X-ray data of the Centaurus cluster, obtained with {it XMM-Newton} for 45 ksec, were analyzed. Deprojected EPIC spectra from concentric thin shell regions were reproduced equally well by a single-phase plasma emission model, or by a two-phase model d
eveloped by {it ASCA}, both incorporating cool (1.7--2.0 keV) and hot ($sim 4$ keV) plasma temperatures. However, EPIC spectra with higher statistics, accumulated over 3-dimentional thick shell regions, were reproduced better by the two-phase model than by the singe-phase one. Therefore, hot and cool plasma phases are inferred to co-exist in the cluster core region within $sim 70$ kpc. The iron and silicon abundances of the plasma were reconfirmed to increase significantly towards the center, while that of oxygen was consistent with being radially constant. The implied non-solar abundance ratios explains away the previously reported excess X-ray absorption from the central region. Although an additional cool ($sim 0.7$ keV) emission was detected within $sim 20$ kpc of the center, the RGS data gave tight upper limits on any emission with a tempeartures below $sim 0.5$ keV. These results are compiled into a magnetosphere model, which interprets the cool phase as confined within closed magnetic loops anchored to the cD galaxy. When combined with so-called Rosner-Tucker-Vaiana mechanism which applies to solar coronae, this model can potentially explain basic properties of the cool phase, including its temperature and thermal stability.
