Is GW190425 Consistent with Being a Neutron Star–Black Hole Merger?

GW190425 is the second neutron star merger event detected by the Advanced LIGO/Virgo detectors. If interpreted as a double neutron star merger, the total gravitational mass is substantially larger than that of the binary systems identified in the Galaxy. In this work we analyze the gravitational-wave data within the neutron star–black hole merger scenario. For the black hole, we yield a mass of and an aligned spin of . As for the neutron star we find a mass of and the dimensionless tidal deformability of . These parameter ranges are for 90% credibility. The inferred masses of the neutron star and the black hole are not in tension with current observations and we suggest that GW190425 is a viable candidate of a neutron star–black hole merger event. Benefitting from the continual enhancement of the sensitivities of the advanced gravitational detectors and the increase of the number of the observatories, similar events are anticipated to be much more precisely measured in the future and the presence of black holes below the so-called mass gap will be unambiguously clarified. If confirmed, the mergers of neutron stars with (quickly rotating) low-mass black holes are likely important production sites of the heaviest r-process elements.

3 . These parameter ranges are for 90% credibility. The inferred masses of the neutron star and the black hole are not in tension with current observations and we suggest that GW190425 is a viable candidate of a neutron star-black hole merger event. Benefitting from the continual enhancement of the sensitivities of the advanced gravitational detectors and the increase of the number of the observatories, similar events are anticipated to be much more precisely measured in the future and the presence of black holes below the so-called mass gap will be unambiguously clarified. If confirmed, the mergers of neutron stars with (quickly rotating) low-mass black holes are likely important production sites of the heaviest r-process elements.

Introduction
The neutron star (NS)-black hole (BH) binary systems, though not directly observed before, have been widely believed to exist in the universe (see Abbott et al. 2018, and the references). In addition to giving rise to strong gravitationalwave (GW) radiation, the NS-BH mergers can also produce electromagnetic transients such as short/long-short gamma-ray bursts (GRBs) and macronovae/kilonovae, as long as the merging neutron stars have been effectively tidally disrupted (e.g., Narayan et al. 1992;Li & Paczyński 1998;Piran 2004;Metzger 2019). In the absence of GW observations, the wellmeasured macronova/kilonova signals in the afterglow of some short/long-short GRBs in principle can shed valuable light on the merger nature (e.g., Hotokezaka et al. 2013;Kyutoku et al. 2020). Indeed, the NS-BH merger model has been adopted to well reproduce the luminous and relatively blue macronova/kilonova signal of the long-short GRB 060614 Yang et al. 2015). According to the macronova/kilonova modeling of a few events, the NS-BH merger rate was estimated to be~´-a few 100 Gpc yr 3 1 and some BHs were speculated to have low masses (Li et al. 2017). These arguments are indirect and more solid evidence for the NS-BH mergers is highly needed. Such a purpose can be achieved in the GW observations. The data of GW170817, the first neutron star merger event, strongly favor the binary neutron star (BNS) merger scenario (Abbott et al. 2017(Abbott et al. , 2019. Though the NS-BH merger possibility has also been examined, the inferred masses of the involved BH and NS are not natural (Coughlin & Dietrich 2019;Hinderer et al. 2019). Very recently, the LIGO/Virgo collaboration reported the detection of GW190425, the second neutron star merger event with a total gravitational mass of (Abbott et al. 2020). These authors concluded that such a massive binary most likely consists of a pair of NSs and has intriguing implications on the stellar evolution. The lack of the detection/identification of such massive binaries in the Galaxy, which is unlikely attributed to their quick merging after birth (Safarzadeh et al. 2020), motivates us to further examine the possible NS-BH merger origin of GW190425. Since the GW data alone are known to be unable to pin down the nature of the two compact objects (Abbott et al. 2020), our main purpose is to check whether the NS-BH merger hypothesis is in agreement with other data or not. LIGO-Hanford (H1) was offline at the time. The signal-to-noise ratio (S/N) of the Advanced Virgo (V1) is low, but it is consistent with the L1 data given the relative sensitivities of the detectors (Abbott et al. 2020). To obtain the source parameters of the GW merger event, we apply the widely used Bayesian parameter inference method. Based on the work of Abbott et al. (2020), we take the cleaned data spanning GPS time ( ) 1240215303, 1240215511 s which are open access and available from the Gravitational Wave Open Science Center 4 (Vallisneri et al. 2015). Due to the low S/N of GW190425, systematic errors caused by the choice of waveform is negligible compared to the large statistical uncertainties. Therefore, we take the spin-aligned waveform template IMRPhenomDNRT Khan et al. 2016;Dietrich et al. 2017Dietrich et al. , 2019 to analyze the data, and use the SEOBNRv4_ROM (Bohé et al. 2017) with added tidal phase correction (Dietrich et al. 2017) to check the result. We do not consider the calibration errors of the detector that will influence the sky localization but has little effect on mass measurements (Abbott et al. 2016). For the noise power spectral density (PSD), we take the files from LIGO Document Control Center. 5 Then the single-detector log-likelihood can be constructed with the GW data d( f ), one-sided PSD S n ( f ), and waveform model

The Data Analysis within the NS-BH Merger Scenario
where we take f min =19.4 Hz and f max =2048 Hz following Abbott et al. (2020). With the likelihood in hand, it is convenient to estimate the posterior probability distributions for the source model parameters using the Bayesian stochastic sampling software; we use the PYCBC Inference Thus the parameters of GW take the form , , , R.A., decl., , , , A., decl., t c , Ψ, and Λ BH (Λ NS ) are chirp mass, mass ratio, aligned spins, luminosity distance, inclination angle, right ascension, declination, geocentric GPS time of the merger, polarization of GW, and dimensionless tidal deformabilities, respectively.
For the NS-BH merger scenario, we set the prior of = q M M NS BH to a uniform distribution in the range of (0.2, 1) (we also set the prior of q −1 into a log-uniform distribution in the range of (1, 5), and find that the results just change slightly), and take Λ BH =0, while Λ NS is assumed to lie in a wide range (0, 10,000) uniformly. Meanwhile, we give a lowspin prior for the component of the spin aligned with the orbital angular momentum of NS | | 10 3 , where k 2 is the tidal Love number (Hinderer 2008;Flanagan & Hinderer 2008), c is the speed of light in vacuum, G is the gravitational constant, and R NS is the radius of NS, respectively. Figure 2(a) shows the result of the tidal deformability and the mass of NS. In comparison to GW170817, the signal of GW190425 has a considerably lower S/N (partly attributing to the non-observation of the H1 detector) and the constraint on Λ is looser (Abbott et al. 2020). Anyhow, the resulting Λ is consistent with the joint constraints set by GW170817, PSR J0030+0451, and the nuclear data . Note that for the signal with a low S/N, the inferred Λ is likely biased to a higher value (Han et al. 2020). Besides, as shown in Figure 2

Testing the Validity of the BH Hypothesis
In principle, the massive BNS merger model can be distinguished from the NS-BH merger model if the macronova/kilonova emission has been well monitored (Kyutoku et al. 2020). However, GW190425 was only poorly localized and the partial volume covered by macronova/kilonova observations just span up to about 40% (Hosseinzadeh et al. 2019, for AT 2017gfo-like macronova/kilonova), which are insufficient to pin down the merger scenario. In this subsection we concentrate on the possibility that the heavy component is a BH.
So far, thanks to the long time radio observations of the massive pulsars, the record of observed maximum mass of NS was broken over and over, e.g., from  Cromartie et al. 2020). Though the maximum mass of a nonrotating neutron star (i.e., M TOV ) still remains unknown, one can statistically estimate the probability distribution of NS's maximum mass (M max , which can be approximated by M TOV since the rotation of all these NSs are slow that do not effectively enhance the gravitational mass) using the mass measurements of dozens of NSs. Such a study has been recently carried out by Alsing et al. (2018), where the authors have found strong evidence for the presence of a maximum mass cutoff. At that moment the mass of PSR J0740 +6620 was unavailable, and their sample included a few massive neutron stars such as PSR J0348+0432, Vela X-1 ( , which is comparable to that suggested in Li et al. (2017, see Table 1 therein) but may be hard to achieve in some population synthesis calculations (e.g., Coté et al. 2017, in which a local NS-BH merger rate of Figure 1. Posterior distributions of the physical parameters, including the source frame masses of the two compact objects (M BH , M NS ), the source frame total mass M tot , the mass ratio q, the dimensionless spins (χ BH , χ NS ), the effective spin parameter χ eff , and the dimensionless tidal deformability of the neutron star Λ NS . The error bars are all for the 90% credible level.
--150 Gpc yr 3 1 has been suggested), implying that some new NS-BH binary formation channels may be present.

Discussion
In this work we have examined the possible NS-BH origin of GW190425, the second neutron star merger event detected by the Advanced LIGO/Virgo detectors. In such a specific scenario, the GW data favor a BH (NS) mass of 3 ). The inferred parameters are not in tension with current observations and we suggest that GW190425 is a viable candidate of an NS-BH merger event. This is different from the case of GW170817, for which the NS-BH modeling yields unnatural masses of the objects (Coughlin & Dietrich 2019;Hinderer et al. 2019). Therefore, GW190425 may be the first detected NS-BH merger event. The current data, however, are insufficient to disfavor the double neutron star merger origin (Abbott et al. 2020) because of the low S/N of the signal and the nondetection of the electromagnetic counterparts. GW190425 was just detected by the LIGO-Livingston Observatory and the Virgo Observatory, but not the LIGO-Hanford Observatory. Currently, the sensitivity of the Virgo Observatory is considerably lower than that of LIGO-Livingston and LIGO-Hanford Observatories, and hence cannot contribute significantly to improving the S/N of the signal (anyhow, the presence of a low-S/N signal in the Virgo Observatory provides a valuable verification). Together with the information reported in https://gracedb.ligo.org/superevents/public/O3/, the duty cycle for each detector of current aLIGO is an important issue. The situation will change substantially in the near future. The Kamioka Gravitational Wave Detector (KAGRA) will join the O3 run of the Advanced LIGO/Virgo network in 2020. The sensitivities of Virgo and KAGRA will be enhanced by a factor of a few in the upcoming O4 run. LIGO-India is anticipated to join in 2025. Therefore, for GW190425-like events taking place in O4 and later runs of LIGO/Virgo/KAGRA, the S/N would be higher by a factor of ∼2-6, benefitting from the enhancement of the sensitivity of the advanced gravitational detectors and the increase of the number of the observatories. With such a high S/N, the GW data will provide much more accurate classification of the compact objects. The joint observation of multiple detectors will improve the localization of the mergers considerably, which is very helpful to catch the macronova/kilonova radiation (and the off-axis afterglow emission), with which the nature of merger can be further revealed.
If the NS-BH merger origin of GW190425-like events has been confirmed, there are some interesting implications: (i) There exists low-mass BH below the so-called mass gap, which favors the formation of BH with a continual mass distribution rather than a gap, as suggested for instance in Ebinger et al. (2019) and Burrows et al. (2020). The continual distribution of the BH masses may lead to the misidentification of a binary BH system with light components (e.g., ∼3M e ) into NS-BH  systems, which consequently brings difficulty for constructing the BH mass function of such systems (Yang et al. 2018;Tang et al. 2020). (ii) The NS-quickly rotating low-mass BH mergers could eject massive subrelativistic neutron-rich outflow (Lattimer & Schramm 1974). In comparison to the BNS merger scenario, very heavy r-process elements are likely easier to form because for the former the huge amount of neutrino emission from the pre-collapse massive neutron stars will make the subrelativistic ejecta less neutron-rich. Therefore, together with a high rate, such mergers can be important sites of the heaviest r-process nucleosynthesis (Yang et al. 2015;Jin et al. 2016;Wehmeyer et al. 2019).
We thank the anonymous referee and Dr. G. Ashton for the helpful suggestions/comments. This work was supported in part by NSFC under grants of No. 11525313 (i.e., Funds for Distinguished Young Scholars), No. 11921003, No. 11933010, and No. 11773078 (Veitch et al. 2015) within LALSUITE (LIGO Scientific Collaboration 2018). Besides, a uniform rather than a loguniform prior for q is adopted in this work, which may also shape the resulting q distribution in view of the low S/N of the current signal).