Gravitational radiation needs a (change of the (mass)) quadrupole moment, and you won't get that from a singleton neutron star. You'd need to break the rough s

Gravitational radiation needs a (change of the (mass)) quadrupole moment, and you won't get that from a singleton neutron star. You'd need to break the rough spherical symmetry pretty extremely and over the course of at least a large fraction of the NS rotation to shed noteworthy gravitational waves. Bumps of realistic matter raised on NSes will tend to flatten out completely on the order of a couple light crossing times of the NS radius, which is super fast (~ 100 microseconds). Getting realistic matter that remains lumpy enough to count as a bump to the top of the crust of an NS is also really hard. More on that in three paragraphs. A binary system has a quadrupole moment (it's like two weights on either end of an imaginary bar rotating about an axis about half way along the bar; GWs are shed with the strength mainly concentrated in the plane of the rotation) but GWs from the (probably binary) system including the main sequence star discussed in the preprint at the top will be low amplitude and low frequency because the orbit is a wide and lasts a couple of years. The merger of a pair of neutron stars will have a strong quadrupole, and consequently will be strong sources of gravitational waves (and typically lots of gamma rays as a signature of various nuclear reactions like the r-process of heavy element generation). Accreting neutron stars can deflagrate with nuclear explosions. They also tend to shed lots of X-rays. Radiation pressure works against the tendency to collapse further; you need quite a dense flow of relatively cold matter onto the NS to turn it into a black hole. The extreme magnetic field of a typical (i.e., not magnetar) neutron star is likely to be relevant too. The lower mass limit on black holes was observational rather than theoretical (Ozel, 2012; Farr, 2011). However, it's no longer really observational. "Masses in the Stellar Graveyard" (Nov. 2021) <https://media.ligo.northwestern.edu/gallery/mass-plot https://media.ligo.northwestern.edu/gallery/mass-plot> (all events through the end of the third LIGO-Virgo-Kagra observing run (O3) with p(astro) > 0.5) and Ethan Siegal's write-up <https://bigthink.com/starts-with-a-bang/mass-gap-dead/ https://bigthink.com/starts-with-a-bang/mass-gap-dead/> has a reasonably accessible summary. That is, there are plenty of BHs close to the ~3 M_sun limit for NSes, although that does not guarantee that they formed by collapse of an isolated NS rather than another massive object scattering off or colliding with an NS progenitor of the BH. It's hard to contrive circumstances in which a slow sparse accretion onto an isolated NS keeps the NS high-frequency-quiet. But it could be hidden in a bunch in otherwise very UV-noisy and gas-rich environments like the central parsecs of galaxies, deep in globular clusters, etc. such that it could be missed by shallow surveys of those using e.g. the Chandra X-Ray observatory.