Modeling the Early Annihilation Radiation Spectrum from INTEGRAL/SPI

We undertake to model the spectrum of the galactic annihilation radiation, using the preliminary data obtained by INTEGRAL/SPI (Jean et al. 2003; Jean et al. 2004; Lonjou et al. 2004). We use the general line spectroscopy approach developed by Guessoum et al. (1991), where the interstellar medium (ISM) is divided into 5 phases: cold, warm neutral, warm ionized, and hot, in addition to a dust grain fraction, each characterized by proper physical parameters. Each phase presents us with particular positron annihilation processes, rates, and line widths. Best-fitting the model spectrum to the SPI data (through a chi-square minimization approach) then allows us to obtain "best values" for the ISM phase fractions (density times filling factor) along with uncertainties. Noting that the (astrophysical) line width determined by SPI (3.0 +/- 0.5 keV in the first set of data and 2.7 +/- 0.3 keV in the most recent analysis; Loujou et al. 2004) is rather significantly larger than the value derived by the TGRS measurements (Harris et al. 1998), i.e. 1.8 +/- 0.5 keV, we submitted the TGRS data to the same analysis. W should also note, however, that the width of the line obtained from the previous Germanium-detector observation missions, namely CRIS and HEXAGONE, were 2.5 +/- 0.4 keV "weighted mean" from 3 observation campaigns; (Leventhal et al. 1993) and 2.66 +/- 0.6 keV (Durouchoux et al. 1993), respectively. The results suggest annihilation in a medium that is warmer or hotter (more fully ionized) and quite devoid of grains, compared to the dustier and hotter (overall) medium suggested by the SPI data. We briefly discuss these results and possible interpretations as well as follow-up analyses and proposed observations.