For the 20 s exposure (see Figs. 7 and 8), this graph shows the upper limits of an absorption edge at 2.67 keV (lower horizontal dash-dotted line) and 3.48 keV (upper horizontal line), compared to the predictions for the K-edge of hydrogen-like ²⁸Si (dotted curve), ³²Ca (solid curve) and ⁵⁶Fe (dashed curve), for abundances as determined from the He0.1 model (pure helium ignition at 0.1 Ṁ_Edd) in Weinberg et al. (2006). The range where the upper limits apply are set by the range of black body color temperatures measured (2–3 keV) multiplied with a maximum possible color correction of 0.6 to arrive at effective temperatures (e.g., Suleimanov et al. 2011). The upper limits would be 3 to 4 times better if the edge energy would be gravitationally redshifted by a factor of 1.31. They are 4.5 times worse for an exposure time of 1 s.