Good things come to those who wait: Watching donor stars evolve towards a mass-transfer instability

Abstract: Unstable mass transfer in binary systems can lead to transients such as luminous red novae (LRNe). Observations of such transients are valuable for understanding and testing models of mass transfer. For donor stars in the Hertzsprung gap, there can be a long phase of mass-transfer evolution before instability sets in. Only few case studies of such delayed dynamical instability (DDI) mass transfer exist. None consider the full pre-instability evolution and the effects thereof on the observable properties of a binary. We systematically analyse detailed models of stable and unstable mass transfer for Hertzsprung-gap donors. We focus on identifying observable evolutionary features characteristic of ultimately unstable mass transfer and not found in stable mass-transfer binaries. Our binary evolution models, calculated with the MESA code, cover initial donor masses between $2.5 M_{\odot}$ and $10 M_{\odot}$ and initial accretor-to-donor mass ratios between $0.1$ and $1$. We find that the pre-instability evolution is qualitatively the same for all DDI donor stars, consisting of a long slow dimming phase followed by a shorter phase of rapid brightening. The latter phase is powered by recombination of hydrogen and accompanied by a strong increase in effective temperature, unique to unstable mass-transfer binaries. We estimate that a significant fraction of the rapid brighteners should be detectable by Gaia throughout the Galaxy. We model the progenitors of LRNe M31-2015 and V838 Mon, find a higher initial donor mass for M31 2015 than past estimates, and propose a new scenario for V838 Mon in which the known tertiary star dominates the pre-outburst photometry and the outburst results from the DDI of a more massive primary star. This work provides a comprehensive framework linking theory to observations of transients and enables improved classification and prediction of mass-transfer events.