Skip to main content

De Sitter Horizon: Surface Tension, Laplace Pressure, and What They Actually Mean

 In the static patch of de Sitter spacetime, the cosmological horizon behaves thermodynamically in ways closely analogous to a physical interface. One can assign it entropy, temperature, and even an effective surface tension. Remarkably, the familiar Young–Laplace pressure relation from surface physics appears naturally at the horizon. 

 


Scope. Everything below is formulated in the static patch of de Sitter spacetime — the causally accessible region for a single inertial observer, covered by static coordinates in which the metric $$ds^2 = -\left(1-\frac{r^2}{L^2}\right)c^2,dt^2 + \left(1-\frac{r^2}{L^2}\right)^{-1}dr^2 + r^2 d\Omega^2$$ is manifestly time-independent. The static patch admits a timelike Killing vector $\partial_t$, and it is this Killing vector that defines the notions of energy, temperature, and thermodynamic equilibrium used throughout. Global de Sitter spacetime has no timelike Killing vector; the thermodynamic framework does not extend beyond the static patch without additional assumptions.

1. De Sitter Horizon Parameters (SI Units)

For the static patch of a 4-dimensional de Sitter spacetime:

Horizon radius:

$$L = \sqrt{3/\Lambda}$$

Horizon area:

$$A = 4\pi L^2$$

Horizon entropy:

$$S = \frac{k_B A}{4 L_P^2} = \frac{\pi k_B L^2}{L_P^2}$$

where

$$L_P^2 = \frac{\hbar G}{c^3}.$$

Horizon temperature (Gibbons–Hawking):

$$T = \frac{\hbar c}{2 \pi k_B L}$$

This is the temperature registered by an inertial observer at $r=0$ in the static patch; it is defined via the periodicity of the Euclidean continuation of the static-patch metric.

Horizon quasilocal energy:

$$E = -\frac{c^4 L}{2 G}$$

The Misner–Sharp mass at the horizon is

$$M_{MS}(r) = \frac{c^2 r}{2G} \left(1 - g^{ab} \partial_a r \partial_b r\right)$$

which gives $M_{MS}(L) = c^2 L/(2G)$ (units of mass). The static-patch energy is $E = -M_{MS},c^2 = -c^4 L/(2G)$ (units of energy), where the negative sign reflects the convention that the observer lies inside the cosmological horizon. The same result follows from the Brown–York or Abbott–Deser quasilocal energy defined with respect to the static-patch Killing vector $\partial_t$.

Smarr-consistent thermodynamic surface tension:

Using the horizon Smarr relation

$$E = 2(TS + \sigma A)$$

we obtain

$$\sigma_{\text{thermo}} = \frac{E - 2TS}{2A} = -\frac{3 c^4}{16 \pi G L}.$$

Negative $\sigma$ → the horizon releases energy when expanding.

This result is also what Chen et al 2017, The modified first laws of thermodynamics of anti-de Sitter and de Sitter space–times derived from the radial Einstein equation rewritten as a thermodynamic identity. We can get the same result also from the membrane paradigm: the stretched horizon stress tensor

$$t_{ab}=\frac{1}{8\pi G}(Kh_{ab}-K_{ab})$$

evaluated for the de Sitter horizon yields the same isotropic tension $\sigma=-\frac{3c^4}{16\pi GL}$, confirming that the thermodynamic and geometric descriptions agree.

A subtlety: for a black hole the stretched membrane sits just outside the horizon, with the outward normal pointing away from the hole. For the cosmological horizon the observer lives in the interior, so the stretched membrane sits just inside $r_C$ and the relevant normal is inward-pointing (toward the observer). This reversal of orientation is the geometric origin of the sign flip $\sigma_{\text{BH}}>0 \to \sigma_{\text{dS}}<0$.


2. Mechanical Laplace Formula

For a spherical interface (classical surface physics)

$$\Delta P_{\text{mech}} = P_{\text{in}} - P_{\text{out}} = \frac{2\sigma}{R}.$$

Applied to the de Sitter horizon:

$$\frac{2\sigma_{\text{thermo}}}{L} = 2 \cdot \frac{-3 c^4}{16 \pi G L} \cdot \frac{1}{L} = -\frac{3 c^4}{8 \pi G L^2}.$$

This gives the effective pressure difference associated with changing the horizon area.


3. Effective Thermodynamic Pressures

For pure de Sitter,

$$S=\frac{\pi k_B L^2}{L_P^2}, \qquad V=\frac{4\pi L^3}{3}$$

where $V$ is the coordinate volume of the static patch (the region $0 \leq r \leq L$). Both are functions of the single parameter $L$. They are not independent thermodynamic variables. Therefore, the pressure cannot be obtained via the standard partial derivative $P=-(\partial E/\partial V)|_S$, which is undefined here because one cannot vary $V$ at fixed $S$.

Instead the pressure must be evaluated along the de Sitter solution curve. With $E$, $S$, and $V$ all functions of $L$ alone, the first law

$$dE = TdS + PdV$$

becomes

$$\frac{dE}{dL} = T\frac{dS}{dL} + P\frac{dV}{dL}$$

which is one equation in one unknown ($P$). Computing the derivatives:

$$\frac{dE}{dL}=-\frac{c^4}{2G},\qquad T\frac{dS}{dL}=\frac{c^4}{G L}\cdot L = \frac{c^4}{G},\qquad \frac{dV}{dL}=4\pi L^2$$

Solving:

$$P_{\text{in}} = -\frac{3c^4}{8\pi G L^2} = -\frac{\Lambda c^4}{8\pi G}.$$

This is exactly the cosmological constant pressure $P_\Lambda = -\rho_\Lambda c^2$.

The fact that this solution-curve pressure coincides numerically with the extended phase space identification $P = -\Lambda c^4/(8\pi G)$ is a result, not a tautology — the two derivations start from different premises and the agreement is a consistency check on the thermodynamic framework.


4. Exterior Pressure

The vacuum stress tensor from $\Lambda$

$$T_{\mu\nu}^{(\Lambda)} = -\frac{\Lambda c^4}{8\pi G},g_{\mu\nu}$$

is homogeneous, so the same vacuum exists on both sides of the horizon. (The horizon itself is observer-dependent — every inertial observer in de Sitter has their own cosmological horizon at distance $L$ — but the vacuum state is global.)

However, the region outside the cosmological horizon is causally disconnected from the static patch. It cannot perform quasi-static thermodynamic work on the system bounded by the horizon.

In the horizon thermodynamic description this is represented by the effective boundary condition

$$P_{\text{out}}=0.$$

To be precise: this does not mean the vacuum pressure vanishes outside, nor is causal disconnection the only physical situation yielding $P_{\text{out}}=0$ (free expansion into vacuum is an ordinary example). The more careful statement is that $P_{\text{out}}=0$ is a thermodynamic boundary condition — the static-patch system has no external agent doing work on it — which is consistent with causal disconnection. The Nariai limit (where BH and cosmological horizons merge) provides a useful test: there both horizons are causally disconnected from each other's exteriors, but neither has $P_{\text{out}}=0$ in the combined system.


5. Laplace Relation

The mechanical relation then becomes

$$P_{\text{in}}-P_{\text{out}} = -\frac{3 c^4}{8 \pi G L^2} = \frac{2\sigma}{L}.$$

The Laplace balance holds exactly.

This relation is not independent. Both $P_{\text{in}}$ and $\sigma$ ultimately arise from the same Einstein equation component

$$G^r_{\ r} + \Lambda, g^r_{\ r} = 0.$$

Equivalently, the Laplace relation is just the Smarr relation

$$E = 2(TS + \sigma A)$$

repackaged as a pressure balance across the horizon.

This deserves emphasis: the Young–Laplace equation applied to horizons has zero independent content beyond the Einstein equations evaluated at the horizon. It is a repackaging, not a new constraint. Interpreting $\sigma < 0$ through the lens of ordinary fluid mechanics (miscible fluids, instability, etc.) therefore reads physical content into what is essentially the same equation dressed in surface-physics language. The physics is general relativity; the surface tension language is a translation, not an extension.


6. Physical Interpretation

The $\Lambda$-vacuum is homogeneous — the universe is not expanding into anything.

$P_{\text{out}}=0$ encodes an effective thermodynamic boundary condition (no external work), consistent with causal disconnection.

$P_{\text{in}}<0$ means increasing the horizon area lowers free energy:

$$F = E - TS.$$

For de Sitter,

$$F = -\frac{c^4 L}{2G} - \frac{c^4 L}{2G} = -\frac{c^4 L}{G},$$

which decreases as $L$ grows.

The horizon therefore expands in the direction of decreasing free energy. (Here "expansion" means: within the family of de Sitter solutions parametrised by $L$, larger $L$ is thermodynamically favoured. In the static patch at fixed $\Lambda$, $L$ is fixed; the statement becomes dynamical only when matter or radiation perturbs the pure dS state.)

The Gibbs free energy provides a complementary check. From the Smarr relation:

$$G = E - TS - \sigma A = -\frac{c^4 L}{2G} - \frac{c^4 L}{2G} + \frac{3c^4 L}{4G} = -\frac{c^4 L}{4G}.$$

This is negative for all $L>0$. In ordinary phase thermodynamics, $G<0$ relative to a reference means the phase is thermodynamically favoured. For de Sitter, this means: the horizon configuration is the preferred state. Combined with $\sigma<0$ (spontaneous expansion) and $F$ decreasing with $L$, the picture is self-consistent — de Sitter space is a thermodynamic attractor.

This connects to the cosmic no-hair theorem (every expanding spacetime with $\Lambda>0$ asymptotes to de Sitter) and suggests a thermodynamic reason for it: de Sitter is the state of minimum free energy within the static patch. Whether this static-patch statement can be promoted to a full thermodynamic derivation of the global no-hair result remains open.


7. Contrast with Black Hole Horizons

Black hole: $\sigma>0$ — work must be done on the horizon to increase area

De Sitter: $\sigma<0$ — the horizon does work as it expands

Thus:

  • Black hole area increases only when matter/energy falls in
  • De Sitter horizon area increases spontaneously (in the sense that the static-patch free energy decreases with increasing $L$)

This is the thermodynamic restatement, within the static patch, of $\Lambda$-driven accelerated expansion.

The sign flip $\sigma_{\text{BH}}>0 \to \sigma_{\text{dS}}<0$ also reflects the reversal of the membrane normal orientation (Section 1), and mirrors the well-known sign flip in the heat capacity and the difficulty of constructing de Sitter holography: the horizon's thermodynamic response functions have the "wrong" sign for a well-defined holographic dual.


8. Wrap

For the de Sitter cosmological horizon as seen from the static patch:

$$\sigma = -\frac{3c^4}{16\pi G L},\qquad P_{\text{in}}=-\frac{3c^4}{8\pi G L^2}.$$

These satisfy

$$P_{\text{in}}-P_{\text{out}}=\frac{2\sigma}{L}.$$

The "Laplace law" for the cosmological horizon is therefore simply another form of the Smarr relation, and ultimately a projection of the Einstein equations onto the horizon surface.

It provides a useful translation between:

  • Horizon thermodynamics
  • Classical surface physics
  • The membrane paradigm of general relativity

But it contains no independent physics beyond what is already in the Einstein equations at the horizon. The entire construction is valid within the static patch; extending it to global de Sitter requires a framework (such as the elliptic de Sitter space or a global Hartle–Hawking state) that goes beyond what is derived here.

Comments

Popular posts from this blog

Blurring the horizon - the quantum width of the cosmic event horizon

A  paper by Zurek applied a random walk argument to a black hole horizon. Credit, Zurek, 2021 Zurek  ( Snowmass 2021 White Paper: Observational Signatures of Quantum Gravity )  called this a blurring of the horizon — a fuzzy, or uncertain horizon — and went through derivations supporting the idea that this length scale is the quantum uncertainty in the position of the black hole horizon: a dynamic quantum width of an event horizon. This is a concept which fundamentally applies to the Universe's own Cosmic Event Horizon (CEH). The Bekenstein-Hawking entropy gives the number of quantum degrees of freedom that can fluctuate. Below, we step out our own cosmic de Sitter derivation of the random walk argument. To do this, let $l_{\Lambda}$ represent the generalised de Sitter horizon scale. Due to the holographic UV/IR correspondence, this scale manifests dually: at the fundamental microscopic limit as $l_{UV} = 2L_p$ (the gravitational/casual limit, aka the Schwarzschild radi...

De Sitter Horizons One Radius to Rule Them All

    Ten perspectives on one geometric object Scope. The following framework is formulated within the static patch of $(3+1)$-dimensional de Sitter spacetime—the causally accessible region for a single inertial observer. The geometry is covered by the metric: $$ds^2 = -\left(1 - \frac{r^2}{l_\Lambda^2}\right)c^2 dt^2 + \left(1 - \frac{r^2}{l_\Lambda^2}\right)^{-1}dr^2 + r^2 d\Omega^2,$$ which admits the timelike Killing vector $\partial_t$. All thermodynamic quantities (temperature, energy, entropy) are defined with respect to this Killing vector. Each section is marked: [T] for theorem/established result, [N] for novel or heuristic connection, [C] for conjecture. No step marked [T] depends on a step marked [C] or [N] . 1. One Scale and the $SL(2,\mathbb{R})$ Sector The static patch is governed by a single macroscopic length: $$l_\Lambda = \frac{c}{H_\Lambda} = \sqrt{\frac{3}{\Lambda}}.$$ Every macroscopic thermodynamic quantity is a function of this scale alone [T] ...

Our cosmic event horizon on a string

A Cosmic Stringy Adventure ! In a previous  post, we showed that a positive cosmological constant $\Lambda$ implies a minimal length  $l_{UV} = 2L_P$, a mass gap (aka minimum GR local "mass") $m_s = M_P/2$, and a scale-invariant maximum force $c^4/(4G)$. We also established the Compton–gravitational duality: every scale $r$ carries two natural masses: $$m_C(r) = \frac{\hbar}{rc}, \qquad m_G(r) = \frac{c^2}{4G}\ r$$ whose product $m_C \  m_G = M_P^2/4$ is scale-independent (UV-IR duality),  and which coincide only at $r = l_{UV}$.  In this post , we explore what happens when McDormand starts swinging that minimal mass on a cosmic string, and we discover a mass-energy flux that doesn't care whether the string is Planck-sized or horizon-sized.  McDormand swinging a cosmic light-like mass-energy $m_{s}$, with a "cosmic string" of radius $l_{\Lambda}$, giving a centripetal force $F_{local}=m_{s} \ c^2/l_{\Lambda}$.    Mass-Energy Flux Along the Stri...