Far-Ultraviolet Emission Line Investigation of Flares on AU Mic

TL;DR

This paper reanalyzes HST/COS FUV spectra of AU Mic flares to search for the Orrall-Zirker signature of non-thermal protons, finding no evidence of red-wing enhancement in Lyα across six high-energy flares. It simultaneously reports a pronounced blue wing enhancement in C II lines during Flare B that cannot be reproduced by standard 1D chromospheric evaporation models, even with enhanced Stark broadening, suggesting alternative mechanisms such as filament eruption. The results constrain the presence of low-energy proton beams in M-dwarf flares and highlight gaps in current flare modeling, motivating multi-dimensional radiative-hydrodynamic simulations and additional simultaneous observations to understand energy transport and its impact on exoplanetary environments. These findings inform flare physics on M dwarfs and have implications for atmospheric responses of short-period exoplanets orbiting active stars.

Abstract

The role of non-thermal proton energy transportation during solar and stellar flares is largely unknown; a better understanding of this physical process will allow us to rectify longstanding deficiencies in flare models. One way to detect the presence of non-thermal protons during flares is through the Orrall-Zirker (OZ) effect, proposed by Orrall & Zirker (1976), whereby an enhanced red wing appears in hydrogen emission lines (e.g., Lyman-$α$ at 1215.67 angstroms). We analyze archival Hubble Space Telescope/Cosmic Origins Spectrograph G130M (1060 - 1360 angstroms) observations of the young M dwarf, AU Mic to search for evidence of OZ effect during the impulsive phase of six stellar flares with $E_\textrm{flare} \approx 10^{30 - 31}$ erg. While we found non-detections of the OZ effect, we note there is a pronounced blue enhancement in several C II and C III emission lines during one of the high-energy flares. We propose that either filament eruptions or chromospheric evaporation could be the mechanism driving this observed blue enhancement. We compare the far-ultraviolet (FUV) spectra to 1D radiative-hydrodynamic stellar flare models, which are unable to reproduce the blue enhancement and broadening in these cool flare lines. By completing a line-by-line analysis of the FUV spectrum of AU Mic, we provide further constraints on the physical mechanisms producing stellar flares on M dwarfs.

Figures (6)

• Figure 1: Identified flares from AU Mic from Visit 3. We present three light curves per visit: white light (top), Si$\;$ ($\lambda_{cen}$ = 1294.55 Å; middle), and C$\;$ ($\lambda_{cen}$ = 1175.64 Å; bottom). Flares are labeled and marked by vertical dashed lines. We identified seven new flares in this visit. The new flare parameters can be found in Table \ref{['tab:Flares_vis3']}. We note that even though these observations were taken a year after Visits 1 and 2, AU Mic shows a similar flaring rate ($\sim 7$ flares visit$^{-1}$). The horizontal black line in each figure is the quiescent flux for each wavelength range.
• Figure 2: An example of a non-detection of the OZ effect in Flare B in Visit 1. Here, an uneven enhancement of the red wing versus the blue wing of the Ly$\alpha$ emission line (1215.67 Å) during the impulsive phase of the flare is not observed. The OZ effect predicts an emission enhancement 5.6 - 17.7 Å red-ward of the core of the Ly$\alpha$ emission line compared to the blue wing kerr23. The region in which an asymmetric red enhancement is expected is highlighted in gray. A similar displacement to the blue is indicated in gray, excluding the emission lines. In this flare we saw significant enhancement of the S$\;$ ($\lambda$ = 1200.97 Å) and Si$\;$ (($\lambda$ = 1206.53 Å or $\lambda$ = 1207.52 Å) emission lines labeled in this figure.
• Figure 3: Light curves of Flare B for the blue wing (blue; top left), the entire wavelength (wavelength from start of blue wing to end of red wing; black; top middle), and the red wing (red; top right) of the C$\;$ emission line ($\lambda_{\text{cen}} = 1335.095$ Å). We highlight the different responses based on the different components of the emission line. In particular, we find that the blue wing has a stronger response to the first peak in Flare B, whereas the red wing has a stronger response to the second peak. This highlights the unique emission line responses that occur during individual flare events.
• Figure 4: Same as Figure \ref{['fig:fig6']} but for the C$\;$ emission line. We see the same behavior in C$\;$ as we do for C$\;$. In particular, note the strong blue asymmetry in the spectrum, whereas the red wing predominantly responds in the second flare peak.
• Figure 5: The observed C$\;$ spectral lines during Flare B at different times throughout the flare 2015RathCarl2001Uitenbroek. Models from the RH code are shown in orange (without damping factor) and blue ($\Gamma \times 30$). The model time step 6.0 seconds appears nearly flat here. However, the spectral line shape is still present, but the flux at 6.0 seconds is only $\sim 5.3 \%$ of the flux at 4.6 seconds. The model surface flux has been scaled to the observations and convolved to the spectral resolution of the COS/HST observations. The models are scaled by optimizing the projected solid angle of AU Mic on the sky to the observations, each model is scaled by the same factor across time.