2. Young supernova remnants
Studies of supernovae from the moment of the explosion up until they dissolve and merge with the general interstellar medium has, in general, been guided by observational opportunities as combined with model interpretations; the interplay of different physical processes, which vary spatially and rapidly within a young supernova remnant, cannot be disentangled from measurements alone. Observational windows are (from low to high energy radiation): Radio maps from electrons synchrotron-emitting in magnetic-field structures around the shocked-gas region; infrared emission from cool and hot dust within dense clumps embedded in the supernova remnant; optical line emission from the edges of dense material embedded in the remnant's hot plasma; X-ray emission from the hot, ionized, gas that has been shocked by the forward blastwave and by the reverse shock traveling inward, respectively; and -rays from long-lived radioactivity such as 44Ti.
Here we are mainly interested in the thermal history of the radioactive 44Ti after it has been ejected in a supernova explosion. 44Ti as well as other iron-group elements are synthesized during the very early moments of the explosion of a massive star in the layers adjacent to the nascent compact object (a neutron star or black hole). Therefore, observations of iron can be used to trace also 44Ti. However, not many good cases are known so far.
SN 1987A is certainly the best studied case. The large Doppler shifts of iron lines observed in the early spectra require that iron moves with velocities much higher than some of the hydrogen which cannot be explained by spherically symmetric explosion models but indicates that it is not homogeneously distributed in an expanding shell but is found in clumpy structures. This conclusion is also supported by the unexpectedly early detection of X- and -rays from the decay of radioactive elements (for summaries, see Woosley & Weaver 1994, Nomoto et al. 1994), which again suggests that these nuclei have been transported far out into the hydrogen-rich shells. Other arguments in favor of this interpretation include the smoothness of the light-curve (e.g., Woosley & Weaver 1994, Nomoto et al. 1994 and references therein), and the time-dependent features of the spectral lines observed soon after the outburst (Utrobin et al. 1995), e.g., in the Bochum event (Hanuschik & Dachs 1987; Phillips & Heathcote 1989). In fact multi-dimensional supernova simulations (e.g., Herant et al. 1992, 1994; Shimizu et al. 1994; Shimizu 1995; Burrows et al. 1995; Janka & Müller 1995, 1996) have demonstrated that the surroundings of the newly-formed neutron star are stirred by hydrodynamic instabilities and that inhomogeneities and clumpiness of the products of explosive nucleosynthesis are likely the consequence. However, it cannot be excluded that SN 1987A is a special case and not comparable with, e.g., Cas A.
44Ti -ray emission is expected from core-collapse supernovae in general (e.g., Woosley & Weaver 1995; Thielemann et al. 1996), yet there is a significant deficit in such -ray line sources in the Galaxy for the inferred Galactic core-collapse supernova rate (Dupraz et al. 1997). The 44Ti detection for Cas A (Iyudin et al. 1994; 1997), despite the low Fe abundance, therefore has given rise to speculations about its exceptional nature, particularly because its optical and X-ray characteristics support the idea of asymmetries and a peculiar circumstellar environment (see below, and Hartmann et al. 1997 and references therein). It was suggested, for example, that the progenitor of Cas A might have been a rapidly spinning star, in which case more 44Ti could have been synthesized (Nagataki et al. 1998).
Recently, a second Galactic 44Ti source has been reported (Iyudin et al. 1998). Its alignment with an also recently discovered X-ray remnant (Aschenbach 1998) suggests that it is a very young and most nearby supernova remnant, with an age around 680y and a distance of 200pc only. This object still provides a puzzle because of the absence of radio and optical emission expected for such a nearby supernova. Yet, if confirmed, it will provide a unique opportunity for the study of supernova-produced 44Ti; in which case it is of some interest to see if the modified decay rate of ionized 44Ti addressed in our paper could also be important.
The 44Ti found in young supernova remnants is probably formed during the -rich freezeout from (near) nuclear statistical equilibrium (Woosley et al. 1973; Woosley & Weaver 1995; Woosley et al. 1995; Thielemann et al. 1996; Timmes et al. 1996; Timmes & Woosley 1997; The et al. 1998). Whereas 44Ti is obviously fully ionized at the time of explosion, 44Ti ions will recombine with electrons during the adiabatic cooling phase, which accompanies the expansion of the exploded star, to become neutral after some 1000 s (e.g., Nomoto et al. 1994), a negligible timescale when compared with the age of a supernova remnant. Therefore, during most of this early phase 44Ti will decay with the laboratory rate. The subsequent evolution of the remnant may however provide conditions for its re-ionization, which is the subject of our modeling effort.
The evolution of young supernova remnants has been modeled extensively, and the spherically-symmetric explosion into homogeneous interstellar surroundings appears well-understood (McKee and Truelove 1995). A free expansion phase is followed by an adiabatic blastwave phase, where interaction with surrounding material produces outward and inward moving shock waves leading to bright X-ray and radio emission, yet being unimportant for the energetics of the remnant. Later phases of significant slowing-down and radiative losses of the expanding remnant lie beyond the early phase where 44Ti still decays. Young remnants such as Cas A are generally understood as being somewhere intermediate between free expansion and the second phase, commonly called "Sedov-Taylor" phase.
The evolution of such idealized young supernova remnants during the ejecta-dominated stage () and the Sedov-Taylor phase () was described in an analytical model by McKee & Truelove (1995). At the interface between ejecta and ambient medium a contact discontinuity occurs, separating the exterior region of shocked and swept-up circumstellar gas behind the outward-moving blastwave from the supernova remnant interior. Inward from the discontinuity, a reverse shock travels through the expanding, cold supernova ejecta, heating up the interior gas. Bright X-ray emission results from the hot plasma on both sides of the contact discontinuity, with higher temperature on the outside heated by the blastwave ( keV), as compared with reverse-shock heated gas on the inside ( keV; e.g. Vink et al. 1996). Although such a two-temperature model appears adequate to describe the X-ray emission of many supernova remnants, considerable uncertainty remains in detail. For example, as the blastwave shock rapidly heats the ions entering the blastwave region, the thermalization times for the X-ray emitting electrons may exceed 100 y, so that non-equilibrium models are needed for a proper description of X-ray and radio emission. The detailed physical conditions within supernova remnants are far from being understood, although the general evolution follows these fairly simple descriptions quite closely, and is controlled by a few parameters, the explosion energy, the mass of ejecta, and the surrounding medium density. According to the McKee & Truelove (1995) model, an explosion energy in the range does not seem unreasonable in the case of Cas A for the possible ejecta mass of (Tsunemi et al. 1986; Vink et al. 1996) if the surrounding ambient gas density is of the order of 20 cm-3 (Tsunemi et al. 1986) and the current blastwave radius is pc for an assumed distance of 3.4 kpc (Anderson & Rudnick 1995; Holt et al. 1994; Jansen et al. 1988; Reed et al. 1995).
Although the general appearance of supernova remnant images in the radio and in X-rays is that of a large-scale shell configuration as expected from the above model, additional prominent clumpy structures appear in some cases (e.g., Anderson & Rudnick 1995; Koralesky et al. 1998). This suggests that the model outlined so far is an oversimplification as far as details are concerned. The gross radio and X-ray emissivity may not be very sensitive to such discrepancy, tracing the electron component in the vicinity of the shock region, but the bulk ejected mass may be inadequately represented by the inferred electron densities and temperatures (e.g., Koralesky et al. 1998). Thus, for the Cas A remnant, prominent structures have been studied in their respective forms of optical knots and filaments, "quasi-stationary flocculi", "fast-moving knots", and "fast-moving flocculi" (e.g., van den Bergh & Kamper 1983; Reed et al. 1995; Peimbert & van den Bergh 1971; Chevalier & Kirshner 1977, 1978, 1979; Reynoso et al. 1997; Lagage et al. 1996). There is overwhelming evidence for dense structures embedded in tenuous material within the entire remnant, and even outside the blastwave shock radius. Obviously the explosion itself produces fragments of material, seen now as fast-moving knots with their abundance patterns supporting an ejecta origin. These clumps might carry heavier elements preferentially as suggested by observations of fast-moving Fe clumps early in SN explosions, e.g. in SN 1987A (e.g., Nomoto et al. 1994, Wooden 1997 and references therein), but a connection between the instabilities and clumpiness early after the supernova explosion and the fragments and "bullets" seen in the remnants has not been established yet.
Given all these uncertainties in the evolution of supernova remnants we do not attempt to model a specific object, such as Cas A, in detail here. We rather shall investigate by means of an admittedly very simple model, varying its parameters within reasonable limits, the potential effects of ionization on the 44Ti abundance estimates obtained from -ray observations.
© European Southern Observatory (ESO) 1999
Online publication: June 17, 1999