Документ взят из кэша поисковой машины. Адрес оригинального документа : http://xmm.vilspa.esa.es/docs/documents/CAL-TN-0184-1-0.pdf Дата изменения: Tue Feb 12 18:07:06 2013 Дата индексирования: Fri Feb 28 23:41:10 2014 Кодировка: Поисковые слова: http www.astronomy.ru forum index.php topic 4644.0.html

XMM-Newton Calibration Technical Note
XMM-SOC-CAL-TN-0184 Testing epspatialcti on clusters of galaxies
Fabio Gastaldello (INAF-IASF Milano, Italy) , Konrad Dennerl (MPE, Germany) January 3, 2013

1

Scop e

This report aims at testing the performance of the SAS task epspatialcti. epspatialcti corrects an EPIC-pn event file for spatially dependent CTI effects, which remain after the standard CTI correction in SAS has been applied. They show up as pixel to pixel variations in the energy scale. These spatially dependent CTI effects are investigated by measuring spectra of bright extended sources. The task and the asso ciated CCF file have been validated using the Al-K (1.5 keV) and Mn-K (5.9 keV) lines from the internal 55 Fe calibration source and the O VII K line (0.6 keV) of the Vela supernova remnant [1]. Here we report the results of the performance of the task when applied to the spectra of bright galaxy clusters.

2

Methodology
K line at the massive galaxy co ordinates for 6-8 keV energy covered by the

We tested the performance of the task epspatialcti on the Fe He-like (Fe XXV) rest frame energy of 6698.6 eV, the strongest line emission in the spectra of bright, clusters. We followed a similar approach to [1] and extracted images in detector singles events with an adaptive binning requiring a fixed number of counts in the band. The boundaries of the regions along the y-axis are aligned with the y ranges 1


XMM-Newton Technical Note

XMM-SOC-CAL-TN-0184

Page:

2

EFF singles FF singles

6.7 keV 14.1 12.0 11.9 9.7

Table 1: Standard deviations before after (in eV) determined in the individual spatial bins after quadratically subtracting the statistical error, before and after running epspatialcti. response matrices. The spectral mo del consists of a power-law to mo del continuum and background emission plus a Gaussian line to mo del the Fe He-like line. In the case of the Perseus cluster an additional line was used to mo del the Fe H-like (Fe XXVI) line at 6973.1 eV rest frame. Both line energies and width of the Fe XXV line were free parameters. We used the c-statistic on the original 5 eV data bins.

3
3.1

Results
The Perseus cluster

We analyzed the long (125 ks on time) Perseus cluster (at a redshift of z=0.0179, so the expected Fe-He line energy is 6586.6 eV) observation (OBSID 0305780101, Extended Full Frame) filtering the observation with a threshold of 0.6 cts s-1 for a light curve in 100 s bins in the 10-12 keV energy band (as in [2]). We required 4000 cts in the 6-8 keV range obtaining 65 regions; the fits were performed in the 6-7.3 keV range. Fits were go o d, with a distribution of cstat/dof with a mean of 1.03 and standard deviation of 0.11 for the spectra before applying epspatialcti and a distribution with a mean of 1.02 and standard deviation of 0.09 for the spectra after applying epspatialcti ; examples are shown in Fig.1. In Fig.2 we show the results for the spatial distribution of the energies before running epspatialcti on the left and on the right after running epspatialcti. Following [1], to quantify the amount of suppression obtained with epspatialcti we computed the standard deviation of the energies derived for the individual areas after quadratically subtracting the statistical error and the results are summarized in Table 1. The result of before = 14.1 eV after = 12 eV can be compared with the result before = 14.2 eV after = 13 eV for EFF singles obtained by [1] using the Mn-K line at 5.9 keV from the internal calibration source. Motivated by the results of the Centaurus cluster (see §3.2) we analyzed an earlier short (51 ks on time) observation of the Perseus cluster taken in Full Frame mo de (OBSID 0085110101). The whole observation is affected by soft proton contamination, therefore no flare filtering has been applied (as detailed in [3]). We required 3000 cts in the 6-8 keV range obtaining 28 regions. Performing fits


XMM-Newton Technical Note
data and folded model 0.05 0.1
-1

XMM-SOC-CAL-TN-0184

Page:

3

data and folded model

normalized counts s-1 keV

0.01

normalized counts s-1 keV

0.02

-1

0.01

5в10

-3

2в10

-3

-1

normalized counts s-1 keV

5в10

-3

normalized counts s-1 keV

-1

10

-3

10

-3

0.02

0

0

-5в10-3 6 6.2 6.4 6.6 Energy (keV) 6.8 7 7.2
gasta 8-Aug-2012 18:09

-0.02 6 6.2 6.4 6.6 Energy (keV) 6.8 7 7.2
gasta 8-Aug-2012 21:09

Figure 1: Examples of fit to spectra of the Perseus cluster (single events). Left panel: fit to the first bin corresponding to RAW-Y 120-139 of the lower part of the detector. Right panel: fit to the sixth bin corresponding to RAW-Y 180-199 of the upper part of the detector. in the same manner as described above we obtained acceptable fits with a distribution of cstat/dof with a mean of 1.10 and standard deviation of 0.11 for the spectra before applying epspatialcti and a distribution with a mean of 1.08 and standard deviation of 0.08 for the spectra after applying epspatialcti. We quantified the amount of suppression as done previously, with results summarized in Table 1 and shown in Fig.3.


XMM-Newton Technical Note

XMM-SOC-CAL-TN-0184

Page:

4

Figure 2: Comparison of the spatially resolved energy determination of the Fe XXV line in the Perseus cluster (expected energy 6586.6 eV at the redshift of the source) for singles in extended full frame mo de, before (left) and after (right) running epspatialcti. The statistical 1 error is 5 eV (mean of the distribution of errors for each area, with a standard deviation of 1.7 eV) in both cases. The individual energy determinations exhibit a mean of 6563 eV with a standard deviation of 15 eV before running epspatialcti and a mean of 6560 eV with a standard deviation of 13 eV. Taking the statistical uncertainties into account, this corresponds to a standard deviation of 14.1 eV before and 12 eV afterwards.


XMM-Newton Technical Note

XMM-SOC-CAL-TN-0184

Page:

5

Figure 3: Comparison of the spatially resolved energy determination of the Fe XXV line in the Perseus cluster (expected energy 6586.6 eV at the redshift of the source) for singles in full frame mo de, before (left) and after (right) running epspatialcti. The statistical 1 error is 5.3 eV with a standard deviation of 1.9 eV and 5.1 eV with a standard deviation of 1.7 respectively. The individual energy determinations exhibit a mean of 6564 eV with a standard deviation of 13 eV before running epspatialcti and a mean of 6561 eV with a standard deviation of 11 eV. Taking the statistical uncertainties into account, this corresponds to a standard deviation of 11.9 eV before and 9.7 eV afterwards.


XMM-Newton Technical Note 3.2 The Centaurus cluster

XMM-SOC-CAL-TN-0184

Page:

6

We analyzed two observations of the Centaurus cluster (at a redshift of z=0.0114, so the expected Fe-He line energy is 6623.1 eV): one (OBSID 0046340101) taken in Full Frame mo de for a go o d exposure after flare cleaning of 34 ks and the other (OBSID 0406200101) taken in Extended Full Frame mo de for a go o d exposure time of 92 ks. The purpose was to explore on the same astrophysical source the two pn mo des; the comparison is somewhat limited by the different statistical quality of the two observations. Given the lower temperature ( 3 keV) of this cluster compared to Perseus we required only 1000 cts in the 6-8 keV range for the spectral extraction, in order to have a decent number of bins in the short exposure. With these requirements we obtained 17 regions for the short FF observations and 43 for the long EFF observation (this number has been reduced from an initial larger number of 60 regions, some of them a posteriori refused because of very low signal). For the long EFF observation fits were acceptable with a distribution of cstat/dof with a mean of 1.11 and a standard deviation of 0.09 for the spectra before applying epspatialcti and a distribution with a mean of 1.12 and a standard deviation of 0.10 for the spectra after applying epspatialcti ; examples are shown in Fig.4. In Fig.5 we show the results for the spatial distribution of the energies before running epspatialcti on the left and on the right after running epspatialcti. Computing the standard deviation of the energies derived for the individual areas after quadratically subtracting the statistical error is probably not meaningful given the same formal level of standard deviation of the distributions and the statistical errors (as detailed in the caption of Fig.5); given the asymmetry of the error distribution (because of the higher signal of the central regions) if we restrict the analysis to the 23 regions where the statistical error is lower than the median of the distribution of the statistical error we obtain a distribution of individual energy determinations with a mean of 6614 eV and standard deviation of 10.7 eV before applying epspatialcti and a distribution with mean 6613 eV and standard deviation 9.5 eV after applying epspatialcti. The 1 error statistical distribution has a mean of 6.5 eV with standard deviation 1.9 eV before epspatialcti and a mean of 6.7 eV with standard deviation 2.1 eV after epspatialcti. This corresponds to before = 8.6 eV after = 6.7 eV. For the short FF observation fits were acceptable with a mean of 1.14 and standard deviation of 0.12 for the fits before epspatialcti and a mean of 1.17 and standard deviation of 0.12 for the fits after epspatialcti. In Fig.6 we show the plot for the results of the spatial distribution of the energies. The standard deviation of the energies derived for the individual areas after quadratically subtracting the statistical error are summarized in Table 2.


XMM-Newton Technical Note

XMM-SOC-CAL-TN-0184

Page:

7

data and folded model

data and folded model

-1

-1

0.01

normalized counts s-1 keV

normalized counts s-1 keV

0.02

5в10-3

0.01

5в10-3

2в10-3

2в10-3

normalized counts s-1 keV

normalized counts s-1 keV

10-3 2в10-3 10
-3

-1

-1

10

-3

5в10-3

0 -10-3 -2в10-3 6 6.2 6.4 6.6 Energy (keV) 6.8 7 7.2
gasta 9-Aug-2012 11:11

0

-5в10

-3

6

6.2

6.4

6.6 Energy (keV)

6.8

7

7.2
gasta 9-Aug-2012 11:17

Figure 4: Examples of fit to spectra observation (OBSID 0406200101). Lef of the lower part of the detector. Right to RAW-Y 180-199 of the upper part

of the Centaurus cluster (single events) for the long EFF t panel: fit to the second bin corresponding to RAW-Y 80-99 panel panel: fit to the fourth bin of the column corresponding of the detector.

EFF singles EFF singles restricted FF singles

6.7 keV 7.4 7.0 8.6 6.7 9.0 3.1

Table 2: Standard deviations before after (in eV) determined in the individual spatial bins after quadratically subtracting the statistical error, before and after running epspatialcti.


XMM-Newton Technical Note

XMM-SOC-CAL-TN-0184

Page:

8

Figure 5: Comparison of the spatially resolved energy determination of the Fe XXV line in the Centaurus cluster (expected energy 6623.1 at the redshift of the source) for singles in extended full frame mo de, before (left) and after (right) running epspatialcti. The statistical 1 error is 12.8 eV (mean of the distribution of errors for each area, with a standard deviation of 9 eV; however the distribution is skewed, the median is in fact 9.5 eV and 10 eV for before and after running epspatialcti respectively) in both cases. The individual energy determinations exhibit a mean of 6608 eV with a standard deviation of 14.8 eV before running epspatialcti and a mean of 6606 eV with a standard deviation of 14.6 eV after running epspatialcti. Taking the statistical uncertainties into account, this corresponds to a standard deviation of 7.4 eV before and 7.0 eV afterwards.


XMM-Newton Technical Note

XMM-SOC-CAL-TN-0184

Page:

9

Figure 6: Comparison of the spatially resolved energy determination of the Fe XXV line in the Centaurus cluster (expected energy 6623.1 at the redshift of the source) for singles in full frame mo de, before (left) and after (right) running epspatialcti. The statistical 1 error distribution has a mean of 8.7 and standard deviation 3.8 eV for the fits before epspatialcti and a mean of 9.4 and standard deviation 4.3 eV for the fits after epspatialcti. The individual energy determinations exhibit a mean of 6610 eV with a standard deviation of 12.5 eV before running epspatialcti and a mean of 6607 eV with a standard deviation of 9.9 eV after running epspatialcti. Taking the statistical uncertainties into account, this correspond to a standard deviation of 9.0 eV before and 3.1 eV afterwards.


XMM-Newton Technical Note

XMM-SOC-CAL-TN-0184 Page:

10

4

C o n c lu s io n s
epspatialcti when applied The improvement on the line emission at 6.7 keV is ernal calibration source.

We provide in this do cument a test of the performance of the SAS task to the spectra of the thermal extended emission of clusters of galaxies. spatial homogeneity of the reconstructed energy of the astrophysical Fe comparable to the one observed in the Mn-K fluorescence line of the int

References
[1] Dennerl, K. & Saxton, R., 2012, XMM-CCF-REL-283 [2] Molendi, S. & Gastaldello, F., 2009, A&A, 493, 13 [3] Gastaldello, F. & Molendi, S., 2004, ApJ, 600, 670