Low-frequency monitoring of flare star binary CR Draconis: long-term electron-cyclotron maser emission
Ray, T. P.; Callingham, J. R.; Shimwell, T. W.; Vedantham, H. K.; Tasse, C.; Lamy, L.; van Weeren, R. J.; Zarka, P.; Röttgering, H. J. A.; Pope, B. J. S.; Best, P. N.; Sabater, J.; Feinstein, A. D.; Veken, K.; Toet, S.
Netherlands, United States, France, South Africa, United Kingdom, Ireland
Abstract
Recently detected coherent low-frequency radio emission from M dwarf systems shares phenomenological similarities with emission produced by magnetospheric processes from the gas giant planets of our Solar System. Such beamed electron-cyclotron maser emission can be driven by a star-planet interaction or a breakdown in co-rotation between a rotating plasma disk and a stellar magnetosphere. Both models suggest that the radio emission could be periodic. Here we present the longest low-frequency interferometric monitoring campaign of an M dwarf system, composed of twenty-one ≈8 h epochs taken in two series of observing blocks separated by a year. We achieved a total on-source time of 6.5 days. We show that the M dwarf binary CR Draconis has a low-frequency 3σ detection rate of 90−8+5% when a noise floor of ≈0.1 mJy is reached, with a median flux density of 0.92 mJy, consistent circularly polarised handedness, and a median circularly polarised fraction of 66%. We resolve three bright radio bursts in dynamic spectra, revealing the brightest is elliptically polarised, confined to 4 MHz of bandwidth centred on 170 MHz, and reaches a flux density of 205 mJy. The burst structure is mottled, indicating it consists of unresolved sub-bursts. Such a structure shares a striking resemblance with the low-frequency emission from Jupiter. We suggest the near-constant detection of high brightness temperature, highly-circularly-polarised radiation that has a consistent circular polarisation handedness implies the emission is produced via the electron-cyclotron maser instability. Optical photometric data reveal the system has a rotation period of 1.984 ± 0.003 days. We observe no periodicity in the radio data, but the sampling of our radio observations produces a window function that would hide the near two-day signal.