Scalar-Wave Separation

The massive scalar equation on Kerr reduces exactly from a partial differential equation in four variables to independent radial and angular ordinary differential equations. Stationarity and axisymmetry remove $t$ and $\phi$; the same hidden symmetry that produces the Carter constant completes the separation in $r$ and $\theta$. This page fixes the separation and boundary-sign conventions but does not compute spectra or Green functions.

Required background. Killing Fields and Isometry Groups supplies the stationary and axial mode labels; Affine and Variational Geodesics supplies the geometric-optics connection to separated rays.

Helpful background. Einstein’s Equation fixes the Kerr background; The Riemann Curvature Tensor supplies covariant derivatives; Exact Solutions of Einstein’s Equations explains the metric and its domain; Frames, Coframes, and Anholonomy prepares the spin-field comparison.

Write $m_K=G M_{\mathrm{phys}}$ for the Kerr gravitational length and $a=J_{\mathrm{phys}}/M_{\mathrm{phys}}$ for its specific angular momentum, with $c=1$. The unadorned integer $m$ below is instead the azimuthal Fourier label.

Separating the Klein–Gordon equation

For a scalar of mass $\mu_s$, the equation in the site signature is

$$(\Box-\mu_s^2)\Phi=0, \qquad \Box\Phi=\frac{1}{\sqrt{-g}}\partial_a \left(\sqrt{-g}\,g^{ab}\partial_b\Phi\right).$$

Use Boyer–Lindquist coordinates, $\Sigma=r^2+a^2\cos^2\theta$, $\Delta=r^2-2m_Kr+a^2$, and the mode convention

$$\Phi=e^{-i\omega t+im\phi}R(r)S(\theta).$$

Single-valuedness around the rotation axis gives integer $m$. The frequency $\omega$ is real for a Fourier decomposition of suitable real-time data; it may later be analytically continued when a spectral problem supplies boundary conditions.

Multiplying the field equation by $\Sigma/\Phi$ leaves a sum of a radial and an angular expression. With angular eigenvalue $A$, the angular equation is

$$\frac{1}{\sin\theta}\frac{d}{d\theta} \left(\sin\theta\frac{dS}{d\theta}\right) +\left[a^2(\omega^2-\mu_s^2)\cos^2\theta -\frac{m^2}{\sin^2\theta}+A\right]S=0.$$

Regularity at both axes selects spheroidal eigenvalues $A=A_{\ell m}(c^2)$, with $\ell\geq |m|$ and $c^2=a^2(\omega^2-\mu_s^2)$. With this sign convention the angular functions are oblate spheroidal for $c^2>0$; for $c^2<0$ they are commonly described as prolate spheroidal, or equivalently as the analytic continuation of the same eigenvalue problem. Define

$$\lambda=A+a^2\omega^2-2am\omega.$$

The radial equation then becomes

$$\frac{d}{dr}\left(\Delta\frac{dR}{dr}\right) +\left[ \frac{\big((r^2+a^2)\omega-am\big)^2}{\Delta} -\mu_s^2r^2-\lambda \right]R=0.$$

Displaying both $A$ and $\lambda$ is essential because the literature calls either one the separation constant. These equations follow from the commuting symmetry operators associated with the principal tensor, as explained in Frolov, Krtouš, and Kubizňák 2017, §§6.2–6.3, pp. 121–127, Open PDF.

The invariant version uses the Carter Killing tensor to form a second-order operator of the schematic form

$$\mathcal K\Phi=\nabla_a\!\left(K^{ab}\nabla_b\Phi\right).$$

For the Kerr Killing–Yano construction, curvature identities ensure that this operator commutes with the scalar wave operator. A separated mode is a simultaneous eigenfunction of $i\partial_t$, $-i\partial_\phi$, and $\mathcal K$. The angular eigenvalue is therefore the field-theory counterpart of the quadratic geodesic integral, with convention-dependent shifts by the manifest mode labels. This operator statement survives a change of separation coordinates and explains why two Killing vectors alone are insufficient.

Boundary behavior without a spectral claim

For a subextremal hole, $r_+=m_K+\sqrt{m_K^2-a^2}$ and $\Omega_H=a/(r_+^2+a^2)$. Introduce the tortoise coordinate by

$$\frac{dr_*}{dr}=\frac{r^2+a^2}{\Delta}.$$

Near the future horizon, local radial factors behave as

$$R\sim e^{\mp i(\omega-m\Omega_H)r_*}.$$

With $e^{-i\omega t}$, the minus sign is ingoing and regular in a future-horizon chart; the plus sign is outgoing there. At asymptotically large $r$, a massless field has $R\sim r^{-1}e^{\pm i\omega r_*}$, whereas a massive field uses wave number $k=(\omega^2-\mu_s^2)^{1/2}$ and becomes evanescent when the chosen branch has $\omega^2<\mu_s^2$.

These local statements do not define quasinormal modes, bound states, or a retarded Green function. Each of those requires a complete domain, endpoint conditions, an analytic prescription, and, for complex $\omega$, careful branch choices. Separation is a geometric reduction; a spectrum is an additional boundary-value problem.

Independent limits and interpretation

When $a\to0$, the angular equation becomes the associated Legendre equation and $A\to\ell(\ell+1)$. The radial equation reduces to the Schwarzschild massive-scalar equation. In the high-frequency WKB limit, writing $R\sim e^{iS_r}$ and $S\sim e^{iS_\theta}$ reproduces the separated null Hamilton–Jacobi equation, linking the wave eigenvalue to the Carter constant at leading order.

The exact separation is not implied merely by the two coordinate symmetries. A generic stationary-axisymmetric metric still permits $e^{-i\omega t+im\phi}$, but the remaining two-variable equation need not split. Commutation of the wave operator with the second-order operator constructed from the Killing tensor supplies the invariant explanation.

Exercise

Take $a=0$ and show that regularity makes $A=\ell(\ell+1)$.

Solution

The angular equation becomes $\sin^{-1}\theta(\sin\theta S')'+[A-m^2/\sin^2\theta]S=0$. With $x=\cos\theta$, this is the associated Legendre equation $(1-x^2)S_{xx}-2xS_x+[A-m^2/(1-x^2)]S=0$. Regularity at $x=\pm1$ terminates the hypergeometric series and requires $A=\ell(\ell+1)$ for integers $\ell\geq|m|$.

References

• Frolov, Valeri P., Pavel Krtouš, and David Kubizňák. “Black Holes, Hidden Symmetries, and Complete Integrability.” Living Reviews in Relativity 20 (2017): 6. DOI; Open PDF.
• Teukolsky, Saul A. “Rotating Black Holes: Separable Wave Equations for Gravitational and Electromagnetic Perturbations.” Physical Review Letters 29 (1972): 1114–1118. DOI.