Light-curve instabilities of Beta Lyrae observed by the BRITE satellites

TL;DR

This study analyzes the photometric instabilities of Beta Lyrae observed by BRITE satellites, revealing stochastic variability likely originating from the invisible companion, with implications for understanding accretion processes in binary systems.

Contribution

First detailed analysis of Beta Lyrae's light-curve instabilities using space-based photometry, characterizing their stochastic nature and potential origin.

Findings

Disturbances occur 3-5 times per orbit with Gaussian amplitude distribution.

Most variability is random and lasts for one or a few orbital revolutions.

De-correlation timescale of variability is approximately 0.88 days.

Abstract

Photometric instabilities of $\beta$ Lyr were observed in 2016 by two red-filter BRITE satellites over more than 10 revolutions of the binary, with $\sim$100-minute sampling. Analysis of the time series shows that flares or fading events take place typically 3 to 5 times per binary orbit. The amplitudes of the disturbances (relative to the mean light curve, in units of the maximum out-of-eclipse light-flux, f.u.) are characterized by a Gaussian distribution with $\sigma=0.0130\pm0.0004$ f.u. Most of the disturbances appear to be random, with a tendency to remain for one or a few orbital revolutions, sometimes changing from brightening to fading or the reverse. Phases just preceding the center of the deeper eclipse showed the most scatter while phases around secondary eclipse were the quietest. This implies that the invisible companion is the most likely source of the instabilities. Wavelet transform analysis showed domination of the variability scales at phase intervals $0.05-0.3$ (0.65--4 d), with the shorter (longer) scales dominating in numbers (variability power) in this range. The series can be well described as a stochastic Gaussian process with the signal at short timescales showing a slightly stronger correlation than red noise. The signal de-correlation timescale $\tau=(0.068\pm0.018)$ in phase or $(0.88\pm0.23)$~d appears to follow the same dependence on the accretor mass as that observed for AGN and QSO masses 5--9 orders of magnitude larger than the $\beta$~Lyr torus-hidden component.