All Figures

Figure 1: Low level emission of Proxima Centauri. Top panel: X-ray light curve. Counts from all three EPIC detectors have been co-added (using counts between 0.18-4.5 keV from the MOS detectors, and 0.15-4.5 keV from the PN). Time resolution is 60 s. The horizontal bar marks the interval that was used for spectral analysis of low-level emission. The vertical lines (also in the other panels) indicate flare-like features. Solid lines refer to clearly detected features (in the X-ray or the optical band), while dotted lines mark borderline cases. The dashed curve until 0.22 d is the MOS-only light curve shown here because the PN was not in operation before 0.18 d (the total light curve has been approximately scaled). Middle panel: Hardness, defined as the ratio of counts between 1-4.5 keV to counts between 0.4-1 keV, at a time resolution of 300 s. Again, counts from all EPIC detectors were co-added. Bottom panel: Optical U-band light curve from OM (on a logarithmic count rate scale). The time resolution is 80 sec per bin. Time gaps are of instrumental origin, specific for the selected observing mode.

Open with DEXTER


In the text

Figure 2: Similar to Fig. 1, with the following exceptions: The top and the middle panels additionally show the PN-only curves (dashed), since only those data were not piled up during the flare peak. And the optical light curve in the bottom panel is shown at a time resolution of 40 s, to illustrate some finer details during high flux. The OM observation was terminated early, but covers the main flare peak. The intervals A-E used for spectroscopy in this paper are marked in the top panel.

Open with DEXTER


In the text

	Figure 3: EPIC PN spectra during various stages of the flare (A-E) and during the pre-flare episode Q. The count rates have been multiplied by various factors to prevent mutual overlap (see factors indicated at lower left). Note the variable high-energy slope, indicative of variable flare temperatures.
Open with DEXTER


In the text

	Figure 4: Fluxed RGS1+2 spectra referring to the intervals Q (pre-flare low level), and B-E (see Fig. 2 for reference). The bin resolution is 50 mÅ except for panel B where a resolution of 78 mÅ has been used to improve the signal-to-noise ratio.
Open with DEXTER


In the text

	Figure 5: O  VII triplets during the low-level episode Q and during various flare intervals. The small number at upper right indicates the flux at the upper border of the plot. The three vertical lines mark the positions of the resonance, the intercombination, and the forbidden line (for increasing wavelength).
Open with DEXTER


In the text

	Figure 6: Same as Fig. 5, but for the Ne  IX triplet. The dashed lines indicate two potentially strong, blending Fe  XIX lines.
Open with DEXTER


In the text

Figure 7: Elemental abundances determined during the low-level episode Q and during flare intervals A-E. All abundances are plotted as ratios to the Fe abundance, and are normalized with the respective solar photospheric abundance ratios according to Anders & Grevesse (1989). For Fe, an updated value given by Grevesse & Sauval (1999) has been used. The open circles refer to solar abundances of O and C reported by Allende Prieto et al. (2001) and Allende Prieto et al. (2002),respectively.

Open with DEXTER


In the text

	Figure 8: Emission measure distributions for the pre-flare interval Q and for flare intervals A-E, using a polynomial reconstruction algorithm that fits the DEM with polynomials of order 6. The lowest panel shows a superposition of all flare DEMs on the same absolute scale.
Open with DEXTER


In the text

	Figure 9: Similar to Fig. 8, but using a regularization inversion of the spectrum to obtain the EM distribution.
Open with DEXTER


In the text

	Figure 10: Comparison of the fluxed RGS spectrum of YY Gem ( bottom panel) with the spectrum obtained during the large flare (combined intervals A, B, C, and E; middle panel) on Proxima Centauri, and the low-level spectrum (Q; top panel). The spectral bin width is 43.74Å.
Open with DEXTER


In the text

Figure 11: Top: Simulation of the large Proxima Centauri flare using a simple 2-R flare model described in the text. The best fit shown here is for n=2, B=2050, t0=4000, q=0.01. The solid line shows the model X-ray luminosity, to be compared with the observations that are plotted by small crosses. The dotted curve indicates the electron density (see scale on the right), and the two solid lines at the bottom of the figure mark the height of the emitting loop (see scale on the right). The model clearly decays too fast beyond t = 1000 s after flare start, which is due to the emergence of a secondary flare. Bottom: Similar to upper figure, but here the light curve synthesized from a time-dependent hydrodynamic simulation of plasma confined in a single footpoint-heated loop is overplotted (solid line). The heating lasts 600 s. The flare start is defined slightly differently to optimize the fit to the rise phase (the precise start of the flare cannot be defined from the soft X-rays unambiguously). Further details are described in Reale et al. (2004).

Open with DEXTER


In the text