Spiral Wave-Driven Ejecta in Neutron-Star Mergers

• Spiral wave-driven ejecta are mass outflows in neutron-star mergers propelled directly by persistent non-axisymmetric spiral density waves.
• Simulations show that these ejecta include an early pulse of ~3×10⁻³ M☉ followed by a quasi-steady wind (~0.05 M☉/s) that enhances r-process nucleosynthesis through effective angular momentum transport.
• Diagnostic methods rely on identifying low-order azimuthal modes (m=1,2,3) to distinguish this hydrodynamics-driven channel from jet-driven and shock-induced ejecta in similar astrophysical settings.

Spiral wave-driven ejecta denotes outflow whose unbinding is directly tied to spiral structure, most explicitly the “spiral-wave wind” or “spiral-wave-driven outflow” identified in long-lived neutron-star merger remnants, where non-axisymmetric spiral density waves launched by the surviving remnant into the disk carry angular momentum outward and drive a secular mass loss (Radice et al., 2023). In broader astrophysical usage, closely related phenomena include spiral-organized, spiral-patterned, or spiral-shell ejecta in which spiral geometry is important but the immediate driver is instead jet pressure, reconnection, Parker-spiral current-sheet reconnection, or binary orbital motion; the literature therefore distinguishes direct spiral-wave launching from systems in which spiral structure primarily organizes or textures the ejecta (Nedora et al., 2019, Schreier et al., 16 Jan 2025, Li et al., 14 Mar 2025, Peterson et al., 2021, Kretschmer et al., 2013, Kim, 2023).

1. Conceptual scope and taxonomy

The most precise use of the term applies to neutron-star mergers that leave a long-lived massive neutron star (MNS). In that setting, the remnant remains strongly non-axisymmetric, excites spiral density waves in the surrounding disk, and the waves “carry angular momentum outwards and drive the mass outflow” (Radice et al., 2023). This is physically distinct from prompt tidal or shock dynamical ejecta, from later viscous secular outflows, and from high-latitude neutrino-driven winds.

A broader but qualified usage appears in several other subfields. In these cases, the spiral pattern is observationally or dynamically important, but the papers do not claim that a classical spiral density wave is itself the sole launching agent.

Setting	Immediate driver	Role of spiral structure
Long-lived neutron-star merger remnant	Spiral density waves from remnant torques	Direct launching of a spiral-wave wind (Radice et al., 2023)
Blue-kilonova interpretation	Spiral density waves plus shock reprocessing	Direct hydrodynamic mass-ejection channel (Nedora et al., 2019)
Milky Way $^{26}$Al ejecta	Forward superbubble blow-out	Spiral arms organize source offsets (Kretschmer et al., 2013)
Common-envelope NS jets	Jet-inflated high-pressure regions and RT instability	Orbit-imprinted spiral ejecta morphology (Schreier et al., 16 Jan 2025)
Solar spiral jet	External reconnection and twist transfer	Spiral jet with torsional Alfvén-wave-like propagation (Li et al., 14 Mar 2025)
Parker-spiral current sheet	Pressure-curvature loss of equilibrium and reconnection	Spiral magnetic topology organizes plasmoid site (Peterson et al., 2021)
Binary circumstellar shells	Orbital motion and hydrodynamic compression	Spiral-shell pattern records anisotropic ejecta kinematics (Kim, 2023)

This taxonomy suggests that “spiral wave-driven ejecta” is exact in merger-disk hydrodynamics, but only approximate or metaphorical in several other environments.

2. Spiral-wave wind in long-lived neutron-star merger remnants

In the merger context, the spiral-wave-driven outflow is a post-merger ejecta channel generated by non-axisymmetric spiral density waves launched by the surviving neutron-star remnant into the disk (Radice et al., 2023). The ejecta history shows an early pulse of unbound material of about ${\sim}3\times 10^{-3}\ M_\odot$ ejected during and shortly after merger, followed by a quasi-steady wind with $\dot{M}\simeq 0.05\ M_\odot\ {\rm s}^{-1}$, which is stated to be “dominated by the spiral-wave wind.” The disk itself forms on a timescale of about ${\sim}20\ {\rm ms}$, direct mass ejection from the central remnant terminates after roughly ${\sim}20{-}30\ {\rm ms}$, and the spiral pattern forms shortly after merger and persists for the entire duration of the ${\sim}100\ {\rm ms}$ simulations.

The remnant origin of the outflow is diagnosed through persistent low-order azimuthal structure. The mode amplitudes are measured through

$$C_m = \int_{z=0} e^{-{\rm i}\, m\, \varphi} \rho\, W\, \sqrt{\gamma}\, {\rm d}^2 x ,$$

with $m=1,2,3$ dominant, while the $m=2$ deformation of the remnant, identified as the bar mode, generates the spiral-wave in the disk (Radice et al., 2023). The associated disk-plane density contrast,

$$\frac{\rho - \langle \rho \rangle}{\langle \rho \rangle},$$

reveals a central ${\sim}3\times 10^{-3}\ M_\odot$0 deformation and connected spiral arms extending through the disk. In this picture, the long-lived remnant is not a passive gravitating center but a continuously forcing non-axisymmetric source.

A related numerical-relativity study described the same mechanism as a new hydrodynamics-driven mass-ejection channel for the blue kilonova: spiral density waves launched by the nonaxisymmetric post-merger remnant propagate into the surrounding disk, transport angular momentum and energy outward, and eject matter from the outer disk layers (Nedora et al., 2019). That work characterized the spiral-wave wind by a mass ${\sim}3\times 10^{-3}\ M_\odot$1, velocity ${\sim}3\times 10^{-3}\ M_\odot$2, and electron fraction mostly distributed above ${\sim}3\times 10^{-3}\ M_\odot$3. Its angular-momentum transport was diagnosed with

${\sim}3\times 10^{-3}\ M_\odot$4

which showed strong outward angular momentum transport associated with the spiral pattern.

3. Geometry, composition, and nucleosynthesis

The spatial structure of the merger ejecta is anisotropic. The spiral-wave-driven outflow is predominantly located close to the disk orbital plane and has a broad distribution of electron fractions, while at higher latitudes a high electron-fraction wind is driven by neutrino radiation (Radice et al., 2023). The angular distribution is summarized by the statement that “Most of the ejecta is equatorial and has a broad range of electron fractions. The polar ejecta is less massive and characterized by significantly larger electron fraction ${\sim}3\times 10^{-3}\ M_\odot$5.” In the gray M1 simulations, the ejecta have a broader distribution, with ${\sim}3\times 10^{-3}\ M_\odot$6, whereas earlier M0 simulations produced ${\sim}3\times 10^{-3}\ M_\odot$7 at all latitudes.

Nucleosynthetically, the 2023 long-lived-MNS calculation concluded that the spiral-wave wind produces a full r-process yield and that the combined ejecta are in good agreement with Solar abundance measurements, with the main discrepancy around ${\sim}3\times 10^{-3}\ M_\odot$8 attributed to the adopted nuclear-mass model rather than the ejecta model itself (Radice et al., 2023). By contrast, the 2019 blue-kilonova interpretation associated the spiral-wave wind with ${\sim}3\times 10^{-3}\ M_\odot$9, partial hydrodynamic shock reprocessing in the expanding arms, and predominantly production up to the second peak, $\dot{M}\simeq 0.05\ M_\odot\ {\rm s}^{-1}$0, while the combination of dynamical ejecta and spiral-wave wind reproduced solar-system $\dot{M}\simeq 0.05\ M_\odot\ {\rm s}^{-1}$1-process abundances (Nedora et al., 2019). This suggests sensitivity to remnant properties, microphysics, and transport treatment.

A third merger study focused on very light elements and strontium in dynamical and spiral-wave wind ejecta (Perego et al., 2020). There, spiral-wave wind ejection rates were reported as $\dot{M}\simeq 0.05\ M_\odot\ {\rm s}^{-1}$2 for BLh_equal and $\dot{M}\simeq 0.05\ M_\odot\ {\rm s}^{-1}$3 for DD2_equal, while the integrated wind strontium mass fraction was about $\dot{M}\simeq 0.05\ M_\odot\ {\rm s}^{-1}$4. Hydrogen was negligible in the spiral-wave wind, helium was of order $\dot{M}\simeq 0.05\ M_\odot\ {\rm s}^{-1}$5 by mass, and strontium was synthesized efficiently enough that a long-lived remnant could produce an amount consistent with, or even larger than, that inferred for GW170817. The paper therefore argued that the remnant in GW170817 likely did not remain in the spiral-wave-wind-producing phase for much longer than about $\dot{M}\simeq 0.05\ M_\odot\ {\rm s}^{-1}$6.

4. Diagnostics and simulation frameworks

The spiral-wave wind has been studied in fully three-dimensional relativistic merger calculations. One implementation used 3D general-relativistic neutrino-radiation hydrodynamics with the gray moment-based M1 code THC_M1, the Z4c spacetime formulation, and 7 AMR levels, with finest spacing $\dot{M}\simeq 0.05\ M_\odot\ {\rm s}^{-1}$7; magnetic fields were neglected (Radice et al., 2023). Ejecta were identified on a sphere at coordinate radius $\dot{M}\simeq 0.05\ M_\odot\ {\rm s}^{-1}$8, including only material with outgoing radial velocity and unbound according to the Bernoulli criterion

$\dot{M}\simeq 0.05\ M_\odot\ {\rm s}^{-1}$9

The remnant and disk were separated by a density threshold ${\sim}20\ {\rm ms}$0.

The earlier spiral-wave-wind kilonova calculations used full ${\sim}20\ {\rm ms}$1 numerical relativity with WhiskyTHC, microphysical equations of state, approximate neutrino transport, and GRLES turbulent viscosity (Nedora et al., 2019). Equal-mass binaries with ${\sim}20\ {\rm ms}$2 were evolved with HS(DD2) and LS220, through merger and remnant phases for at least ${\sim}20\ {\rm ms}$3, up to ${\sim}20\ {\rm ms}$4 depending on the model. Seven refinement levels were used, with finest spacings ${\sim}20\ {\rm ms}$5, ${\sim}20\ {\rm ms}$6, and ${\sim}20\ {\rm ms}$7, and ejecta were measured on coordinate spheres at ${\sim}20\ {\rm ms}$8. Dynamical ejecta were computed with the geodesic criterion, whereas wind ejecta were computed with the Bernoulli criterion.

The light-element and strontium study used the same broad merger framework—WhiskyTHC, 3+1 Z4c evolution, finite-temperature composition-dependent EOSs, approximate neutrino transport with absorption in optically thin conditions, and GRLES viscosity—and extended the model set to BLh_equal, BLh_unequal, and DD2_equal binaries targeted to GW170817 (Perego et al., 2020). For the wind component, the asymptotic speed was defined as

${\sim}20\ {\rm ms}$9

with ${\sim}20{-}30\ {\rm ms}$0 for geodetic extraction and ${\sim}20{-}30\ {\rm ms}$1 for Bernoulli extraction. The nucleosynthesis trajectories were evolved with SkyNet.

5. Related spiral-organized ejecta in other astrophysical settings

Outside merger disks, spiral structure often remains central to ejecta morphology or kinematics, but the causal chain differs. In the Milky Way ${\sim}20{-}30\ {\rm ms}$2Al study, radial velocities vary systematically with Galactic longitude, reaching approximately ${\sim}20{-}30\ {\rm ms}$3 in the first quadrant and ${\sim}20{-}30\ {\rm ms}$4 in the fourth, with the average bulk velocity being ${\sim}20{-}30\ {\rm ms}$5 larger than the Galactic-rotation prediction; the preferred explanation is spiral-arm source geometry plus a forward superbubble blow-out with ${\sim}20{-}30\ {\rm ms}$6, not direct acceleration by a spiral density wave (Kretschmer et al., 2013). In that interpretation, spiral structure organizes where massive stars form relative to dense gas, and the offset toward the arm leading edge shapes the direction of ejecta blow-out.

In jetting common-envelope evolution, three-dimensional hydrodynamical simulations of a neutron star spiraling inside a rotating red-supergiant envelope found that envelope rotation leads to more prominent spiral structures of the ejecta than in the non-rotating case, while the ejecta themselves are accelerated by jet-inflated high-pressure volumes and fragmented by Rayleigh–Taylor instability (Schreier et al., 16 Jan 2025). The most obvious difference among models with ${\sim}20{-}30\ {\rm ms}$7, ${\sim}20{-}30\ {\rm ms}$8, and ${\sim}20{-}30\ {\rm ms}$9 is that the faster the rotation is, the more pronounced the spiral structure in planes parallel to the orbital plane is. The paper explicitly states that the jet-inflated high-pressure volumes around the neutron star accelerate the envelope, a process prone to Rayleigh–Taylor instability.

In solar-coronal spiral jets, a data-constrained 3D MHD model found that a tiny spiral jet formed when an unstable mini-flux rope erupted in a fan–spine magnetic topology, the pre-existing magnetic null collapsed into a curved 3D current sheet, and external reconnection transferred both twist and cool material onto the outer spine (Li et al., 14 Mar 2025). The twist then propagated upward at about ${\sim}100\ {\rm ms}$0, close to a local estimated Alfvén speed of about ${\sim}100\ {\rm ms}$1, which the authors interpreted as consistent with a nonlinear torsional Alfvén wave. Here again, the spiral character is real, but the eruption is primarily reconnection-driven.

A laboratory Parker-spiral current-sheet analogue provided a different near-analogue: quasi-periodic plasmoids form at the tip of a streamer structure when a high-${\sim}100\ {\rm ms}$2 region undergoes pressure-curvature-driven dynamic loss of equilibrium, and a simple heuristic model extrapolated to solar streamers yielded plasmoid periods of about ${\sim}100\ {\rm ms}$3 minutes at ${\sim}100\ {\rm ms}$4 (Peterson et al., 2021). The “spiral” in that case is the Parker-spiral background magnetic topology rather than a spiral wave that directly launches ejecta.

Binary circumstellar outflows offer a kinematic analogue. Reanalysis of 3D hydrodynamic models for spiral-shell patterns induced by the orbital motion of a mass-losing post-main-sequence star found that the variation of the transverse wind velocity is as large as half the average wind velocity over the entire three dimensional domain in the simulated models (Kim, 2023). The paper argued that the often-adopted isotropic wind assumption and the binary hypothesis as the underlying origin for circumstellar multilayered shells are mutually incompatible, so the spiral-shell pattern should be read as a direction-dependent record of ejecta kinematics.

6. Dependencies, caveats, and conceptual limits

Within neutron-star mergers, the prominence of spiral-wave-driven ejecta depends on remnant survival. The mechanism is a hallmark of long-lived MNS remnants and is strongly suppressed if the remnant collapses promptly to a black hole, because the non-axisymmetric stellar core that launches the waves disappears (Radice et al., 2023). The calculations also remain limited to the early secular phase, typically the first ${\sim}100\ {\rm ms}$5, so quoted wind rates characterize that interval rather than the full seconds-long evolution. Additional caveats include omitted magnetic fields in the THC_M1 calculation, approximate neutrino transport in both the M1 and earlier leakage+M0 schemes, model dependence on EOS, compactness, total mass, and mass ratio, and the explicit caution that the Cartesian grid may artificially enhance the ${\sim}100\ {\rm ms}$6 mode (Radice et al., 2023, Nedora et al., 2019, Perego et al., 2020).

Across the broader literature, the phrase itself requires care. The ${\sim}100\ {\rm ms}$7Al paper does not show spiral density waves directly accelerating ejecta by ${\sim}100\ {\rm ms}$8; it supports a mediated claim in which spiral/bar geometry and source offsets create preferential superbubble breakout (Kretschmer et al., 2013). The common-envelope jet simulations do not identify a coherent spiral wave mode; they instead show jet-driven, pressure-driven ejecta with a spiral morphology imposed by inspiral and enhanced by envelope rotation (Schreier et al., 16 Jan 2025). The Parker-spiral current-sheet work attributes ejecta to pressure buildup, bad magnetic curvature, and reconnection/plasmoid formation rather than to a propagating spiral wave (Peterson et al., 2021). The circumstellar-shell analysis shows that even when a spiral pattern is the dominant observable, pattern propagation in the sky plane should not automatically be interpreted as a unique bulk ejecta expansion speed (Kim, 2023). In the solar-jet case, the torsional Alfvén-wave-like disturbance is embedded within a reconnection-driven eruption rather than replacing it as the primary trigger (Li et al., 14 Mar 2025).

Taken together, these studies establish two linked meanings of the subject. In the strict sense, spiral wave-driven ejecta refers to the equatorially concentrated, broad-${\sim}100\ {\rm ms}$9, post-merger wind unbound when persistent non-axisymmetric spiral density waves torque a merger disk (Radice et al., 2023, Nedora et al., 2019). In the broader comparative sense, spiral structure can also organize where ejecta are launched, how they are collimated, and how their kinematics are recorded, even when the immediate driver is jet pressure, reconnection, or orbital-motion-induced hydrodynamic shaping.