The maximum energy of electrons in supernova remnant (SNR) shocks is typically limited by radiative losses, where the synchrotron cooling time equals the acceleration time. The low speed of shocks in a dense medium increases the acceleration time, leading to lower maximum electron energies and fainter X-ray emissions. However, in Kepler's SNR, an enhanced electron acceleration, which proceeds close to the Bohm limit, occurs in the north of its shell, where the shock is slowed by a dense circumstellar medium (CSM). To investigate whether this scenario still holds at smaller scales, we analyzed the temporal evolution of the X-ray synchrotron flux in filamentary structures, using the two deepest Chandra/ACIS X-ray observations, performed in 2006 and 2014. We examined spectra from different filaments, we measured their proper motion and calculated the acceleration to synchrotron time-scale ratios. The interaction with the turbulent and dense northern CSM induces competing effects on electron acceleration: on one hand, turbulence reduces the electron mean free path enhancing the acceleration efficiency, on the other hand, lower shock velocities increase the acceleration time-scale. In most filaments, these effects compensate each other, but in one region the acceleration time-scale exceeds the synchrotron time-scale, resulting in a significant decrease in nonthermal X-ray emission from 2006 to 2014, indicating fading synchrotron emission. Our findings provide a coherent understanding of the different regimes of electron acceleration observed in Kepler's SNR through various diagnostics.

Characterizing the ejecta in young supernova remnants is a requisite step towards a better understanding of stellar evolution. In Cassiopeia A the density and total mass remaining in the unshocked ejecta are important parameters for modeling its explosion and subsequent evolution. Low frequency (<100 MHz) radio observations of sufficient angular resolution offer a unique probe of unshocked ejecta revealed via free-free absorption against the synchrotron emitting shell. We have used the Very Large Array plus Pie Town Link extension to probe this cool, ionized absorber at 9 arcseconds and 18.5 arcseconds resolution at 74 MHz. Together with higher frequency data we estimate an electron density of 4.2 electrons per cubic centimeters and a total mass of 0.39 Solar masses with uncertainties of a factor of about 2. This is a significant improvement over the 100 electrons per cubic centimeter upper limit offered by infrared [S III] line ratios from the Spitzer Space Telescope. Our estimates are sensitive to a number of factors including temperature and geometry. However using reasonable values for each, our unshocked mass estimate agrees with predictions from dynamical models. We also consider the presence, or absence, of cold iron- and carbon-rich ejecta and how these affect our calculations. Finally we reconcile the intrinsic absorption from unshocked ejecta with the turnover in Cas A's integrated spectrum documented decades ago at much lower frequencies. These and other recent observations below 100 MHz confirm that spatially resolved thermal absorption, when extended to lower frequencies and higher resolution, will offer a powerful new tool for low frequency astrophysics.