Origin of the stellar Fe Kα line clarified with FUV and X-ray observations of a superflare on the RS Canum Venaticorum-type Star UX Arietis

TL;DR

This work tackles the longstanding ambiguity in the origin of the Fe Kα line in stellar flares by combining simultaneous FUV and soft X-ray observations of a UX Arietis superflare with detailed spectral modeling and 3D radiative-transfer simulations. The Fe Kα line peaks in tandem with the thermal X-ray emission and exhibits an equivalent width of $67^{+28}_{-20}$ eV at $5.3\sigma$, supporting a photoionization origin rather than collisional excitation by non-thermal electrons. Radiative-transfer simulations with SKIRT show the observed line strength is consistent with photoionization across plausible flare loop heights $l_{\mathrm{SY}} = 1.0$–$3.0\,R_{\odot}$ and varying inclinations, demonstrating that the Fe Kα line can constrain flare geometry (loop size and latitude) and potentially CME direction. This establishes Fe Kα as a quantitative diagnostic for stellar flare geometry and motivates future high-resolution spectroscopy with XRISM to refine abundance and structural parameters.

Abstract

Fluorescence line diagnostics of the Fe Kα line at $\sim 6.4$ keV observed in both solar and stellar flares can constrain the latitude and size of the flare loop, even in the absence of imaging observations. However, they are hampered by the unresolved origin of stellar Fe Kα lines: i.e., it is unclear which of the two mechanisms-photoionization by hard X-ray photons or collisional ionization by non-thermal electrons-is the dominant process. We present clear evidence for the photoionization origin based on simultaneous far ultraviolet (FUV) and soft X-ray observations of a superflare on the RS Canum Venaticorum-type Star UX Arietis with Extreme ultraviolet spetrosCope for ExosphEric Dynamic (EXCEED; 900$-$1480 Å) onboard Hisaki and Neutron Star Interior Composition Explorer (NICER; 0.2$-$12 keV). The flare started at 22:50 UT on 2018 November 15 and released $2 \times 10^{36}$ erg in the 900$-$1480 Å band and $3 \times 10^{36}$ erg in the 0.3$-$4 keV band. The FUV emission, a proxy for non-thermal activity, peaked approximately 1.4 hours before the soft X-rays. In contrast, the Fe Kα line, detected at a statistical significance of $5.3 σ$ with an equivalent width of $67^{+28}_{-20}$ eV, peaked simultaneously with the thermal X-ray maximum rather than the non-thermal FUV peak-strongly supporting the photoionization hypothesis. Radiative transfer calculations, combined with the observed Fe Kα line intensity, further support the photoionization scenario and demonstrate the potential of this line to provide the flare geometry.

Figures (7)

• Figure 1: Light curves of the superflare of UX Ari on 16 November 2018. (a) The $0.3-4$ keV NICER count rate (cps; $\mathrm{counts \: s^{-1}}$) binned at 64 sec of UX Ari and $0.3-4$ keV luminosity binned at $\sim90$ min ISS orbit. The time origin of JD 2458438 corresponds to 12:00 UT 2018 November 15. The one standard deviation statistical uncertainties are smaller than the symbol size. The black dashed and dash-dot lines indicate the 4th order Chebyshev polynomial and exponential function fitted to the impulsive and decay phases of the count rate light curve, respectively. Interval numbers in our spectral analysis are also shown. (b) Background-subtracted $900-1480$$\mathrm{\AA}$ luminosity of UX Ari measured by Hisaki. Black dash-dot-dot line shows the time derivative of the 4th order Chebyshev polynomial in panel (a). The derivative curve of the Chebyshev function is normalized to the FUV peak luminosity. (c) Time variation of the Fe line luminosity in X-rays. The blue circles, green diamonds and red squares are the luminosity of the Fe K$\alpha$, Fe XXV He$\alpha$, and Fe XXVI Ly$\alpha$ lines, respectively. The red and blue shaded area show the X-ray and FUV flare peaks, respectively. The down arrow of the Fe K$\alpha$ line means the 90% upper limit and the errorbars indicate 90% confidence level.
• Figure 2: Background-subtracted and response-uncorrected X-ray spectra at each interval shown in Figure \ref{['Figure1']}a. Panels (a-d) are spectra during the flare, whereas panel (e) is the spectrum during the quiescence. (a-d) The best-fit curves for two-temperature vapec models with the fixed quiescence component are shown by solid black lines. Green dotted, red dash-dot, blue dash-dot-dot, lines represent the quiescence, high temperature, and low temperature flare component, respectively. The right upper inset panels show the enlarged spectrum around the Fe K$\alpha$ at 6.4 keV, Fe XXV He$\alpha$ at 6.7 keV, and Fe XXVI Ly$\alpha$ at 6.9 keV. (e) Orange dash-dot-dot, green dashed, and blue dash-dot lines represent the high-, medium-, and low-temperature components, respectively.
• Figure 3: Schematic diagram of the setup of the SKIRT simulation.
• Figure 4: Comparison between the observed Fe K$\alpha$ line equivalent width and expectations as a function of $\theta_{f}$ obtained by SKIRT. Blue, orange, and green lines correspond to $l_{\mathrm{SY}} = 1.0R_{\odot}$, $2.0R_{\odot}$, and $3.0R_{\odot}$ cases, respectively. The observed equivalent width with 90% confidence level is shown as gray-shaded area. Note that $n_{\mathrm{p}}$, $\Delta r$, and $Z_{\mathrm{p}}$ are fixed to $10^{16}$ cm$^{-3}$, $0.01R_{\odot}$, and $Z_{\odot}$, respectively, in this calculation.
• Figure 5: Time variation of the luminosity of (a-g) FUV emission lines and (h) continuum in 900$-$1480 Å. Black dash-dot-dot lines show the time derivative of the 4-th order Chebyshev function fitted for the impulsive phase of the soft X-ray light curve (Figure \ref{['Figure1']}). The derivative curves of the Chebyshev function are normalized to the peak luminosity of each line.