Existing analysis based on spectra from the Reflection Grating Spectrometer (RGS) on board XMM-Newton already shows that the G-ratio of the O VII Heα triplet in the inner bulge of M31 is too high to be consistent with a pure optically thin thermal plasma in collisional ionization equilibrium (CIE). Different processes that may affect properties of diffuse hot plasma were proposed, such as resonance scattering (RS) and charge exchange (CX) with cold gas. To determine which physical process(es) may be responsible for this inconsistency, we present a systematic spectroscopic analysis based on 0.8 Ms XMM- Newton/RGS data, together with complementary Chandra/ACIS-S images. The combination of these data enables us to reveal multiple non-CIE spectroscopic diagnostics, including but not limited to the large G-ratios of Heα triplets (O VII, N VI, and Ne IX) and the high Lyman series line ratios (O VIII Lyβ/Lyα and Lyγ/Lyα, and N VII Lyβ/Lyα), which are not expected for a CIE plasma, and the high iron line ratios (Fe XVIII 14.2 Å/Fe XVII 17 Å and Fe XVII 15 Å/17 Å), which suggest much higher temperatures than other line ratios, as well as their spatial variations. Neither CX nor RS explains all these spectroscopic diagnostics satisfactorily. Alternatively, we find that an active galactic nucleus (AGN) relic scenario provides a plausible explanation for virtually all the signatures. We estimate that an AGN was present at the center of M31 about half a million years ago and that the initial ionization parameter ξ of the relic plasma is in the range of 3─4.
