Nariai and Ultracold Limits

When horizons coalesce, simply substituting the degenerate parameters into a static chart loses the region one wants to study. A simultaneous parameter and coordinate scaling can instead retain a finite geometry. The uncharged black-hole–cosmological merger gives the Nariai product $dS_2\times S^2$. In charged de Sitter solutions, different pairs can merge, and a triple root produces an ultracold limit with a locally flat two-dimensional factor. Root multiplicity determines which limit is meant.

Helpful background. Exact Solutions of Einstein’s Equations supplies product and Reissner–Nordström–de Sitter geometries; Causality and Global Structure supplies limiting observer domains; The Einstein Equation supplies the Einstein–Maxwell–$\Lambda$ equations; and Black-Hole Geometry and Mechanics supplies degenerate Killing horizons.

Root multiplicity labels the limiting family

For Schwarzschild–de Sitter,

$$f(r)=1-\frac{2m}{r}-\frac{\Lambda r^2}{3}.$$

The Nariai point satisfies $f(r_0)=f'(r_0)=0$, giving

$$r_0=\frac1{\sqrt\Lambda}, \qquad m_0=\frac1{3\sqrt\Lambda}.$$

The black-hole and cosmological roots merge, and the surface gravity of the original static generator vanishes. The latter statement uses the fixed normalization in the approach to the double root; surface gravity always scales with its generator, as emphasized in Bardeen, Carter, and Hawking 1973, pp. 164–166.

Charge makes the multiplicity diagram richer. With geometric charge length $q$, define

$$f(r)=1-\frac{2m}{r}+\frac{q^2}{r^2}-\frac{\Lambda r^2}{3}.$$

A cold limit merges the inner and outer black-hole horizons. A charged Nariai limit merges the outer black-hole and cosmological horizons. An ultracold point merges all three positive horizons. These names record which roots coincide; they should not be inferred merely from whether one reported surface gravity is zero. Romans classifies the charged branches and their parameter ranges in Romans 1992, §§2–3.

The Nariai blow-up

The parameter path is part of the limit, not an optional detail. Choose $\epsilon\gt0$ by

$$9m(\epsilon)^2\Lambda=1-3\epsilon^2.$$

Then

$$m(\epsilon)=\frac1{3\sqrt\Lambda} -\frac{\epsilon^2}{2\sqrt\Lambda}+O(\epsilon^4),$$

so the family approaches $m_0$ at the rate needed to balance the shrinking root separation. Introduce the blown-up coordinates

$$r=r_0+\frac{\epsilon}{\sqrt\Lambda}\cos\chi, \qquad t=\frac{\psi}{\epsilon\sqrt\Lambda},$$

To see the cancellation directly, set $\rho=\sqrt\Lambda r=1+\epsilon\cos\chi$ and $\mu=\sqrt\Lambda m=1/3-\epsilon^2/2+O(\epsilon^4)$. Then

$$f=1-\frac{2\mu}{\rho}-\frac{\rho^2}{3} =\epsilon^2\left(1-\cos^2\chi\right)+O(\epsilon^3) =\epsilon^2\sin^2\chi+O(\epsilon^3).$$

Then the apparent zero-width interval is magnified, and the metric has the finite limit

$$ds^2_{\mathrm N}=\frac1\Lambda \left(-\sin^2\chi\,d\psi^2+d\chi^2+d\Omega_2^2\right).$$

This is the static patch of $dS_2\times S^2$, with both curvature radii equal to $1/\sqrt\Lambda$. The two horizons lie at $\chi=0$ and $\chi=\pi$ rather than at two distinct values of the old areal coordinate. Ginsparg and Perry carry out the near-coalescence scaling in Ginsparg and Perry 1983, pp. 250–254.

The finite proper separation is visible before taking the limit:

$$\int_{r_b}^{r_c}\frac{dr}{\sqrt{f(r)}} \longrightarrow\frac{\pi}{\sqrt\Lambda}.$$

Coordinate separation tends to zero, but proper distance does not. Likewise, $T=\partial_t$ tends to zero relative to $\partial_\psi$ because $T=\epsilon\sqrt\Lambda\,\partial_\psi$. The vanishing of $\kappa[T]$ and the finite acceleration scale of the rescaled Nariai generator are therefore compatible.

The construction is a scaled or pointed limit of the inter-horizon region, not the result of inserting $m=m_0$ into the old chart and declaring that chart complete. Other parts of the original spacetime need not survive this blow-up with the same global interpretation.

Solving the ultracold triple-root conditions

It is convenient to remove the pole at $r=0$:

$$P(r)=r^2f(r)=r^2-2mr+q^2-\frac{r^4}{\ell^2}, \qquad \ell^2=\frac3\Lambda.$$

At an ultracold point,

$$P(r_0)=P'(r_0)=P''(r_0)=0.$$

Since

$$P''(r)=2-\frac{12r^2}{\ell^2},$$

the triple root has

$$r_0^2=\frac{\ell^2}{6}=\frac1{2\Lambda}.$$

The equations $P'=0$ and $P=0$ then give

$$m_0=\frac{2\ell}{3\sqrt6} =\frac{\sqrt2}{3\sqrt\Lambda}, \qquad q_0^2=\frac{\ell^2}{12}=\frac1{4\Lambda}.$$

Finally, $P'''(r_0)=-24r_0/\ell^2\ne0$, confirming multiplicity three rather than four. Every power of $\Lambda$ is fixed by dimensional analysis: $r_0$, $m_0$, and $q_0$ are lengths, while $q_0^2$ is an area.

Approaching this point through the charged family and scaling the shrinking region yields an ultracold product whose two-dimensional factor is locally Minkowski space and whose sphere has radius $r_0$. Different coordinates can display the flat factor as an inertial plane or a Rindler wedge; these are not distinct local geometries. By contrast, the cold double-root branch has an $AdS_2$ factor and the charged Nariai branch has a $dS_2$ factor. The sign of the two-dimensional curvature distinguishes the three cases.

Curvature and area checks

The uncharged Nariai product passes the trace equation immediately. Each two-dimensional factor has scalar curvature $2\Lambda$, so the four-dimensional scalar curvature is $R=4\Lambda$, as required by vacuum Einstein equations. Both horizon cross-sections have

$$\mathcal A=\frac{4\pi}{\Lambda}.$$

For the ultracold product, the flat factor contributes zero scalar curvature. The sphere with $r_0^2=1/(2\Lambda)$ contributes $2/r_0^2=4\Lambda$, again matching the trace equation because Maxwell stress-energy is traceless. This check catches the common erroneous value $r_0^2=1/\Lambda$ for the triple root.

The general Kerr–de Sitter coalescence construction also illustrates why a double root can retain finite proper distance; see Gibbons, Lü, Page, and Pope 2005, §5, especially eqs. (5.6)–(5.9). Rotation changes the fibration and identifications, so the spherical formulas above should not be imported unchanged into a rotating limit.

These limits classify local geometry near a tuned root structure. They do not show that generic evolution reaches the tuned parameters, that the product geometry is nonlinearly stable, or that a quantum state is regular on every component. Formation, stability, and semiclassical backreaction are separate questions. The controlled conclusion here is narrower: after a specified scaling, the exact classical family has a finite limiting metric with the stated root multiplicity and curvature.

Diagnose the zero-width argument

A calculation sets $r_b=r_c=r_0$ directly in static Schwarzschild–de Sitter coordinates and concludes that the inter-horizon geometry has zero spatial size and no finite horizon scale. Identify both mistakes. Then verify that the ultracold parameters really describe a triple root.

Solution

The first mistake equates coordinate separation with proper distance. Near coalescence, $f$ vanishes at the same rate as $(r_c-r_b)^2$, so the integral of $dr/\sqrt f$ remains finite. The radial blow-up replaces the collapsed coordinate interval by $0\lt\chi\lt\pi$.

The second mistake holds the degenerating generator $\partial_t$ fixed. Since $t=\psi/(\epsilon\sqrt\Lambda)$, the finite Nariai generator is proportional to $\partial_t/\epsilon$. Surface gravity rescales with it.

For the charged ultracold data, substitution of $r_0^2=1/(2\Lambda)$, $m_0=\sqrt2/(3\sqrt\Lambda)$, and $q_0^2=1/(4\Lambda)$ makes $P$, $P'$, and $P''$ vanish. Because $P'''$ is nonzero, the root is exactly triple.

Charges and First Laws in de Sitter next tracks how horizon orientation and energy convention affect variational identities.

References

• Bardeen, James M., Brandon Carter, and Stephen W. Hawking. “The Four Laws of Black Hole Mechanics.” Communications in Mathematical Physics 31 (1973): 161–170. DOI.
• Gibbons, Gary W., Hong Lü, Don N. Page, and Christopher N. Pope. “The General Kerr–de Sitter Metrics in All Dimensions.” Journal of Geometry and Physics 53 (2005): 49–73. DOI. Open manuscript.
• Ginsparg, Paul, and Malcolm J. Perry. “Semiclassical Perdurance of de Sitter Space.” Nuclear Physics B 222 (1983): 245–268. DOI.
• Romans, Larry J. “Supersymmetric, Cold and Lukewarm Black Holes in Cosmological Einstein–Maxwell Theory.” Nuclear Physics B 383 (1992): 395–415. DOI. Open manuscript.