As extended objects with high temperatures and a potentially complex set of abundances and ionization conditions, supernova remnants (SNRs) are prime objects for studies with imaging X-ray telescopes. With the superb angular resolution provided by Chandra, along with temporal and spectral information, unprecedented details of SNR shock structure and the presence and nature of associated compact objects are being revealed. This article presents a summary of Chandra results in SNR research. Space prohibits anything close to a complete review, even at this early juncture in the mission; hopefully my colleagues in the SNR community will not feel slighted if some of their excellent work has not fit into this short review of selected interesting results.
Except where otherwise noted, the images in this article are ``X-ray color'' images in which red, green, and blue are used to represent low, medium, and high energy X-rays. It is important to note, however, that energy band selections, scaling, and image stretch vary from one image to the next.
G292.0+1.8 is a member of the ``oxygen-rich'' class of SNRs with optical emission revealing the presence of ejecta whose tell-tale composition points to a massive star progenitor. Previous studies have shown that the X-ray emission is dominated by ejecta, and have revealed a region of hard emission indicative of a pulsar-driven nebula. Chandra observations have opened the curtain to clearly reveal an unprecedented example of a composite SNR showing both such components (Hughes et al. 2001, ApJ, 559, L153; Park et al. 2002, ApJ, 564, L39). Figure 1 (left) shows an image of the remnant that reveals structure and spectral variations on both large and small scales. The bulk of the interior emission is dominated by ejecta, and spectra from discrete regions suggest large-scale variations in ionization state and elemental abundance within the ejecta. The bar-like structure running across the center, along with the thin filamentary structure encircling most of the SNR, has solar abundances indicative of material swept up by the forward shock - in this case, perhaps the contents of an equatorially-enhanced stellar wind bubble. Slightly offset from the center is a compact source surrounded by faint, hard emission (dark blue in the figure). The hard-band image in Figure 1 (right) reveals this as a neutron star with a surrounding pulsar wind nebula, similar in size to the Crab. Most recently, observations with the Parkes radio telescope (Camilo et al. 2002, astro-ph/0201384) have identified a young pulsar with a 135 ms rotation period that appears to be the counterpart, thus establishing G292.0+1.8 as the poster-child for SNRs as we thought they would be. This SNR is not alone, however. Kes 75 houses a fast pulsar, and Chandra observations have revealed the surrounding wind nebula as well as a shell whose spectrum may be composed of both ejecta and swept up material (Collins, Gotthelf, & Helfand 2001, astro-ph/0112037). Ongoing studies are sure to resolve these various components of more SNRs and provide us with a rich set against which to constrain models for their formation and evolution.
The stellar ejecta revealed in young SNRs like G292.0+1.8 allow us to probe the details of the explosive nucleosynthesis and the resulting distribution and mixing of the enriched material. In Cas A, for example, Chandra observations show a complex structure of clumpy ejecta (Figure 2, left). Spectral investigations reveal discrete emission regions whose composition traces their origins to different layers in the progenitor core (Hughes et al. 2000, ApJ, 528, L109). Iron-rich regions (which appear red in Figure 2, due to the prominent soft Fe-L emission) which were synthesized nearest the stellar core are found prominently in the outermost regions of the remnant, a discovery confirmed by larger scale equivalent-width maps (Hwang et al. 2000, ApJ, 537, L119).
In 1E 0102.2-7219 (Figure 2, right), an O-rich SNR in the LMC, an inner ring of ejecta is observed which contains several solar masses of oxygen, as well as significant amounts of neon and magnesium (Canizares et al. 2000, astro-ph/0105060), but with curiously little evidence of iron, prompting speculation that the bulk of the iron synthesized in the progenitor core remains in an unseen compact object created in the collapse (although XMM-Newton RGS observations reveal some weak Fe-L emission; Rasmussen et al. 2001, A&A, 365, L231).
On a much larger scale, it has been suggested that structures outside of the nearby Vela SNR correspond to fragments of ejecta that have overrun and bypassed the SNR shock. Chandra observations of Vela ``Shrapnel A'' (Miyata, E. et al. 2001, ApJ, 559, L45) indicate a very high Si/O abundance ratio and ionization conditions far from equilibrium, supporting an interpretation that this is a large clump of ejecta thrown out in the explosion of the progenitor star. Meanwhile, the nature of Vela ``Bullet D'' is less clear. For a plasma in ionization equilibrium, the abundances of O, Ne, and Mg are enhanced over solar values, consistent with an ejecta interpretation, while a nonequilibrium model yields roughly solar values indicative of swept-up material (Plucinsky et al. 2001, astro-ph/0112287). The morphology of X-ray emission revealed with Chandra appears more consistent with an interpretation wherein the primary shock has broken through a low density region and forged ahead of the SNR shell to ultimately interact with dense interstellar/circumstellar material.
Chandra's ability to provide spectra on small angular scales has been particularly important in studies of SNR shock structure. The basic picture for the formation of a young SNR shell is that the expanding ejecta drives a shock into the surrounding medium. This shock sweeps up and heats ambient material, creating an X-ray emitting shell with abundances typical of the ISM or CSM. As the shock slows down, the increased pressure drives a reverse shock into the ejecta, heating it as well. This creates an inner shell with abundances that reflect the composition of the stellar ejecta. This is seen clearly in Figure 2 (of 1E 0102.2-7219), where a hot (blue) outer shell is observed to have normal (SMC) abundances, while the somewhat cooler (white) inner shell is dominated by ejecta.
Because SNRs are relatively young (lasting only of order 50 kyr in X-rays) with relatively low densities, the ionization state typically lags behind the observed electron temperature. This effect is observed clearly in 1E 0102.2-7219. ACIS spectra from the ejecta ring show that the O VII emission peaks inward of the O VIII emission, consistent with an ionizing reverse shock propagating inward (Gaetz et al. 2000, ApJ, 543, L47). Similarly, HETG observations which provide images of the entire SNR in individual emission line, show that the radius of the O VII image is smaller than than in O VIII (Flanagan et al. 2001, astro-ph/0105040).
In the simplest picture, the passage of material through the SNR shock results in electrons and ions being boosted to the velocity of the shock. Because of the mass difference, this means that the electrons and ions are not initially in temperature equilibrium. The maximum timescale for equilibration is that provided by Coulomb interactions, but plasma processes may reduce this considerably. The state of equilibration is important, however, because while the dynamics of SNR evolution are dominated by the ions (which carry the bulk of the momentum), it is the electrons that produce the X-ray emission we observe. Thus, when temperature measurements are used to infer the shock velocity, for example, the state of temperature equilibration is exceedingly important. HETG observations of SN 1987A (Michael et al. 2001, astro-ph/0112261) yield spectra consistent with an electron temperature of 2.6 keV. However, the broadened line profiles indicate a blast wave speed of which corresponds to a post-shock temperature of 17 keV, providing evidence for incomplete electron-ion temperature equilibration. Similarly, the blast wave speed inferred from an expansion study comparing an ACIS image of 1E 0102.2-7219 with high resolution images taken with Einstein and ROSAT indicates a post-shock temperature which is much higher than the observed electron temperature (Hughes, Rakowski, & Decourchelle 2001, ApJ, 543, L61). In this case, the discrepancy appears to be larger than can be accounted for assuming Coulomb equilibration, suggesting that a significant fraction of the shock energy went into cosmic ray acceleration rather than thermal heating of the postshock gas.
The Crab Pulsar has (too) often been viewed as the prototypical compact object we should expect to be formed in the collapse of a massive star. But searches for similar objects in young SNRs have more often turned up either no obvious counterparts or objects whose luminosities and spin-down properties pale in comparison to the Crab. Indeed, Chandra observations of Cas A - possibly the youngest SNR we know of in our Galaxy - uncovered a compact source whose properties couldn't be less like the Crab; there is no synchrotron nebula, the spectrum is not well-described by a hard power law, to date no pulsations have been detected, and the luminosity is orders of magnitude lower than that of the Crab. Compact objects clearly different from the ``prototype'' are observed in RCW 103 and other SNRs as well. A summary of the properties of these objects, as well as the anomalous X-ray pulsars, is beyond what can be covered here. Suffice it to say that we now know that the parameter space occupied by young neutron stars spans a considerably broader range than that represented by the Crab.
That said, Chandra observations have now shown us that the Crab does have its close cousins. In the most general terms, the basic structure of such pulsar-driven nebula can be described as follows: A pulsar injects an energetic equatorial wind into its surroundings, in addition to particle jets along the pulsar spin axis, forming an expanding magnetic bubble filled with synchrotron-emitting particles. The power law spectrum of the nebula steepens with radius as the more energetic particles suffer greater synchrotron losses. The confinement of the bubble by the pressure of swept-up ejecta defines an outer boundary condition; a termination shock is formed in the far interior, where the pulsar wind is decelerated to gradually match the nebular flow. An extended synchrotron nebula in an SNR thus indicates the presence of a pulsar that powers the system, and the position of the wind termination shock depends on the internal pressure in the nebula and on the rate at which energy is delivered into the wind from the spin-down of the pulsar. For the Crab, optical wisps mark this termination shock at a distance of pc (10 arcsec) from the pulsar. For weaker pulsars, the shock may be at smaller radii.
Chandra observations of G21.5-0.9 (Slane et al. 2000, ApJ, 533, L29; Safi-Harb et al. 2001, 561, 308), a Crab-like nebula, clearly reveal the tell-tale spectral steepening with radius indicative of a central particle injection site. In addition, a compact core is observed. However, with Chandra's exquisite resolution, the core is resolved; it is not a point-like pulsar but rather a fuzzy object whose radius is consistent with the expected size of the wind termination shock. Similarly, in the Crab-like remnant 3C 58, the compact core is resolved by Chandra to reveal a non-spherical extended core. In this object, Chandra also reveals the young pulsar with a period of ms (Murray et al. 2001, astro-ph/108489); the spin-down energy predicts a termination shock consistent with the size of the extended core. Probing to yet smaller sizes, limits on emission from the surface of the neutron star itself yield a temperature less than K which, given the age of 3C 58 (believed to be the counterpart to SN 1181), is well below standard models for the cooling of neutron stars, indicating enhanced neutrino production in the interior perhaps associated with exotic material such as pion condensates (Slane, Helfand, & Murray 2002).
Figure 3 (left) shows a Chandra image of G11.2-0.3, a young SNR containing a 65 ms pulsar (Kaspi et al. 2001, ApJ, 560, 371). The pulsar is surrounded by an extended wind nebula that is clearly non-spherical, possibly indicating outflows from the pulsar. Such jets are observed in the Crab and Vela pulsars, and Chandra observations of other pulsars are indicative of such structure as well (Gotthelf et al. 2001, astro-ph//0105128). Rivaling the Crab itself is the complex structure observed in Chandra observations of PSR 1509-58 and its associated wind nebula (Gaensler et al. 2001, astro-ph/0110454). Shown in Figure 3 (right), the nebula is dominated by an elongated feature interpreted as a relativistic jet from the pulsar, and is surrounded by discrete knots and an arc of emission interpreted as an equatorial outflow. Along the northern extent of the nebula is a region of thermal emission (green in the Figure) which may be associated with the surrounding SNR.
Studies of SNRs and their compact objects have yielded what are arguably some of the most spectacular images produced by Chandra. But beneath the pretty pictures lies a wealth of information on the physics of these objects, and on their structure and evolution. They provide the ability to identify stellar ejecta, study the SNR shock structure, and identify the associated compact objects, their spin periods and emission properties, as well as the structure of the nebulae they produce. This is resulting in ground-breaking advances in our understanding of stellar explosions and their aftermath.