Spectroscopic detection of a 2.9-hour orbit in a long-period radio transient

Abstract

Long-period radio transients (LPTs) are a mysterious new class of radio transients pulsating on periods of minutes to hours. To date, nine LPTs have been discovered predominantly at low Galactic latitudes, and yet their nature remains unknown. Here I present the first phase-resolved optical spectroscopy of the 2.9 h LPT GLEAM-X J0704–37, acquired with the 10 m Keck I telescope. Radial velocity (RV) shifts of 189 ± 3 km s^{−1 of an M5-type star in a binary system are detected on a period nearly equal to the radio period. Weak Hα emission is also present, with some of it possibly originating from outside of the M dwarf. Based on the RV amplitude, and assuming a typical M dwarf mass, the companion mass must be M ≥ 0.22 M_{⊙. Calibrating the spectra with space-based Gaia photometry reveals that the system is nearly four times closer than previously reported, at d ≈ 400 pc, suggesting that more systems could be nearby and amenable to optical characterization. The optical spectrum between 3500–10 000 Åis well modeled by a binary comprised of a massive white dwarf (WD; Teff ≈ 7300 K, M ≈ 0.8 − 1.0 M⊙) and an M dwarf (Teff ≈ 3000 K, M ≈ 0.14 M⊙). Radio pulses arrive when the WD is at nearly maximum blueshift and the M dwarf at nearly maximum redshift, in contrast to what has been reported for a similar LPT, ILT J1101+5521. GLEAM-X J0704–37 is now the second LPT with an orbital period nearly equal to the radio period, establishing a class of LPTs associated with WD + M dwarf binaries; other LPTs are likely related to WD and/or neutron star spins. This work demonstrates that the precise localization of LPTs, which enables optical follow-up, will be key in uncovering the mechanism(s) that power this new class of phenomena.}}