Polarization from Rapidly Rotating Massive Stars

TL;DR

The paper investigates how rapid rotation in massive stars shapes their surface geometry and temperature distribution, and how these effects imprint a continuum polarization signal, especially in the far ultraviolet. It models the rotating star using a Roche geometry and compares two gravity-darkening prescriptions, von Zeipel and ELR, to predict polarization across spectral types. The study finds that near-critical rotation can produce detectable far-UV polarization, with ELR generally yielding lower amplitudes than von Zeipel, yet still observable with future UV instruments. These results suggest UV spectropolarimetry as a powerful tool to constrain angular momentum transport and mass-transfer limits in Be/Bn star systems and related evolutionary pathways.

Abstract

Stellar rotation has long been recognized as important to the evolution of stars, by virtue of the chemical mixing it can induce and how it interacts with binary mass transfer. Binary interaction and rapid rotation are both common in massive stars and involve processes of angular momentum distribution and transport. An important question is how this angular momentum transport leads to the creation of two important classes of rapidly rotating massive stars, Be stars defined by disklike emission lines, and Bn stars defined by rotationally broadened absorption lines. A related question is what limits this rotation places on how conservative the mass transfer can be. Central to addressing these issues is knowledge of how close to rotational break-up stars can get before they produce a disk. Here we calculate diagnostics of this rotational criticality using the continuum polarization arising from a combination of rotational stellar distortion (i.e., oblateness) and redistribution of stellar flux (i.e., gravity darkening), and compare polarizations produced in the von Zeipel approximation with the approach of Espinosa Lara & Rieutord (ELR). Both produce similar photospheric polarizations that rise significantly in the far ultraviolet (FUV) for B stars, with a stronger signal in the von Zeipel case. For early main-sequence and subgiant stars, it reaches a maximum of ~1% at 140 nm for stars rotating at 98% of critical, when seen edge-on. Rotational rates above 80% critical result in polarizations of several tenths of a percent, at high inclination. These predicted stable signal strengths indicate that determinations of near-critical rotations in B stars could be achieved with future spectropolarimetric instrumentation that can reach deep into the FUV, such as CASSTOR, the Polstar mission concept, or the POLLUX detector design.

Figures (8)

• Figure 1: The shapes of rapidly rotating stars with $W$ ranging from 0.0 (non-rotating and circular) to 1.0 (largest equatorial extent) in steps of 0.1. The coordinates are normalized to the polar radius which is assumed not to change with rotation.
• Figure 2: Isolating the effect of rotational distortion for stellar polarization. These models assume the temperature and gravity are the same as the non-rotating B1IV star model everywhere but now for non-spherical shape from rotational distortion. The polarization contributions are actually negative at short wavelengths with $q<0$ but plotted here as positive definite polarization $p_\lambda = |q|$. Each panel is for the rotation rate $W$ as labeled. Colors are for inclinations of $20^\circ-90^\circ$ in $10^\circ$ increments, with edge-on giving the highest polarization. As will be seen, for this spectral class, the geometrical distortion accounts for roughly half of the polarization signal.
• Figure 3: Compares polarization using von Zeipel against ELR. The diagonal dashed line is where the two approaches produce the same polarization. For a nonrotating star of type B1IV, the different colors correspond to rotations rates of $\Omega=0.60$, 0.80, 0.90, 0.95, 0.98, and 0.99 in order of increasing polarization. The circles are for different wavelengths, starting from 120 nm for the greatest negative polarizations, and declining toward optical wavelengths at the upper right.
• Figure 4: Luminosity distributions with $L_\lambda = L_I$ for rotation $W$ and viewing inclination $i$ for a B1IV star. Each panel is labeled for $W$. Colors are for inclinations as in Fig. \ref{['fig:oblate']}. The spectra have been scaled by the brightest value of all models, and each panel has the same vertical axis scale.
• Figure 5: Polarization for the B1IV star model, highlighting shorter wavelengths below 300 nm, with $\Omega=0.80$, 0.90, and 0.95 as labeled. Colors are for inclinations as in Figs. \ref{['fig:oblate']} and \ref{['fig:B1IVflx']}.