GW190814: Landmark Compact Binary Merger

• GW190814 is a gravitational-wave event involving a 23M☉ black hole and a 2.6M☉ compact object with the most unequal mass ratio observed to date.
• The event’s waveform reveals significant higher multipole components and stringent general relativity tests that refine mass and distance estimates.
• The merger challenges conventional compact-object classification and formation theories while serving as a high-value dark siren for cosmology.

GW190814 is a gravitational-wave transient from the coalescence of a $23.2^{+1.1}_{-1.0}\,M_\odot$ black hole with a $2.59^{+0.08}_{-0.09}\,M_\odot$ compact object, detected by LIGO and Virgo on 2019-08-14 at 21:10:39 UTC. It is distinguished by a network matched-filter signal-to-noise ratio of 25, a 90% localization area of $18.5\,{\rm deg}^2$, a luminosity distance $D_L = 241^{+41}_{-45}\,{\rm Mpc}$, and the most unequal mass ratio yet measured in a gravitational-wave compact-binary merger, $q=0.112^{+0.008}_{-0.009}$. Its secondary lies in the lower mass gap or ambiguous transition region between known neutron stars and black holes, so GW190814 immediately became central to debates on compact-object taxonomy, binary formation, and multimessenger follow-up; no electromagnetic counterpart has been confirmed to date (Collaboration et al., 2020).

1. Detection and inferred source properties

GW190814 was observed in the three-detector network of LIGO Livingston, LIGO Hanford, and Virgo, although Hanford was not in nominal observing mode at the trigger time and its data were incorporated after follow-up checks established that the relevant data segment was usable (Collaboration et al., 2020). The source-frame component masses inferred from waveform models that include spin precession and higher multipoles are

$$m_1 = 23.2^{+1.1}_{-1.0}\,M_\odot,\qquad m_2 = 2.59^{+0.08}_{-0.09}\,M_\odot,$$

with total mass

$M = 25.8^{+1.0}_{-0.9}\,M_\odot,$

and chirp mass

$\mathcal{M} = 6.09^{+0.06}_{-0.06}\,M_\odot.$

The standard mass-ratio definition,

$q=\frac{m_2}{m_1}\le 1,$

gives $q = 0.112^{+0.008}_{-0.009}$, making the system much more asymmetric than previously observed LIGO/Virgo compact-binary mergers (Collaboration et al., 2020).

The spin information is unusually restrictive for the primary and weak for the secondary. The effective inspiral spin,

$2.59^{+0.08}_{-0.09}\,M_\odot$0

is measured as $2.59^{+0.08}_{-0.09}\,M_\odot$1, while the effective precession parameter satisfies $2.59^{+0.08}_{-0.09}\,M_\odot$2 and the primary spin magnitude satisfies $2.59^{+0.08}_{-0.09}\,M_\odot$3 at 90% credibility (Collaboration et al., 2020). The folded inclination is $2.59^{+0.08}_{-0.09}\,M_\odot$4, and the inferred remnant black hole has $2.59^{+0.08}_{-0.09}\,M_\odot$5 and $2.59^{+0.08}_{-0.09}\,M_\odot$6 (Collaboration et al., 2020).

The event’s classification problem follows directly from these measurements. At $2.59^{+0.08}_{-0.09}\,M_\odot$7, the lighter component is either the heaviest neutron star or the lightest black hole ever observed in a double compact-object system (Collaboration et al., 2020).

2. Waveform structure, higher multipoles, and tests of general relativity

GW190814 is one of the clearest higher-multipole detections in the compact-binary catalog. The discovery analysis reported very strong evidence for subdominant harmonic content beyond the dominant $2.59^{+0.08}_{-0.09}\,M_\odot$8 mode, with $2.59^{+0.08}_{-0.09}\,M_\odot$9 favoring higher-multipole models over pure-quadrupole models, and $18.5\,{\rm deg}^2$0 for the $18.5\,{\rm deg}^2$1 multipole alone. The orthogonal optimal SNR of the $18.5\,{\rm deg}^2$2 mode was $18.5\,{\rm deg}^2$3 (Collaboration et al., 2020). This structure materially improved the distance–inclination and mass-ratio inference.

Tests of general relativity found no measurable deviations from the theory. Residual analyses after subtracting best-fit waveforms were statistically consistent with surrounding off-source noise, parameterized inspiral tests found no significant deviation, and the source’s higher-multipole content was itself consistent with relativistic binary dynamics (Collaboration et al., 2020). In a complementary “binary black hole spectroscopy” analysis that treated GW190814 as a BBH and compared source parameters inferred from the $18.5\,{\rm deg}^2$4 and $18.5\,{\rm deg}^2$5 harmonics, all results were consistent with GR. For GW190814 specifically, the $18.5\,{\rm deg}^2$6-mode chirp-mass deviation was constrained to $18.5\,{\rm deg}^2$7 when masses, spins, and phase were all allowed to vary, and to $18.5\,{\rm deg}^2$8 when only masses and phase were allowed to vary (Capano et al., 2020).

A recurrent misconception is that the secondary’s nature can be read directly from the waveform. The published analyses do not support that claim. Tidal effects were not measurably detected, but the event’s extreme mass ratio makes a neutron-star tidal signature intrinsically weak, so the waveform alone does not definitively identify the lighter object as either a neutron star or a black hole (Collaboration et al., 2020).

3. The secondary component and the lower-mass-gap problem

The core interpretive issue in GW190814 is whether a compact object of mass $18.5\,{\rm deg}^2$9 can be a neutron star under current constraints on dense matter. Several equation-of-state studies favored a binary-black-hole interpretation. Using a nuclear-physics–multimessenger astrophysics framework, one study found that GW190814 was a BBH with probability $D_L = 241^{+41}_{-45}\,{\rm Mpc}$0 when the GW170817-based upper bound on the maximum nonrotating neutron-star mass was imposed, and still $D_L = 241^{+41}_{-45}\,{\rm Mpc}$1 when that upper bound was relaxed (Tews et al., 2020). A separate hybrid-EoS Bayesian study found $D_L = 241^{+41}_{-45}\,{\rm Mpc}$2 for a nonrotating neutron-star interpretation, rising to about $D_L = 241^{+41}_{-45}\,{\rm Mpc}$3 if the secondary were assumed to be a uniformly rotating neutron star at 716 Hz; conditioning on a rapidly rotating neutron-star interpretation yielded an inferred spin frequency $D_L = 241^{+41}_{-45}\,{\rm Mpc}$4 Hz, which would make the secondary the fastest rotating neutron star ever observed (Biswas et al., 2020).

These inferences sharpen when additional microphysics is included. A hypernuclear compact-star analysis concluded that hypernuclear stars are inconsistent with a stellar-nature interpretation of the light companion even for maximally rotating Keplerian configurations, implying that GW190814 involved two black holes rather than a neutron star and a black hole if hyperonization occurs in dense matter (Sedrakian et al., 2020). By contrast, a phenomenological anisotropic-neutron-star model argued that the GW190814 secondary could be an anisotropic neutron star compatible with LIGO/Virgo constraints if the radius lies in the range $D_L = 241^{+41}_{-45}\,{\rm Mpc}$5 and the anisotropic core boundary density lies in the range $D_L = 241^{+41}_{-45}\,{\rm Mpc}$6 (Roupas, 2020).

The literature therefore did not converge on a single ontology for the secondary, but it did converge on the point that a neutron-star interpretation, if retained, requires restrictive conditions. In one study those conditions were a very stiff high-density EoS or very large spin (Biswas et al., 2020); in another they included $D_L = 241^{+41}_{-45}\,{\rm Mpc}$7 for the maximum sound speed in the unlikely neutron-star scenario (Tews et al., 2020). This suggests that GW190814 is less a routine NS–BH candidate than a stress test for the upper end of the compact-object mass distribution.

4. Electromagnetic follow-up and the significance of the null result

GW190814 triggered one of the most extensive electromagnetic follow-up programs of the O3 run, enabled by its unusually small localization region. The overall result was negative: no compelling optical, infrared, radio, or spectroscopically confirmed counterpart was found (Vieira et al., 2020). That negative result, however, does not by itself imply that the merger was a BBH.

Deep optical imaging with CFHT/MegaCam covered a mean total integrated probability of 67.0% of the localization region, reaching $D_L = 241^{+41}_{-45}\,{\rm Mpc}$8 at 1.7 days and $D_L = 241^{+41}_{-45}\,{\rm Mpc}$9 and $q=0.112^{+0.008}_{-0.009}$0 at 3.7 and 8.7 days, respectively. No compelling candidate transient counterparts were found. Interpreting the event as NS–BH, the non-detection implied $q=0.112^{+0.008}_{-0.009}$1 for a blue kilonova with $q=0.112^{+0.008}_{-0.009}$2 and $q=0.112^{+0.008}_{-0.009}$3 for a red kilonova with $q=0.112^{+0.008}_{-0.009}$4 (Vieira et al., 2020). DES/DECam subsequently imaged the entire 90 percent confidence level localization area at 0, 1, 2, 3, 6, and 16 nights after merger, rejected all candidates, and used simulated three-component kilonova light curves to show that if a kilonova occurred, configurations with ejected matter greater than $q=0.112^{+0.008}_{-0.009}$5, lanthanide abundance less than $q=0.112^{+0.008}_{-0.009}$6, and velocity between $q=0.112^{+0.008}_{-0.009}$7 and $q=0.112^{+0.008}_{-0.009}$8 are disfavored at the $q=0.112^{+0.008}_{-0.009}$9 level (Morgan et al., 2020).

Other follow-up campaigns reinforced the same conclusion at different wavelengths and cadences. The Gravity Collective analyzed 189 optical transients over $$m_1 = 23.2^{+1.1}_{-1.0}\,M_\odot,\qquad m_2 = 2.59^{+0.08}_{-0.09}\,M_\odot,$$0 and 94.6% of the two-dimensional localization region and concluded that none were credible counterparts; an AT 2017gfo-like transient was ruled out at only $$m_1 = 23.2^{+1.1}_{-1.0}\,M_\odot,\qquad m_2 = 2.59^{+0.08}_{-0.09}\,M_\odot,$$1 confidence, while all known types of short gamma-ray burst afterglows with viewing angles $$m_1 = 23.2^{+1.1}_{-1.0}\,M_\odot,\qquad m_2 = 2.59^{+0.08}_{-0.09}\,M_\odot,$$2 were strongly ruled out under the adopted jet assumptions (Kilpatrick et al., 2021). A late-time galaxy-targeted VLA search spanning $$m_1 = 23.2^{+1.1}_{-1.0}\,M_\odot,\qquad m_2 = 2.59^{+0.08}_{-0.09}\,M_\odot,$$3 days found no radio counterpart in 75 galaxies, corresponding to about 32% of the final localization-volume stellar luminosity, and ruled out a typical short-GRB-like Gaussian jet with $$m_1 = 23.2^{+1.1}_{-1.0}\,M_\odot,\qquad m_2 = 2.59^{+0.08}_{-0.09}\,M_\odot,$$4 in media with $$m_1 = 23.2^{+1.1}_{-1.0}\,M_\odot,\qquad m_2 = 2.59^{+0.08}_{-0.09}\,M_\odot,$$5 for a viewing angle of $$m_1 = 23.2^{+1.1}_{-1.0}\,M_\odot,\qquad m_2 = 2.59^{+0.08}_{-0.09}\,M_\odot,$$6, assuming $$m_1 = 23.2^{+1.1}_{-1.0}\,M_\odot,\qquad m_2 = 2.59^{+0.08}_{-0.09}\,M_\odot,$$7 and $$m_1 = 23.2^{+1.1}_{-1.0}\,M_\odot,\qquad m_2 = 2.59^{+0.08}_{-0.09}\,M_\odot,$$8 (Alexander et al., 2021).

The null result therefore has a precise but limited meaning. It disfavors bright on-axis afterglows and only the most luminous kilonova models. It does not exclude a weak-EM NS–BH merger, because a $$m_1 = 23.2^{+1.1}_{-1.0}\,M_\odot,\qquad m_2 = 2.59^{+0.08}_{-0.09}\,M_\odot,$$9 primary with low spin can swallow a neutron star whole with little ejecta, and it does not exclude the possibility that the lighter object was itself a low-mass black hole (Vieira et al., 2020).

5. Formation channels and competing astrophysical interpretations

The discovery analysis concluded that binaries with mass ratios similar to GW190814 are unlikely to have formed in globular clusters and that the combination of mass ratio, component masses, and inferred merger rate challenges all current models for the formation and mass distribution of compact-object binaries (Collaboration et al., 2020). Subsequent work diversified rather than eliminated the list of viable channels.

Under the assumption that GW190814 was a BBH formed through isolated evolution, one detailed reanalysis argued that it is likely to have formed through the classical common-envelope channel. In that picture, the immediate progenitor after common-envelope ejection was a $M = 25.8^{+1.0}_{-0.9}\,M_\odot,$0 BH plus a helium star with $M = 25.8^{+1.0}_{-0.9}\,M_\odot,$1 at solar metallicity or $M = 25.8^{+1.0}_{-0.9}\,M_\odot,$2 at 10% solar metallicity, with an initial orbital period around 1.0 day; the inferred low spin of the secondary was then used to argue that the progenitor metallicity should not be too low, e.g. $M = 25.8^{+1.0}_{-0.9}\,M_\odot,$3 (Lyu et al., 2023). A distinct isolated-binary interpretation instead attributed GW190814-like systems to explodability fluctuations of massive stellar cores, arguing that fallback CCSNe can place a sizeable fraction of remnants in the $M = 25.8^{+1.0}_{-0.9}\,M_\odot,$4 range and thereby convert otherwise ordinary BBH progenitors into asymmetric mergers such as GW190814; the corresponding merger-rate density was estimated to be $M = 25.8^{+1.0}_{-0.9}\,M_\odot,$5 of the total BBH merger rate, namely $M = 25.8^{+1.0}_{-0.9}\,M_\odot,$6 (Antoniadis et al., 2021).

Dynamical and hierarchical channels remain active alternatives. One large suite of $M = 25.8^{+1.0}_{-0.9}\,M_\odot,$7-body scattering calculations found that GW190814-like mergers are likely to form in young, metal-rich clusters and estimated $M = 25.8^{+1.0}_{-0.9}\,M_\odot,$8, consistent with the observationally inferred $M = 25.8^{+1.0}_{-0.9}\,M_\odot,$9, provided the remnant mass spectrum contains a small excess of objects in the $\mathcal{M} = 6.09^{+0.06}_{-0.06}\,M_\odot.$0 range (Sedda, 2021). A hierarchical-triple model proposed that GW190814 was a second-generation merger in which the $\mathcal{M} = 6.09^{+0.06}_{-0.06}\,M_\odot.$1 object was itself the remnant of a previous binary-neutron-star merger, predicting, among other signatures, a secondary spin of 0.6 to 0.7 and a narrow secondary-mass peak between $\mathcal{M} = 6.09^{+0.06}_{-0.06}\,M_\odot.$2 and $\mathcal{M} = 6.09^{+0.06}_{-0.06}\,M_\odot.$3 in future populations (Lu et al., 2020).

More speculative reanalyses have extended the range of proposals. A line-of-sight-acceleration interpretation reported $\mathcal{M} = 6.09^{+0.06}_{-0.06}\,M_\odot.$4 and $\mathcal{M} = 6.09^{+0.06}_{-0.06}\,M_\odot.$5, interpreting GW190814 as a BBH merging near a third compact object; the same study noted that this claim rests on a simplified constant-acceleration waveform model and would benefit from independent confirmation with alternative waveform families (Han et al., 2024). A collapsar-disk-fragment scenario proposed that GW190814 originated from a compact fragment in a gravitationally unstable, neutrino-cooled collapsar disk and identified the Type Ib supernova candidate SN2019npv, inside the GW190814 credible volume about 60 days before coalescence, as a possible electromagnetic precursor; the reported host association remained tentative at roughly $\mathcal{M} = 6.09^{+0.06}_{-0.06}\,M_\odot.$6 (Baibhav et al., 22 Jun 2026).

The present state of the subject is therefore plural. GW190814 is not a solved formation problem but a benchmark against which isolated-binary, dynamical, hierarchical, and more exotic channels are all being tested.

6. Cosmology and longer-term significance

Because GW190814 had the second-smallest localization volume after GW170817 and no confirmed counterpart, it became one of the best single-event dark sirens available for cosmology. A DES-based statistical standard-siren analysis using photometric redshift PDFs and the refined higher-modes skymap obtained

$\mathcal{M} = 6.09^{+0.06}_{-0.06}\,M_\odot.$7

for GW190814 alone, and

$\mathcal{M} = 6.09^{+0.06}_{-0.06}\,M_\odot.$8

from the combination of GW190814, GW170814, and GW170817, both quoted as 68% highest-density intervals for a flat prior on $\mathcal{M} = 6.09^{+0.06}_{-0.06}\,M_\odot.$9 between 20 and 140 $q=\frac{m_2}{m_1}\le 1,$0 (Palmese et al., 2020). An independent GLADE-based statistical-siren analysis found

$q=\frac{m_2}{m_1}\le 1,$1

for $q=\frac{m_2}{m_1}\le 1,$2, and

$q=\frac{m_2}{m_1}\le 1,$3

for $q=\frac{m_2}{m_1}\le 1,$4, with a combined GW170817+GW190814 constraint of

$q=\frac{m_2}{m_1}\le 1,$5

tightening to

$q=\frac{m_2}{m_1}\le 1,$6

when full posterior samples were used (Vasylyev et al., 2020).

This cosmological role is independent of the unresolved astrophysical identity of the secondary. The event’s value as a dark siren derives from its precise sky localization and informative luminosity distance, not from whether the lighter object was a neutron star or a black hole. A later collapsar-based interpretation went further and argued that if SN2019npv were the host, GW190814 would become a bright standard siren with

$q=\frac{m_2}{m_1}\le 1,$7

but that result is conditional on a precursor-host association that remains tentative (Baibhav et al., 22 Jun 2026).

GW190814 accordingly occupies a rare position in gravitational-wave astronomy. It is simultaneously a precision-measured strong-field test of GR, a compact-object classification problem centered on the lower mass gap, a stringent null result in multimessenger follow-up, a challenge case for formation theory, and a high-value dark siren for cosmology. The continuing significance of the event lies in that conjunction rather than in any single settled interpretation.