Multi-band Aperture Polarimetry of Betelgeuse during the 2019–20 Dimming

Betelegeuse (α Orion) is a red supergiant (M2 Iab) and a semi-regular long period variable (SRc) with periods prominent around 400 and 2100 days (Montargès et al. 2016). Its most recent dimming event was first noticed by Guinan et al. (2019), who subsequently determined the minimum magnitude to be the lowest on record (Guinan et al. 2020), and whose reporting catalyzed other observations. New VLT-SPHERE images, when compared to those from a year earlier, reveal a darkened lower hemisphere (Montargès et al. 2020). Levesque & Massey (2020) analyzed their own spectrophotometry in the context of those images, finding the photosphere temperature to be nearly unchanged and attributing Betelgeuse's dimming to episodic mass-loss and a subsequent increase in large grain circumstellar dust.

Aperture polarimetry was used extensively in the study of late-type supergiant variability until the 1980s (Clarke 2010). Betelgeuse, like similar stars, can be highly polarized at blue wavelengths reducing ∝λ⁻⁴ to a minimum at red and reversing in the infrared; it undergoes slow changes in both polarization magnitude (p) and position angle (ζ), on similar timescales to, but not necessarily correlated with its photometric periods. The most popular explanations relate to absorption and scattering from either convective cells (hotspots), circumstellar dust or both. Aurière et al. (2016) describe a scenario where near-surface features induce a net polarization by asymmetrically depolarizing the continuum. The polarization behavior then depends on feature geometry and composition. Recent polarimetric imaging with SPHERE-ZIMPOL provides justification for both mechanisms (Montargès et al. 2016); such measurements provide superior detail, but the resources required restrict their frequency.

To support VAMPIRES observations, like those of Norris et al. (2012), we observed Betelgeuse during three runs spanning the recent dimming event. The observations used the HIPPI-2 aperture polarimeter (Bailey et al. 2020) mounted on the 3.9 m AAT at Siding Spring Observatory (2019 December) and the 60 cm telescope at WSU's Penrith Observatory (2019 October, 2020 February). On bright stars at WSU HIPPI-2 has a precision of about 10 parts-per-million (ppm, 10⁻⁶) depending on the passband and is substantially better at the AAT. We used the filters, detectors and standard observation and reduction techniques described in Bailey et al. (2020).

The calibrated observations of Betelgeuse are shown in Figures 1(b) and (c) in terms of q and u in Equatorial co-ordinates; different bands are represented by symbols colored according to effective wavelength, λ_eff.

Zoom In Zoom Out Reset image size

In Figure 1(f) we follow the approach of Tinbergen et al. (1981) to calculate the interstellar polarization. The interstellar polarization and intrinsic polarization position angle, ${\zeta }_{\unicode{x02605}}$, are assumed to be the same in g' and r'[B], whereby lines drawn to pass through those two points from each run⁹ should intersect at (q_i, u_i). Subtracting these from q and u gives the intrinsic components.

The interstellar subtraction process should be regarded as approximate, especially given the difference to Tinbergen et al. (1981)'s value—though it is likely dust supplied by the star to the ISM has changed the interstellar polarization. Nevertheless it appears in this case the polarigenics and photometrics are correlated, and we can make two comments: (1) Polarization at blue wavelengths decreases approaching photometric minimum. This is consistent with the idea that the polarization comes from the stellar photosphere (as argued by Aurière et al. 2016) and not from circumstellar dust scattering (which would be expected to increase the polarization). Together with the SPHERE imaging, it is also consistent with something obscuring part of the star and therefore changing the detailed asymmetry pattern. (2) If the interstellar subtraction is accurate then the polarization in the red is greater than in g' near photometric minimum (squares in Figure 1(g)). This is different to the usual p(λ) behavior seen in the earlier runs (circles and diamonds in Figure 1(g)) and bodes well for using disk resolved polarimetry to reveal character and grain size distribution to confirm Levesque & Massey (2020)'s conclusions.

We acknowledge with thanks the variable star observations from the AAVSO International Database contributed by observers worldwide and used in this research. We thank the AAT staff, the SSO Director A/Prof. Chris Lidman, and Prof. Miroslav Filipovic for providing access to Penrith Observatory. UNSW provided funding for the construction of HIPPI-2.