Automated x-ray spectral classification of young stellar sources detected in Chandra Orion Ultradeep Project observations

Abstract

We are developing an automated classification technique to group X-ray sources detected in CCD imaging into clusters based on X-ray spectral attributes. The algorithm is non-parametric and uses established multivariate statistics and pattern recognition techniques. We have applied this spectral classification technique to a sample of sources selected from among the 1616 X-ray sources detected in Chandra Orion Ultradeep Project (COUP) imaging. The COUP dataset was compiled from an 850 ks Chandra observation of the Orion Nebula Cluster (ONC) that was obtained in January 2003 with the Advanced CCD Imaging Spectrometer (ACIS), and represents by far the most sensitive and comprehensive description of X-ray emission from a pre-main sequence star cluster (Getman et al. 2005, ApJS, in press; astro-ph/0410136). We describe preliminary classification results for a subset of approximately 440 COUP sources for which high-quality ACIS spectra can be extracted. These results demonstrate that our spectral clustering technique can be used to efficiently identify pre-main sequence X-ray sources that lack optical or near-infrared counterparts and that display strong Fe Kα line emission and/or large-amplitude, impulsive flares. Within the COUP dataset, such sources likely represent the youngest protostars in the ONC.