Astron. Astrophys. 324, 683-689 (1997)

5. The search for sources of ⁴⁴ Ti line emission

Various estimations (Tammann, Löffler & Schröder 1994, van den Bergh & McClure 1994) suggest a supernova rate -3.0 events per century in the Milky Way. About 7-9 supernovae are expected to have exploded since Cas A around 1680. However no such event has been attested and no candidate supernova remnant is known.

If supernovae are distributed like Ia : Ib : II 1 : 1.5 : 7.5 (van den Bergh & Tammann 1991), most of them are core-collapse events possibly embedded in the dense clouds that gave birth to their massive progenitors. Column densities are common within molecular clouds and along the line of sight through the Molecular Ring of the inner Galaxy. With (Bohlin, Savage & Drake 1978), the combined foreground, local, and even circumstellar (Sect. 4.3) obscuration may well exceed 10-15 mag. Beside distance effects, large visual extinction can easily hide supernovae to the observer's eye.

The Milky Way is transparent to gamma rays. Though invisible at optical wavelengths, young SNR s could be betrayed through ⁴⁴ Ti line decay. As exemplified by Cas A, COMPTEL has the potential for detecting these sources. We examine any excess left in the likelihood map of Fig. 2. As explained in Sect. 2.3, we adopt ( significance level) as our criterion for serendipitous source detection. By far, no excess in the map actually fulfils this requirement. We conclude that, with the sole exception of Cas A, no Galactic source of Ti line emission is detected by COMPTEL.

With , the excess located at and is worth a mention. Although far off the Galactic plane, it is associated with a star-formation complex in Perseus which includes the molecular cloud IC 348 (Ungerechts & Thaddeus 1987) as well as a young cluster of intermediate-mass stars at the edge of the Perseus OB 2 association (Herbig & Jones 1983). Assuming a distance of (Ungerechts & Thaddeus 1987) and a fiducial ⁴⁴ Ti yield , the COMPTEL line flux of might be interpreted as a hint for a supernova event in Perseus some 700 years ago. More data are obviously required before any definite statement can be made about this excess.

Whether COMPTEL is expected to detect a supernova depends on the distance D and age t of the event [Eq. (2)]. Let us take a flux threshold as the instrument sensitivity (Sect. 4.1). Assuming an initial ⁴⁴ Ti mass of , we build the age-distance diagram of Fig. 3. Events of the past hundred years appear detectable by COMPTEL up to . This includes the Galactic Centre and represents about half of the Milky Way. For 19th-century events, the accessible domain still encompasses the forefront part of the molecular ring. With -3.0 per century, we then expect COMPTEL to detect about 3 SNR s from the last 200 years.

From the HEAO 3 and SMM measurements, important constraints have been put on the Galactic rate () and nucleosynthesis models () of supernovae (Hartmann et al. 1993, Leising & Share 1994). Our negative survey strengthens these constraints. None of the 3 sources expected on statistical grounds is clearly seen by COMPTEL. If supernova occurence follows Poisson's statistics, this is consistent at the less than 5% confidence level with the canonical rate Galactic events per century and typical yield .

As emphasised by Leising (1994), this stringent result is puzzling since the Galactic abundance of ⁴⁴ Ca imposes a contradictory conclusion. Simple arguments (Leising & Share 1994) lead to the expectation that about of ⁴⁴ Ca are produced per average century. This value must be equal to if ⁴⁴ Ti is the only natural parent of ⁴⁴ Ca. Do detonating helium white dwarfs, very rare SN Ip events which release as much as of ⁴⁴ Ti (Woosley, Taam & Weaver 1986), synthesise most Galactic ⁴⁴ Ca? Or has this side of the Milky Way been excessively supernova-quiet over more than two centuries?

So far, Cassiopeia A is the only ⁴⁴ Ti source detected by COMPTEL. It should probably be left to X-ray (XTE and SAX) or future gamma-ray (INTEGRAL) missions to unveil other supernova remnants either older (SN 1572, SN 1604) or located farther away in the Galaxy. Yet more discoveries may await COMPTEL at , triggered by further observations of sky regions where the current exposure is still poor. The Perseus excess seems the most promising target for further study.

© European Southern Observatory (ESO) 1997

Online publication: May 26, 1998

helpdesk.link@springer.de