# PACS Photometer Data release of the 'Mass-loss of Evolved StarS (MESS)' KPGT (Groenewegen et al.)

"The circumstellar environment in post-main-sequence objects"

Martin Groenewegen, Royal Observatory of Belgium, Brussels
, Mike Barlow, University College London, Franz Kerschbaum, University of Vienna, Joris Blommaert,
Instituut voor Sterrenkunde, K.U. Leuven, Pedro Garcia Lario, ESA-ESAC,
J. Cernicharo, CSIC, Madrid, Oliver Krause , Max-Planck-Institut fuer
Astronomie Heidelberg, Angela Baier, Univ. Vienna, Jeroen Bouwman,
Max-Planck-Institut fuer Astronomie, Martin Cohen, Berkeley, USA, Leen
Decin, K.U. Leuven, Thomas Henning, Max-Planck-Institut fuer
Astronomie, Damien Hutsemekers, IAGL, Liege, Rob Iveson, UK ATC, Royal
Observatory, Edinburgh, Djazia Ladjal, K.U. Leuven, Tanya Lim,
Rutherford Appleton Laboratory, Goeran Olofsson, Stockholm University,
Thomas Posch, Univ. Vienna, Gregor Rauw, IAGL, Pierre Royer,
K.U. Leuven, Bruce Sibthorpe, Cardiff University, Bruce Swinyard,
Rutherford Appleton Laboratory, Toshiya Ueta, Univ. Denver, Bart
Vandenbussche, K.U. Leuven, Griet Van de Steene, Royal Observatory of Belgium,
Brussels, Hans Van Winckel, K.U. Leuven, Eva Verdugo, ESA-ESAC,
Christoffel Waelkens, K.U. Leuven

The Proposal:
Mass-loss is one of the most fundamental properties of post-main sequence evolution. The mass-loss process leads to the formation of circumstellar shells containing dust and molecules. Although the mass-loss phenomenon has been studied since the 1960s, and important results have been obtained with the IRAS, ISO and Spitzer space missions, the details of the mass-loss process and the formation and evolution of the circumstellar shells are still not well understood. With its improved spatial resolution compared to ISO and Spitzer, better sensitivity, the extension to longer and unexplored wavelength regions, and medium resolution spectrometers, the combination of PACS and SPIRE observations will lead to a significant improvement in our understanding of the phenomena of mass-loss and dust formation. The main aims of this programme are three-fold: (1) to study the time dependence of the mass-loss process, via a search for shells and multiple shells around a wide range of evolved objects, in order to quantify the total amounts of mass lost at the various evolutionary stages of low to high-mass stars, (2) to study the dust and gas chemistry as a function of progenitor mass, and (3) to study the properties and asymmetries of evolved star envelopes. To this end, a sample of 103 Asymptotic Giant Branch and Red Super Giants, post-AGB and Planetary Nebulae, Luminous Blue Variables and Wolf-Rayet stars, and 5 Supernovae remnants will be imaged with PACS at 70+170 micron, and a sub-set of 32 stars will be imaged at all 3 wavelengths with SPIRE. In spectroscopy, a sample of 55 stars will be observed over the full wavelength range of PACS and, 23 stars will be observed with the SPIRE FTS. The sample of AGB stars has been selected to cover all chemical types (M-, S-, C-stars), variability types (irregular, semi-regular, Miras) and periods, and mass-loss rates. Stars have been selected to have high IRAS fluxes and low background levels. The spectroscopic targets are typically the brightest of the mapping targets.

Results:
A project overview of the MESS programme is given in the paper by
Groenewegen et al. (2011). First results were presented in the A&A
special edition (e.g., Kerschbaum et al. 2010), followed by dedicated
object studies such as, e.g., Decin et al. (2011). A morphological
overview and classification scheme of the AGB sub-sample was presented
by Cox et al. (2012). Detailed analyses on the different morphological
types of AGB-stars were conducted, focusing on binary-shaped
environments (e.g., Mayer et al. 2013), as well as on spherically
symmetric detached shell objects (Mecina et al. 2014). Studies on
post-AGB phases, such as planetary nebulae in the MESS sample can be
found in, e.g., van Hoof et al. (2013). Results on dust in young Supernova are discussed in e.g. Gomez et al. (2012).

The Data Processing:
This release contains PACS photometer data obtained in scanmap mode
within the MESS KPGT.  The data consist of typically a blue (70µm) and red (160µm) image. 
Several targets were also observed and are available in the green filter (100µm).

The data were processed using 
the latest version (12/2014) of the PACS photometer pipeline script up to level 1
in HIPE (version 13.0.3930). 
Additionally, we implemented a reconstructed astrometry, taking into account pointing
information from the gyroscopes (method provided by the 'Herschel
Pointing Working Group'). This mitigated some severe pointing errors,
responsible for smeared images, as well as improved the overall
astrometric quality. In addition, we were also able to merge
observations of the same target taken at different epochs and thus
combine all available data into one map.

The mapping was done following the JScanam routine within HIPE, which
is a implementation of Scanamorphos (Roussell 2013). In contrast to
the standard pipeline products, we projected our final maps with a
resolution of 1 arcsec/pixel and 2 arcsec/pixel for the blue/green and
red bands, respectively, thus oversampling the detector pixels by a
factor of 3.2. This allowed us to better assess the fine details
present in our extended structures. Further, the drizzling was set to a pixel fraction of 0.1 for each detector. Remaining parameters were left as set by default.

Despite the inferior background rendering at the map borders, we chose
JScanam over the inverse methods (MADMAP, UNIMAP), because with the
latter we were not able to obtain completely artefact-free (crosses on
brights point sources) results (see, Ottensamer et al. 2011; Mecina et al. 2014b).

As several targets were observed twice, i.e., having four OBSIDs, we
also merged all available datasets of each source to get the best
signal-to-noise ratio possible and to mitigate potential scanning artefacts. These combined maps are also available.

In some cases, e.g., binary dominated morphology (Mayer et al. 2013, 2014),  deconvolution was applied on the final maps to better disentangle real structures at the resolution limit from artefacts due to the complex point spread function. For this task a maximum entropy method was used, similar to the one described by Hollis et al. (1992). As the input PSF model we used the maps of AFGL 190, reduced with the JScanam task. All steps were conducted within HIPE. The further processed images are also available within this release. 

References:
Cox et al. 2012 A&A...537A..35C
Decin et al. 2011 A&A...534A...1D
Gomez et al. 2012 ApJ...760...96G
Groenewegen et al. 2011 A&A...526A.162G
Hollis et al. 1992 ApJ...386...293H
van Hoof et al. 2013 A&A...560A...7V
Kerschbaum et al. 2010 A&A...518L.140K
Mayer et al. 2013 A&A...549A..69M
Mayer et al. 2014 A&A...570A.113M
Mecina et al. 2014a A&A...566A..69M
Mecina et al. 2014b SPIE.9152E..2TM
Ottensamer et al. 2011 ASPC..445..625O
Roussel 2013 PASP..125.1126R