Precessing Frames and Mode Mixing

Precession makes a nearly simple radiation pattern sweep across an inertial sky. A time-dependent rotation into a coprecessing frame can concentrate that motion into the frame itself, but it cannot remove intrinsic asymmetry or nonlinear source evolution. The key distinction is algebraic: a rigid rotation mixes $m$ values unitarily at fixed $\ell$, whereas differentiating a time-dependent rotation adds angular-velocity terms to the news. This page constructs that separation with declared Wigner-$D$ indices and the minimal-rotation condition. It does not prescribe source precession dynamics or infer physical parameters.

Required background. Spacetime Symmetries and Killing Fields supplies rotation generators, angular momentum operators, and group composition.

Helpful background. Einstein Equations supplies the radiative tensor field; Waves, Green Functions, and Geometric Optics supplies time-dependent basis expansions; and Gravitational Radiation and Waveforms supplies spin-$-2$ strain and news modes.

Choosing a radiation axis

Let $H(u,\boldsymbol n)$ be the asymptotic complex strain in a fixed Bondi inertial frame. One convention-explicit radiation-axis prescription uses the symmetric angular tensor

$$Q_{ab}(u) =\frac{ \int_{S^2}\bar H\,L_{(a}L_{b)}H\,d\Omega }{ \int_{S^2}\lvert H\rvert^2d\Omega },$$

where $L_a$ are the rotation generators in the declared spin-weighted representation. A principal eigenvector $\boldsymbol a(u)$ defines an axis when its largest eigenvalue is nondegenerate. This construction is based on radiation, not on an assumed orbital vector. Near degeneracies, low amplitude, or an insufficient mode set, the axis can become ill conditioned.

Choose an active rotation $R(u)$ that maps the coprecessing $z$ axis to $\boldsymbol a(u)$ in the inertial frame. The sign $\boldsymbol a\leftrightarrow-\boldsymbol a$ and the rotation about $\boldsymbol a$ remain to be fixed. Boyle and collaborators develop radiation-axis and rotating-frame choices and examine which asymmetries survive them in Boyle et al. 2014, §§II–VI, PDF.

Mode mixing under the rotation

With the active Wigner convention used in this volume,

$$H^I_{\ell m}(u) =\sum_{m'=-\ell}^{\ell} D^\ell_{m m'}(R(u))H^C_{\ell m'}(u),$$

where $I$ denotes the inertial frame and $C$ the coprecessing frame. The inverse relation uses $D^\ell(R)^\dagger$. Unitarity gives

$$\sum_m\lvert H^I_{\ell m}\rvert^2 =\sum_m\lvert H^C_{\ell m}\rvert^2$$

at every retarded time. A pure rotation mixes only $m$ at fixed $\ell$. This norm and selection rule independently detect an index reversal, a nonunit quaternion, or an accidental passive transformation.

The news requires more care. Differentiating gives

$$N^I_{\ell m} =\sum_{m'} \left[ \dot D^\ell_{m m'}(R)H^C_{\ell m'} +D^\ell_{m m'}(R)\dot H^C_{\ell m'} \right].$$

The first term is frame motion. Simply differentiating coprecessing coefficients and rotating the result omits it, so the reconstructed inertial luminosity is wrong. Equivalently, time differentiation at fixed inertial angle differs from differentiation at fixed rotating coordinate.

Fixing the residual rotation

Mapping $z$ to $\boldsymbol a$ leaves a rotation about that axis. The minimal-rotation condition sets the frame angular velocity along $\boldsymbol a$ to zero:

$$\boldsymbol\varpi_{\rm frame}\!\cdot\boldsymbol a=0.$$

For the declared active $z$–$y$–$z$ Euler decomposition

$$R=R_z(\alpha)R_y(\beta)R_z(\gamma),$$

this becomes

$$\dot\gamma=-\dot\alpha\cos\beta.$$

One way to see the result is to add the angular velocities of the three Euler rotations. Their projections along the final axis are $\dot\alpha\cos\beta$, zero, and $\dot\gamma$, respectively. Setting their sum to zero gives the displayed condition. The coordinate-independent derivation and Euler-angle form are in Boyle, Owen, and Pfeiffer 2011, §III, p. 5, eqs. (10)–(18), PDF.

An initial constant in $\gamma$ remains and must be reported. Euler angles are coordinate singular when the axis approaches a pole, so the geometric condition—not a particular angle chart—is the definition. A corotating frame deliberately adds an axial rotation that follows waveform phase; it is useful for compression but is not the minimal-rotation coprecessing frame.

Frame-induced and intrinsic structure

If a simple coprecessing mode set produces many inertial $m$ modes exactly through the Wigner matrix, that mixing is kinematic. It can be undone by the inverse rotation. Structure that remains after a well-defined inverse rotation may be intrinsic, including unequal symmetry-related modes, mode-dependent phase evolution, or radiation-axis ambiguity.

No rotating frame generically restores a nonprecessing reflection relation. Boyle and collaborators show that precessing radiation can break reflection symmetry across an orbital plane even in parity-invariant general relativity Boyle et al. 2014, §§III–VI, PDF. This is a property of the source and its waveform, not parity violation in the gravitational action. Conversely, a translation or center-of-mass drift can also produce apparent mixing, but that is a BMS transformation rather than an $SO(3)$ rotation. Boyle’s full asymptotic transformation rules distinguish these cases Boyle 2016, §§III–VI, pp. 9–19, PDF.

The nonprecessing limit is decisive: if $\boldsymbol a$ is constant, choose constant $R$ and the frame-motion term vanishes. A prescription that continues to generate time-dependent mixing in that limit has not fixed its residual axial gauge consistently.

As a final check, rotate the reconstructed coprecessing field back to the inertial sky and compare the full angular field, not only its dominant mode.

Exercise

Let the coprecessing frame contain only

$$H^C_{22}=Ae^{-i\omega u},$$

and take $R(u)=R_z(\alpha(u))$. Find $H^I_{22}$ and $N^I_{22}$ in the active convention. Check the strain norm and identify the frame contribution.

Solution

Because $D^2_{22}(R_z)=e^{-2i\alpha}$,

$$H^I_{22} =Ae^{-i\omega u-2i\alpha(u)}.$$

Thus

$$N^I_{22} =-i\left(\omega+2\dot\alpha\right)H^I_{22}.$$

The strain power remains $\lvert H^I_{22}\rvert^2=\lvert A\rvert^2$. The term $-2i\dot\alpha H^I_{22}$ comes from $\dot D H^C$ and would be missed by rotating only $\dot H^C$. When $\alpha$ is constant, that term vanishes and the nonprecessing rigid-rotation limit is recovered.

References

• Boyle, Michael, Lawrence E. Kidder, Serguei Ossokine, and Harald P. Pfeiffer. “Gravitational-Wave Modes from Precessing Black-Hole Binaries.” Physical Review D 90 (2014): 124010. DOI. Open PDF.
• Boyle, Michael, Robert Owen, and Harald P. Pfeiffer. “Geometric Approach to the Precession of Compact Binaries.” Physical Review D 84 (2011): 124011. DOI. Open PDF.
• Boyle, Michael. “Transformations of Asymptotic Gravitational-Wave Data.” Physical Review D 93 (2016): 084031. DOI. Open PDF.