... CIAO

Chandra Interactive Analysis of Observations. http://cxc.harvard.edu/ciao/

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... data

And from XMM2, we only make use of the EPIC pn data, since it has a better calibration below 0.8 keV than the EPIC MOS (see http://xmm.vilspa.esa.es/docs/documents/CAL-TN-0018-2-4.pdf). No significant differences are seen between MOS and pn data, and the conclusions of the analysis can be easily applied to the MOS data as well.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... dB.

Decibans (tenths of a power of 10), is a common unit to represent weights of evidence.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... XSTAR

Version 2.1kn6. See http://heasarc.gsfc.nasa.gov/docs/software/xstar/xstar.html.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... abundance

Atomic abundances are entered relative to solar abundances asdefined in Grevesse et al. (1996), with 1.0 being defined as the solar.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... unacceptable

$P_{289,339.7}[{\rm Model~A};{\it XMM}$- ${\it Newton}]=2 \times 10^{-2}$ and $P_{183,241.3}[{\rm Model~A};{\it Chandra}]=2 \times 10^{-3}$, where $P_{\nu,\chi^2}$ is the probability of exceeding $\chi ^2$ for $\nu$ degrees of freedom.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... spectra

i.e., not taking into account radiative transfer effects.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... data

We made the analysis of best-fit velocity by inspecting the evaluation of $\chi ^2$ at every outflow velocity point of our grid of velocities, while the other parameters of interest are varied as usual. Then, we proceed to adopt the velocity-model with the minimum $\chi ^2$.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... component)

This is the best high-velocity component we found able to fit both sets of data with high $\chi ^2$-probability.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... fluxes

For completeness we have computed the observed flux of the $\sim$16 ks XMM-Newton observation (XMM1) and the result is: $F_{(0.2-10~{\rm keV})}[{\rm XMM1}]=9.6\pm 0.6 \times 10^{-13}$ erg cm^-2 ${\rm s}^{-1}$.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... band

We verified that the data is sensitive to an increase of oxygen abundance. We take the best-fit two-absorbers model and compute models with the oxygen abundance at 1.5, 3 and 5 $\times$ solar oxygen abundance. The fit get worse with $\chi^2({\rm O=1.5})=304$, $\chi^2({\rm O=3})=333$ and $\chi^2({\rm O=5})=388$. The conclusions are: 1) the data is sensitive to the Fe/O ratio. 2) This ratio must be $\approx$$3 \pm 1$, in order to produce acceptable fits (and produce good description of the low-energy band).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... luminosity

This is the unabsorbed intrinsic (1-50) keV rest-frame luminosity of the source. From the XMM-Newton data $L_{\rm x}({\it XMM}$- ${\it Newton})=1.07(\pm0.03)\times 10^{47}k^{-1}$ erg  ${\rm s}^{-1}$, from the Chandra data $L_{\rm x}({\it Chandra})=1.17(\pm0.05)\times 10^{47}k^{-1}$ erg  ${\rm s}^{-1}$. We are taking the average of both.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.