4. Optical identification of soft sources

From the list of sources that are found in the fields of Table 1 (click here), I selected all sources that are compatible with a soft, thermal source spectrum. Paper I showed that HR2 is a better estimator for the softness of a source spectrum because it suffers less from intervening absorption. Sources with a black body spectrum and a source temperature of 70eV or less will have a hardness ratio HR2 of -0.8 or less, if they penetrate less than . Neutron stars accreting from the interstellar medium are not expected to show harder source spectra (e.g. Blaes & Madau 1993). Some authors, e.g. Zamperi et al. (1995), have argued that the emergent spectrum of accreting neutron stars might have a hard component but the bulk of the emission () is still equivalent to a soft thermal spectrum.

4.1. Selection criteria

A conservative selection criterion for soft sources would require that the hardness ratio plus its error is less than some threshold, . However, this criterion is too stringent for my sample. Soft sources that are only detected in the soft band and are not detected in the hard band are assigned by the standard processing an artificially large error in the hardness ratio. Applying the above selection criterion would exclude all these sources. Instead, I define a filter based on source brightness. I have selected all sources with a rate and at least 20 detected counts. A minimum of 20 counts gives a reasonable confidence in the accuracy of the hardness ratio, and the count rate limit guarantees an area coverage of better than 92%.

 

X-ray Coordinates Rate Cnts HR1 HR2 ID Type m_V Name

RX J0550.6+0005 05 : 50 : 37.7 +00 : 05 : 56 0.866 430 -0.98 -1.00 GD 257 DA1 15.10

RX J0605.0-0707 06 : 04 : 58.8 -07 : 07 : 06 0.038 20 0.79 -1.00 spec AGN, z=0.233

RX J0614.0-0714 06 : 13 : 58.1 -07 : 14 : 50 0.092 24 0.06 -1.00 HD43066 B9 6.7

RX J0620.2-0904 06 : 20 : 11.5 -09 : 04 : 26 0.096 29 -0.01 -1.00 spec dMe

RX J0629.6+0829 06 : 29 : 37.0 +08 : 29 : 29 0.110 51 -0.25 -1.00 HD45759 F8 7.5

RX J1156.0-7700 11 : 55 : 57.7 -77 : 00 : 40 0.065 21 -0.67 -1.00 HD103673 G5V 8.6

RX J1623.9-3312 16 : 23 : 54.7 -33 : 12 : 28 0.060 23 -0.09 -0.82 HR6097 A0V+ 7.0

RX J1856.6-3754 18 : 56 : 34.4 -37 : 54 : 26 3.456 567 -0.73 -0.92 NS candidate

RX J1917.7+2226 19 : 17 : 39.3 +22 : 26 : 28 0.034 21 0.89 -1.00 HD180939 B5V+ 6.9

RX J1920.9+1903 19 : 20 : 51.4 +19 : 03 : 44 0.070 40 -0.27 -1.00 1E1918.7+1857 dMe

RX J1922.7+0950 19 : 22 : 40.7 +09 : 50 : 55 0.202 43 1.00 -1.00 spec G star

RX J1947.4+3045 19 : 47 : 23.3 +30 : 45 : 58 0.873 197 -0.95 -1.00 spec hot white dwarf?

RX J1958.6+7218 19 : 58 : 34.1 +72 : 18 : 24 0.062 52 -0.03 -0.85 probably stellar

RX J2011.5+3824 20 : 11 : 28.3 +38 : 24 : 12 0.087 33 -0.02 -1.00 HD192020 G8V 7.93

RX J2019.6+4048 20 : 19 : 38.8 +40 : 48 : 25 0.230 63 0.85 -1.00 faint red object

RX J2114.6+4607 21 : 14 : 33.6 +46 : 07 : 44 0.135 39 0.96 -0.86 faint red object

Table 2: All soft sources in the dark cloud sample that match the brightness criterion (see text). Left to right columns: ROSAT name; X-ray source coordinates in epoch 2000.0; ROSAT count rate; total number of detected counts; hardness ratio 1 and 2; SIMBAD name of proposed optical counterpart (spec: optical spectrum presented in this paper); spectral type and optical magnitude of counterpart or remark  

Out of 2270 source in my dark cloud sample only 16 sources pass this selection filter. Table 2 (click here) lists all 16 sources with their ROSAT name, coordinates and their X-ray parameters. I have approximated the boundaries of the dark clouds with a wide rectangle. Therefore some of the sources in this areas might not be in front of a cloud and extragalactic objects might shine through gaps between individual cloudlets. A reduced stellar density close to the source position is a good indicator for the presence of a dark cloud.

4.2. The soft sub-sample

I have used the digitized sky-survey to create finding charts for each of the 16 sources, Fig. 3 (click here). A circle with a radius of 30'' is centered in each figure at the X-ray position. Because most sources in the sample are faint, their uncertainty in the X-ray position is dominated by systematics. 30'' is generally considered a conservative choice and supported by recent identification programs (see Paper I for more details). For some fields capital letters identify objects that have been further studied with optical spectroscopy.

     
Figure 3: Finding charts for soft sources. The charts have been produced from the digitized Palomar Sky-Survey. The circles are centered at the X-ray position and have a radius of 30 arcseconds. North is up and East to the left. Capital letters identify the brightest objects next to them  

All source positions were searched in the SIMBAD and NED catalogs for known counterparts. The result of this search is summarized in Table 2 (click here). Eight sources are identified with bright stars. For four additional fields, I conducted optical spectroscopy at the Palomar 60-inch and the Palomar 200-inch telescope. Low resolution spectra of candidates are shown in Fig. 4 (click here).

 
Figure 4: Identification spectra for 4 objects. RX : M dwarf with emission. RX : faint stellar object RX : AGN with [OIII] and broad emission at z=0.2. RX : G star. The X-axes are in Angstrom and the Y-Axes in arbitrary units  

 
Figure 5: CCD frames of the field of RX J2019.6+4048. The images were taken with 60 seconds integration time at the Palomar 60-inch telescope. The images are in the blue Kron-Cousin B (left) and R filter (right). An arrow in the read image marks a faint red object, probably an extragalactic object. The circles are centered at the X-ray source position and have a radius of 30 arcseconds. North is up and East to the left  

RX is an AGN with a redshift of 0.233. RX is a M dwarf star with emission; the second object within the error circle RX shows a stellar spectrum without emission lines. I therefore suggest the M dwarf to be the source of RX . RX is identified as G star. No spectroscopic data is available for RX . However, Fig. 5 (click here) shows CCD frames in the Kron-Cousin B and R filter of this field. In the R image, a faint red object close to the X-ray position is apparent. An inspection of the R and B plate of the second generation of the Palomar Sky Survey (POSSII) yielded another red object coincident with RX J2114.6+4607, see mark on finding chart in Fig.3 (click here). Pending identification through spectroscopy, I suggest that those two red objects are the likely extragalactic counterparts to RX J2019.6+4048 and RX J2114.6+4607.

Unfortunately there is no spectroscopic data available for RX J1958.6+7218. However, due to the moderately hard spectrum, the low count rate of and the presence of object A inside the error circle I expect this source to be due to coronal emission from A.

4.3. The brightest sources

The three brightest sources in the sub-sample, RX J0550.6+0005, RX J1856.6-3754 and RX J1947.4+3045, are also the sources with the softest HR1 values. This indicates that they are affected very little by absorption. The brightest source RX J1856.6-3754 with has been indeed suggested by Fred Walter et al. (1996) to be a nearby isolated neutron star, accreting from the interstellar medium. RX J0550.6+0005 is the known hot white dwarf star GD257 (Bergeron et al. 1994) with an effective surface temperature of also visible as EUV source (Marsh et al. 1997).

This left the intriguing case of RX J1947.4+3045 with very similar X-ray properties as GD257 but a much fainter optical counterpart (m_V>19 compared to m_V=15.1 for GD257). A comparison of the B and I plate of POSSII showed a very blue object within the X-ray error circle, see mark on finding chart in Fig.3 (click here). I was able to obtain a spectrum of the source taken with the Palomar Hale telescope and the double spectrograph in June 1997, see Fig. 6 (click here). The spectrum is similar to that of a hot white dwarf star. It shows a hot continuum, no emission lines, very weak Balmer lines in absorption and broad absorption bands that resemble Swan C₂ bands, due to molecular carbon. The hot continuum and the molecular bands give conflicting informations on the source temperature, demanding high and low surface temperatures simultaneously. This suggests that we are either looking at a binary system or absorption from circumstellar material. None of the models is particular conclusive. I am therefore not yet able to describe all features in the spectrum with a consistent model. This interesting object will be discussed in detail in a forthcoming letter.

 
Figure 6: Spectrum of the blue object inside the error box of RX J1947.4+3045. The gap in the spectrum around 5300 Å is caused by the lack of overlap between the red and the blue side of the spectrograph. The spectrum immediately next to the gap is heavily affected by effects of the dichroic that can not be calibrated. The Y-Axis has arbitrary units