arXiv:1005.1132v3 [gr-qc] 29 Jun 2010 Tolman mass, generalized surface gravity, and entropy bounds Gabriel Abreu and Matt Visser School of Mathematics, Statistics, and Operations Research; Victoria University of Wellington, Wellington, New Zealand. (Dated: 7 May 2010; 25 May 2010; 28 June 2010; L A T E X-ed October 28, 2018) In any static spacetime the quasi-local Tolman mass contained within a volume can be reduced to a Gauss-like surface integral involving the flux of a suitably defined generalized surface gravity. By introducing some basic thermodynamics and invoking the Unruh effect one can then develop elementary bounds on the quasi-local entropy that are very similar in spirit to the holographic bound, and closely related to entanglement entropy. Tolman mass is one of the standard notions of quasi- local mass in common use in general relativity [1]. Using classical general relativity, this quasi-local Tolman mass can, in any static spacetime (either with or without a black hole region), be reduced to a flux integral of (gener- alized) surface gravity across the boundary of the region of interest. (This is closely related to the classical laws of black hole mechanics [2].) General relativistic thermo- dynamics, together with a minimal appeal to quantum physics as embodied in the Unruh effect [3], is then suf- ficient to develop elementary but powerful bounds on a suitably defined notion of quasi-local entropy — bounds very similar in spirit to the holographic bound [4–8], and closely related to entanglement entropy [9]. In a static spacetime where the metric is taken to be of the form ds 2 = e dt 2 + g ij dx i dx j , (1) the Tolman mass contained in a region Ω is defined in terms of the orthonormal components of stress-energy by first taking ρ = T ˆ 0 ˆ 0 and p = 1 3 tr{T ˆ ıˆ }; and then setting m T (Ω) = Ω g 4 {ρ +3p} d 3 x. (2) The Einstein equations then imply the purely geometrical statement m T (Ω) = 1 4π Ω g 4 R ˆ 0 ˆ 0 d 3 x. (3) The Tolman mass is intimately related to the Komar mass [10], though we will not be phrasing any of the dis- cussion below in terms of Killing vectors. It is a very old result, going back at least as far as Landau–Lifshitz [11] that in any stationary metric R 0 0 = 1 g 4 i ( g 4 g 0a Γ i a0 ) . (4) (Here a ∈{0, 1, 2, 3}; i ∈{1, 2, 3}.) Adopting the mani- fest static coordinates of equation (1), and then going to an orthonormal basis, this is more simply phrased as R ˆ 0 ˆ 0 = 1 g 4 i ( g 4 g 00 Γ i 00 ) . (5) * gabriel.abreu@msor.vuw.ac.nz matt.visser@msor.vuw.ac.nz To get a clean physical interpretation of this formula, consider a fiducial observer (FIDO) with 4-velocity V a = |g 00 |;0, 0, 0 . (6) By definition the 4-acceleration of these FIDOs is A a =( V V ) a = V b b V a = V 0 ( 0 V a a c0 V c ) = |g 00 | Γ a 00 |g 00 | = |g 00 | Γ a 00 . (7) But then, since V is 4-orthogonal to A, we have A 0 = 0; A i = |g 00 | Γ i 00 ; (8) where A i are the 3 spatial components of 4-acceleration. Therefore in any static spacetime, in the region outside the horizon, the Landau–Lifshitz result is R ˆ 0 ˆ 0 = 1 g 4 i ( g 4 A i ) . (9) Then for any 3-volume Ω (if a horizon is present then for convenience we confine ourselves to a region that lies out- side the horizon) we can use ordinary partial derivative integration by parts to deduce Ω g 4 R ˆ 0 ˆ 0 d 3 x = Ω i ( g 4 A i ) d 3 x = Ω i ( g 3 {e Ψ A i } ) d 3 x = Ω {e Ψ A i } ˆ n i g 2 d 2 x, (10) where ˆ n is the unit normal, (defined in terms of the 3- metric g ij ), and g 2 is the induced area measure on Ω. Define a (generalized) surface gravity (3-vector) and its norm by κ i = e Ψ A i ; κ = g ij κ i κ j = e Ψ g ij A i A j . (11) This is just the “redshifted” 4-acceleration of the FIDOs, and is a natural generalization of surface gravity, not just for any event horizon that might be present, but also applying to FIDOs skimming along the boundary Ω. In terms of this generalized surface gravity we now have m T (Ω) = 1 4π Ω κ i ˆ n i g 2 d 2 x = 1 4π Ω κ· ˆ n dA . (12)