Listening to the Standard Candles: A Survey-Scale Search for Phase-Coded Beacons in Cepheids, RR Lyrae and $\beta$ Cephei Stars

Abstract — v1

Learned et al. (2008) proposed that a civilization could use Cepheids and RR Lyrae stars, the most closely watched clocks in astronomy, as beacons, firing pulsations early with a particle beam to write a message into the timing residuals. The one search to date examined a single Kepler star. We describe Star SETI, an automated pipeline that selects Cepheids, RR Lyrae and $\beta$ Cephei stars from Gaia DR3, measures per-cycle (TESS, Kepler) or block-averaged (ASAS-SN, ZTF) observed-minus-computed timing residuals, tests them for non-white structure, decodes them as a bit stream under the cumulative (“staircase”) modulation of a one-sided trigger, and matches the words against catalogs of recognizable sequences. A resonant antineutrino beam grazing the base of the driving zone needs hundreds to tens of thousands of times less power than a central beam, and driving an RR Lyrae star takes most of its light. Injections into real light curves show that only RR Lyrae stars are audible near the amplitude Learned et al. envisaged, with half of the codes recovered at 0.2–0.4 per cent of a period per bit in TESS; the ground-based surveys hear only codes held for tens of cycles per bit and read none. We time 97,725 light curves of 77,596 stars and vet 38,230 timing anomalies; 6823 carry timing structure the vetting cannot attribute to a known cause, all but eight of them on block-averaged timing to which only part of the vetting applies. The highest-ranked are non-radial modes in first-overtone RR Lyrae stars, Blazhko modulation and mode beating in $\beta$ Cephei stars, and no decoded word survives a false-discovery correction. We find no beacon. That bounds only beacons: a perfectly compressed message is indistinguishable from noise, so a civilization communicating efficiently rather than advertising could leave residuals that pass every test we can apply, in stars we cannot single out.