functions
- III. The renormalized stress tensor
- IV. Stress tensor correlators and fluctuations
- V. Discussions
SLIDE 3
- I. Introduction
- 1. Fluctuations of the stress energy tensor can backreact onto
the background spacetime. In the theory of stochastic gravity, this is in the form of a stochastic force on the right hand side
- f the Einstein equation.
- 2. Fluctuations of Hawking radiation (Wu and Ford (1999)):
How big are they?
- 3. Fluctuations of the quantum field near the horizon: Are they
divergent?
- 4. Sizable fluctuations might induce instability and invalidate the
semi-classical approximation.
- 5. Renormalization is needed to obtain finite quantities. Here we
adopt the point-splitting method.
- 6. There are a lot of simplifications in two dimensions. In
particular, two dimensional spacetimes are all conformal to the Minkowski spacetime.
SLIDE 4
- II. Black hole vacua and mode functions
In two dimensions, take the black hole metric as ds2 = − ( 1 − 2M r ) dt2 + ( 1 − 2M r )−1 dr2 = ( 1 − 2M r ) ( −dt2 + dx2) where x = r + 2M ln ( r 2M − 1 ) is the tortoise coordinate.
SLIDE 5
It is conformal to Minkowski spacetime. Using the null coordinates, u = t − x and v = t + x, one has ds2 = ( 1 − 2M r ) du dv The mode functions for a massless minimally coupled scalar are just 1 √ 4휋휔 e−i휔u ; 1 √ 4휋휔 e−i휔v This is the Schwarzschild coordinates.
SLIDE 6
One can also use the Kruskal coordinates, U = −4Me−u/4M ; V = 4Mev/4M then the metric becomes ds2 = 2M r e−r/2MdU dV which is well-defined (like Minkowski) at the horizon, r = 2M. The mode functions are 1 √ 4휋휔 e−i휔U ; 1 √ 4휋휔 e−i휔V
SLIDE 7
Choosing different mode functions corresponds to choosing different vacua: Boulware vacuum, 1 √ 4휋휔 e−i휔u and 1 √ 4휋휔 e−i휔v Hartle-Hawking vacuum, 1 √ 4휋휔 e−i휔U and 1 √ 4휋휔 e−i휔V Unruh vacuum, 1 √ 4휋휔 e−i휔U and 1 √ 4휋휔 e−i휔v
SLIDE 8
- III. The renormalized stress tensor
For a massless minimally coupled scalar field 휙, the stress tensor T휇휈 = ∇휇휙∇휈휙 − 1 2g휇휈∇휌휙∇휌휙 ⟨T휇휈(x)⟩ is divergent. Point-splitting regularization, ⟨T휇휈(x)⟩ = lim
x′→x,x′′→x
1 2 ( g휇훼′g휈훽′′ + g휇훽′′g휈훼′ + g휇휈g훼′훽′′) ∇훼′∇훽′′G +(x′, x′′) where G +(x′, x′′) = ⟨휙(x′)휙(x′′)⟩ is the Wightman function. Usually one take x′ = x + 휖 and x′′ = x − 휖 along a geodesic with 휖 the geodesic distance. The limit means that 휖 → 0.
SLIDE 9
In this two dimensional setting, the renormalized stress tensor was given by Davies and Fulling (1977) ⟨T휇휈⟩ren = 휃휇휈 + 1 48휋Rg휇휈 where the state dependent tensor 휃uu = − 1 12휋 ( C 1/2∂2
uC −1/2)
휃vv = − 1 12휋 ( C 1/2∂2
vC −1/2)
휃uv = and the Ricci scalar R = − 4 C [∂u∂vC C − (∂uC)(∂vC) C 2 ]
SLIDE 10
Boulware vacuum 휂, C = 1 − 2M/r. T 휂
uu
= 1 24휋M2 (3M4 2r4 − M3 r3 ) = T 휂
vv
T 휂
uv
= 1 24휋M2 ( 1 − 2M r ) ( −M3 r3 ) T 휂
tt
= 1 24휋M3 (7M4 r4 − 4M3 r3 ) ∼ 1 r3 as r → ∞ T 휂
tr
= T 휂
rr
= 1 24휋M2 ( 1 − 2M r )−2 ( −M4 r4 ) ∼ 1 r4 as r → ∞ In a local frame, T 휂
ˆ tˆ t and T 휂 ˆ rˆ r ∼ (1 − 2M/r)−1 as r → 2M.
SLIDE 11
Hartle-Hawking vacuum 휈, C = 2Me−r/2M/r, T 휈
uu
= 1 24휋M2 (3M4 2r4 − M3 r3 + 1 32 ) = T 휈
vv
T 휈
uv
= 1 24휋M2 ( 1 − 2M r ) ( −M3 r3 ) = T 휂
uv
T 휈
tt
= 1 24휋M2 (7M4 r4 − 4M3 r3 + 1 16 ) ∼ 휋 6 ( 1 8휋M )2 as r → ∞ T 휈
tr
= T 휈
rr
= 1 24휋M2 ( 1 − 2M r )−2 ( −M4 r4 + 1 16 ) ∼ 휋 6 ( 1 8휋M )2 as r → ∞ This corresponds to a thermal gas with temperature T = 1/8휋M.
SLIDE 12
In a local frame, as r → 2M, T 휈
ˆ tˆ t
∼ − 1 96휋M2 T 휈
ˆ rˆ r
∼ 1 96휋M2 The stress tensor is finite in this near horizon limit. The Hartle-Hawking vacuum is defined with respect to the Kruskal coordinates which are well-defined at the horizon.
SLIDE 13
Unruh vacuum 휉 T 휉
uu
= T 휈
uu =
1 24휋M2 (3M4 2r4 − M3 r3 + 1 32 ) T 휉
uv
= T 휂
uv =
1 24휋M2 ( 1 − 2M r ) ( −M3 r3 ) T 휉
vv
= T 휂
vv =
1 24휋M2 (3M4 2r4 − M3 r3 ) T 휉
tr
= 1 24휋M2 ( 1 − 2M r )−1 ( − 1 32 ) ∼ − 휋 12 ( 1 8휋M )2 as r → ∞ This represents an out-going flux of Hawking radiation with temperature T = 1/8휋M.
SLIDE 14
- IV. Stress tensor correlators and fluctuations
Define the correlation, ΔT 2
휇휈훼′훽′(x, x′) = ⟨T휇휈(x)T훼′훽′(x′)⟩ − ⟨T휇휈(x)⟩⟨T훼′훽′(x′)⟩
Using point-splitting regularization, one arrives at the expression ΔT 2
휇휈훼′훽′(x, x′)
= [ ∇휇∇훼′G +(x, x′) ] [ ∇휈∇훽′G +(x, x′) ] + [ ∇휇∇훽′G +(x, x′) ] [ ∇휈∇훼′G +(x, x′) ] −g휇휈 [ ∇휌∇훼′G +(x, x′) ] [ ∇휌∇훽′G +(x, x′) ] +g훼′훽′ [ ∇휇∇휎′G +(x, x′) ] [ ∇휈∇휎′G +(x, x′) ] +1 2g휇휈g훼′훽′ [ ∇휌∇휎′G +(x, x′) ] [ ∇휌∇휎′G +(x, x′) ]
SLIDE 15
In the Schwarzschild coordinates (Boulware vacuum), G +(x, x′) = − 1 4휋ln(ΔuΔv) The nonzero correlators are ( ΔT 2
uuu′u′
)휂 = 1 8휋2(Δu)4 ( ΔT 2
vvv′v′
)휂 = 1 8휋2(Δv)4 They are well-defined when x and x′ are non-coincident. Here we consider only non-null separation.
SLIDE 16
Similarly in the Kruskal coordinates (Hartle-Hawking vacuum), ( ΔT 2
UUU′U′
)휈 = 1 8휋2(ΔU)4 ( ΔT 2
VVV ′V ′
)휈 = 1 8휋2(ΔV )4 In the Unruh vacuum, ( ΔT 2
UUU′U′
)휉 = 1 8휋2(ΔU)4 ( ΔT 2
vvv′v′
)휉 = 1 8휋2(Δv)4
SLIDE 17
To study the fluctuations we have to take the coincident limit x′ → x which is divergent. We again use the point-splitting regularization and we obtain ( ΔT 2
휇휈훼훽(x)
)
ren
= ( 휃휇훼휃휈훽 + 휃휇훽휃휈훼 − g휇휈휃훼휌휃휌
훽 − g훼훽휃휇휌휃휌 휈 + 1
2g휇휈g훼훽휃휌휎휃휌휎 ) + R 48휋 (g휇훼휃휈훽 + g휇훽휃휈훼 + g휈훼휃휇훽 + g휈훽휃휇훼 −2g휇휈휃훼훽 − 2g훼훽휃휇휈 + g휇휈g훼훽휃휌
휌
) + ( R 48휋 )2 (g휇훼g휈훽 + g휇훽g휈훼 − g휇휈g훼훽)
SLIDE 18
For the Boulware vacuum, we have ( ΔT 2
tttt
)휂
ren
= 4 ( 1 24휋M2 )2 (41M8 4r8 − 11M7 r7 + 3M6 r6 ) ( ΔT 2
rrrr
)휂
ren
= 4 ( 1 24휋M2 )2 ( 1 − 2M r )−4 × (41M8 4r8 − 11M7 r7 + 3M6 r6 ) As r → ∞, ( ΔT 2
tttt
)휂
ren ∼
( ΔT 2
rrrr
)휂
ren ∼ 12
( 1 24휋M2 )2 (M6 r6 ) Note that in the same limit, T 휂
tt ∼ 1/r3 and T 휂 rr ∼ 1/r4.
SLIDE 19
For the Hartle-Hawking vacuum, we have ( ΔT 2
tttt
)휈
ren
= 4 ( 1 24휋M2 )2 (41M8 4r8 − 11M7 r7 + 3M6 r6 +3M4 32r4 − M3 16r3 + 1 1024 ) ( ΔT 2
rrrr
)휈
ren
= ( 1 − 2M r )−4 ( ΔT 2
tttt
)휈
ren
SLIDE 20
As r → ∞, we have ( ΔT 2
tttt
)휈
ren ∼
( ΔT 2
rrrr
)휈
ren ∼
1 256 ( 1 24휋M2 )2 Since in the same limit, T 휈
tt ∼ T 휈 rr ∼ 1
16 ( 1 24휋M2 ) Hence, we have √( ΔT 2
tttt
)휈
ren
T 휈
tt
∼ √ (ΔT 2
rrrr)휈 ren
T 휈
rr
∼ 1
SLIDE 21 As r → 2M, in a local frame, ( ΔT 2
ˆ tˆ tˆ tˆ t
)휈
ren ∼
( ΔT 2
ˆ rˆ rˆ rˆ r
)휈
ren ∼ 1
8 ( 1 24휋M2 )2 In the same limit, T 휈
ˆ tˆ t ∼ −1
4 ( 1 24휋M2 ) ; T 휈
ˆ rˆ r ∼ 1
4 ( 1 24휋M2 ) we have again
( ΔT 2
ˆ tˆ tˆ tˆ t
)휈
ren
(T 휈
ˆ tˆ t)2
∼ √( ΔT 2
ˆ rˆ rˆ rˆ r
)휈
ren
(T 휈
ˆ rˆ r)2
∼ √ 2
SLIDE 22 For the Unruh vacuum, ( ΔT 2
trtr
)휉
ren
= 2 ( 1 24휋M2 )2 ( 1 − 2M r )−2 × (9M8 2r8 − 6M7 r7 + 2M6 r6 + 3M4 32r4 − M3 16r3 + 1 1024 ) As r → ∞, ( ΔT 2
trtr
)휉
ren ∼
1 512 ( 1 24휋M2 )2 Hence, we have
( ΔT 2
trtr
)휉
ren
( T 휉
tr
)2 ∼ √ 2
SLIDE 23
- V. Discussions
- 1. The fluctuations of the Hartle-Hawking vacuum, for both the
density and the pressure, are of order 1. The same applies to the fluctuations near the horizon.
- 2. Fluctuations of the Hawking flux in the Unruh vacuum are
also of order 1.
- 3. The results show that fluctuations are sizable and they might
induce passive spacetime metric fluctuation to invalidate the semi-classical approximation. This is true even for static spacetimes.
- 4. Results in two dimensions should only be taken as an
- indication. Much more work has to be done in four
dimensions.