GW190425: High-Mass Binary Neutron Star Merger

• GW190425 is a gravitational-wave event from a compact binary merger featuring a total mass (~3.4 M⊙) that challenges conventional binary neutron star models.
• Parameter estimation revealed a chirp mass of 1.44 M⊙ and ambiguous tidal signatures, complicating its classification between BNS, NSBH, and low-mass gap systems.
• Extensive electromagnetic follow-up failed to detect a counterpart, underscoring the need for enhanced detector sensitivity and coordinated multi-messenger strategies.

GW190425 is a compact binary coalescence detected on 2019 April 25 at 08:18:05 UTC and reported as the second confident gravitational-wave signal consistent with the coalescence of a binary neutron star. Its source-frame chirp mass, $\mathcal{M}=1.44^{+0.02}_{-0.02}\,M_{\odot}$, and total mass, $3.4^{+0.3}_{-0.1}\,M_{\odot}$, are substantially larger than those of known Galactic binary neutron star systems. At the same time, no confirmed electromagnetic counterpart was identified, so the event remains central to discussions of compact-object taxonomy, prompt-collapse thresholds, kilonova detectability, and binary formation physics (Collaboration et al., 2020).

1. Detection, localization, and gravitational-wave inference

GW190425 was recorded when only the LIGO Livingston and Virgo detectors were collecting science-mode data; LIGO Hanford was offline for maintenance. The event was first identified as a single-detector candidate in Livingston by the GSTLAL matched-filter pipeline with a matched-filter signal-to-noise ratio of 12.9, while Virgo had signal-to-noise ratio $\approx 2.5$ and contributed primarily to parameter estimation and sky localization rather than detection significance. The false-alarm rate assigned by GSTLAL was 1 per 69,000 yr, and comparison with $\sim 169.5$ days of O1-O2 and 50 days of O3 background showed no noise triggers as loud (Collaboration et al., 2020).

Parameter estimation used the standard compact-binary quantities

$$\mathcal{M}=\frac{(m_1m_2)^{3/5}}{(m_1+m_2)^{1/5}}, \qquad \chi_{\rm eff}=\frac{m_1\chi_{1z}+m_2\chi_{2z}}{m_1+m_2},$$

together with tidal parameters. In the discovery analysis, the source-frame component masses were $m_1=1.60$–$1.87\,M_\odot$ and $m_2=1.46$–$1.69\,M_\odot$ under the low-spin prior, and $m_1=1.61$–$3.4^{+0.3}_{-0.1}\,M_{\odot}$0 and $3.4^{+0.3}_{-0.1}\,M_{\odot}$1–$3.4^{+0.3}_{-0.1}\,M_{\odot}$2 under the high-spin prior. The luminosity distance was $3.4^{+0.3}_{-0.1}\,M_{\odot}$3, corresponding to $3.4^{+0.3}_{-0.1}\,M_{\odot}$4, and the effective aligned spin was consistent with zero; under the high-spin prior, $3.4^{+0.3}_{-0.1}\,M_{\odot}$5, but the Bayes factor was $3.4^{+0.3}_{-0.1}\,M_{\odot}$6 when comparing spinning versus non-spinning models (Collaboration et al., 2020).

Sky localization was unusually broad because the network effectively operated with one high-S/N detector. Subsequent follow-up papers quoted $3.4^{+0.3}_{-0.1}\,M_{\odot}$7 for the initial alert, $3.4^{+0.3}_{-0.1}\,M_{\odot}$8 for the 24 hr update, and $3.4^{+0.3}_{-0.1}\,M_{\odot}$9 for the final GWTC-2 map. This large area strongly conditioned every later electromagnetic-search strategy (Keinan et al., 2024).

The event’s mass scale made it a pronounced outlier relative to Galactic double neutron stars. Among the $\approx 2.5$0 Galactic binary neutron stars with total-mass measurements, the mean total mass is $\approx 2.5$1 with $\approx 2.5$2; GW190425 lies $\approx 2.5$3 above that mean, and its chirp mass lies $\approx 2.5$4 above the known range. That discrepancy immediately motivated both alternative source interpretations and alternative evolutionary channels (Collaboration et al., 2020).

2. Source classification: BNS, NSBH, and low-mass compact-object ambiguity

A central issue is whether GW190425 should be interpreted as a binary neutron star, a neutron star-black hole system, or a merger involving compact objects in the lower mass gap. In low latency, its chirp mass $\approx 2.5$5 was noted to lie in an “ambiguous interval,” roughly $\approx 2.5$6, within which a detected binary can be either NS-NS or BH-NS depending on the equation of state, tidal sector, and mass ratio (Barbieri et al., 2020).

Astrophysically constrained reanalyses sharpened the BNS interpretation by imposing priors motivated by neutron-star masses. Under the assumption that both components were neutron stars, one study obtained $\approx 2.5$7 and $\approx 2.5$8; for a “BNS with $\approx 2.5$9 Milky Way non-recycled NS,” it found $\sim 169.5$0 and $\sim 169.5$1; and for a nearly equal-mass BNS it found $\sim 169.5$2 and $\sim 169.5$3. In that framework, all credible BNS solutions favored prompt collapse and a redder, fainter counterpart than GW170817 (Foley et al., 2020).

A distinct reanalysis treated the heavier component explicitly as a black hole and recovered a viable NSBH solution with $\sim 169.5$4, $\sim 169.5$5, $\sim 169.5$6, and $\sim 169.5$7, all at 90% credibility. The corresponding black-hole mass interval, $\sim 169.5$8–$\sim 169.5$9, falls below the traditional lower mass gap of $$\mathcal{M}=\frac{(m_1m_2)^{3/5}}{(m_1+m_2)^{1/5}}, \qquad \chi_{\rm eff}=\frac{m_1\chi_{1z}+m_2\chi_{2z}}{m_1+m_2},$$0–$$\mathcal{M}=\frac{(m_1m_2)^{3/5}}{(m_1+m_2)^{1/5}}, \qquad \chi_{\rm eff}=\frac{m_1\chi_{1z}+m_2\chi_{2z}}{m_1+m_2},$$1 and was therefore taken as evidence that a low-mass black hole could not be excluded by the gravitational-wave data alone (Han et al., 2020).

The classification problem remains statistically difficult for GW190425-like signals because tidal information is weak at network signal-to-noise ratio $$\mathcal{M}=\frac{(m_1m_2)^{3/5}}{(m_1+m_2)^{1/5}}, \qquad \chi_{\rm eff}=\frac{m_1\chi_{1z}+m_2\chi_{2z}}{m_1+m_2},$$2. A later case study emphasized that the posterior on the effective tidal deformability peaks at zero but has 90-percentile width $$\mathcal{M}=\frac{(m_1m_2)^{3/5}}{(m_1+m_2)^{1/5}}, \qquad \chi_{\rm eff}=\frac{m_1\chi_{1z}+m_2\chi_{2z}}{m_1+m_2},$$3, so present evidence is not decisive. In forecasts, a network with A+ sensitivity cannot classify a GW190425-like event with sufficient confidence; HLI# can do so only if neutron stars follow a relatively stiff equation of state; and next-generation observatories such as Einstein Telescope and Cosmic Explorer recover the tidal signature even for soft, compact stars (Khadkikar et al., 10 Jul 2025).

More speculative low-mass-compact-object interpretations have also been advanced. A primordial-black-hole scenario associated the event with a QCD-era peak in the mass function at $$\mathcal{M}=\frac{(m_1m_2)^{3/5}}{(m_1+m_2)^{1/5}}, \qquad \chi_{\rm eff}=\frac{m_1\chi_{1z}+m_2\chi_{2z}}{m_1+m_2},$$4–$$\mathcal{M}=\frac{(m_1m_2)^{3/5}}{(m_1+m_2)^{1/5}}, \qquad \chi_{\rm eff}=\frac{m_1\chi_{1z}+m_2\chi_{2z}}{m_1+m_2},$$5 and argued that the inferred rate $$\mathcal{M}=\frac{(m_1m_2)^{3/5}}{(m_1+m_2)^{1/5}}, \qquad \chi_{\rm eff}=\frac{m_1\chi_{1z}+m_2\chi_{2z}}{m_1+m_2},$$6 can be reproduced if primordial black holes are sufficiently clustered (Clesse et al., 2020). This does not displace the astrophysical NS-NS and NSBH interpretations, but it illustrates how GW190425 became a test case for compact-object demographics below $$\mathcal{M}=\frac{(m_1m_2)^{3/5}}{(m_1+m_2)^{1/5}}, \qquad \chi_{\rm eff}=\frac{m_1\chi_{1z}+m_2\chi_{2z}}{m_1+m_2},$$7.

3. Merger outcome, prompt collapse, and numerical-relativity constraints

For binary-neutron-star interpretations, the remnant fate is usually discussed in terms of the threshold for prompt collapse,

$$\mathcal{M}=\frac{(m_1m_2)^{3/5}}{(m_1+m_2)^{1/5}}, \qquad \chi_{\rm eff}=\frac{m_1\chi_{1z}+m_2\chi_{2z}}{m_1+m_2},$$8

where $$\mathcal{M}=\frac{(m_1m_2)^{3/5}}{(m_1+m_2)^{1/5}}, \qquad \chi_{\rm eff}=\frac{m_1\chi_{1z}+m_2\chi_{2z}}{m_1+m_2},$$9 is the maximum non-rotating neutron-star mass and $m_1=1.60$0 is the radius of a $m_1=1.60$1 star. With $m_1=1.60$2 and $m_1=1.60$3, one obtains $m_1=1.60$4. Since all credible BNS solutions have $m_1=1.60$5, prompt black-hole collapse was estimated to be $m_1=1.60$6 probable (Foley et al., 2020).

This expectation is consistent with a large set of numerical-relativity calculations. One study simulated 28 BNS mergers with fixed chirp mass $m_1=1.60$7 and mass ratio $m_1=1.60$8, using finite-temperature, composition-dependent equations of state and neutrino radiation. In those simulations, prompt collapse to a black hole occurred in all cases, the total energy emitted in gravitational waves was $m_1=1.60$9, the dynamical ejecta ranged from $1.87\,M_\odot$0 to $1.87\,M_\odot$1, and the disc mass ranged from $1.87\,M_\odot$2 to $1.87\,M_\odot$3. The resulting kilonovae were relatively dim compared with GW170817, with peak AB magnitudes always dimmer than $1.87\,M_\odot$4 in the $1.87\,M_\odot$5, $1.87\,M_\odot$6, and $1.87\,M_\odot$7 bands for distances compatible with GW190425 (Camilletti et al., 2022).

A second numerical-relativity study used new BNS and BHNS simulations with $1.87\,M_\odot$8 to ask whether the electromagnetic non-detection could constrain the source parameters. Under the explicit assumptions of sufficient sky coverage and face-on orientation, that analysis found GW190425 incompatible with an unequal-mass BNS merger with mass ratio $1.87\,M_\odot$9 for stiff or moderately stiff equations of state, and it also found a nonspinning APR4 BHNS comparison model with $m_2=1.46$0 disfavored because it would have produced a bright, relatively face-on kilonova. The paper also emphasized that these bounds weaken substantially if the orientation was edge-on or the true sky position was not adequately covered (Dudi et al., 2021).

An important counterpoint is that prompt collapse is not formally unavoidable in every equation of state consistent with current constraints. Using numerical relativity with the “Big Apple” equation of state, another study found that binaries with baryonic masses $m_2=1.46$1–$m_2=1.46$2 for $m_2=1.46$3 remain below the uniformly rotating support limit and can form supramassive or differentially rotating remnants stable on secular timescales, with only the $m_2=1.46$4 case undergoing prompt black-hole collapse. That same analysis, however, concluded that a stable remnant would have produced a bright kilonova in tension with ZTF upper limits at the proposed FRB localization and that the ejecta would have been optically thick to radio emission for days to months, leaving only a narrow window for a long-lived remnant if the relevant sky region escaped coverage (Radice et al., 2023).

4. Electromagnetic follow-up and kilonova detectability

The electromagnetic search for GW190425 spanned ultraviolet to infrared wavelengths and became a benchmark study in large-area multimessenger follow-up. A comprehensive reanalysis combined new data with all publicly reported imaging, covering $m_2=1.46$5 of unique area and $m_2=1.46$6 of the LIGO/Virgo-assigned localization probability. Follow-up began $m_2=1.46$7 hr post-merger and extended to $m_2=1.46$8 weeks across 14 facilities. Under the assumption that no counterpart was found in these data, the combined dataset implied a $m_2=1.46$9 chance of detecting an AT 2017gfo-like counterpart and a detection probability of only $1.69\,M_\odot$0 for a red kilonova with $1.69\,M_\odot$1 (Coulter et al., 2024).

That study introduced the galaxy-weighted localization-redistribution algorithm teglon, which uses galaxy catalogs and luminosity completeness to reweight the sky map. For GW190425, teglon reduced the 90th-percentile area from $1.69\,M_\odot$2 to $1.69\,M_\odot$3, a $1.69\,M_\odot$4 improvement, and produced an order-of-magnitude efficiency boost for instruments with fields of view $1.69\,M_\odot$5. The same analysis identified 28 candidate optical counterparts that could not be ruled out as associated with GW190425, with 4 such counterparts discovered within the localization volume and within 5 days of merger exhibiting luminosities consistent with a kilonova. The paper nevertheless judged these 28 transients almost certainly unrelated, while noting that four deserved continued scrutiny precisely because their luminosities and rise times were not immediately inconsistent with kilonova expectations (Coulter et al., 2024).

The detectability question is model dependent. In one set of astrophysically informed calculations, almost all reported observations were judged too shallow or too blue to detect the red, faint kilonova expected for GW190425, and only very deep $1.69\,M_\odot$6 or near-infrared imaging at $1.69\,M_\odot$7 in the first 2–3 days would have had a reasonable chance (Foley et al., 2020). By contrast, other models, especially those with more substantial disk-wind ejecta or certain BHNS configurations, predicted optical magnitudes of $1.69\,M_\odot$8–$1.69\,M_\odot$9 at $m_1=1.61$0 day and therefore a materially better chance of detection (Kyutoku et al., 2020).

The event also exposed structural inefficiencies in community follow-up. A later audit of reported observations counted 8,258 unique pointings by $m_1=1.61$1 facilities and found that the uncoordinated search reached $m_1=1.61$2 days for unique coverage of 50% of the final GWTC-2 probability, never reached 90% unique coverage within 5.65 days, and repeatedly imaged some northern patches more than 100 times while leaving other regions untouched. In an idealized coordinated strategy using footprint sharing and tile ranking, the same resources could have achieved $m_1=1.61$3, $m_1=1.61$4, and $m_1=1.61$5 (Keinan et al., 2024).

5. Formation channels and population origin

Because its total mass is inconsistent with the known Galactic binary-neutron-star population, GW190425 has been used to motivate several distinct evolutionary channels. The discovery paper already suggested either an unobserved fast-merging population of tight binary neutron stars or alternative formation channels, noting that isolated binary evolution with a Case-BB common-envelope phase may produce ultra-tight neutron-star binaries with orbital periods $m_1=1.61$6 hr that inspiral too quickly to be seen as radio pulsars (Collaboration et al., 2020).

One explicit realization of that idea invoked unstable “case BB” mass transfer. In this picture, the progenitor is a neutron star of mass $m_1=1.61$7 with a helium-star companion of total mass $m_1=1.61$8–$m_1=1.61$9, followed by a second common-envelope phase, an ultra-stripped supernova, and a tight post-supernova double neutron star with merger time $3.4^{+0.3}_{-0.1}\,M_{\odot}$00. The same work argued that such a short merger time helps explain why similar systems are not seen in Galactic radio surveys (Romero-Shaw et al., 2020).

A different channel, “fallback supernova assembly,” started from a massive stripped helium star that remains compact after common envelope and therefore avoids post-common-envelope mass transfer onto the first-born neutron star. Three-dimensional SPH simulations in that model gave a stripped-star mass at collapse of $3.4^{+0.3}_{-0.1}\,M_{\odot}$01, a fallback-driven mass increase of the second-born neutron star of $3.4^{+0.3}_{-0.1}\,M_{\odot}$02, and a final second neutron-star mass of $3.4^{+0.3}_{-0.1}\,M_{\odot}$03, producing a system with final total mass $3.4^{+0.3}_{-0.1}\,M_{\odot}$04. The same channel predicts negligible accretion onto the first-born neutron star, high post-supernova eccentricity $3.4^{+0.3}_{-0.1}\,M_{\odot}$05–$3.4^{+0.3}_{-0.1}\,M_{\odot}$06, and long spin periods in the $3.4^{+0.3}_{-0.1}\,M_{\odot}$07–$3.4^{+0.3}_{-0.1}\,M_{\odot}$08 regime (Vigna-Gómez et al., 2021).

Another proposed route requires super-Eddington stable Case BB mass transfer onto the first-born neutron star. Detailed binary evolution modeling found that a post-common-envelope binary containing a $3.4^{+0.3}_{-0.1}\,M_{\odot}$09 neutron star and a helium-rich star with $3.4^{+0.3}_{-0.1}\,M_{\odot}$10–$3.4^{+0.3}_{-0.1}\,M_{\odot}$11 and initial orbital period $3.4^{+0.3}_{-0.1}\,M_{\odot}$12–$3.4^{+0.3}_{-0.1}\,M_{\odot}$13 days can reproduce GW190425-like systems if the neutron star accretes at $3.4^{+0.3}_{-0.1}\,M_{\odot}$14. In that scenario, the first-born neutron star gains $3.4^{+0.3}_{-0.1}\,M_{\odot}$15–$3.4^{+0.3}_{-0.1}\,M_{\odot}$16, ends with $3.4^{+0.3}_{-0.1}\,M_{\odot}$17–$3.4^{+0.3}_{-0.1}\,M_{\odot}$18, and merges with a $3.4^{+0.3}_{-0.1}\,M_{\odot}$19–$3.4^{+0.3}_{-0.1}\,M_{\odot}$20 second-born neutron star after $3.4^{+0.3}_{-0.1}\,M_{\odot}$21–$3.4^{+0.3}_{-0.1}\,M_{\odot}$22 (Zhang et al., 2023).

A more recent extension examined stable Case BB/BC mass transfer with internal differential rotation and tidal coupling. It found that immediate post-common-envelope systems consisting of a primary $3.4^{+0.3}_{-0.1}\,M_{\odot}$23 or $3.4^{+0.3}_{-0.1}\,M_{\odot}$24 neutron star and a secondary helium-rich star with initial mass $3.4^{+0.3}_{-0.1}\,M_{\odot}$25–$3.4^{+0.3}_{-0.1}\,M_{\odot}$26 or $3.4^{+0.3}_{-0.1}\,M_{\odot}$27–$3.4^{+0.3}_{-0.1}\,M_{\odot}$28, in close binaries with initial period $3.4^{+0.3}_{-0.1}\,M_{\odot}$29–$3.4^{+0.3}_{-0.1}\,M_{\odot}$30 days or $3.4^{+0.3}_{-0.1}\,M_{\odot}$31–$3.4^{+0.3}_{-0.1}\,M_{\odot}$32 days, can reproduce GW190425-like BNS events. In those models, tidal spin-up can leave the second-born neutron star with rotational energy reaching $3.4^{+0.3}_{-0.1}\,M_{\odot}$33, enabling magnetar-driven ultra-stripped, broad-line Type Ic, or superluminous supernova outcomes before the eventual compact-binary merger (Qin et al., 2024).

6. Eccentricity, FRB association, and remaining open questions

The orbital eccentricity of GW190425 has been analyzed because it can encode formation history. A full Bayesian study with TaylorF2Ecc templates found a 90% credible upper limit of $3.4^{+0.3}_{-0.1}\,M_{\odot}$34 with a uniform prior on eccentricity, while switching to a log-uniform prior tightened the bound to $3.4^{+0.3}_{-0.1}\,M_{\odot}$35. The strong prior dependence was emphasized as evidence that the signal carries limited information about eccentricity, although the result disfavors highly eccentric dynamical-capture channels at the 90% level (Lenon et al., 2020). A separate reanalysis motivated by unstable case BB evolution reported $3.4^{+0.3}_{-0.1}\,M_{\odot}$36 with 90% confidence and likewise concluded that present detectors cannot use such small residual eccentricities to decide among realistic isolated-binary channels (Romero-Shaw et al., 2020).

A second unresolved line of inquiry concerns the suggested association with FRB 190425 or FRB 20190425A. Pan-STARRS and ATLAS began wide-field optical coverage within 1.36 hr and 0.80 hr of the merger, respectively, and later papers highlighted the candidate host galaxy UGC 10667 at $3.4^{+0.3}_{-0.1}\,M_{\odot}$37. Pan-STARRS obtained an $3.4^{+0.3}_{-0.1}\,M_{\odot}$38-band frame at $3.4^{+0.3}_{-0.1}\,M_{\odot}$39 d with limit $3.4^{+0.3}_{-0.1}\,M_{\odot}$40, while ATLAS obtained $3.4^{+0.3}_{-0.1}\,M_{\odot}$41 at $3.4^{+0.3}_{-0.1}\,M_{\odot}$42 d and $3.4^{+0.3}_{-0.1}\,M_{\odot}$43 at $3.4^{+0.3}_{-0.1}\,M_{\odot}$44 d; no optical emission was detected. Under the FRB-host hypothesis, these limits ruled out magnetar-enhanced kilonova models and disfavored, but did not disprove, the FRB-GW link (Smartt et al., 2023).

The numerical-relativity study of a possible long-lived remnant strengthened the case against association. For ejecta masses and velocities appropriate to a stable remnant, it obtained dispersion measures $3.4^{+0.3}_{-0.1}\,M_{\odot}$45 and $3.4^{+0.3}_{-0.1}\,M_{\odot}$46 at $3.4^{+0.3}_{-0.1}\,M_{\odot}$47 hr, far above the observed near-source value $3.4^{+0.3}_{-0.1}\,M_{\odot}$48, together with free-free optical depths $3.4^{+0.3}_{-0.1}\,M_{\odot}$49 for the dynamical ejecta and $3.4^{+0.3}_{-0.1}\,M_{\odot}$50 for the disk wind. That analysis concluded that radio photons could not escape until $3.4^{+0.3}_{-0.1}\,M_{\odot}$51 days-months and that FRB 20190425A and GW190425 are not associated, while still noting that incomplete electromagnetic coverage prevents a complete exclusion of a long-lived remnant (Radice et al., 2023).

Several methodological lessons recur across the literature. One is that prompt public release of source parameters, especially chirp mass and refined sky maps, is crucial for optimizing electromagnetic follow-up and for using kilonova constraints to probe the equation of state (Radice et al., 2023). Another is that future detector networks will likely shift GW190425 from an ambiguous case to a calibrator of low-mass compact-object classification, tidal physics, and possibly dark-matter-induced compact-object collapse scenarios (Khadkikar et al., 10 Jul 2025). As a result, GW190425 remains both an astrophysical event and a methodological reference point: a single merger whose unusually high mass, poor localization, and missing counterpart continue to shape inference strategies across gravitational-wave astronomy.