Inspiral Phase and Amplitude Structure

A quasi-circular inspiral waveform is organized by a slowly evolving orbital scale, a rapidly accumulated carrier phase, and a hierarchy of angular modes and amplitude corrections. The total mass fixes the time and length scale; the symmetric mass ratio fixes the leading chirp; mass asymmetry and dimensionless spins control further structure. This page derives the leading waveform scalings while treating higher-order orbital dynamics as an accepted input.

Helpful background. Einstein Equations supplies the field dynamics behind energy balance; Waves, Green Functions, and Geometric Optics supplies outgoing-wave phase conventions; Relativistic Two-Body Dynamics supplies the binding energy and orbital evolution used here; Gravitational Radiation and Waveforms supplies radiative moments and flux; and Computational Relativity explains how the late waveform is obtained when the inspiral expansion fails.

Quasi-circular source invariants

Let source-frame component masses satisfy $m_1\ge m_2$, and define

$$M=m_1+m_2,\qquad \mu=\frac{m_1m_2}{M},\qquad \nu=\frac{\mu}{M},\qquad \delta=\frac{m_1-m_2}{M}.$$

The symmetric mass ratio obeys $0\lt\nu\le1/4$, with equality for equal masses, while $\delta^2=1-4\nu$. For quasi-circular orbital angular frequency $\Omega$, the invariant velocity parameter is

$$x=(GM\Omega)^{2/3}.$$

The dimensionless spins are $\boldsymbol\chi_A=\boldsymbol S_A/(Gm_A^2)$. In a nonprecessing system their projections along the orbital angular momentum are constant on the orbital timescale; in a precessing system the full vectors and a rotating frame are required.

This page assumes an adiabatic hierarchy

$$T_{\mathrm{orb}}\ll T_{\mathrm{rr}},$$

where radiation reaction changes $x$ little during one orbit. The derivation of the binary equations of motion is not treated here, and observational parameter estimation is a separate downstream problem.

Carrier phase and mode amplitudes

At large source-frame areal radius $r$, define the distance-rescaled complex strain

$$H(t,\theta,\phi) =r\left(h_+-ih_\times\right) =\sum_{\ell=2}^{\infty}\sum_{m=-\ell}^{\ell} H_{\ell m}(t)\,{}_{-2}Y_{\ell m}(\theta,\phi),$$

with unit-normalized spin-weight $-2$ harmonics:

$$\int d\Omega\, {}_{-2}Y_{\ell m}\, {}_{-2}\overline{Y}_{\ell' m'} =\delta_{\ell\ell'}\delta_{mm'}.$$

The global time convention is $e^{-i\omega t}$. For a nonprecessing quasi-circular source, a mode can be organized as

$$H_{\ell m} =GM\nu\,x\,\widehat H_{\ell m} (x,\nu,\delta,\boldsymbol\chi_A)\, e^{-im\psi},$$

where $\dot\psi=\Omega$ at leading order and tail propagation produces higher-order logarithmic shifts between $\psi$ and a coordinate orbital phase. The notation $\widehat H_{\ell m}$ collects the relative amplitude series; it is not assumed to be the same for every $(\ell,m)$.

In the stated harmonic convention, the dominant Newtonian mode is

$$H_{22}^{\mathrm N} =-8\sqrt{\frac{\pi}{5}}\, GM\nu x\,e^{-2i\psi}.$$

Here the mode formula uses the source-frame mass $M=M_{\rm src}$. For a cosmological source, propagation converts it to the detector-facing scale

$$\frac{(1+z)H_{22}^{\rm src}}{D_L} \sim \frac{GM_z\nu x}{D_L}, \qquad M_z=(1+z)M_{\rm src},$$

where $x=(GM_{\rm src}\Omega_{\rm src})^{2/3}=(GM_z\Omega_{\rm det})^{2/3}$ is redshift invariant. Thus $H$ has dimensions of length and $h$ is dimensionless; nearby sources recover the scale $GM\nu x/D$. The $(2,2)$ wave angular frequency is approximately $2\Omega$, so its cyclic frequency is $f_{22}=\Omega/\pi$ in the same frame as $\Omega$.

Blanchet 2024, §3.4.5, pp. 212–219 gives the convention-explicit mode construction and higher-order amplitude organization.

Leading chirp from balance

The Newtonian circular binding energy and quadrupole luminosity are accepted dynamical and radiative inputs:

$$E_{\mathrm N}(x)=-\frac12M\nu x, \qquad \mathcal F_{\mathrm N}(x) =\frac{32}{5G}\nu^2x^5.$$

Energy balance, $dE/dt=-\mathcal F$, gives

$$\frac{dx}{dt} =\frac{64}{5}\frac{\nu}{GM}x^5.$$

Since $\Omega=x^{3/2}/(GM)$,

$$\dot\Omega =\frac{96}{5} (G\mathcal M)^{5/3}\Omega^{11/3}, \qquad \mathcal M=M\nu^{3/5}.$$

$\mathcal M$ is the chirp mass. In terms of the dominant cyclic frequency,

$$\dot f_{22} =\frac{96}{5}\pi^{8/3} (G\mathcal M)^{5/3}f_{22}^{11/3}.$$

The sign is positive: a bound circular inspiral loses energy, $x$ grows, and the carrier chirps upward. The balance construction and high-order separation between energy, flux, phase, and modes are developed in Blanchet 2024, §§3.4.3–3.4.5, pp. 206–219.

Why does energy loss make the wave faster? The binding energy is negative: losing energy makes the binary more tightly bound, decreases its separation, and increases its orbital speed. At fixed distance the leading strain amplitude grows as $x\propto f_{22}^{2/3}$, while the time between successive wave cycles shrinks. These two trends produce the characteristic chirp.

Integrating the leading frequency equation from $f$ to a higher frequency $f_2$ gives the elapsed source-frame time

$$\Delta t =\frac{5}{256\pi^{8/3}(G\mathcal M)^{5/3}} \left(f^{-8/3}-f_2^{-8/3}\right).$$

The formal limit $f_2\to\infty$ defines the Newtonian integration constant $t_c-t$. It is a useful early-inspiral timescale, not a claim that this approximation predicts the merger. Equivalently, $x\propto(t_c-t)^{-1/4}$ and $f\propto(t_c-t)^{-3/8}$ while the approximation remains controlled. For observed frequencies, replace $\mathcal M$ by the redshifted chirp mass and use detector time; the source/detector dictionary derives that replacement.

What controls the hierarchy

The figure shows the two parts of the chirp together: shrinking cycle spacing and a growing amplitude envelope. It integrates the displayed Newtonian equations for $\nu=1/4$, starting at $x_0=0.06$ with $u=0$ and $\psi=0$, and stopping at $x=0.08$. The upper curve is the normalized real mode

$$q(u)\equiv \frac{-\operatorname{Re}H_{22}}{8\sqrt{\pi/5}\,GM\nu x_0} =\frac{x(u)}{x_0}\cos[2\psi(u)],$$

while the lower curve is the dimensionless angular wave frequency $GM\omega_{22}=2x^{3/2}$. Read both against the same source-frame time axis; neither curve continues into merger.

Quantitative leading-order Newtonian chirp for an equal-mass, nonspinning circular binary, with $\nu=1/4$ and $0.06\le x\le0.08$. Dashed lines mark the upper panel’s envelopes $\pm x/x_0$; the lower panel gives $GM\omega_{22}$ and the shared time coordinate is $u/(1000GM)$. The interval ends at $u/(GM)\simeq4120.815$. These are solutions of the leading inspiral equations, not numerical-relativity data or a merger prediction; the diagram supplies no estimate of neglected PN corrections.

Different invariant combinations enter different parts of the waveform:

• $GM$ sets the overall time scale and, together with distance, the amplitude scale.
• $\mathcal M$ controls the leading secular phase acceleration.
• $\nu$ controls finite-mass-ratio corrections and the overall leading amplitude.
• $\delta$ changes sign under exchange $1\leftrightarrow2$ and therefore multiplies exchange-antisymmetric contributions, including leading odd-$m$ structure in the nonspinning circular limit.
• spin projections enter phase and amplitude corrections; in-plane spins additionally drive precessional rotations.
• $x$ orders the slow-motion series. Amplitude and phase series must state their truncation separately.

Phase is especially demanding because a small fractional error in $\dot\Omega$ accumulates over many cycles. A “restricted” waveform may retain a high-order phase while keeping only the leading amplitude; a full amplitude model retains the harmonics and corrections required at the declared order. These are different approximations, even when their carrier phase is identical.

The hierarchy requires both $x\ll1$ and slow radiation reaction; neither condition supplies a universal numerical cutoff. A truncated PN waveform may lose the accuracy required for a particular application well before $x$ reaches unity. Compare successive orders, independent approximants, and numerical or perturbative waveforms in their overlap. Merger structure requires controlled strong-field input rather than extrapolating the displayed power series without an error estimate.

Common pitfalls

Calling $\psi$ the detector phase without a frame declaration. $\psi$ is a source waveform phase. Detector arrival time, sky position, polarization basis, and cosmological redshift are separate mappings.

Using total mass where chirp mass belongs. $M$ sets the time unit, but the leading frequency derivative depends on $M^{5/3}\nu=\mathcal M^{5/3}$.

Assuming every harmonic has frequency $2\Omega$. A quasi-circular $(\ell,m)$ mode has leading phase $m\psi$ and angular frequency approximately $m\Omega$. Eccentricity and precession add further frequencies.

Exercise

Starting only from $E_{\mathrm N}(x)$, $\mathcal F_{\mathrm N}(x)$, and $x=(GM\Omega)^{2/3}$, derive $\dot\Omega$ and show that its mass dependence can be written entirely through $\mathcal M$.

Solution

$dE_{\mathrm N}/dx=-M\nu/2$. Balance therefore gives $-(M\nu/2)\dot x=-(32/5G)\nu^2x^5$, or $\dot x=64\nu x^5/(5GM)$. Differentiating $\Omega=x^{3/2}/(GM)$ yields $\dot\Omega=(3/2)x^{1/2}\dot x/(GM)=96\nu x^{11/2}/[5(GM)^2]$. Replacing $x$ by $(GM\Omega)^{2/3}$ gives $\dot\Omega=(96/5)(GM)^{5/3}\nu\Omega^{11/3}$. Because $\mathcal M^{5/3}=M^{5/3}\nu$, this is the stated chirp-mass formula. Every factor is positive, so the frequency increases monotonically within the adiabatic regime.

References

• Blanchet, Luc. “Post-Newtonian Theory for Gravitational Waves.” Living Reviews in Relativity 27 (2024) 4. DOI. Open PDF.