Astronomy & Astrophysics

All Figures

$\begin{figure} \par\includegraphics[angle=-90,width=8.8cm,clip]{0421_f1.ps} \end{figure}$ Figure 1: Light curve of a typical sky field observation. Events are extracted from the in FOV region in the energy range 0.5-12 keV; time bin is 30 s. The second part of the observation is affected by intense soft proton flares, the peak count rate is more than 200 times higher than the quiescent one. The selected threshold, in units of sigma of the quiescent count rate distribution (see text and Fig. 2), is marked with a dashed line.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[angle=-90,width=8.8cm,clip]{0421_f2.ps} \end{figure}$ Figure 2: Histogram of the count rate distribution for a typical sky field observation. The corresponding light curve is shown in Fig. 1. The peak corresponds to the quiescent count rate Poisson distribution; points falling to the right correspond to the soft proton flares. The GTI threshold, in units of sigma of the quiescent count rate distribution, is marked with a dashed line.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=8.8cm,clip]{0421_f3.eps} \end{figure}$ Figure 3: The cosmic X-ray background spectrum in the 2-8 keV range is displayed, folded with the instrumental response. MOS1 data are represented in grey, MOS2 in black. The best fit model is overplotted. The lower panel shows the residuals in units of sigma.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[angle=-90,width=8.8cm,clip]{0421_f4.ps} \end{figure}$ Figure 4: The best fit photon index of the CXB is plotted as a function of the selected GTI threshold (in units of sigma of the quiescent count rate distribution - see Sect. 3.2). Errors (at 90% confidence level for a single interesting parameter) include the uncertainty on the renormalization of the quiescent NXB spectrum. The adopted 3.3 $\sigma$ threshold is marked by the vertical solid line. No significant variations in the CXB photon index are obtained as a result of different choices of the GTI threshold within the range $\sim$ $2.9{-}3.8\sigma$ , marked by the dotted vertical lines.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[angle=-90,width=8.8cm,clip]{0421_f5.ps} \end{figure}$ Figure 5: Same as Fig. 4, the best fit CXB normalization is shown. Units are photons cm^-2 s^-1 sr^-1 keV^-1 at 3 keV. The contributions from out-of-field scattered light have not been subtracted. Errors (at 90% confidence level for a single interesting parameter) include the uncertainty on the renormalization of the quiescent NXB spectrum. The adopted GTI threshold is marked by the vertical dashed line. It is evident that no significant variations are obtained as a result of small changes of the selected GTI threshold (within the range $\sim$ $2.9{-}3.8\sigma$ , marked by the dotted vertical lines). See text for further details.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[angle=-90,width=8.8cm,clip]{0421_f6.ps} \end{figure}$ Figure 6: Best fit CXB photon index as a function of the maximum allowed ratio for the 8-12 keV surface brightness $\Sigma _{{\rm IN}}/\Sigma _{{\rm OUT}}$ , a probe of the presence of residual soft proton NXB. The MOS1 case is shown (MOS2 is very similar). Errors (at 90% confidence level for a single interesting parameter) include the uncertainty on the renormalization of the quiescent NXB spectrum. No significant changes are obtained. The case of the CXB normalization is markedly different (see Fig. 7).
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[angle=-90,width=8.8cm,clip]{0421_f7.ps} \end{figure}$ Figure 7: Same as Fig. 6, the best fit CXB normalization (in units of photons cm^-2 s^-1 sr^-1 keV^-1 at 3 keV) is plotted here. The stray light correction have not been applied. Errors (at 90% confidence level for a single interesting parameter) include the uncertainty on the renormalization of the quiescent NXB spectrum. When dataset having large values of the ratio $\Sigma _{{\rm IN}}/\Sigma _{{\rm OUT}}$ are included, the normalization is systematically higher owing to the presence of a significant residual SP NXB component (see text). The adopted maximum allowed ratio ( $R_{{\rm MAX}}=1.3$ ) is marked by the vertical line. No significant changes are obtained within small changes of $R_{{\rm MAX}}$ . See Appendix B for further details.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[angle=-90,width=8.8cm,clip]{0421_f8.ps} \end{figure}$ Figure 8: Best fit CXB photon index as a function of the off-axis angle. Each point represents an independent measure obtained from the analysis of a selected portion of the detector plane (see text). Errors (at 90% confidence level for a single interesting parameter) include the uncertainty on the renormalization of the quiescent NXB spectrum. The MOS1 case is shown (MOS2 is very similar). The constant trend is a result of the correct vignetting function calibration.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[angle=-90,width=8.8cm,clip]{0421_f9.ps} \end{figure}$ Figure 9: Same as Fig. 8, the case of the CXB normalization is presented here. Units are photons cm^-2 s^-1 sr^-1 at 3 keV. The corrections for the light coming from out-of-field angles have not been applied. Errors (at 90% confidence level for a single interesting parameter) include the uncertainty on the renormalization of the quiescent NXB spectrum. The variations are always smaller than the uncertainty, confirming the overall correctness of the vignetting curve calibration.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[angle=-90,width=8.8cm,clip]{0421_f10.ps} \end{figure}$ Figure 10: CXB intensity measurements. The flux in the 2-10 keV band is represented as a function of the epoch of the experiment. The plot is an update of Fig. 3 of Moretti et al. (2003) including the results of the present work. From left to right the CXB values are from Marshall et al. (1980) with HEAO-1 data; McCammon et al. (1983) with a rocket measurement; Gendreau et al. (1995) with ASCA SIS data; Miyaji et al. (1998) with ASCA GIS; Ueda et al. (1999) with ASCA GIS and SIS; Vecchi et al. (1999) with BeppoSAX LECS/MECS; Lumb et al. (2002) with XMM-Newton EPIC/MOS; Kushino et al. (2002) with ASCA GIS. Finally, the hollow star mark our own CXB measurement with the EPIC MOS cameras onboard XMM-Newton. All the uncertainties are at 1 $\sigma$ level. The horizontal dotted lines mark the range of CXB intensity constrained by our measurement after including an extra 3.2% uncertainty (1 $\sigma$ confidence level, rescaled from the 5% uncertainty quoted in the text, which corresponds to 90% confidence level) on the absolute flux calibration of the EPIC cameras (it is indeed a measure of the cross-calibration accuracy among different instruments). Such range represents the most robust estimate of the true absolute sky surface brightness in the 2-8 keV band.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[angle=-90,width=8.8cm,clip]{0421_f11.ps} \end{figure}$ Figure A.1: Quiescent NXB spectrum for the MOS1 camera. From top to bottom: data from closed observations (black), in FOV region; from the sky merged dataset (red), out FOV region; from closed obs.(green), out FOV region. The closed out FOV spectrum has been rescaled by a factor of 3 to ease its visibility; no renormalization was performed on the other spectra. Total exposure time is of $\sim$ 430 ks for the closed spectra, $\sim$ 1.15 Ms for the sky data. The labels identify the strongest internal fluorescence lines. Significant differences in the fluorescence lines intensities are evident between the in FOV and the out FOV spectra.

In the text

$\begin{figure} \par\includegraphics[angle=-90,width=8.8cm,clip]{0421_f12.ps}\par\end{figure}$ Figure A.2: Ratio between the spectra extracted from the out FOV region for the sky merged dataset and for the closed dataset. Data are from the MOS1 camera; the MOS2 case is very similar. A linear fit to the range 2-12 keV is fully consistent with a constant.

In the text

$\begin{figure} \par\includegraphics[width=8.8cm,clip]{0421_f13.ps} \end{figure}$ Figure A.3: Secular variation of the surface brightness (units of counts cm^-2 s^-1) in the 8-12 keV range for MOS1 (red points) and MOS2 (black points): out FOV regions in sky fields observations (empty squares) and in FOV region in closed observation (filled squares). A smooth trend having a broad minimum around revolution $\sim$ 150 is seen; a high scattering, due to different particle radiation conditions, is also evident.

In the text

$\begin{figure} \par\includegraphics[width=8.8cm,clip]{0421_f14.ps} \end{figure}$ Figure A.4: The correlation between the surface brightness (units of counts cm^-2 s^-1) in the 1.4-1.6 keV range (dominated by the internal Al-K fluorescence line - see Fig. A.1) and the surface brightness in the range 2.5-5 keV (continuum). Data are for closed observations; MOS1 red and MOS2 black. A rather high scatter is evident.

In the text

$\begin{figure} \par\includegraphics[width=8.8cm,clip]{0421_f15.ps} \end{figure}$ Figure A.5: Same as Fig. A.4; the surface brightness (units of counts cm^-2 s^-1) in the energy ranges (2.5-5 keV) and (8-12 keV) show a better correlation.

In the text

$\begin{figure} \par\includegraphics[angle=-90,width=8.8cm,clip]{0421_f16.ps} \end{figure}$ Figure A.6: Ratio of the spectra from the "high intensity'' and "low intensity'' samples of closed observations (see text) for the MOS1 camera. The MOS2 case is very similar. Although deviations are seen in correspondence of the fluorescence lines, the overall shape in the 2-12 keV range is clearly constant: the shape of the quiescent NXB spectrum is stable in spite of high variations in the flux of high energy particles.

In the text

$\begin{figure} \par\includegraphics[width=8.8cm,clip]{0421_f17.ps} \end{figure}$ Figure B.1: The correlation between the 8-12 keV surface brightness (counts cm^-2 s^-1) in FOV and out FOV is shown for the sky fields observations. Red and black points represent MOS1 and MOS2 observations, respectively. A high scatter is clearly evident.

In the text

$\begin{figure} \par\includegraphics[width=8.8cm,clip]{0421_f18.ps} \end{figure}$ Figure B.2: Same of Fig. B.1 for the case of the closed observations. The correlation is much better.

In the text

$\begin{figure} \par\includegraphics[width=8.8cm,clip]{0421_f19.ps} \end{figure}$ Figure B.3: The ratio of the 8-12 keV surface brightness ( in FOV)/( out FOV) is plotted as a function of the surface brightness in FOV (counts cm^-2 s^-1). Data are from sky fields observations, red points correspond to MOS1 and black points to MOS2. A definite correlation is absolutely evident.

In the text

$\begin{figure} \includegraphics[width=8.8cm,clip]{0421_f20.ps} \end{figure}$ Figure B.4: The ratio $\Sigma _{{\rm IN}}/\Sigma _{{\rm OUT}}$ (8-12 keV) is plotted here as a function of $\Sigma _{{\rm OUT}}$ (counts cm^-2 s^-1). No correlation can be seen in this case, markedly different from the result presented in Fig. B.3.

In the text

$\begin{figure} \par\includegraphics[angle=-90,width=8.8cm,clip]{0421_f21.ps} \end{figure}$ Figure B.5: The light curve of this observation looks "typical'' at first glance, with intense soft proton flares superposed on a quiescent count rate. However it is found that the quiescent count rate is much higher than the value expected for a typical blank sky field (marked by the dashed line) and shows a slow modulation. The whole observation is affected by a "long flare'' of low-energy particles; the analysis of the 8-12 keV surface brightness after GTI screening clearly indicates an anomalous ratio $\Sigma _{{\rm IN}}/\Sigma _{{\rm OUT}}$ .

In the text

$\begin{figure} \par\includegraphics[angle=-90,width=8.8cm,clip]{0421_f22.ps}\end{figure}$ Figure B.6: Light curve of an observation badly affected by soft proton flares during all of its duration. The expected count rate for a typical sky field is marked by the dashed line. The automated GTI filtering algorithm used in our study cannot properly work in such conditions (see text).

In the text

$\begin{figure} \par\includegraphics[angle=-90,width=8.8cm,clip]{0421_f23.ps} \end{figure}$ Figure B.7: The ratio of the 8-12 keV surface brightness $\Sigma _{{\rm IN}}/\Sigma _{{\rm OUT}}$ is plotted as a function of $\Sigma _{{\rm IN}}$ (counts cm^-2 s^-1) for the sky fields (blue points) and for the closed observations (green points). A linear fit to the two distributions is overplotted. It is evident that the sky fields observations show higher values of $\Sigma _{{\rm IN}}/\Sigma _{{\rm OUT}}$ even in the less contaminated cases.

In the text

$\begin{figure} \par\includegraphics[angle=-90,width=8.8cm,clip]{0421_f24.ps} \end{figure}$ Figure B.8: Spectra extracted from the sky fields observations having different values of the ratio R of surface brightness $\Sigma _{{\rm IN}}/\Sigma _{{\rm OUT}}$ : R < 1.05 (green - dataset "a''), 1.05 < R < 1.30 (red - dataset "b''), R > 1.30 (black - dataset "c'').

In the text

$\begin{figure} \par\includegraphics[angle=-90,width=8.8cm,clip]{0421_f25.ps} \end{figure}$ Figure B.9: Best fit CXB parameters computed from the three independent dataset corresponding to different values of $\Sigma _{{\rm IN}}/\Sigma _{{\rm OUT}}$ (see text and Fig. B.8 for the color code. The normalization is in photons cm^-2 s^-1 sr^-1 keV^-1 at 3 keV. The stray light contribution have not been subtracted. Errors (at 90% confidence level for a single interesting parameter) include the uncertainty on the renormalization of the quiescent NXB spectrum. The normalization and the photon index computed for the dataset c) are uncompatible with the values for dataset a) and b), which are found, conversely, to be in good agreement.

In the text

$\begin{figure} \par\includegraphics[angle=-90,width=8.8cm,clip]{0421_f2.ps} \end{figure}$	Figure 2: Histogram of the count rate distribution for a typical sky field observation. The corresponding light curve is shown in Fig. 1. The peak corresponds to the quiescent count rate Poisson distribution; points falling to the right correspond to the soft proton flares. The GTI threshold, in units of sigma of the quiescent count rate distribution, is marked with a dashed line.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=8.8cm,clip]{0421_f3.eps} \end{figure}$	Figure 3: The cosmic X-ray background spectrum in the 2-8 keV range is displayed, folded with the instrumental response. MOS1 data are represented in grey, MOS2 in black. The best fit model is overplotted. The lower panel shows the residuals in units of sigma.
Open with DEXTER