Gravitational Memory Effects Overview

• Gravitational memory effects are permanent spacetime deformations caused by gravitational-wave bursts, manifesting as displacement, velocity, and spin changes.
• They are described using asymptotic Bondi–Sachs frameworks that capture nonlinear radiative dynamics and multipole moment variations, which are observable via detector 'kicks' and displacements.
• These effects provide critical insights into modified gravity theories and BMS symmetries, offering new avenues for understanding gravitational-wave physics and testing general relativity.

Gravitational memory effects are persistent, non-oscillatory changes in the spacetime configuration of test particles, fields, or geometric structures induced by the passage of gravitational radiation. In essence, a gravitational-wave memory effect is any permanent change remaining after a gravitational wave burst has traversed a region—even after local spacetime returns to Minkowskian form. These effects encode fundamental information about the nonlinear and infrared structure of gravity, asymptotic symmetries, and radiation content, and are directly tied to the observable consequences of gravitational-wave bursts or impulsive events on detectors, ranging from static displacements to velocity changes and gyroscopic precessions.

1. Classification and Mathematical Framework

Gravitational-memory phenomena are naturally classified by their physical manifestation and spacetime origin. The dominant categories are:

• Displacement Memory: A permanent change in the relative displacement of initially comoving test masses (also called “DC offset” in the strain). The archetype is the Christodoulou (nonlinear) memory, reflecting the net energy flux of gravitational waves at null infinity (Favata, 2010, Grant et al., 2022).
• Velocity (Kick) Memory: A permanent change in the relative velocity between test masses, e.g., the “velocity memory effect” observed for impulsive or sandwich plane waves (Zhang et al., 2017, Zhang et al., 2017, Divakarla et al., 2021).
• Spin (Gyroscopic) Memory: The net time delay (or precession) experienced by counter-orbiting null or nearly null geodesics; physically encodes fluxes of angular momentum (Mao et al., 2018, Seraj et al., 2022, Grant et al., 2022).
• Center-of-Mass (CM) Memory: A higher-order effect associated with boosts and translations in the infinite-dimensional BMS symmetry group, appearing as time-weighted integrals of the Bondi shear (Hou et al., 2021).

The detailed metric formalism is based on asymptotic expansions in Bondi–Sachs coordinates $(u,r,x^A)$, with the shear $C_{AB}(u,x^C)$ and its time derivative (“Bondi news” $N_{AB}$) encoding radiative data. The key observable,

$$\Delta h_{AB} \equiv C_{AB}(+\infty) - C_{AB}(-\infty),$$

is the net strain change, whose $E$-mode (gradient) and $B$-mode (curl) decompositions determine the polarization characteristics and physical sky pattern (Mädler et al., 2016).

For impulsive or plane wave cases in Brinkmann coordinates $(X^i,U,V)$, the memory manifests as nontrivial asymptotic behavior in transverse geodesics: $\Delta v^i = -\int C^i_j(U')\,dU' \, X^j_0$ and, for impulsive profiles, as instantaneous velocity kicks (Zhang et al., 2017).

2. Linear and Nonlinear Memory: Physical Origins

Two principal physical mechanisms underlie gravitational-memory phenomena:

• Ordinary (Linear) Memory: Arises from a net change in the radiative multipole moments due to mass ejection or radiation escaping to infinity. This effect is captured by solutions to the linearized Einstein equations with unbound sources,

$$\Delta h_{ij}^{TT} = \Delta \sum_A \frac{4 M_A}{R \sqrt{1 - v_A^2}}\left[\frac{v_A^i v_A^j}{1 - \mathbf{v}_A \cdot \mathbf{N}}\right]^{TT}$$

(Favata, 2010, Mädler et al., 2016).

• Nonlinear (Christodoulou) Memory: Generated by the gravitational-wave stress–energy itself, producing an effect built from the time integral of the radiative energy flux at null infinity,

$$\delta h_{jk}^{TT}(T_R) = \frac{4}{R} \int_{-\infty}^{T_R} dt'\left[ \int d\Omega' \frac{dE^{gw}}{dt' d\Omega'} \frac{n'_j n'_k}{1 - \mathbf{n}' \cdot \mathbf{N}} \right]^{TT}$$

Appearing formally at high PN order in the waveform, this contribution grows secularly through the inspiral/merger of compact binaries, entering the Newtonian strain amplitude through a hereditary time integral (Favata, 2010, Divakarla et al., 2021).

In the Bondi–Sachs framework, these are reflected as flux-balance laws for the mass- and angular–momentum aspects (Grant et al., 2022), with the non-linear memory arising from the square of the news tensor $C_{AB}(u,x^C)$0 integrated over retarded time.

3. Memory in Modified Theories of Gravity

Gravitational memory effects are sensitive to the field content and symmetries of the underlying gravitational theory:

• Scalar–Tensor Theories (Brans–Dicke, DEF): In Brans–Dicke (BD) theory and its Damour–Esposito–Farèse (DEF) extension, an additional scalar news field $C_{AB}(u,x^C)$1 contributes to both the Bondi mass-loss formula and the observable memory. The displacement memory in BD/DEF is

$C_{AB}(u,x^C)$2

with the scalar “breathing mode” leading to isotropic offsets (Hou, 2020, Tahura et al., 2021, Tahura et al., 13 Jan 2025).

• Chern–Simons Gravity: Asymptotic analyses show no new leading-order memory effects arise from the pseudoscalar $C_{AB}(u,x^C)$3 coupling; the memory structure (displacement, spin, CM) is identical to GR at $C_{AB}(u,x^C)$4 (Hou et al., 2021).
• Generalized Proca and Massive Gravity: New propagating spin–1 and spin–0 degrees of freedom (either due to vector fields or a graviton mass) yield additional channels for memory. The effective stress–energy tensors and their group velocities modify both the amplitude and the angular profile of the displacement memory, introducing Yukawa suppression and discrete trace modifications in the massive limit (Kilicarslan et al., 2018, Heisenberg et al., 28 Aug 2025). For group velocities differing from light speed, memory observables depend explicitly on the aberration factor and can exhibit enhanced or reduced signal in specific directions.
• Tachyon Gravity: Non-minimal scalar couplings generate breathing-mode memory, with isotropic expansion/contraction distinct from the quadrupolar GR pattern, and an associated scalar “soft theorem” (Ilkhchi et al., 2 Jul 2025).

4. Hierarchy, Subleading Effects, and Asymptotic Symmetries

Recent work has revealed an infinite hierarchy or “tower” of gravitational-wave memory effects, each tied to subleading orders in the $C_{AB}(u,x^C)$5 expansion and to conserved charges in the extended BMS symmetry group:

• Subleading Memory and “Bulk” Observables: Working in Newman–Unti gauge or carrying subleading $C_{AB}(u,x^C)$6 terms to higher order, one finds memory effects that probe the deeper Coulombic and multipolar content of the gravitational field, such as the relative radial displacement at order $C_{AB}(u,x^C)$7 corresponding to changes in the subleading Bondi charges (Mirzaiyan, 2020).
• Spin and Center-of-Mass (CM) Memory: These effects are connected to the magnetic-parity (spin) and electric-parity (CM) pieces of the Bondi shear, corresponding to fluxes of angular momentum and supertranslation/boost charges (Mao et al., 2018, Grant et al., 2022).
• BMS Symmetries and Vacuum Degeneracy: At null infinity, memory effects are interpreted as vacuum transitions in the infinite-dimensional space of metric and scalar configurations, parameterized by supertranslations and, in scalar–tensor theories, by independent scalar vacuum moduli (Hou, 2020, Tahura et al., 13 Jan 2025). Memory effects thus realize large diffeomorphisms (soft graviton theorems) in the classical regime (Mädler et al., 2016).

5. Special Cases: Impulsive, Plane Waves, and Non-Radiative Memory

In physically idealized or analytic backgrounds, memory manifests with technical clarity:

• Plane and Impulsive Waves: Exact plane wave spacetimes in Brinkmann or Baldwin–Jeffery–Rosen coordinates exhibit closed-form solutions for geodesic deviation. For generic sandwich/impulsive profiles $C_{AB}(u,x^C)$8, particles experience instantaneous velocity “kicks” and discontinuities in affine coordinates, constituting sharp memory effects unattainable in smooth profiles (Zhang et al., 2017, Zhang et al., 2017, Datta et al., 2022).
• Holonomy Formulation: Memory observables (displacement, kick, spin/gyroscopic) can be encoded in the gauge holonomy (Wilson loop for EM, Poincaré holonomy for gravity) around an appropriate loop in phase space, splitting naturally into translational (displacement/kick) and Lorentz (gyroscopic) components (Seraj et al., 2022).
• Non-Radiative (Scalar/Coulombic) Memory: Even in the absence of radiative quadrupolar GWs, secular changes in charge-like multipoles (e.g., monopole mass emission) induce scalar velocity memory, scaling as $C_{AB}(u,x^C)$9 and providing a “kick” rather than a displacement, though unobservably small for astrophysical sources (Leandro et al., 2020).

6. Memory in Observational Context and Future Prospects

Memory effects are directly, but subtly, observable via their non-oscillatory DC signatures in gravitational-wave detectors and astrophysical timing arrays:

• Signal Characteristics and Detection: Displacement memory yields a step in arm length, typically $N_{AB}$0 for stellar-mass binaries at hundreds of Mpc, with spin memory $N_{AB}$1 and scalar/Proca deviations far smaller under current constraints (Grant et al., 2022, Favata, 2010, Heisenberg et al., 28 Aug 2025). For LISA-band supermassive mergers, memory could yield observable SNRs (Favata, 2010).
• Population Stacking and Event Rates: Second– and third–generation interferometers can reach detection thresholds for the displacement memory by stacking populations or observing loud events. The spin memory is predicted to require event stacking or next-generation sensitivity (Grant et al., 2022).
• Modeling and Waveform Systematics: Accurate modeling of the memory requires incorporating all waveform modes, PN corrections, merger-ringdown matching, and, in modified theories, scalar/vector channels and group-velocity effects (Favata, 2010, Tahura et al., 2021, Heisenberg et al., 28 Aug 2025).
• Novel Regimes and Growing Memory: Slow decay of metric/extrinsic-curvature initial data, or non-compact massless radiation (e.g., neutrino flows), can generate diverging electric and magnetic (B-mode) memory (Bieri, 2020). B-mode memory is otherwise only possible from primordial or homogeneous wave backgrounds.

7. Summary Table: Memory Effect Types and Physical Manifestations

Type	Physical Observable	Leading Source
Displacement (E-mode)	Permanent separation	GW energy flux, unbound mass ejection
Spin (B-mode)	Time delay/precession/CM shift	GW angular momentum flux (“soft subleading”)
Velocity (kick)	Permanent velocity difference	Plane or impulsive GW, initial velocity
Breathing (scalar)	Isotropic ring expansion/contraction	Scalar–tensor, tachyon, Proca theories
Longitudinal (null)	Shift in radial/affine parameter	Subleading BMS charge, higher multipoles
Magnetic (divergent)	Diverging B-mode memory	Slow decay, neutrino/dark matter flow

These effects collectively probe the nonlinear and symmetry content of the gravitational sector, testing predictions of general relativity, sensitivity to compact-object microphysics, and the viability of alternative metric theories. Only the dominant displacement (Christodoulou) memory is expected to become a robust experimental probe in the near-term, but the full hierarchy carries profound implications for gravitational-wave physics, infrared structure, and the cosmic history of the Universe.