The Mass-Inflation Mechanism

Mass inflation is a dynamical amplification mechanism, not the growth of a black hole’s externally measured ADM mass. In spherical charged models, the quasi-local mass associated with an interior symmetry sphere can diverge near the Cauchy horizon while the sphere’s areal radius approaches a finite nonzero value. Counterstreaming fluxes and the inner-horizon blueshift make those two facts compatible.

Helpful background. Black-Hole Geometry and Mechanics supplies horizon geometry and quasi-local mass ideas; Black-Hole Perturbation Theory supplies exterior tails; Formulations of General Relativity supplies characteristic evolution; Causality and Global Structure supplies Cauchy horizons and extensions; Computational Relativity supplies convergence and invariant diagnostics.

Hawking and renormalized mass in double-null form

For this calculation use geometrized variables, $G=c=1$, so the mass functions, $Q$, $r$, $u$, and $v$ have dimensions of length. Work in spherical symmetry with

$$ds^2=-\Omega^2(u,v)\,du\,dv+r^2(u,v)d\Omega_2^2,$$

where $g_{uv}=-\Omega^2/2$. Define the geometric Hawking mass $m_H$ and, for total charge $Q$, the charge-renormalized mass $\varpi$ by

$$m_H=\frac r2\left(1-g^{ab}\nabla_a r\nabla_b r\right), \qquad \varpi=m_H+\frac{Q^2}{2r}.$$

The charged identity is therefore

$$1-\frac{2\varpi}{r}+\frac{Q^2}{r^2} =g^{ab}\nabla_a r\nabla_b r =-\frac{4}{\Omega^2}\,\partial_u r\,\partial_v r.$$

This identity is geometric: $r$ is fixed by the area $A=4\pi r^2$, and the right-hand side is the squared norm of its gradient on the two-dimensional quotient. Solving for $\varpi$ gives

$$\varpi=\frac r2\left( 1+\frac{Q^2}{r^2} +\frac{4\,\partial_u r\,\partial_v r}{\Omega^2} \right).$$

Equivalently,

$$m_H=\frac r2\left( 1+\frac{4\,\partial_u r\,\partial_v r}{\Omega^2} \right).$$

In exterior-normalized double-null coordinates, approaching a Cauchy horizon can give $r\to r_{\rm CH}\gt0$ while $\Omega^2\to0$. The vanishing lapse alone is not an invariant singularity: it also occurs in exact Reissner–Nordström, where the derivative product cancels it. Unless $\partial_u r\,\partial_v r$ vanishes rapidly enough, however, the ratio diverges and so does $\varpi$. Because $Q^2/(2r)$ stays finite when $r\to r_{\rm CH}\gt0$, $m_H$ diverges if and only if $\varpi$ does. Finite size is controlled by $r$, whereas the internal mass depends on the invariant derivative combination.

Restoring dimensions, the physical Hawking mass is $M_{\rm H}=c^2m_H/G$ and its gravitational radius is $G M_{\rm H}/c^2=m_H$. Thus mass inflation is a divergence of $M_{\rm H}$ at finite areal radius, not an enlargement of the exterior ADM mass or horizon area.

The charged Einstein–scalar equations and the monotonicity identities used to control $r$, $m_H$, and $\varpi$ are set out in Dafermos 2005, §§2–5, pp. 8–22, PDF; the mass-inflation region and blowup conclusion are developed in Dafermos 2005, §§9–12, pp. 33–51, PDF.

Counterstreaming supplies the feedback

One flux alone can be strongly blueshifted, but the characteristic mass-inflation picture requires both ingoing and outgoing streams. Late ingoing radiation crosses outgoing radiation generated by earlier collapse or backscattering. Their relative Lorentz factor grows without bound near the would-be inner horizon. In the spherical scalar model, the null energy fluxes are proportional to $(\partial_v\phi)^2$ and $(\partial_u\phi)^2$.

Take a neutral, minimally coupled massless scalar with canonical stress tensor

$$T^{\phi}_{ab}=\nabla_a\phi\nabla_b\phi-\tfrac12g_{ab}(\nabla\phi)^2.$$

For conserved Maxwell charge $Q$ and $G=1$, the exact evolution equation is

$$\partial_v \varpi =\frac{8\pi r^2}{\Omega^2} \bigl(-\partial_u r\bigr)(\partial_v\phi)^2,$$

with the equation obtained by exchanging $u$ and $v$ for the other stream. This follows from the spherical Einstein mass identity, with the Maxwell contribution removed by $Q^2/(2r)$; see Dafermos 2005, “Spherically Symmetric Spacetimes,” §1.2, Eqs. (3)–(4), p. 3, PDF. A scalar rescaling $\psi=\sqrt{4\pi}\phi$ changes the coefficient to $2$, as in common mathematical conventions. A null-coordinate reparametrization changes all derivative and lapse factors together and does not change that fixed coefficient.

In the trapped region $\partial_u r\lt0$, so the mass increases along $v$ when scalar flux is nonzero. Counterstreaming changes the geometric weight $-\partial_u r/\Omega^2$; integrating it against decaying incoming flux can make $\varpi$ diverge. The algebraic mass identity alone does not establish that this dynamical growth occurs.

Poisson and Israel first isolated this nonlinear feedback in a crossflow model, showing that the effective internal mass parameter grows exponentially near the inner horizon Poisson and Israel 1990, §§II–IV, pp. 1797–1807. Ori then constructed an explicit model in which the mass diverges on a null singularity while the areal radius remains finite Ori 1991, pp. 789–792.

Curvature grows before size collapses

An exact spherical curvature identity is more informative than importing a stationary Kretschmann formula. In orthonormal directions tangent to a symmetry sphere, the warped-product curvature gives

$$R_{\hat\theta\hat\phi\hat\theta\hat\phi} =\frac{1-g^{ab}\nabla_a r\nabla_b r}{r^2} =\frac{2m_H}{r^3}.$$

Thus $m_H\to\infty$ at $r\to r_{\rm CH}\gt0$ forces this geometrically selected angular curvature to diverge. A dynamical scalar field also produces Ricci curvature, so a term proportional to $\varpi^2/r^6$ is not a universal leading asymptotic for the Kretschmann scalar. At the same time, the double proper-time integral of the measured tidal matrix can remain finite. That is why mass inflation naturally produces a candidate weak null singularity rather than an immediate spacelike crushing singularity.

The conclusion is model-sensitive. Spherical symmetry removes gravitational radiation and angular structure; charge stands in for the inner-horizon structure that rotation supplies in Kerr. In nonsymmetric vacuum, no scalar Hawking mass captures all degrees of freedom, and one instead tracks null connection coefficients, curvature components, and geometric flux norms.

Discriminating checks

• Verify $r$ from the area, rather than from a coordinate label.
• Verify mass growth through the invariant relation for $\nabla r$, not through a fitted metric coefficient alone.
• Resolve both null fluxes. Suppressing the outgoing stream can erase the nonlinear feedback.
• Separate divergence of $\varpi$, $m_H$, or curvature from the regularity of the metric itself.
• Keep the asymptotic regime subextremal; the $\kappa_-\to0$ extremal limit is singular.

Mass inflation therefore answers one question sharply—how internal energy and curvature can grow at finite radius—while leaving the extension class and the ultimate global shape of the singular boundary to further analysis.

Exercise

Define $X=4\partial_u r\,\partial_v r/\Omega^2$. Suppose along an outgoing ray $r\to r_0\gt0$, $Q$ is fixed, and $X\to+\infty$. Determine the leading mass and angular curvature, and explain why $\Omega^2\to0$ alone would not suffice.

Solution

The mass identity gives $m_H\sim r_0X/2$, while the finite charge correction implies $\varpi\sim m_H$. Therefore $R_{\hat\theta\hat\phi\hat\theta\hat\phi}\sim X/r_0^2$. Under independent reparametrizations of $u$ and $v$, the derivative product and $\Omega^2$ acquire the same factors, leaving $X$ unchanged. A vanishing lapse alone proves nothing because the numerator could vanish at the same rate, as at the smooth stationary inner horizon.

References

• Dafermos, Mihalis, “The Interior of Charged Black Holes and the Problem of Uniqueness in General Relativity”, Communications on Pure and Applied Mathematics 58, 445–504, 2005. Open PDF.
• Dafermos, Mihalis. “Spherically Symmetric Spacetimes with a Trapped Surface.” Classical and Quantum Gravity 22 (2005): 2221–2232. doi:10.1088/0264-9381/22/11/019.
• Ori, Amos, “Inner Structure of a Charged Black Hole: An Exact Mass-Inflation Solution”, Physical Review Letters 67, 789–792, 1991.
• Poisson, Eric, and Werner Israel, “Internal Structure of Black Holes”, Physical Review D 41, 1796–1809, 1990.