Vector-Field and Weighted-Energy Methods

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.

Vector-field methods turn spacetime symmetries and approximate symmetries into energy, decay, and integrated-flux estimates. The basic identity is exact; the art lies in choosing multipliers and commutators whose boundary fluxes are coercive and whose deformation errors are integrable. Weighted $r^p$ estimates then move energy through outgoing null hypersurfaces and recover decay without relying on a Fourier decomposition.

Required background. Analysis for Mathematical Relativity supplies Sobolev and trace estimates; The Einstein Equation supplies the nonlinear application; PDEs and Green Functions for Gravitation supplies wave energy identities.

Helpful background. Causality and Global Structure clarifies null fluxes and completeness.

The multiplier identity

For a scalar field satisfying

\Box_g\phi=F,

define the stress tensor

T_{\mu\nu}[\phi] =\nabla_\mu\phi\nabla_\nu\phi -\frac12g_{\mu\nu} \nabla^\alpha\phi\nabla_\alpha\phi.

Given a vector field $X$ , let

J^X_\mu=T_{\mu\nu}X^\nu, \qquad {}^{(X)}\pi_{\mu\nu} =\frac12(\nabla_\mu X_\nu+\nabla_\nu X_\mu).

Direct differentiation gives

\nabla^\mu J^X_\mu =T^{\mu\nu}{}^{(X)}\pi_{\mu\nu} +F\,X\phi.

Integrating over a spacetime slab converts the divergence into fluxes across its spacelike, timelike, or null boundary. If $X$ is future timelike and the dominant energy condition holds for $T[\phi]$ , the spacelike energy flux is nonnegative. If $X$ is Killing, ${}^{(X)}\pi=0$ , so a homogeneous wave has a conserved $X$ -energy.

For Einstein’s equations, one applies analogous identities to wave-gauge metric components, curvature through the Bianchi equations, or gauge-invariant combinations. The scalar identity is the model, not a claim that gravitational energy has a local tensor $T_{\mu\nu}$ .

Commutators and pointwise decay

Minkowski space has translations, rotations, boosts, and the scaling vector field. Commuting a linear wave with these fields produces another wave, exactly for Killing fields and up to a controlled multiple for scaling. Energies of the commuted fields control weighted derivatives of $\phi$ .

The Klainerman–Sobolev inequality converts those energies to pointwise decay. Schematically in $3+1$ dimensions,

\lvert\partial\phi(t,x)\rvert \lesssim \frac{1}{(1+t+r)(1+\lvert t-r\rvert)^{1/2}} \sum_{\lvert I\rvert\le3}\lVert\partial Z^I\phi(t)\rVert_{L^2},

where $Z$ ranges over the commutation fields. The extra factor away from the light cone is what makes many quadratic errors time-integrable. Klainerman derives the invariant vector-field decay mechanism in Klainerman 1985, pp. 321–332.

On a perturbed spacetime these fields are only approximately Killing. Their deformation tensors generate error integrals, and commutation with $\Box_g$ introduces derivatives of $g$ . A nonlinear proof assigns different weights to good tangential derivatives and bad transversal derivatives so that these errors do not lose the hierarchy.

Morawetz and integrated local energy

A radial multiplier $X=f(r)\partial_r$ is not Killing. Its deformation tensor produces a bulk term. After adding a lower-order correction current and choosing $f$ appropriately, one obtains an integrated local-energy or Morawetz estimate: energy cannot remain indefinitely concentrated in a fixed compact region.

The raw sign is not automatically positive. In black-hole exteriors, trapping creates a degeneracy or derivative loss near the photon region. Even in flat space, the correction term is needed to balance radial and angular pieces. Thus “use a radial multiplier” is not a proof until the full bulk density and boundary terms have been checked.

The rᵖ hierarchy

The flat-space calculation fixes the powers of $r$ that persist in curved far-field estimates. Set $u=t-r$ , $v=t+r$ , so that $r=(v-u)/2$ , and put $\psi=r\phi$ . Let $\not\nabla$ and $\Delta_{S^2}$ refer to the unit round sphere, with $\Delta_{S^2}Y_{\ell m}=-\ell(\ell+1)Y_{\ell m}$ . The equation $\Box_m\phi=0$ becomes

-4\partial_u\partial_v\psi+r^{-2}\Delta_{S^2}\psi=0.

Multiply by $r^p\partial_v\psi$ and integrate by parts on $S^2$ . With $d\omega$ its area element, the resulting exact identity is

\begin{aligned} &\partial_u\!\int_{S^2}2r^p\lvert\partial_v\psi\rvert^2\,d\omega +\partial_v\!\int_{S^2}\frac12r^{p-2} \lvert\not\nabla\psi\rvert^2\,d\omega\\ &\quad+\int_{S^2}\left[ p\,r^{p-1}\lvert\partial_v\psi\rvert^2 +\frac{2-p}{4}r^{p-3}\lvert\not\nabla\psi\rvert^2 \right]d\omega=0. \end{aligned}

Thus the bulk measure is $du\,dv\,d\omega$ , and the unit-sphere angular term has two fewer powers of $r$ than the outgoing-derivative term. If one instead uses the physical angular gradient $r^{-1}\not\nabla$ , the two displayed weights can both be written $r^{p-1}$ . Keeping track of that choice prevents a dimensional error. On curved exteriors, additional lower-order terms and inner-boundary fluxes must be estimated.

Both bulk coefficients are nonnegative for

0\le p\le2.

At the endpoints one term vanishes: strict bulk control of both terms uses $0\lt p\lt2$ . Comparing successive weighted fluxes and combining them with integrated local energy yields polynomial decay. Dafermos and Rodnianski develop this hierarchy and its black-hole applications in Dafermos–Rodnianski 2010, §§2–5.

For nonlinear Einstein evolution, the hierarchy must be coupled to constraint propagation, asymptotic mass corrections, and commuted curvature estimates. The Minkowski stability theorem uses these ingredients in a global bootstrap rather than as isolated estimates.

The hypotheses behind the displayed identities include a sufficiently regular solution on a globally hyperbolic slab, finite boundary fluxes, and decay or compact support that removes uncontrolled flux at spatial infinity. Sources must lie in the dual spacetime norms used by the energy estimate. A central limitation is that the flat-space commutators cease to be exact symmetries on a curved background; their deformation and commutator errors must be estimated rather than discarded. Trapping can also prevent a globally coercive radial bulk term.

Solved diagnostic: three multiplier signs

For a homogeneous Minkowski wave, compare the bulk terms generated by $\partial_t$ , by $f(r)\partial_r$ with $f\ge0$ , and by the $r^p$ identity. Does each choice automatically give a positive bulk term?

Solution

For the homogeneous Minkowski wave:

With $X=\partial_t$ , the deformation tensor vanishes, so the bulk term is zero and the time-translation energy is conserved.
With a radial $X=f(r)\partial_r$ , the deformation tensor does not vanish. No sign follows from $f\ge0$ alone; one must inspect $f'$ , $f/r$ , and the correction current. This diagnoses why a naive radial-energy proof is incomplete.
In the displayed normalization of the $r^p$ identity, $p=0$ leaves an angular coefficient $1/2$ , while $p=2$ leaves an outgoing coefficient $2$ . For $p\gt2$ , the angular coefficient is negative; for $p\lt0$ , the outgoing coefficient is negative. The interval $[0,2]$ follows from the bulk sign, but the endpoints do not control both derivatives.

The next page explains why the nonlinear source terms produced by commutation are integrable only after using the null structure of Einstein’s equations.

References

Dafermos, Mihalis, and Igor Rodnianski. “A New Physical-Space Approach to Decay for the Wave Equation with Applications to Black Hole Spacetimes.” In XVIth International Congress on Mathematical Physics, 421–432. World Scientific, 2010. arXiv:0910.4957.
Klainerman, Sergiu. “Uniform Decay Estimates and the Lorentz Invariance of the Classical Wave Equation.” Communications on Pure and Applied Mathematics 38 (1985): 321–332. DOI.