A&A 384, 190-194 (2002)
DOI: 10.1051/0004-6361:20011837

CHANDRA detection of 16 new X-ray sources

A. Küpcü Yoldas - S. Balman

Department of Physics, Middle East Technical University, Inönü Bulvari, 06531 Ankara, Turkey

Received 6 November 2001 / Accepted 20 December 2001

Abstract
We have detected 18 sources over 6 $\sigma$ threshold within two regions 8 $.\mkern-4mu^\prime$3$\times$16 $.\mkern-4mu^\prime$9 and 8 $.\mkern-4mu^\prime$3 $\times$ 33 $.\mkern-4mu^\prime$6 in the vicinity of the point with $\alpha=$ 03$^{\rm h}$31$^{\rm m}$02 $.\!\!^{\rm s}$45 (J2000) and $\delta=$ +43 $\hbox{$^\circ$ }$47 $\hbox{$^\prime$ }$58 $.\!\!^{\prime\prime}$5 (J2000) using a CHANDRA ACIS (S+I) observation. Two of the sources were detected before with ROSAT HRI and one source could be closely identified with a star in the optical catalog, USNO A-2. We have also studied source spectra applying four spectral models to the data. Most of the sources can be classified as Cataclysmic Variable, Low Mass X-ray Binary or single star candidates due to their spectral characteristics and luminosities. We also searched for the extragalactic origin for these 18 sources. The source count rates vary between 5.8 $\times 10^{-4}$-4.7 $\times 10^{-3}$ counts/s. Due to low count rates temporal characteristics of the sources can not be studied effectively.

Key words: X-rays: stars - X-rays: galaxies - X-rays: binaries - stars: cataclysmic variables, neutron


1 Introduction

The imaging capability and high sensitivity of the Chandra X-Ray Observatory offers significant advantages for source detection. The good instrumental response for photon energies up to 8 keV allows detection of faint sources even if the source is subject to high absorption. The main goal of this search was to detect faint and Super Soft X-ray Sources (SSS) using a 95 ks observation obtained by the CHANDRA Advanced CCD Imaging Spectrometer (ACIS) detector. The original target of the observation was a classical nova Persei 1901 (GK Per; principal investigator = S. Balman). The scientific aim of the proposal was deriving and studying the spectrum of the first classical nova shell resolved and detected in the X-ray wavelengths (Balman & Ögelman 1999). The CHANDRA data reveals for the first time a classical nova evolving like a young, miniature supernova remnant. The scientific results of the CHANDRA observation can be found in Balman (2001, 2002).

As expected most of the sources we detected were faint sources, but we did not find any SSS above 5$\sigma$ confidence level. We analyzed the spectra of these new X-ray sources detected in two regions 8 $.\mkern-4mu^\prime$3$\times$16 $.\mkern-4mu^\prime$9 and 8 $.\mkern-4mu^\prime$3$\times33$ $.\mkern-4mu^\prime$6 around the vicinity of the point with $\alpha=03^{\rm h}$31$^{\rm m}$02 $.\!\!^{\rm s}$45 (J2000) and $\delta=+43\hbox{$^\circ$ }47\hbox{$^\prime$ }58\hbox{$.\!\!^{\prime\prime}$ }54$ (J2000) within an energy range of 0.3-10 keV in order to identify their nature.

Details of the observation and the analysis methods are given in Sect. 2. Section 3 discusses the detection properties of the sources and Sect. 4 is on the spectral analysis and discussion.


 

 
Table 1: Identification table of the sources (all coordinates are epoch = J2000).
Src Source$^\dagger$ RA DEC Counts/s $\times 10^{-3}$ $\sigma$$^\ddagger$
1 CXOAYSB J033025.9+434522.3 03 30 25.92 +43 45 22.32 1.588 $\pm$ 0.451 7
2 CXOAYSB J033037.0+434443.3 03 30 36.99 +43 44 43.33 0.812 $\pm$ 0.395 6
3 CXOAYSB J033017.9+435604.3 03 30 17.98 +43 56 04.32 1.478 $\pm$ 0.411 8
4 CXOAYSB J033121.0+434002.2 03 31 21.04 +43 40 02.16 3.444 $\pm$ 0.439 7
5 1RXH J033136.5+434213$^\star$ 03 31 36.46 +43 42 11.59 1.672 $\pm$ 0.430 6
6 1RXH J033102.3+434757$^\star$ 03 31 02.45 +43 47 58.54 2.159 $\pm$ 0.422 13
7 CXOAYSB J033056.1+434824.2 03 30 56.11 +43 48 24.22 0.940 $\pm$ 0.406 8
8 CXOAYSB J033128.6+435021.2 03 31 28.62 +43 50 21.24 0.962 $\pm$ 0.394 7
9 CXOAYSB J033122.2+435646.8 03 31 22.24 +43 56 46.77 1.513 $\pm$ 0.420 10
10 CXOAYSB J033117.7+435221.6 03 31 17.66 +43 52 21.56 1.516 $\pm$ 0.409 8
11 CXOAYSB J033108.3+435751.4 03 31 08.28 +43 57 51.41 1.332 $\pm$ 0.428 8
12 CXOAYSB J033045.0+435822.4 03 30 44.99 +43 58 22.42 1.577 $\pm$ 0.473 10
13 CXOAYSB J033131.6+435648.8 03 31 31.59 +43 56 48.78 0.841 $\pm$ 0.400 7
14 CXOAYSB J033105.2+435808.1 03 31 05.18 +43 58 08.05 0.584 $\pm$ 0.408 6
15 CXOAYSB J033057.1+435750.4 03 30 57.12 +43 57 50.36 1.312 $\pm$ 0.411 7
16 CXOAYSB J033118.3+435235.0 03 31 18.26 +43 52 35.01 1.330 $\pm$ 0.411 8
17 CXOAYSB J033113.3+435246.7 03 31 13.26 +43 52 46.68 1.672 $\pm$ 0.402 8
18 CXOAYSB J033106.0+440328.8 03 31 06.03 +44 03 28.78 4.743 $\pm$ 0.452 18


$^\star$ ROSAT Complete Results Archive Sources for the HRI.
$^\dagger$ The acronym CXOAYSB is registered in the IAU Registry, see http://cdsweb.u-strasbg.fr/cgi-bin/Dic?CXOAYSB.
$^\ddagger$ Detection sigma above the background.


2 The data and analysis

The classical nova remnant of GK Per was observed with the CHANDRA ACIS between 2000-02-10 and 2000-02-11.

ACIS contains 10 planar, 1024 $\times$ 1024 pixel CCDs; four arranged in a 2$\times$2 array (ACIS-I) used for imaging, and six arranged in a 1$\times$6 array (ACIS-S) used either for imaging or as a grating readout. Two CCDs are back-illuminated (BI) and eight are front-illuminated (FI). 6 CCDs (2, 3, 5, 6, 7, 8) have been used during our observation. GK Per is on CCD number 7. The level 2 data were processed using CIAO software (version 2.0). Bad pixels were removed and the latest suitable calibration files were used. In addition, we cleaned the data for the flaring effects of the ACIS-S3, and the exposure was reduced to 81 ks.

We ran the CELLDETECT[*] algorithm (or the "sliding cell'' method), a source detection algorithm for X-ray data. It was developed for use with the Einstein Observatory images and also employed by the standard processing of ROSAT data (Harnden et al. 1984). This method was tailored to optimize the detection of unresolved sources and had two variants, "local detect'' and "map detect''. In the local detect that is used for our data, the background is estimated in a frame around the detect cell. At each point where a cell is placed, a signal-to-noise ratio of source counts to background counts is computed. If this ratio is above the detection threshold, a candidate source is recorded. The CELLDETECT method is good for faint point sources, outside crowded fields.

The CELLDETECT algorithm, run on the exposure corrected 0.3-10 keV images with exposure maps derived separately for five of the CCDs (2, 3, 5, 6, 7), yielded 20 sources over 6$\sigma$ detection threshold (6$\sigma$ above the background). Since CELLDETECT divides extended sources into multiple point sources and thus requires fine tuning of the parameters, we have chosen different values for the fixedcell parameter for different CCDs. The findpeaks parameter was set to "yes'' so that the adjacent detections were recognized as a single source and the cell with the largest S/N appeared in the source list. We have omitted three of the sources and include one source detected with CELLDETECT algorithm over 12 $\sigma$ threshold at CCD 8. We have increased the threshold sigma of the detection algorithm up to 12 $\sigma$ for CCD 8 because of the excess number of bad pixels. Exposure maps for each of the CCDs were prepared using the single-chip exposure map thread of CIAO 2.0 and used to eliminate the false detections near the edges.

3 Source properties

The results of our analysis with the CELLDETECT algorithm yield 18 sources over 6 $\sigma$ detection threshold. Table 1 displays all 18 sources including the source designations and count rates. The right ascension and declination errors determined from the CELLDETECT algorithm are between 0 $.\!\!^{\prime\prime}$5 and 0 $.\!\!^{\prime\prime}$2. The spatial resolution of CHANDRA is 0 $.\!\!^{\prime\prime}$49 pixel-1. The source count rates vary in a range 0.5$\times$10-3-5$\times$10-3 counts/s.

After searching the HEASARC data archive[*] including multi-wavelength catalogs and mission catalogs, two of the sources were found to be detected before with ROSAT HRI and one source (Src 9) can be closely identified (with 0 $.\!\!^{\prime\prime}$3 offset) with the source in the optical USNO A-2 Catalog (Monet et al. 1998). The red and blue magnitude of the USNO source associated with Src 9 is $m_{\rm r}=$ 14.7, $m_{\rm b}=$ 16.2. For offset radii up to 2 $.\!\!^{\prime\prime}$5 we obtain 5 more possible candidate associations with the stars in USNO A-2 Catalog. However, we exclude these possibilities since the error in RA and DEC determined by CELLDETECT is lower than 2 $.\!\!^{\prime\prime}$5.

  \begin{figure} \par\mbox{ \includegraphics[height=3.70cm,angle=-90,width=4cm,cli... ...ludegraphics[height=3.70cm,angle=-90,width=4cm,clip]{h3378s18.ps} } \end{figure} Figure 1: Fitted spectra of the 18 sources using the data from the CHANDRA ACIS detector between 0.3-10 keV energy range.
Open with DEXTER


  
Table 2: Spectral parameters obtained using fits with blackbody and power law emission models.


\begin{displaymath}\begin{tabular}{l\vert lllll\vert lllll} \hline\hline \multic... ...41.44$^{+87.00}_{-50.42}$ & 45.9 & 1.58 \\ \hline \end{tabular}\end{displaymath}


1 $\times 10^{-7}$, 2 $\times 10^{-15}$, 3 Fits that have $\chi^2_\nu >$ 2 are excluded, 4 $N_{\rm H}$ is fixed at the given value.



  
Table 3: Spectral parameters obtained using fits with thermal bremsstrahlung and VMEKAL models.


\begin{displaymath}\begin{tabular}{l\vert lllll\vert lllll} \hline\hline \multic... ...3.50}_{-28.60}$ & 41.8 & 1.60 & & & & & \\ \hline \end{tabular}\end{displaymath}


1 Solar abundances are assumed, 2 $\times 10^{-7}$, 3 $\times 10^{-15}$, 4 Fits that have $\chi^2_\nu\,>\,2$ are excluded.


4 Spectral analysis and discussion

The source spectra and insturment responses are generated using CIAO (version 2.0) and analyzed using XSPEC (version 11.0.1) applying four models to the source data between 0.3 and 10 keV: blackbody, bremsstrahlung, powerlaw and VMEKAL with absorption. Spectra were grouped to have 5-10 counts per energy bin. The absorbed fluxes are found to be in a range 1$\times$10-16-3 $\times10^{-14}$ erg cm-2 s-1.

Table 2 shows the spectral parameters of the blackbody and power law models and Table 3 shows the spectral parameters of the bremsstrahlung and VMEKAL models. The errors on the spectral parameters are at 2$\sigma$ confidence level. Figure 1 shows the fitted source spectra of all the 18 sources that has the best $\chi^2_\nu$s. The residuals in the figures are omitted since they were in a range of 2$\sigma$--2$\sigma$, and did not show significant fluctuations.

Three of the sources (Src 4, 12, 16) are fitted using a power law model with photon indicies between 2.5-4 and have blackbody temperatures around 0.5 keV resembling the spectra of the Anomolous X-Ray Pulsars (AXPs) (Israel et al. 1999). At a distance of 10 kpc, the luminosities of these three sources are found to be around 1032 ergs/s, and thus we exclude the possible AXP connection because the X-ray luminosities of AXPs are around 1035 ergs/s. Assuming a source distance of 10 kpc the luminosities of all the 18 sources are calculated to be around 1032-1033 ergs/s. Such luminosities are consistent with galactic quiescent CV and quiescent LMXB origin. Two of the sources (Src 11, 13) show evidence for line emission with high absorbtion ( $N_{\rm H} >$ 3$\times$ 1021 cm-2) and the best fit is the absorbed VMEKAL model. These sources are strong Cataclysmic Variable and quiescent LMXB candidates (Warner 1995; Verbunt et al. 1997; Guseinov et al. 2000). In addition, luminosities around 1032 ergs/s could be attained by type O, B or giant stars as a consequence of shocks in the stellar winds or coronal emission (Cassinelli et al. 1981; Schmitt et al. 1993). When we exclude the sources with bremsstrahlung temperatures above 1 keV (since almost all of O, B type or giant stars have temperatures below 1 keV), we are left with 3 candidates for galactic stars; Src 6, 9 and 12. In general, we reject an HMXB origin for our sources, since their luminosities are low compared with an HMXB where the luminosity is $\sim$1036 ergs/s (Guseinov et al. 2000). The luminosities of the sources are around 1037 ergs/s at 4 Mpc in agreement with the luminosities of the X-ray Binaries in other local galaxies, however we do not observe any host galaxy at those directions (Bauer et al. 2001). As noted in Sect. 1, none of these sources can be SSSs. The luminosities of the SSSs are around Eddington Luminosity (1038 ergs/s) and the blackbody temperatures are between 10-60 eV with almost no emission above 1 keV (Kahabka & van den Heuvel 1997). None of our 18 sources have such spectral characteristics. For the case of Dim Thermal Neutron stars (DTNs) or cooling neutron stars, the spectra of DTNs and cooling neutron stars are very soft and almost all of them are within 100 pc distance (Alpar 2001; Ögelman 1995). A neutron star with a surface temperature around 106 K and a 10 km radius has an expected luminosity around 1033 ergs/s. The observations on cooling neutron stars depend on the sensitivity of the X-ray telescopes. Among our 18 sources, the softest two (Src 6, 11) have blackbody temperatures of 0.19 keV. This temperature is relatively high for a DTN or a cooling neutron star, however we do not exclude the possible DTN or cooling neutron star connection since the sensitivity of CHANDRA allows us to observe fluxes around 10-15 ergs cm-2 s-1 consistent with observation of a neutron star at 10 kpc with a temperature $\sim$0.1 keV. Due to the relative hardness of the spectra of these 18 sources, we can also say that none of these sources are isolated hot white dwarf candidates since the temperatures of the hot white dwarfs are ultrasoft; a few eVs (Vennes 1999; Finley et al. 1997).

We searched for the extragalactic origins of all the 18 sources. The rest frame luminosities of AGNs and galaxies are known to be between 1039 and 1045 ergs/s, and their spectra are best fitted with power law models having photon indicies around 1.7-2 (Brandt et al. 2001; Ishisaki et al. 2001). The rest frame luminosities of nearly all of the 18 sources are calculated to be around 1042-1044 ergs/s after fitting zmodels using XSPEC. Thus, we can not exclude the possibility of extragalactic origin for any of the sources. Two highly absorbed sources Src 3 and 5 have rest frame luminosities around 1042 ergs/s (redshift = 0.206) and 1043 ergs/s (redshift = 0.500) respectively. The photon indicies of these two sources are 1.61 and 1.89 respectively. Hence, we might categorize these two sources as strong AGN and galaxy candidates. Additionally, the $N_{\rm H}$ values may be further evidence to exclude the extragalactic connection. The sources with $N_{\rm H}$ higher than the value of the galactic $N_{\rm H}$ in the direction of GK Per are more likely to be of extragalactic origin than the sources having the same order of $N_{\rm H}$ with the galactic value. We may say that none of these sources are clusters since they are not extended and the rest frame luminosities of the clusters are usually higher than these values (Schindler 1999).

In addition, we searched for temporal characteristics of the new sources. We performed power spectrum analysis on three of the sources with the highest count rates (Src 4, 6, 18). However we could not find any significant periods. The 3$\sigma$-4$\sigma$ detection threshold of power is around 40 in our data, and the power upper limit of the three sources is found to be 20.

Drawing conclusions about the classification of all the 18 sources is difficult using the X-ray data at hand. We can not study their temporal characteristics due to low count-rates, and we do not have detailed spectra. We have planned further deep observations of this field in the optical wavelengths with the 1.5 m telescope of the National Observatory at Antalya, Turkey to ensure optical identification. Complementary observations in other wavelengths are necessary for proper classification of these 18 sources.

References

 
Copyright ESO 2002