5 Discussion

There is observational evidence for the clumpy structure of SN ejecta in the remnants of different types and ages. Observations are available for the core-collapsed SNRs (e.g. Cas A) and also for the remnants thought to be of Type Ia (see e.g. the recent XMM-Newton observations of Tycho by Decourchelle et al. 2001). Some of the SNRs are young (e.g. Cas A) and others are older (e.g. the Vela SNR).

The observation of the ejecta clumps composition would provide a valuable test to study the details of SN phenomena. The simulations of SN nucleosynthesis (e.g. Woosley & Weaver 1995; Thielemann et al. 1996) predict the production of $\sim$ $0.1~\mbox{$M_{\odot}$ }$ of ⁵⁶Ni and ²⁸Si and $\ga$ $0.003~\mbox{$M_{\odot}$ }$ of ³⁶Ar in a core-collapsed SN. The metal distribution through the ejecta at the postexplosion stage is still poorly known. There is conclusive evidence for large-scale macroscopical mixing of Ni in the SN 1987A ejecta (e.g. McCray 1993). However the modeling of the ⁵⁶Fe distribution through the ejecta at the microscopic level is rather a complicated task. This is because of the energy deposition effect from ⁵⁶Ni decay (see e.g. Wang & Chevalier 2001; Blondin et al. 2001 for recent discussions). The X-ray line emission in our model is less sensitive to the details of the element distribution in a fragment because of the high penetrating ability of energetic electrons responsible for the line excitation. Even if the Fe (or other element) atoms are locked in the dust grains we would still have X-ray line emission excited by fast electrons.

We show that the X-ray line luminosities and spectra are different for the SN ejecta fragments in a dense molecular cloud and for that in a tenuous medium. The efficiency of X-ray line production is higher in the fragments of lower ionization state. Strong MHD shocks could transform a sizable fraction of the kinetic energy into nonthermal particles thus reducing the heating of the shocked gas and increasing its compression. The column density of the shocked gas could reach $\sim$ $10^{21} \rm ~cm^{-2}$ and the ion temperature just behind the shock transition region is below 5 $\times 10^6$ K. Thus, the fast fragments (of velocity $v_{\rm k} \la1500 \rm ~km~s^{-1}$) moving through molecular clouds of density > $10^3 \rm ~cm^{-3}$ would drive a radiative bow shock wave. Hollenbach & McKee (1989) have modeled the spectra of the radiative shocks. They found rich spectra of H₂ emission as well as atomic fine-structure lines with strong [OI] (63 $\rm\mu m$) [OIII], [Fe II], [C II] (158 $\rm\mu m$), [Si II] (35 $\rm\mu m$), [Ne II] (12.8 $\rm\mu m$) lines. This implies a correlation of IR and optical emission of the fragment with the X-ray line emission dominated by the lines of relatively low ion charge states. The soft thermal X-ray continuum emission from the postshock gas of T $\la10^7$ K is not expected to be observable in that case. On the other hand nonthermal bremsstrahlung emission with the hard spectrum of a typical photon indexes <1.5 is predicted in our model. The continuum bremsstrahlung emission is rather sensitive to the ionization state of the fragment. Our simulations show that the X-ray line luminosity is $\la$10^-4 of the kinetic energy dissipation rate in the fragment's bow shock.

The lifetime of the X-ray-line-emitting fragments in a dense molecular cloud is expected to be a few times less than the fragment deceleration time. Such a fragment should appear as a nonthermal source showing variability on a timescale of years. The variability is most important for radio emission. Synchrotron radio emission at a level about 100 mJy arcsec^-2at 100 MHz could be expected from the fragments interacting with a dense molecular cloud if the ambient magnetic field is about 0.1 mG. Radio emission from a fragment propagating through a tenuous medium is much fainter. However, MHD type instabilities may greatly enhance the magnetic fields along the fragment boundary providing substantial radio emission, see Anderson et al. (1994); Jones et al. (1996).

The X-ray line luminosity $L_{\rm x} \sim$ 10 $^{31}~ \rm ~erg~s^{-1}$ (per 10 $^{-4}~\mbox{$M_{\odot}$ }$ of Si, S, Ar, Ca, Fe) is predicted from an individual SN fragment in a molecular cloud. It could be detected from a few kpc distance with the current detectors aboard Chandra and XMM-Newton. An obvious candidate to study is IC 443: an SNR interacting with a molecular cloud. The recent XMM-Newton observation of IC 443 by Bocchino & Bykov (2001) has resolved the hard X-ray source 1SAX J0618.0+2227 correlated with a molecular cloud into two sources one of which is generally consistent with the ejecta fragment interpretation. A dedicated observation of 1SAX J0618.0+2227 is desirable to check the interpretation.

An ensemble of unresolved SN fragments could contribute substantially to the observed diffuse iron line emission. Diffuse iron line emission has been found by ASCA from the Galactic Centre region on a scale of about a degree along the galactic plane (Koyama et al. 1996) and from a more extended region of the Galactic Ridge (e.g. Tanaka 2002). He-like and H-like Fe-K lines analyzed by Tanaka (2002) are significantly broadened corresponding to a velocity dispersion of a few thousand  $\rm ~km~s^{-1}$. The high dispersion can be naturally explained if the lines are due to unresolved faint SN fragments. Alternatively, Tanaka (2002) has suggested that the high dispersion is due to charge-exchange processes of low-energy cosmic rays providing the high velocities of the line emitting ions.

A molecular cloud irradiated by a hard X-ray source is expected to be observable in fluorescent X-ray lines as X-ray reflection nebulae - a new class of X-ray sources (e.g. Sunyaev et al. 1993; Sunyaev & Churazov 1998). The observed iron line emission from the Sgr B2 (Murakami et al. 2001) and Sgr C (Murakami et al. 2001a) complexes was considered by the authors as the X-ray reflection nebulae irradiated by a bright source in the GC which was active in the past. Clumps of X-ray line emission of neutral and highly ionized iron of equivalent widths $\sim$keV and absorption corrected 2-10 keV fluxes $\sim$ $10^{-12} \rm ~erg~s^{-1}~cm^{-2}$ were found by Chandra in the Galactic center region (Bamba et al. 2002). The photon indices of 2-10 keV continuum emission - while not too constrained - are broadly consistent with the photon indices $\leq$1.5 predicted by the model of SN fragments emission considered above. If a substantial fraction of the iron ejected by an SN in a molecular cloud would reside in the fast moving fragments we could expect an iron line luminosity of $L_{\rm x} \sim$ 10 $^{34}~ \rm ~erg~s^{-1}$ per SN which is consistent with that observed from Sgr C complex. The line emission from the Sgr B2 is somewhat brighter and would require contributions from more than one SNe. Infrared line observations from radiative shocks that accompany SN fragments in molecular clouds would help to distinguish the SN fragments contributions. A correlation between SiO ( $J=2\rightarrow 0$) emission morphology and 6.4 keV Fe line emission on the Galactic Center large scale and also within Sgr A and B regions found recently by Martin-Pintado et al. (2000) could be consistent with the multiple SN fragments contribution. ISO observations of 18 molecular clouds in the Galactic Center region were performed by Martin-Pintado et al. (2000a). Towards most of the clouds they have detected the "ionized bubbles'' with the fine structure line emission of ionized species: SIII, NeII, and in some cases also NeIII, NII, NIII, OIII. The "bubbles'' could be relevant to the fragment-cloud interactions.

Nonthermal hard X-ray continuum emission from the Galactic ridge requires a population of accelerated electrons (e.g. Valinia et al. 2000; Dogiel et al. 2002). Hard continuum of the SN fragments could contribute to the observed emission. The galactic diffuse emission mapping with forthcoming INTEGRAL mission will address the issue.

Acknowledgements

I thank the referee for very constructive comments. The work was supported by INTAS-ESA 99-1627 grant.