Astron. Astrophys. 331, 726-736 (1998)

3. Where do the EPs come from?

The above overview allows one to draw an important conclusion: the EPs interacting in Orion represent an energetic component distinct from the ordinary Galactic cosmic rays (GCRs). Indeed: i) they are much richer in C and O, ii) they have a different energy spectrum, with a characteristic energy of order 30 MeV/n (instead of 1 GeV/n for the GCRs), and iii) their energy density is at least ten times higher (Ramaty, 1996), and probably much more (see below, Sect. 4). So, what is the origin of the Orion EPs?

3.1. Internal source?

Since the gamma-ray emission follows the Orion clouds, it may seem natural that the EPs are accelerated inside, but we show here that this is unlikely. The point is that the EPs cannot travel far from their injection site. The lifetime of C and O nuclei of a few tens of MeV/n against ionisation energy losses is rather small in a dense media. For initial energies in the energy range of interest (10-100 MeV/n), we calculated the time spent by the different nuclei above the nuclear excitation thresholds, , and obtained the following fit:

where is the Orion density () and equals 1/3 for ¹² C nuclei, and 1/4 for ¹⁶ O nuclei. Consequently, the EPs have only a few years (or years if ) to diffuse away from their injection site, and cannot go further than a distance:

where is the diffusion coefficient in the Orion clouds.

Since the Orion gamma-ray source is extended, one can conclude from this low value of the diffusion length that if the EPs originate from an internal source, then there must be actually several sources. Moreover, according to Eq. (5), all of these sources have to be recent (a few thousand years). Now this seems very difficult to achieve within the standard paradigm. Indeed, the energetics is very constraining. As shall be shown in Sect. 4, the EPs lose more than ergs every second in Orion. This means that some mechanism provides as much power either to re-inject EPs at energies above the nuclear thresholds, or to re-accelerate the initial EPs so as to maintain their energy above the thresholds. Large energy release is known to accompany the activity of massive stars (strong winds of WR stars, SN explosions), but the time scales implied by Eq. 5are too short, as there is no evidence for a SN explosion in Orion in the last years - especially not several of them ! Neither is any currently active WR star.

Note that if reacceleration takes place within the Orion clouds, so that the EP energy never drops below the excitation thresholds, the effective time scales for nuclear interactions can be lengthened. However the lifetime of energetic C and O nuclei cannot exceed their destruction time, , which is given in Table 1 for different values of the ion energy (using inelastic cross-sections from Silberberg & Tsao, 1990).

Table 1. Destruction time of C and O nuclei in a medium of mean density ().

These time scales are still short if one needs several distinct sources, and again, there is not enough power in the magnetic turbulence or in stellar winds ( ; Brown et al., 1994) to ensure reacceleration. One should therefore think of other sources of energy, such as gravitation. Accretion on black holes or neutron stars, for example, are known to involve powers comparable to the Eddington luminosity (). Several of these objects could therefore be at the origin of the Orion gamma-ray emission, as recently suggested by Bykov and Bloemen (1997).

3.2. External source?

Apart from re-accelerating the EPs or re-injecting new ones, another possibility is to dispose of a huge reservoir of EPs to draw on. As seen from the Orion clouds, this is actually a kind of re-injection, but from the outside. One advantage of this situation is that the external source doesn't have to be active now, which would raise again the problem of energetics. It suffices that enough energy was released in the past, and that the EPs keep on reaching the Orion clouds, at the rate imposed by their diffusion from the actual source. Of course, in order that the EPs survive against energy losses and nuclear destruction until they interact with the cloud material, the reservoir has to have a low density. Now such a reservoir exists close to Orion: it is the Orion-Eridanus superbubble (see below).

An other appeal of the external source scenario is that it accounts naturally for the extension of the gamma-ray source. Indeed, if the pool of EPs is as large as (or larger than) the Orion clouds A and B, the irradiation takes place over their whole surface, resulting in the observed global correlation with the CO contours (see Fig. 1). Moreover, the detailed anti-correlation mentioned in Sect. 2.1may find an explanation in that the small diffusion length discussed above is translated here into a small penetration length. This implies that the EPs encounter the Orion clouds only superficially. As a consequence, the gamma-ray emission is (in this scenario) proportionnal to the irradiated surface, while the CO observations are related to the total amount of matter, i.e. to the volume. The gamma-ray/CO anti-correlation would thus result from the basic surface/volume (limb/center) anti-correlation of clouds. Besides, this effect should be significanty enhanced in Orion, because of the very clumpy and filamentary structure of the molecular complex (Gentzel & Stutzki, 1989).

The surface irradiation of the Orion clouds is also required from an independent argument: the volume deposition of a power as high as a few times would indeed imply an ionisation rate higher than allowed by astrochemistry (Bloemen et al., 1994; Ramaty et al., 1996; Dogiel, 1996).

Finally, Ramaty et al. (1997a) have argued that the detailed structure of the Orion gamma-ray emission spectrum may be explained by an irradiation geometry in reservoir the EP source is located outside the molecular clouds, and in front of them in the line of sight.

3.3. The Orion-Eridanus bubble

All the above arguments against an internal origin of the EPs as well as in favour of a large pool outside, and more precisely in front of the Orion clouds, find a fortunate convergence in the evocation of the Orion-Eridanus (super)bubble. It was first observed in (Sivan, 1974), with shell structures extending all the way from the

arnard's loop, around the Orion clouds A and B, to 50 degrees below the Galactic plane, in the Eridanus constellation. It consists of a cavity filled with hot ionised gas and surrounded by an expanding shell of neutral hydrogen, most certainly related to the strong stellar wind and SN activity of the Orion OB1 association (Reynolds & Ogden, 1979; Burrows et al., 1993; Brown et al., 1994; Brown et al., 1995). The bubble is located in front of the Orion clouds A and B, at about 350 pc from the Sun, and pc below the Galactic plane. Its present size ( pc) and expanding velocity () allow one to estimate its age, years, and energy, ergs, in very good agreement with theoretical expectations based on stellar evolution calculations.

So we dispose of a lot of energy at the right place (close to, and in front of the Orion complex), in a medium with sufficiently low density () for the possible EPs to survive until they interact with the Orion cloud material, and we dispose in addition of a natural acceleration mechanism. Indeed, superbubbles are known to convert a significant fraction of their free energy into low-energy cosmic rays, with spectra very similar to that required by the observations recalled above (Bykov & Fleishman, 1992a; Bykov & Fleishman, 1992a, Sect. 2.4).

Finally, the Orion-Eridanus bubble may provide a natural explanation for the last ingredient of the standard paradigm of the Orion gamma-ray emission: the EP composition. As discussed in detail in Parizot et al. (1997a), the potentially accelerable material in the superbubble is made of the combined wind and SN ejecta from massive stars in the OB1 association. The resulting mean-wind composition is rich in C and O, and was shown to satisfy the various observational constraints, in particular the most relevant band ratio upper limit (1-3 MeV)/(3-7 MeV). This adds to the list of converging arguments pointing to the Orion-Eridanus bubble as the acceleration site of the Orion EPs. However, we still have to turn to the question of energetics, and try to gather further informations on the timescales of the different processes, as will be needed for purposes of generalisation to other astrophysical sites and possible Galactic emission.

© European Southern Observatory (ESO) 1998

Online publication: February 16, 1998
helpdesk@link.springer.de