Search

Understanding morphological types of galaxies is a key parameter for studying their formation and evolution. Neural networks that have been used previously for galaxy morphology classification have some disadvantages, such as not being inherently invariant under rotation. In this work, we studied the performance of Capsule Network (CapsNet), a recently introduced neural network architecture that is rotationally invariant and spatially aware, on the task of galaxy morphology classification. We designed two evaluation scenarios based on the answers from the question tree in the Galaxy Zoo project. In the first scenario, we used CapsNet for regression and predicted probabilities for all of the questions. In the second scenario, we chose the answer to the first morphology question that had the highest user agreement as the class of the object and trained a CapsNet classifier, where we also reconstructed galaxy images. We achieved promising results in both of these scenarios. Automated approaches such as the one introduced here will play a critical role in the upcoming large sky surveys. [ABSTRACT FROM AUTHOR]

We present high signal-to-noise ratio Gemini and MMT spectroscopy of the optical afterglow of the gamma-ray burst (GRB) 130606A at redshift z = 5.913, discovered by Swift. This is the first high-redshift GRB afterglow to have spectra of comparable quality to those of z ≈ 6 quasars. The data exhibit a smooth continuum at near-infrared wavelengths that is sharply cut off blueward of 8410Å due to absorption from Lyα at redshift z ≈ 5.91, with some flux transmitted through the Lyα forest between 7000 and 7800 Å. We use column densities inferred from metal absorption lines to constrain the metallicity of the host galaxy between a lower limit of [Si/H] ≳ -1.7 and an upper limit of [S/H] ≲ -0.5 set by the non-detection of S ii absorption. We demonstrate consistency between the dramatic evolution in the transmission fraction of Lyα seen in this spectrum over the redshift range z = 4.9-5.85 with that previously measured from observations of high-redshift quasars. There is an extended redshift interval of Δz = 0.12 in the Lyα forest at z = 5.77 with no detected transmission, leading to a 3σ upper limit on the mean Lyα transmission fraction of ≲0.2% (or τ eff GP (Lyα) > 6.4). This is comparable to the lowest-redshift Gunn-Peterson troughs found in quasar spectra. Some Lyβ and Lyγ transmission is detected in this redshift window, indicating that it is not completely opaque, and hence that the intergalactic medium (IGM) is nonetheless mostly ionized at these redshifts. We set a 2σ upper limit of 0.11 on the neutral fraction of the IGM at the redshift of the GRB from the lack of a Lyα red damping wing, assuming a model with a constant neutral density. GRB 130606A thus for the first time realizes the promise of GRBs as probes of the first galaxies and cosmic reionization. [ABSTRACT FROM AUTHOR]

For gamma-ray burst 100901A, we have obtained Gemini-North and Very Large Telescope optical afterglow spectra at four epochs: 1 h, 1 d, 3 d and 1 week after the burst, thanks to the afterglow remaining unusually bright at late times. Apart from a wealth of metal resonance lines, we also detect lines arising from fine-structure levels of the ground state of Fe II, and from metastable levels of Fe II and Ni II at the host redshift (z = 1.4084). These lines are found to vary significantly in time. The combination of the data and modelling results shows that we detect the fall of the Ni II 4F9/2 metastable level population, which to date has not been observed. Assuming that the population of the excited states is due to the ultraviolet radiation of the afterglow, we estimate an absorber distance of a few hundred pc. This appears to be a typical value when compared to similar studies. We detect two intervening absorbers (z = 1.3147, 1.3179). Despite the wide temporal range of the data, we do not see significant variation in the absorption lines of these two intervening systems. [ABSTRACT FROM AUTHOR]

The recent long-duration GRB 100316D, associated with supernova SN 2010bh and detected by Swift, is one of the nearest gamma-ray burst (GRB)-supernovae (SNe) ever observed (z = 0.059). This provides us with a unique opportunity to study the explosion environment on ~kpc scale in relation to the host galaxy complex. Here we present spatially resolved spectrophotometry of the host galaxy, focusing on both the explosion site and the brightest star-forming regions. Using these data, we extract the spatial profiles of the relevant emission features (Hα, Hβ, [O III])λ5007, and [N II]λ6584) and use these profiles to examine variations in metallicity and star formation rate (SFR) as a function of position in the host galaxy. We conclude that GRB 100316D/SN2010bh occurred in a low-metallicity host galaxy, and that the GRB-SN explosion site corresponds to the region with the lowest metallicity and highest SFR sampled by our observations. [ABSTRACT FROM AUTHOR]

We comment on recent spectropolarimetric studies that compare the observed polarization position angles (PAs) of type 1 Seyfert galaxies near Hα with the observed orientations of their radio-source axes on the sky. For a Seyfert galaxy in which scattering occurs mainly in an equatorial scattering region, the polarization PA is expected to be parallel to the radio axis, while for a case in which light scatters predominantly in the polar regions, the Hα polarization PA should be perpendicular to the radio axis. In practice, these correlations are difficult to establish because a Galactic interstellar polarization contribution can introduce a significant uncertainty into the polarization PA determination, even when the magnitude of interstellar polarization is small. We show how such uncertainties may affect the analysis of PA alignments, and present spectropolarimetric observations of a probe star along the line of sight to the type 1 Seyfert galaxy Mrk 871 that allow us to assess the intrinsic Hα polarization and PA of Mrk 871. These results suggest that spectropolarimetric observations of such probe stars should form an integral part of future polarization studies of Seyfert galaxies. [ABSTRACT FROM AUTHOR]

SN 2002es is a peculiar subluminous Type Ia supernova (SN Ia) with a combination of observed characteristics never before seen in an SN Ia. At maximum light, SN 2002es shares spectroscopic properties with the underluminous SN 1991bg subclass of SNe Ia, but with substantially lower expansion velocities (∼6000 km s–1) more typical of the peculiar SN 2002cx subclass. Photometrically, SN 2002es differs from both SN 1991bg-like and SN 2002cx-like supernovae. Although at maximum light it is subluminous (MB = –17.78 mag), SN 2002es has a relatively broad light curve (Δm15(B) = 1.28 ± 0.04 mag), making it a significant outlier in the light-curve width versus luminosity relationship. We estimate a 56Ni mass of 0.17 ± 0.05 M☼ synthesized in the explosion, relatively low for an SN Ia. One month after maximum light, we find an unexpected plummet in the bolometric luminosity. The late-time decay of the light curves is inconsistent with our estimated 56Ni mass, indicating that either the light curve was not completely powered by 56Ni decay or the ejecta became optically thin to γ-rays within a month after maximum light. The host galaxy is classified as an S0 galaxy with little to no star formation, indicating that the progenitor of SN 2002es is likely from an old stellar population. We also present a less extensive data set for SN 1999bh, an object which shares similar photometric and spectroscopic properties. Both objects were found as part of the Lick Observatory Supernova Search, allowing us to estimate that these objects should account for 2.5% of SNe Ia within a fixed volume. Current theoretical models are unable to explain the observed characteristics of SN 2002es. [ABSTRACT FROM AUTHOR]