If Dark Stars made the first black hole seeds, pulsar timing arrays are already counting them

Sohan Ghodla and Cosmin Ilie of Colgate University have published a letter in Physical Review D arguing that the low-frequency gravitational-wave background now detected by pulsar timing arrays could be dominated by the descendants of supermassive Dark Stars, a proposed class of primordial star powered by dark matter heating rather than nuclear fusion ^[1][2][6]. The consequence is not the exotic star; it is that a signal usually read as a census of supermassive black hole binaries in the relatively recent universe becomes an upper limit on how many massive black hole seeds the early universe was allowed to make ^[3][11].

The mechanics are unglamorous. Pulsar timing arrays use rapidly rotating neutron stars as clocks, and passing gravitational waves shift the arrival times of their radio pulses; monitoring many pulsars over years has produced evidence for a stochastic background at nanohertz frequencies ^[4]. The leading astrophysical explanation is a population of inspiraling supermassive black hole binaries, with systems above roughly a billion solar masses combined doing most of the work at those frequencies ^[5]. That pushes the question backwards: those objects had to start as something, and observations with the James Webb Space Telescope and Chandra have found massive black holes uncomfortably early in cosmic history ^[8].

Ghodla and Ilie considered two seed channels, direct-collapse black holes and collapsing supermassive Dark Stars ^[9]. In the WIMP dark matter scenario they assume, a Dark Star stays comparatively cool and extended while it keeps accreting, potentially reaching a million solar masses or more before collapse ^[6]. The authors then followed the cosmological evolution of the resulting black holes, modeled their host halos, computed merger rates, and predicted the resulting background ^[7]. Their finding: at a Dark Star remnant number density of order 10^-3 per cubic megaparsec, the descendants can make a major and potentially dominant contribution to the PTA signal ^[10]. The direct-collapse population they consider is expected to be far rarer, around 10^-6 per cubic megaparsec, and contributes much less ^[12] - a thousandfold difference in space density before any other physics is applied ^[13].

The useful part is the falsifiability, which Ilie states plainly: produce too many massive seeds and you overproduce the PTA-detected signal; produce too few and you need other sources to assemble supermassive black holes efficiently ^[11]. That is a two-sided constraint on a quantity nobody can otherwise measure. It also imposes an accounting burden on the Dark Star channel itself. A million-solar-mass remnant has to grow by at least a factor of about a thousand to reach the billion-solar-mass regime that matters at PTA frequencies ^[14], which the paper handles by modeling growth alongside host galaxies rather than by assumption ^[7].

Two caveats belong in the same paragraph as the result. The claim is conditional on a specific dark matter candidate, WIMPs, and on a seed abundance that is currently a free parameter rather than a measurement ^[6][10]. And it is a single letter from two authors ^[1][2].

What to watch: whether the amplitude and spectral shape of the nanohertz background tighten enough to distinguish a seed-dominated contribution from ordinary late-time binaries ^[4][5], and whether the early massive black holes already seen by JWST and Chandra can be counted precisely enough to test the 10^-3 per cubic megaparsec figure independently ^[8][10].