Skip to content

Extensions and Controlled Limits of the C-Metric

Scrollable formula. Use Left and Right arrow keys to reveal hidden content.

Scrollable table. Use Left and Right arrow keys to reveal additional columns.

Scrollable figure. Use Left and Right arrow keys to inspect hidden content.

Several important C-metric limits are smooth as geometries but singular in the original polynomial coordinates. The Schwarzschild limit sends yy to infinity while rescaling time; the massless limit is flat spacetime in an accelerated wedge; a double-horizon limit magnifies a shrinking interval. A controlled limit specifies which geometric quantities remain finite and transforms coordinates before setting the parameter to its endpoint.

Required background. Killing Fields and Isometry Groups supplies boost, stationary, and horizon generators.

Helpful background. The Einstein Equation supplies the limiting field equations; The Riemann Curvature Tensor supplies invariant convergence tests; Exact Solutions of Einstein’s Equations places the family; Global Structure develops maximal extensions; and Frames, Coframes, and Anholonomy supplies regular limiting frames.

Zero acceleration: keep physical coordinates fixed

Section titled “Zero acceleration: keep physical coordinates fixed”

The factorized chart contains the prefactor A−2(x−y)−2A^{-2}(x-y)^{-2}, so substituting A=0A=0 there is meaningless. Instead set

x=cos⁡θ,y=−1Ar,τ=At,x=\cos\theta, \qquad y=-\frac1{Ar}, \qquad \tau=At,

before taking the limit. The metric becomes

ds2=1(1+Arcos⁡θ)2[−Qdt2+dr2Q+r2(dθ2P+Psin⁡2θ dϕ2)],ds^2=\frac1{(1+Ar\cos\theta)^2} \left[-Qdt^2+\frac{dr^2}{Q} +r^2\left(\frac{d\theta^2}{P}+P\sin^2\theta\,d\phi^2\right)\right],

with P=1+2mAcos⁡θP=1+2mA\cos\theta and

Q=(1−A2r2)(1−2m/r).Q=(1-A^2r^2)(1-2m/r).

At fixed (t,r,θ,ϕ)(t,r,\theta,\phi),

A→0:ds2⟶−(1−2mr)dt2+dr21−2m/r+r2dΩ2.A\to0: \qquad ds^2\longrightarrow -\left(1-\frac{2m}{r}\right)dt^2 +\frac{dr^2}{1-2m/r}+r^2d\Omega^2.

Meanwhile the acceleration horizon r=A−1r=A^{-1} recedes to infinity. The black-hole root remains at r=2mr=2m. This simultaneous scaling is the coordinate transformation that connects the polynomial root y=−1/(2mA)y=-1/(2mA) to a finite Schwarzschild radius.

Set m=0m=0 and choose a regular angular identification K=1K=1. Then

ds2=1Ω2[−(1−A2r2)dt2+dr21−A2r2+r2dΩ22],Ω=1+Arcos⁡θ.ds^2=\frac1{\Omega^2} \left[-(1-A^2r^2)dt^2 +\frac{dr^2}{1-A^2r^2}+r^2d\Omega_2^2\right], \qquad \Omega=1+Ar\cos\theta.

Define cylindrical Minkowski coordinates by

T=1−A2r2AΩsinh⁡(At),Z=1−A2r2AΩcosh⁡(At),X=rsin⁡θΩcos⁡ϕ,Y=rsin⁡θΩsin⁡ϕ.\begin{aligned} T&=\frac{\sqrt{1-A^2r^2}}{A\Omega}\sinh(At),\\ Z&=\frac{\sqrt{1-A^2r^2}}{A\Omega}\cosh(At),\\ X&=\frac{r\sin\theta}{\Omega}\cos\phi, \qquad Y=\frac{r\sin\theta}{\Omega}\sin\phi. \end{aligned}

Direct differentiation gives ds2=−dT2+dX2+dY2+dZ2ds^2=-dT^2+dX^2+dY^2+dZ^2. The static Killing field becomes

∂t=A(Z∂T+T∂Z),\partial_t=A\left(Z\partial_T+T\partial_Z\right),

the Lorentz-boost generator. Its norm vanishes at r=A−1r=A^{-1}, which is therefore a Rindler horizon in the massless limit. If K≠1K\ne1, the spacetime remains locally flat away from the axis but retains a conical quotient; zero mass does not automatically erase a fixed global defect.

This calculation explains the term “boost-rotation symmetry.” In the static wedge the boost generator looks like time translation. Extending through the acceleration horizon reveals regions where it is spacelike, just as ordinary Rindler time does in Minkowski space. The massive C-metric deforms this boost-adapted geometry rather than possessing a globally timelike stationary field.

Charge or rotation can make the inner and outer black-hole roots coincide. Acceleration can also bring the outer black-hole root to an acceleration root. These operations must distinguish the throat of an already extremal metric from a parameter limit in which two distinct roots coalesce. Let QϵQ_\epsilon be a family whose relevant root separation is O(ϵ)O(\epsilon), and require the scaled expansion

Qϵ(r0+ϵρ)=ϵ2Q^(ρ)+O(ϵ3).Q_\epsilon(r_0+\epsilon\rho) =\epsilon^2\widehat Q(\rho)+O(\epsilon^3).

The coordinate interval collapses in the unscaled variables. Introduce

r=r0+ϵρ,t=t^ϵ,r=r_0+\epsilon\rho, \qquad t=\frac{\hat t}{\epsilon},

and take ϵ→0\epsilon\to0 while tuning the parameters along the declared family. The radial-time sector approaches

−Q^(ρ)dt^2+dρ2Q^(ρ),-\widehat Q(\rho)d\hat t^2 +\frac{d\rho^2}{\widehat Q(\rho)},

up to the finite angular conformal factor at r0r_0. This displayed patch is static only where Q^>0\widehat Q>0; across a root, the coordinate roles change. For an exact double root of a fixed metric, Q=c(r−r0)2+O((r−r0)3)Q=c(r-r_0)^2+O((r-r_0)^3) gives Q^=cρ2\widehat Q=c\rho^2. A coalescing pair generally retains a detuning term—for example, Q^=C(ρ02−ρ2)\widehat Q=C(\rho_0^2-\rho^2)—and can instead produce a locally de Sitter two-factor, as in a Nariai-type limit. Without the inverse scaling of tt, the time term vanishes and the limiting metric is degenerate. Charged extremality, black-hole/acceleration coincidence, and cosmological coincidence can therefore have different transverse geometries and must not all be called the same near-horizon limit.

An invariant check is to compare curvature in an orthonormal frame after scaling. A finite limiting Kretschmann scalar and nondegenerate frame metric show that a genuine spacetime limit has been obtained. Merely observing a repeated polynomial root does not.

Limits can fail to commute when global identifications vary. For example, regularizing one axis sets K=1±2mAK=1\pm2mA in the vacuum family. Taking m→0m\to0 then sends K→1K\to1 and removes the defect. Holding a nonzero string tension instead requires a different correlated scaling, leaving a flat conical spacetime. Likewise, taking A→0A\to0 at fixed mm differs from holding mAmA fixed, which forces mm to diverge.

Each limit should state what is fixed: local coordinates, conserved charges or metric parameters, axial period, cosmological radius, and the selected root interval. The classic extension analysis that established the accelerating interpretation is Kinnersley and Walker 1970, pp. 1359–1370.

Substituting a limiting parameter before transforming coordinates. A divergent prefactor or disappearing interval may be a chart failure. Identify finite invariant scales and rescale first.

Ignoring the angular identification. Local curvature can converge while a conical quotient survives. A spacetime limit includes topology and periodicity, not just metric coefficients.

Verify from the massless transformation that, away from its conformal-infinity corner, the acceleration horizon maps to a Minkowski null surface.

Solution

Introduce the Minkowski null coordinates U=Z−TU=Z-T and V=Z+TV=Z+T. The transformation gives

U=1−A2r2AΩe−At,V=1−A2r2AΩeAt,UV=Z2−T2=1−A2r2A2Ω2.\begin{aligned} U&=\frac{\sqrt{1-A^2r^2}}{A\Omega}e^{-At},\\ V&=\frac{\sqrt{1-A^2r^2}}{A\Omega}e^{At},\\ UV&=Z^2-T^2=\frac{1-A^2r^2}{A^2\Omega^2}. \end{aligned}

Fix θ≠π\theta\ne\pi, so that Ω→1+cos⁡θ\Omega\to1+\cos\theta remains finite and nonzero at r=A−1r=A^{-1}. At finite tt, taking this limit sends both UU and VV to zero. This reaches the bifurcation two-surface T=Z=0T=Z=0, not a generic point of either null branch. To reach the future horizon U=0U=0 with finite nonzero VV, take r→A−1r\to A^{-1} and t→+∞t\to+\infty at fixed θ≠π\theta\ne\pi while holding 1−A2r2 eAt/(AΩ)\sqrt{1-A^2r^2}\,e^{At}/(A\Omega) fixed. The past horizon is obtained with t→−∞t\to-\infty and the opposite exponential fixed, giving V=0V=0 with finite nonzero UU. The branches bounding the static wedge are therefore the null half-planes U=0,V>0U=0,V>0 and V=0,U>0V=0,U>0; the analytic Minkowski extension supplies the complete null planes Z=±TZ=\pm T. The excluded corner r=A−1,θ=πr=A^{-1},\theta=\pi also has Ω=0\Omega=0 and belongs to conformal infinity, so it is not covered by this horizon limit. At the same coordinate root away from that corner, g(∂t,∂t)=−(1−A2r2)/Ω2g(\partial_t,\partial_t)=-(1-A^2r^2)/\Omega^2 vanishes, so it is exactly the Rindler Killing horizon while all Minkowski curvature invariants remain zero.

  • William Kinnersley and Martin Walker, “Uniformly Accelerating Charged Mass in General Relativity,” Physical Review D 2, 1359–1370 (1970). DOI.