In the summers of 1991 and 1992 the Andromeda galaxy (M31) was observed in a mosaic-like way with two deep ROSAT PSPC surveys implementing different pointing modes. The first survey consisted of six, partly overlapping deep single observations arranged along the major axis of the galaxy, whereas the second survey consisted of 80 raster-pointings with low exposure leading to a highly homogeneous image of the Andromeda galaxy. The total exposure of 200,000 secondsof each survey and a total field of view much greater than the size of M31 represent the first complete look at M31 in X-rays at such a high sensitivity. In combination with the high spectral resolution provided by ROSAT PSPC, this survey allowed us to obtain qualitatively and quantitatively new results of 31 far beyond the morphological structure detected with the Einstein Observatory.

Figure 1 shows the complete map of M31 in the hard X-ray energy band (0.5 - 1.5 keV) obtained from the data of the homogenious second survey. This is an exposure corrected count rate image with a technique applied to point out single point sources. Additionally, the D₂₅-ellipse of M31 is sketched in Fig. 1 to show the size of the galaxy.

Figure 1

The analysis of both surveys uncovered 560 single X-ray sources in the region of M31 with luminosities in the range of ~ 10³⁵ to 10³⁸ erg s^-1. Figure 2 shows the positions of all 560 sources as a projection on an optical image of the Andromeda galaxy taken from Mount Palomar Sky Survey. The high source density in the bulge region (< 5' resp. < 1 kpc around the optical center) together with the spatial resolution of the ROSAT PSPC of ~ 25' leads to source confusion. The 31 point sources detected in this region could therefore be increased up to 48 resolved sources with the help of a ROSAT HRI observation.

Figure 2

A spatial deprojection of the source distribution in M31 yielded a spiral arm-like-structure. Investigations of the radial profiles of the source density of all sources revealed a single power law distribution over the entire galaxy. The power law index of this model calculated for all sources is the same as for sources associated with globular clusters. This can be interpreted as an indication of most sources beeing low mass X-ray binaries (LMXBs).

For individual identifications of single bright sources correlation with known source catalogs were made. To this end, a source correlation method was developed to enable the determination of source completeness. An application of this method to the correlation between the 560 ROSAT sources and the 108 sources found with the Einstein Observatory in M31 resulted in a confirmation level of 90% (on a 2-sigma significance level). Additionally, a large amount of new hitherto unknown X-ray sources have been discovered. The application of transient criteria enabled us to determine the fraction sources which are variable. Extensive correlations with optical catalogs revealed assignments of the X-ray sources to known objects. 55 X-ray sources were identified with foreground stars; 33 X-ray sources were assigned to globular clusters in M31; 23 supernova remnants could be correlated, and 10 individual background galaxies were found to be X-ray emitters (among them the elliptical dwarf galaxy M32 in the immediate neighbourhood of M31). The remaining 439 non-identified X-ray sources were statistically analysed (see below) to determine their source characteristics.

Spectral investigations of the brightestsources relvealed typical spectra for sources of different object classes. Figure 3 shows a color representation of the energy distribution drawn from the data of the homogenious second survey. Red indicates X-ray flux in the soft energy band from 0.1 to 0.4 keV, green in the range of 0.5 to 0.9 keV, and blue in the hard energy band from 0.9 to 2.0 keV. Additionally, the D₂₅-ellipse of M31 is sketched in Fig. 3 to show the size of the galaxy. The main part of all sources have hard spectradue to the relatively high foreground absorption by galactic hydrogen along line of sight to M31 (N_H ~ 6 x 10²⁰ cm^-2). Actually, most of the red sources in Fig. 3 were identified as foreground stars. For the group of hard sources spectral fits allowed the calculation of hydrogen column densities which agree with local values in M31 known from radio observations.

Figure 3

The (morphological) comparison of the Andromeda galaxy with the Milky Way was of special interest for the investigations. These two objects are similar in their Hubble types (Sb), size, mass, and age. Because of these similarities, and based on current models of galaxy evolution, similar globular cluster (GC) luminosity functions in both galaxies were expected. Actually, an unexplained discrepancy between both functions had been found in the past. These luminosity functions indicate that GC-sources seem to be more numerous and/or more luminous in M31 than in the Milky Way. This result, mainly based on the data of the Einstein Observatory, therefore implied a morphological (and perhaps evolutional) difference between both galaxies for which an explanation had to be found. To help in solving this problem, our collaborating group at MIT performed a deep optical survey of M31. The comparison of their optical data with the improved X-ray data by ROSAT (compared with the data from the Einstein Observatory) together with the adoption of statitical methods to handle censored data now enabled us to solve the GC-problem. Now the GC-luminosity function of M31 turned out to be comperable to that of our Milky Way within the statistical errors (with a Kolmogorov-Smirnov possibility of 86%). Therefore, this result confirms the morphological analogy of M31 with the Milky Way and supports our knowledge about the general structure of spiral galaxies.

A further morphological aspect is the distibution of gas and dust masses within M31 and their influence on X-ray spectra and luminosity of single sources as well as for the whole galaxy. Not only is the attenuation of sources in M31 by interstellar material important, but it can also obscure sources behind M31 shining through the galaxy (called background sources). Taking into account the absorption properties of M31 allows us to make statistical statements about the group of the 439 non-identified X-ray sources. Using results of radio observations of M31, the galaxy was divided into different regions according to their absorption properties. Published elemental abundances were used in the absorption calculations to achieve accurate results. With the arrived transmission values the source population of M31 could be divided into foreground, background, and real M31 sources. As a result, ~ 1/7 of all sources are found to be foreground objects, ~ 1/5 background sources, and only 2/3 are sources in M31 itself. Since this analysis has been performed in the logN-logS space integrating the flux of all M31 sources has enabled us for the first time to calculate the total X-ray luminosity of M31,

in the 0.1-2.4 keV energy band. The diffuse emission of hot gas in the central region typical for spiral galaxies was confirmed for M31. After the substraction of luminosity due to point sources a rest luminosity of

for the diffuse emission was found for the bulge region. The emission cannot be explained as a population of point sources below the detection threshold. On the other hand, an explanation of the diffuse emission in terms of thermal bremsstrahlung from hot gaseous with kT = 5 keV as a spectral model and foreground absorption by a hydrogren column density of N_H = 6 x 10²⁰ cm^-2 yields an upper limit of

for the gas mass.

Possible absorption of the diffuse background X-ray radiation by the interstellar material in M31 has been theoretically discussed in the past but never observed. The shadowing effect should be especially prominent in the HI ring region in the middle of the disk where the column density is particularly high. Because this effect is a superposition of diffuse galactic foreground emission and absorption it can only be detected with an instrument with high sensitivity together with a low intrinsic (particle) background. Such emission was, therefore, not detectable with the Einstein Observatory, and it was even near the detection limit of ROSAT. ROSAT has succeeded in qualitatively detecting this effect for the first time and giving us an order of magnitude estimate of the extragalactic diffuse X-ray emission , based on a semi-quantitative analysis. Similar values have been found for the diffuse X-ray emission within the Ursa Major region, which is more easily accessible. This strengthens the interpretation that the decreased count rate density in th ring region is a shadowing effect of the diffuse X-ray background by M31, an effect that had been unsuccessfully looked for in the past.