Документ взят из кэша поисковой машины. Адрес оригинального документа : http://xmm.vilspa.esa.es/docs/documents/CAL-TN-0096.pdf Дата изменения: Wed Dec 19 14:16:13 2012 Дата индексирования: Mon Feb 4 08:16:14 2013 Кодировка:

XMM-Newton Calibration Technical Note
XMM-SOC-CAL-TN-0096 The stability of the EPIC-pn camera
Matteo Guainazzi, Martin Stuhlinger, Nathan Dickinson, Andrew Pollo ck (ESA-ESAC, Vil lafranca del Castil lo, Spain) November 23, 2012

History
Version 1.0 Date Novemb er 23, 2012 Editor M.Guainazzi Note First public version

1


XMM-Newton Technical Note

XMM-SOC-CAL-TN-0096

Page:

2

1

Scop e

In this rep ort we investigate the stability of the EPIC-pn camera on b oard XMM-Newton (Struder Ё et al., 2001). "Stability" as measured in this document reflects all and each of optics plus detector system elements, and their calibration: effective area, filter transmission, quantum efficiency, and redistribution. This document discusses on-axis (within 15") observations only.

2

RXJ1856-3754

RXJ1856.6-3754 is one of the seven known isolated neutron stars (INSs) identified in the ROSAT All Sky Survey (Treves et al. 2000). Their X-ray sp ectrum is extremely soft, and can b e generally well modelled by a blackb ody with temp eratures kT50-100 eV (Burwitz et al. 2003, 2001; Trump er et Ё al. 2004), although one might exp ect deviations from a simple blackb ody shap e due to atmospheres of either heavy elements (due to debris from the progenitor) or light elements (gravitational setting or accretion; Ho et al. 2007). Interestingly enough, RXJ1856.6-3754 lacks broad sp ectral atmospheric features seen in other INSs such as RXJ0720.4-3125 (Hab erl et al. 2004; Kaplan & van Kerkwijk 2005; Hambaryan et al. 2009). Starting from 2004, RXJ1856-3754 has b een observed approximately twice p er year in the framework of the XMM-Newton Routine Calibration Plan (Guainazzi 2012). Tab. 1 rep orts the list of Table 1: EPIC-pn configuration for the observations discussed in this document Obs.# 0106260101 0165971601 0165971901 0165972001 0165972101 0412600101 0412600201 0415180101 0412600301 0412600401 0412600601 0412600701 0412600801 0412600901 0412601101 0412601301 0412601401 0412602301 Date (YYYY-MM-DDTHH:MM:SS) 2002-04-08T16:21:27 2004-09-24T01:42:13 2005-03-23T08:34:42 2005-09-24T07:58:13 2006-03-26T15:40:29 2006-10-24T00:33:51 2007-03-14T20:50:01 2007-03-25T05:36:47 2007-10-04T05:48:49 2008-03-13T18:49:45 2008-10-05T01:00:58 2009-03-19T21:30:04 2009-10-07T15:20:44 2010-03-22T02:50:25 2010-09-28T23:09:10 2011-03-14T00:47:48 2012-04-13T07:12:25 2012-09-20T11:23:08 RAWX 38 38 37 38 37 38 38 36 38 37 39 36 38 37 36 37 36 40 RAWY 191 190 191 191 191 190 191 168 190 191 190 191 190 191 192 179 179 189 T exp (ks) 39 22 16 22 48 50 21 17 14 33 43 47 38 50 47 48 31 50 P H A 1 -1 -1 2 4 3 -3 1 -1 3 0 4 1 3 4 1 3 0


XMM-Newton Technical Note

XMM-SOC-CAL-TN-0096

Page:

3

observations discussed in this document together with their instrumental configuration. All observations discussed in this document were taken in Smal l Window (SW) with the Thin optical blocking filter.

2.1

Work published on the EPIC-pn stability, and our approach

Sartore et al.(2012) discuss the stability of the EPIC-pn camera using the same source (and most of the data) used in this document. Employing standard sp ectral fitting techniques, they conclude that no evidence of sp ectral or flux variations is detected from March 2005 to Octob er 2011, once observations taken at the same detector p osition are considered. The 3 upp er limit on the 0.15-1.2 keV flux variation is 3%. In this document, we present the result of a study aiming at the same goal. We follow a complementary approach based on the temp oral evolution of background-subtracted count rates, to avoid astrophysical uncertainties in the conversion b etween instrumental counts and fluxes. Moreover, we apply a self-calibration of the energy reconstruction in PHA space (see Sect. 4, which allows us to make use of all on-axis data, notwithstanding the source p osition in detector coordinates .

3

Sp ectral fitting

In this Section we present for completeness the results of a standard sp ectral fitting. Data reduction followed the procedure describ ed in Stuhlinger et al. (2011). Sp ectra were extracted from 37.5" radius circular regions around the X-ray source centroid. Background sp ectra were extracted from source-free b oxes at the edge of the SW reading area. The sp ectrum of each camera and observation was fit in the 0.15-0.80 keV1 energy band with a model constituted by a single blackb ody comp onent seen through a screen of neutral photoelectric absorption (in compact XSPEC jargon: tbabs*bbody), with all the parameters free to vary. While more complex models could b e astrophysically justified (Burwitz et al. 2003), no consensus currently exists on the correct astrophysical model for RXJ18563754. Fig. 1 shows sp ectra and residuals against the b est-fit model for all observations in Tab. 1. Readers are referred to Sartore et al. (2012) for a discussion of the sp ectral analysis results.

4

Count rate stability

Small column-by-column variations of the instrumental gain can substantially affect the measured count rates. Simulations indicate that a shift of ±5 eV translates into a ±5% systematic error on the background subtracted count rate in the 0.15-0.8 keV band. In order to minimise this effect, we applied a relative calibration method to all the RXJ1856.63754 sp ectra discussed in this document. We have chosen the sp ectrum with the longest exp oWe intentionally here use an energy range extending at energies lower than the nominal recommended bandpass for illustration purp oses.
1


XMM-Newton Technical Note

XMM-SOC-CAL-TN-0096

Page:

4

RXJ1856.6-3754 - EPN

10 Cts/s/keV

1

Rev.#1883 Rev.#1800 Rev.#1699 Rev.#1513 Rev.#1432 Rev.#1330 Rev.#1259 Rev.#1153 Rev.#1061 Rev.#968 Rev.#427

1.2 1.1 1 0.9 0.8 0.2 Energy (keV) 0.5

Figure 1: Sp ectra (upper panel) and residuals in units of data/model ratio (lower panel) for all the EPIC-pn observations of RXJ1856.6-3754 discussed in this document. sure time (Obs.#041260010, R(P H A)) as reference. For each of the other sp ectra S (P H A), we calculated the sp ectral shift P H A which minimises the quantity: i [Si (P H A + P H A) - Ri (P H A)]2 /Ri (P H A). In Tab. 1 we rep ort the P H A values applied to each observation. In Fig. 2 we show the distribution of the background-subtracted count rates in the PHA range b etween 40 and 200 (approximately corresp onding to the nominal 0.2-0.8 keV energy band). The standard deviation of the distribution is 1.70% once normalized to its average (2.0% if no gain selfcorrection is applied). There is a marginal correlation b etween count rates and time (cf. Fig. 3; R(t)=(5.584 ± 0.011) - [(1.9 ± 0.2) в 10-6 ]t). However, the scatter of the data p oints is dominated by a time-indep endent comp onent. The relative energy-scale self-calibration removes any dep endence of the count rate on the p osition of the source centroid in detector coordinates (Fig. 4). On the other hand, the count rate during each of the observations is constant within statistics. We extracted source and background light curves from the same regions as background/source sp ectra with a binning time of 100 s. In Fig. 5 we show the distribution of the differences b etween the background subtracted count rates and their average during the observation, normalised to their statistical error of each measurement. The distributions are consistent with the statistical fluctuations.

Data/model


XMM-Newton Technical Note

XMM-SOC-CAL-TN-0096

Page:

5

Figure 2: Distribution of the background-subtracted, time-averaged Count Rates (C R) in the 40­ 200 PHA range for all the on-axis EPIC-pn exp osures of RXJ1856.6-3754 p erformed so far. The inset indicates the average, standard deviation, and standard deviation normalised to the mean of the distribution. The energy-scale self-corrected EPIC-pn count rates are correlated with the normalisation of the simultaneous RGS exp osures (Fig. 6). The correlation is mainly driven by the single data p oint with the largest count rate/normalisations. However, the data p oints are affected by a scatter larger than the statistical error bars. It is still unknown whether this scatter is due to instability of the EPIC-pn effective area or energy reconstruction, or to residual mis-calibration of the RGS contamination time evolution, or b oth.

5

Conclusions

We provide in this document a metric of the stability of the EPIC-pn camera, using the observations of the isolated neutron star RXJ1856.6-3754 p erformed so far (Tab. 1). We use the count rate in a PHA range nominally corresp onding to the 0.2­0.8 keV energy range. After applying a relative self-calibration of the energy scale, to minimise any effect due to uncertainties in the column-bycolumn gain calibration, the normalised standard deviation of the count rate distribution is 1.70% (Fig. 2). The scatter or the data p oints in the count rate versus time plane is dominated by a


XMM-Newton Technical Note

XMM-SOC-CAL-TN-0096

Page:

6

Figure 3: Background-subtracted, time-averaged C R in the 40­200 PHA range as a function of observation start time. The dotted line represent the b est linear fit, the dashed line the envelop e of the linear relation corresp onding to a ±1 uncertainty on the b est-fit parameters. The error bars indicate the Poissonian error calculated according to the Gehrels (1986) prescription. time-indep endent comp onent.


XMM-Newton Technical Note

XMM-SOC-CAL-TN-0096

Page:

7

Figure 4: Upper panel: background-subtracted, time-averaged C R in the 40­200 PHA range as a function of the b est-fit source centroid column (RAWX) for all the on-axis EPIC-pn exp osures of RXJ1856.6-3754 p erformed so far.


XMM-Newton Technical Note

XMM-SOC-CAL-TN-0096

Page:

8

Figure 5: Distributions of the difference b etween the background subtracted count rates in the 40­ 200 PHA range in temp oral bins of 100 seconds and their mean during the observation, normalised by the statistical error of each measurement (black lines). The red histogram is the fit of the observed distribution with a Gaussian function. In the insets: the average count rate, and the standard deviation of the fitted distribution.


XMM-Newton Technical Note

XMM-SOC-CAL-TN-0096

Page:

9

Figure 6: background-subtracted, time-averaged Count Rate (C R) in the 40­200 PHA range as a function of the blackb ody sp ectral comp onent normalisation (in relative units) in the simultaneous RGS exp osures: black: RGS 1: red: RGS 2. The dotted lines indicate the linear b est-fits; the dashed lines indicate the envelop e corresp onding to the ±1 uncertainties on the b est-fit parameters.


XMM-Newton Technical Note

XMM-SOC-CAL-TN-0096

Page:

10

References
Burwitz V. et al., 2001, A&A, 379, L35 Gehrels N., 1986, ApJ, 303, 336 Hab erl F., et al., 2004, A&A, 419, 1077 Ho W.C.G., et al., 2007, Ap&SS, 308, 279 Sartore M., et al., 2012, A&A, 541, 66 Struder L., et al., 2001, A&A, 365, L18 Ё Trump er J.E., 2004, NuPHS, 132, 560 Ё Burwitz V. et al., 2001, A&A, 399, 1109 Guainazzi M. (Ed.), 2012, XMM-SOC-CAL-PL-00012 Hambaryan V., et al., 2009, A&A, 497, L9 Kaplan D.L. & van Kerkwijk M.H., 2005, ApJ, 628, L45 Stuhlinger M., et al., 2011, XMM-SOC-CAL-TN-00523 Treves A., et al., 2000, PASP, 112, 297