5. Broad band integrated gamma-ray fluxes
The high resolution gamma-ray spectrometer SPI to be launched onboard INTEGRAL in the early 2000s should teach us much about the detailed nuclear processes responsible for the gamma-ray emission in Orion, and hopefully in other regions (Winkler 1997, Mandrou et al. 1997, Gehrels et al. 1997). However the broad band analysis available from the COMPTEL data can already provide some interesting information, and is also relevant to the INTEGRAL's imager IBIS (Ubertini et al. 1997).
5.1. The 0.2-1 MeV band
As we shall see in Sect. 7, apart from the 10 B spallation line at 0.717 MeV which dominates only for high values of , the main contribution to the 0.2-1 MeV band emission is the 7 Li-7 Be feature around 0.450 MeV. Since it is mainly produced through - interactions, the emissivity is greater for EP compositions richer in particles. As a consequence, the highest emission in this band is obtained for the SS composition, and the lowest for the late-WC one. In Fig. 3 we show the fluxes calculated for some intermediate EP compositions, namely our mean-wind and mean-OB compositions, at both solar () and twice solar () metallicity.
It can be seen that the (0.2-1 MeV)/(3-7 MeV) band ratio is enhanced by a factor of 1.5-2 at twice the solar metallicity. As a consequence, we predict a slight difference between the gamma-ray emission from Orion (solar metallicity) and that of the central radian of the Galaxy, dominated by the `4 kpc ring' where the metallicity is enhanced.
5.2. The 1-3 MeV band
Most of the nuclear gamma-ray lines, except those from 12 C and 16 O, are between 1 and 3 MeV. The (1-3 MeV)/(3-7 MeV) band ratio therefore depends mainly on the (C O)/He abundance ratio in the EPs. Some typical results are shown in Fig. 4. Again, we see that the 1-3 MeV band emission is higher, or if one prefers, the 3-7 MeV band emission is lower by a factor of 1.5-3 for . This is due to the fact that the 12 C and 16 O nuclei (producing the 3-7 MeV flux) have lower abundances in the winds of massive stars with higher initial metallicity, as explained in Sect. 2.2. This effect could also be of interest for future observations of diffuse gamma-ray line emission as a function of galactocentric longitude. Note in passing that the stellar evolutionary models using an enhanced mass loss rate during the MS phase lead to compositions richer in 1 H and 4 He (or poorer in 12 C and 16 O), and in turn to a slightly higher gamma-ray emission between 1 and 3 MeV.
The same conclusions can be drawn from Fig. 5, where we present essentially the same results in a slightly different way, in order to compare our calculations with the observations made by COMPTEL in Orion, as well as with previous calculations by Ramaty and co-workers (Ramaty 1996, RKL96). Indeed, the COMPTEL data set a upper limit of 0.13 on the band ratio R defined as:
The results that we obtain depend on the way we deal with the so-called `unresolved gamma-ray lines', discussed in RKL79. Nuclear interactions between energetic particles and complex nuclei (with ) produce numerous gamma-ray lines resulting from transitions between many high-lying nuclear levels populated by both direct excitations and spallation reactions. The cross sections for the production of these lines have not been measured individually (hence the term `unresolved'), but Zobel et al. (1968) measured the total production cross sections of gamma-rays of energies greater than 0.7 MeV. Now the value of the (1-3 MeV)/(3-7 MeV) band ratio depends on the distribution in energy of these unresolved gamma-rays. Estimations of this distribution for different nuclei are given in RKL79 (based on experimental data), showing a peak between 1 and 2 MeV. This peak is sharper for heavier nuclei. This is intuitively in agreement with the idea that high-lying levels are rather close to one another.
The results obtained with this prescription for the unresolved gamma-ray lines are shown in Fig. 5a. They are in very good agreement with the calculations of Ramaty and co-workers for any of our common compositions. The main result here is that all our mean-wind and mean-OB compositions lead to R band ratios compatible with available COMPTEL observations. This is important in itself since the R ratio proves to be very constraining in the case of Orion, excluding the most `natural' EP compositions, such as SS or CRS (e.g. RKL96). In the same time, it justifies the investigation of the `superbubble model' for the Orion gamma-ray line emission, discussed in Parizot (1997).
In order to stress the importance of the unresolved gamma-ray lines, we also calculated the R ratio taking into account only the resolved lines. The results are presented in Fig. 5b. They show that, in this case, no EP composition can be excluded anymore on the basis of the sole R ratio. This points out the major role of these (up to now) unresolved gamma-rays, and the astrophysical interest that would represent the precise measurement of the corresponding cross sections - at least to confirm the current estimates on the gamma-ray energy distribution. However, as argued by Ramaty (private communication), the same estimates have been used to model the solar flare emission, and the data were fit quite well (see Ramaty et al. 1997). This suggests that the estimates are actually good.
© European Southern Observatory (ESO) 1997
Online publication: March 24, 1998