BPS Black-Hole Solutions

A BPS black-hole solution is obtained most reliably by setting the fermion supersymmetry variations to zero, not by guessing a metric and checking only Einstein’s equation. The resulting Killing-spinor equations turn part of the second-order field problem into a first-order radial flow. We develop that method for static four-dimensional $\mathcal N=2$ supergravity and explain which checks are still independent.

Required background. Gauge Theory and String-Theory Preparation supplies gauge charges, spinors, and supersymmetry transformations. Gravitational Actions and Boundary Terms supplies the variational and asymptotic framework used to define the mass and conserved charges.

Helpful background. Quantum Gravity places supergravity in its effective-field-theory regime, Quantum Information helps interpret protected sectors, and Black-Hole Geometry and Mechanics supplies horizons, extremality, and surface gravity.

Static supersymmetric ansatz

Take ungauged four-dimensional $\mathcal N=2$ supergravity coupled to Abelian vector multiplets, with scalar coordinates $z^i$ on a special Kähler manifold and a fixed symplectic charge vector

$$\Gamma=(p^I,q_I).$$

We assume spherical symmetry, asymptotic flatness, a single center, and two-derivative equations. In isotropic radius $r$ the metric may be written

$$ds^2=-e^{2U(r)}dt^2+e^{-2U(r)}\bigl(dr^2+r^2d\Omega_2^2\bigr).$$

The superalgebra contains an energy-normalized central charge $Z_E(z,\bar z;\Gamma)=e^{K/2}(q_I X^I-p^I F_I)$ in a common special-geometry normalization. For the radial geometry define

$$Z_{\rm geo}=G_4Z_E, \qquad M_{\rm geo}=G_4M_{\rm ADM}.$$

Both have units of length. A different normalization of the symplectic section rescales intermediate formulae, so $Z_E$, the gauge kinetic matrix, and the charges must be translated together. The original construction and its conventions are given in Ferrara, Kallosh, and Strominger 1995, §2, pp. 2–4, Eqs. (1)–(14).

Killing-spinor equations as radial flow

Set the gravitino and gaugino variations to zero for a nonzero spinor satisfying the radial BPS projector. With inverse radius $\tau=1/r$ and the scalar-metric normalization used here, the equations take the gradient-flow form

$$\frac{dU}{d\tau}=-e^U|Z_{\rm geo}|, \qquad \frac{dz^i}{d\tau}=-2e^U g^{i\bar j}\partial_{\bar j}|Z_{\rm geo}|,$$

with the displayed sign fixed by $\tau$ increasing inward. It makes $e^{-U}$ grow toward a regular extremal throat. Ferrara, Kallosh, and Strominger derive the corresponding equations directly from the fermion variations in Ferrara, Kallosh, and Strominger 1995, §3, pp. 4–5, Eqs. (15)–(19).

The first-order system implies a monotonicity relation:

$$\frac{d|Z_{\rm geo}|}{d\tau} =-4e^U g^{i\bar j}\partial_i|Z_{\rm geo}|\,\partial_{\bar j}|Z_{\rm geo}|\leq0$$

for the displayed orientation and positive scalar metric. The factor four comes from differentiating through both $z^i$ and $\bar z^{\bar i}$; omitting the conjugate contribution would miss a factor two. Monotonicity alone does not prove that a regular endpoint exists: the flow could reach a singular locus or run to the boundary of moduli space. If it reaches a regular finite-area attractor horizon, $\partial_i|Z_{\rm geo}|=0$ fixes the attracted vector-multiplet scalars in terms of $\Gamma$. Hypermultiplet scalars, which do not enter $Z_E$ in ungauged two-derivative $\mathcal N=2$ supergravity, need not be fixed by this argument.

Harmonic functions and a concrete limit

In symmetric models the solution can often be packaged in harmonic functions

$$H^I(r)=h^I+\frac{p^I}{r}, \qquad H_I(r)=h_I+\frac{q_I}{r},$$

with algebraic stabilization equations relating $(X^I,F_I)$ to $(H^I,H_I)$. The constants $h$ encode asymptotic moduli and must satisfy the asymptotic-flatness normalization. Near $r=0$, the charge terms dominate, so the horizon data lose dependence on $h$ whenever the attractor is regular.

The simplest check is extremal Reissner–Nordström embedded in supergravity with constant scalars. In isotropic coordinates,

$$e^{-U}=1+\frac{M_{\rm geo}}{r}, \qquad M_{\rm geo}=|Z_{\rm geo}|=\sqrt{Q_{\rm geo}^2+P_{\rm geo}^2}$$

in a normalization with unit gauge coupling, where $Q_{\rm geo}$ and $P_{\rm geo}$ are the corresponding charge lengths. The surface $r=0$ is not zero area: its areal radius is $e^{-U}r\to M_{\rm geo}$. The near-horizon geometry is $\mathrm{AdS}_2\times S^2$ with equal radii $M_{\rm geo}$. This limit checks both the sign of the flow and the double zero of the lapse.

Multi-charge branes use the same logic but generally have several harmonic functions. Peet works through the BPS harmonic-function rule and the relation among tension, charge, and supersymmetry in Peet 2000, §§2.3 and 3.2, pp. 19–21 and 29–31. Smearing, compactification, and frame changes must be performed before interpreting a lower-dimensional horizon area.

Independent equations and regularity checks

Killing-spinor equations are necessary, but a trustworthy solution also passes four separate tests.

1. The Bianchi identities and Maxwell equations fix the $1/r$ residues and charge normalization.
2. The scalar and Einstein equations must follow from the first-order system, including any integrability conditions. A supersymmetric ansatz can omit an equation if a source or singular locus is present.
3. The horizon area must be positive and curvature invariants finite in a freely falling frame. A formal $r=0$ limit with vanishing area is a “small” black hole at two derivatives, not a regular macroscopic horizon.
4. The asymptotic supercharge algebra must reproduce $M_{\rm ADM}=|Z_E(z_\infty;\Gamma)|$, equivalently $M_{\rm geo}=|Z_{\rm geo}(z_\infty;\Gamma)|$, and the number of independent Killing spinors must match the claimed supersymmetry fraction.

The method is controlled within classical two-derivative supergravity when curvatures are small in string units and the effective string coupling is small. Higher-derivative terms modify the fermion variations and stabilization equations. Gauging adds a scalar potential and can change the asymptotics. Rotation and multi-center solutions introduce angular-momentum one-forms and additional integrability conditions. None should be inferred from the static single-center equations above.

Exercises

Derive the near-horizon radius of the extremal Reissner–Nordström embedding from $e^{-U}=1+M_{\rm geo}/r$, and verify that the horizon area is finite although the isotropic radius vanishes.

Solution

The coefficient of $d\Omega_2^2$ is $e^{-2U}r^2=(r+M_{\rm geo})^2$, so the areal radius tends to $M_{\rm geo}$ as $r\to0$ and $A_H=4\pi M_{\rm geo}^2$. Set $r=M_{\rm geo}^2/\rho$ near the horizon. Keeping the leading terms gives $ds^2\simeq -(M_{\rm geo}^2/\rho^2)dt^2+(M_{\rm geo}^2/\rho^2)d\rho^2+M_{\rm geo}^2d\Omega_2^2$ after a harmless radial redefinition, displaying $\mathrm{AdS}_2\times S^2$ with radius $M_{\rm geo}$. Therefore the coordinate value $r=0$ does not determine the geometric size. This check also confirms that the vanishing lapse and finite sphere radius describe an extremal horizon rather than an ordinary coordinate origin.

References

• Ferrara, Sergio, Renata Kallosh, and Andrew Strominger. “N=2 Extremal Black Holes.” Physical Review D 52 (1995): R5412–R5416. DOI. Open manuscript.
• Peet, Amanda W. “TASI Lectures on Black Holes in String Theory.” arXiv:hep-th/0008241 (2000). Open manuscript.