We present deep images of the SN 1006 remnant (G327.6+14.6) with the ACIS instrument on the Chandra X-Ray Observatory. Two regions have been observed, the synchrotron-dominated northeast limb and the thermally dominated northwest limb, as well as a substantial portion of the interior of the remnant shell. The brightest features in the X-ray images correspond closely to radio features in the northeast and to Balmer-dominated filaments in the northwest. The spectra of the brighter filaments in the northeast are harder, with less prominent line emission than those in the northwest. In addition to highly elongated filaments, both images show enhanced clumps of emission well inside of the shock front that appear to be dominated by emission from oxygen. These probably arise from shock-heated ejecta, based on analysis of their X-ray spectra. We find no firm evidence for a halo of X-ray emission outside the shock to the northeast, as predicted by the Fermi shock-acceleration picture, in which relativistic electrons should be diffusing ahead of the shock. Our limits on upstream emission are less than 1.5% of the postshock levels in regions where the supernova remnant is brightest. This strongly suggests that the bright rims are flattened sheets nearly perpendicular to the plane of the sky and that the magnetic field strength jumps at the shock by a factor significantly larger than 4, as has been proposed if the shock puts significant energy into accelerating nonthermal ions. The spectra obtained of the northwest rim are all dominated by the helium-like ions of O, Ne, Mg, and Si expected from shocks with ionization (n e t) parameters of order 100 cm À3 yr and electron temperatures of 0.5-1 keV, far lower than the postshock ion temperature implied by estimates of the shock speed obtained from the shape of the H line.

Supernova remnants may exhibit both thermal and nonthermal X-ray emission. In a previous study with ASCA data, we found that the middle-aged supernova remnant RCW 86 showed evidence for both processes, and predicted that observations with much higher spatial resolution would distinguish harder X-rays, which we proposed were primarily synchrotron emission, from softer, thermal X-rays. Here we describe Chandra observations which amply confirm our predictions. Striking differences in the morphology of X-rays below 1 keV and above 2 keV point to a different physical origin. Hard X-ray emission is correlated fairly well with the edges of regions of radio emission, suggesting that these are the locations of shock waves at which both short-lived X-ray emitting electrons, and longer-lived radio-emitting electrons, are accelerated. Soft X-rays are spatially well-correlated with optical emission from nonradiative shocks, which are almost certainly portions of the outer blast wave. These soft X-rays are well fit with simple thermal plane-shock models. Harder X-rays show Fe Kα emission and are well described with a similar soft thermal component, but a much stronger synchrotron continuum dominating above 2 keV, and a strong Fe Kα line. Quantitative analysis of this line and the surrounding continuum shows that it cannot be produced by thermal emission from a cosmic-abundance plasma; the ionization time is too short, as shown both by the low centroid energy (6.4 keV) and the absence of oxygen lines below 1 keV. Instead, a model of a plane shock into Fe-rich ejecta, with a synchrotron continuum, provides a natural explanation. This requires that reverse shocks into ejecta be accelerating electrons to energies of order 50 TeV. We show that maximum energies of this order can be produced by radiation-limited diffusive shock acceleration at the reverse shocks.In an Appendix, we demonstrate that an explanation of the continuum as due to these predictions, we observed RCW 86 with the Chandra X-ray Observatory in February 2001. We report here on these observations, which not only confirm our predictions, but allow us to associate synchrotron X-ray emission with Fe-rich SN ejecta seen through Fe Kα line emission, and identify all high-energy continuum as synchrotron emission.

We present new Very Large Array observations at five frequencies, from 1.4 to 22 GHz, of the extremely low-metallicity blue compact dwarf SBS 0335−052. The radio spectrum shows considerable absorption at 1.49 GHz, and a composite thermal+non-thermal slope. After fitting the data with a variety of models, we find the best-fitting geometry to be one with free-free absorption homogeneously intermixed with the emission of both thermal and non-thermal components. The best-fitting model gives an an emission measure EM ∼ 8×10 7 pc cm −6 and a diameter of the radio-emitting region D ≈ 17 pc. The inferred density is n e ∼ 2000 cm −3 . The thermal emission comes from an ensemble of ∼ 9000 O7 stars, with a massive star-formation rate (≥ 5 M ⊙ ) of 0.13-0.15 yr −1 , and a supernova rate of 0.006 yr −1 . We find evidence for ionized gas emission from stellar winds, since the observed Brα line flux significantly exceeds that inferred from the thermal radio emission. The non-thermal fraction at 5 GHz is ∼0.7, corresponding to a non-thermal luminosity of ∼ 2×10 20 W Hz −1 . We derive an equipartition magnetic field of ∼ 0.6-1 mG, and a pressure of ∼ 3 × 10 −8 − 1 × 10 −7 dyne cm −2 . Because of the young age and compact size of the starburst, it is difficult to interpret the non-thermal radio emission as resulting from diffusion of supernova (SN) accelerated electrons over 10 7 − 10 8 yr timescales. Rather, we attribute the nonthermal radio emission to an ensemble of compact SN remnants expanding in a dense interstellar medium. If the radio properties of SBS 0335−052 are representative of star formation in extremely low-metallicity environments, derivations of the star formation rate from the radio continuum in high redshift primordial galaxies need to be reconsidered. Moreover, photometric redshifts inferred from "standard" spectral energy distributions could be incorrect.

The remnant of SN 1006 has an X-ray spectrum dominated by nonthermal emission, and pre-ASCA observations were well described by a synchrotron calculation with electron energies limited by escape. We describe the results of a much more stringent test: fitting spatially integrated ASCA GIS (0.6-8 keV) and RXTE PCA (3-10 keV) data with a combination of the synchrotron model sresc newly ported to XSPEC and a new thermal shock model vpshock. The new model can describe the continuum emission above 2 keV well, in both spatial distribution and spectrum. We find that the emission is dominantly nonthermal, with a small but noticeable thermal component: Mg and Si are clearly visible in the integrated spectrum. The synchrotron component rolls off smoothly from the extrapolated radio spectrum, with a characteristic rolloff frequency of 3.1 × 10 17 Hz, at which the spectrum has dropped about a factor of 6 below a powerlaw extrapolation from the radio spectrum. Comparison of TeV observations with new TeV model images and spectra based on the X-ray model fits gives a mean postshock magnetic field strength of about 9 µG, implying (for a compression ratio of 4) an upstream magnetic field of 3 µG, and fixing the current energy content in relativistic electrons at about 7 × 10 48 erg, resulting in a current electron-acceleration efficiency of about 5%. This total energy is about 100 times the energy in the magnetic field. The X-ray fit also implies that electrons escape ahead of the shock above an energy of about 30 TeV. This escape could result from an absence of scattering magnetohydrodynamic waves above a wavelength of about 10 17 cm. Our results indicate that joint thermal and nonthermal fitting, using sophisticated models, will be required for analysis of most supernova-remnant X-ray data in the future. Recent models of nonlinear shock acceleration (e.g., Berezhko Yelshin & Ksenofontov Berezhko