3 Comparison with other surveys

 

3.1 VLA 20 cm surveys

The observed field is included in the VLA 20 cm galactic plane survey (ZGS; Zoonematkermani 1990) and in the NVSS (Condon et al. 1998). Both these surveys used the VLA at 20 cm (1.4 GHz), however, the ZGS used the moderately extended B array and has a typical peak flux density sensitivity of 25 mJy/beam and a synthesized beam of $\sim 5^{\prime\prime}$ (FWHM), while the NVSS used the most compact D array with a flux density limit of $\sim 2.5$ mJy/beam ($\sim 0.5$ mJy/beam rms) and an angular resolution of $\sim 45^{\prime\prime}$.

Given the relatively low sensitivity, and the similar (u,v) sampling with our 6 cm observations, we expect to detect all the ZGS sources in our maps (see also Becker et al. 1994). On the other hand, due to the much higher sensitivity of the NVSS and its ability to detect extended structures, many of the fainter 20 cm sources with non-thermal spectral indexes and/or sizes greater than 10$^{\prime\prime}$will not be detectable in our observations. In Fig. 4b we show the positions of all the ZGS (11 - pluses) overlaid on the contour plot of the NVSS image of our survey region.

In Table 2 the results of the correlation between our catalogue and the 20 cm surveys is presented. The relevant parameters (names, positions, flux densities and sizes) of the 20 cm sources are from the published catalogues (Zoonematkermani et al. 1990 for the ZGS and the deconvolved data from the fits catalogue available on the World Wide Web at http://www.nrao.edu in October 1998 for the NVSS). The matching criterion used is positional coincidence: ZGS sources are considered to be associated with our sources if the positional difference is less than half a beamwidth for point sources, or if the source position falls within the boundary of one of our extended sources; NVSS sources are considered to be associated if one of our point source falls inside of, or if the boundaries of one of our extended sources overlap with, the deconvolved size of the 20 cm source. As expected, all the ZGS sources in our surveyed field do have a counterpart at 6 cm. In one case (source #32 in our list), we considered two ZGS sources as being part of the same (extended) 6 cm source. In Table 2, Cols. 1 and 2 report the numbers and names of our sources from Table 1, Cols. 3 to 6 the names, peak and integrated flux densities, and sizes of the ZGS sources, Cols. 7 to 10 the names, integrated flux densities and deconvolved sizes of the NVSS sources, and Col. 11 the IRAS sources names (see Sect. 3.4).

In general, given the higher sensitivity of the NVSS and its ability to detect extended sources that might be resolved out in the ZGS, we expect that all the ZGS sources in our field should be detected in the NVSS as well. The only possible exception is that of very compact high surface brightness sources close or inside large low surface brightness sources with high integrated flux. There are 3 ZGS sources without an NVSS counterpart, one (045.129+0.131, associated to our #17) is indeed inside the bright complex shown in Fig. A4, and thus may be missing from the NVSS catalogue due to confusion. Similarly, the one associated with our #19 could be undetected in the NVSS due to its proximity to the extended source #33. Both #17 and #19 have thermal spectral indexes (see below and Table 3) and we do not expect them to be variable at 20 cm. On the other hand, the ZGS source associated with #29 should have been detected in the NVSS, thus for this source, given also its non-thermal spectral index, the only viable explanation for the NVSS non-detection is variability at 20 cm.

Finally, there is a very bright ($\sim$280 mJy), unresolved, NVSS source which is undetected in the ZGS and in our survey. This source (clearly visible in Fig. 4b at $l\sim 45.35$, $b\sim -$0.22) is the high energy source G1915+105 (Mirabel & Rodríguez 1994). At radio wavelengths is known to be highly variable, with flux densities at 1.4 GHz that can be as high as $\sim 1$ Jy at the peak of radio bursts and below the mJy level during quiescence (Rodríguez & Mirabel 1999).

 

Table 2:  Associated ZGS, NVSS and IRAS-PSC2 sources

In Table 3, Cols. 2 to 6, we report the radio continuum spectral indexes ($\alpha$, defined as $F_\nu\sim\nu^\alpha$)as calculated from our integrated flux densities and the ZGS and NVSS integrated flux densities. It should be noted that all extended sources are probably partially resolved out in the higher resolution surveys, particularly in our 3.6 cm images, and thus some of the measured spectral indexes are probably lower limits due to the missing flux at high frequency.

 

Figure 5:  Comparison between the high and low frequency spectral indexes. Top panel: sources in our survey detected at 20 cm in the NVSS. Bottom panel: sources detected in the ZGS. In both panels the dotted line represent equal spectral indexes

In Fig. 5 we compare the high frequency spectral indexes (those calculated between 3.6 and 6 cm) with the low frequency ones (calculated between 6 and 20 cm), only the sources inside the area observed both at 3.6 and 6 cm have been considered. A 10% error has been assumed for all ZGS and NVSS integrated flux densities (this may be a slight underestimate of the true error for the faintest sources in these surveys). In the upper panel we show the comparison for sources detected in the NVSS and in the lower panel that for sources detected in the ZGS. We find very good agreement between the high frequency and the low frequency spectral indexes for ZGS sources. This is probably due to the matched beams of the observations. In contrast, for NVSS sources, the spread between low and high frequency spectral indexes is much greater. There are two possible explanations for this: 1) the increased sensitivity to extended structures of the NVSS and 2) the greater sensitivity of the NVSS with respect to the ZGS. The increased sensitivity allows for the detection in the NVSS of some thermal sources that are optically thick at low frequency and become optically thin at high frequency (this is probably the case for #9 and #34).

 

Table 3:  Radio continuum spectral indexes and IRAS fluxes for the detected sources

3.1.1 NVSS sources undetected at high frequency

Most of the NVSS sources in our field (48) are not detected at 6 and/or 3.6 cm. We believe that in most cases the negative spectral index, rather than the different (u,v) coverage between the observations, is the main reason for the non-detection at high frequency. The most plausible explanation is that a large fraction of these NVSS sources are extragalactic objects, with a possible contamination from faint planetary nebulae.

 

Figure 6:  Top panel: differential luminosity functions of all (All; thin continuous line) and not detected at high frequency (NHF; thick continuous line) NVSS sources inside our field, and of NVSS sources from two 0.652 sq. deg. areas close to the northern (NGP; dotted line) and southern (SGP; dashed line) galactic poles. Bottom panel: cumulative luminosity functions for the same sources shown in the upper panel

To check whether the 20 cm flux distribution and source count for the NVSS sources not detected at high frequency are consistent with the population of extragalactic radio sources, we extracted from the NVSS the sources in two areas toward the galactic poles, each of the two with the same extent of our surveyed region. The number of sources extracted toward the northern and southern galactic poles are 36 and 27, respectively, these numbers compare relatively well with the 37 NVSS sources without high frequency counterpart in our field. As additional check, in Fig. 6, we show the differential and cumulative luminosity functions for the sources in our field and those in the areas toward the galactic poles. The luminosity function of all the sources in our field (thin line) show an excess of bright sources with respect to the galactic poles, this excess disappears if we plot only the sources without a high frequency counterpart (thick line). This effect is more clear in the cumulative luminosity function plot (Fig. 6, lower panel). More quantitatively, the Kolmogorov-Smirnov test on the cumulative luminosity functions gives a probability lower than 40% that the NVSS sources in the Galactic poles samples and those in our field are drawn from the same distribution. This probability rises above 80% if we remove from our sample the sources detected at high frequency and the well known galactic high energy source G1915+105.

3.2 Effelsberg 5 GHz survey

  As mentioned in Sect. 1, our surveyed region has been covered by the Altenhoff et al. (1978) 5 GHz (6 cm) single dish survey. The names and peak flux densities of the seven single dish sources inside or partially within our survey boundaries are listed in Table 4. In the same table, for each source, we report the integrated flux densities of our VLA 6 cm sources within the Effelsberg beam (2.6$^\prime$).

For one of the single dish sources (44.786-0.490) the peak is outside our survey area. We do not detect this source at either 6 or 3.6 cm, probably because it is resolved out in our interferometric observations. The last two sources in Table 4 are in the region covered only at 3.6 cm, they have been detected at this wavelength and correspond to sources [18+19+33] and 32 in Table 1. The other four sources have been detected in our 6 cm observations, and our integrated flux densities are in reasonable agreement with the single dish ones, except for 45.341-0.370 (our source 26) which is known to be a highly variable source (see e.g. Harmon et al. 1997). Somewhat surprisingly, in our VLA 6 cm images we recover almost all the single dish flux for the extended complexes 45.066-0.135 and 45.125+0.136, while about half of the single dish flux is recovered for 45.202-0.411.

 

Table 4:  Comparison between single dish and VLA 5 GHz sources

^a) This source is known to be variable (e.g. Harmon et al. 1997).

3.3 IRAS Point Sources Catalogue

To search for far infrared (FIR) counterparts to our detected radio sources, we extracted from the IRAS-PSC2 (Beichman et al. 1988) catalogue all the sources inside our survey area. In Fig. 4c we show the positions of all (43) IRAS point sources inside the observed field. We could find an IRAS counterpart within 100$^{\prime\prime}$ only for 5 of our 3.6 and/or 6 cm sources. In all five cases, the IRAS error ellipse contains the radio continuum source or overlaps with the boundaries of the extended radio sources. In fact, in all five cases the distance from the peak of the radio continuum source and the nominal IRAS position is less than 30$^{\prime\prime}$.The FIR fluxes of these five sources are reported in Table 3, Cols. 7 to 10.

The study of the IRAS color-color diagram is a powerful tool to investigate the nature of the FIR sources. Different types of objects tend to populate different parts of the color-color planes. In Fig. 7 we show three of the color-color diagrams that can be constructed using the four IRAS fluxes, and that have been shown to be able to separate different types of galactic sources (e.g. Eder et al. 1988; Pottasch et al. 1988; WC89 White, Becker & Helfand 1991). In each diagram the contour plots represent the normalized surface density of the colors ([$\lambda_i$,$\lambda_j]\equiv \log_{10}(F_{\lambda_i}/F_{\lambda_j})$) of IRAS-PSC2 sources within the inner galactic plane, defined as: $\vert l\vert\le 90^\circ$, $\vert b\vert\le0^\circ\!\!.65$.

 

Figure 7:  In each diagram: [$\lambda_i$,$\lambda_j]\equiv \log_{10}(F_{\lambda_i}/F_{\lambda_j})$;the contour plots represent the normalized surface density of IRAS-PSC2 sources in the region $\vert l\vert\le 90^\circ$, $\vert b\vert\le0^\circ\!\!.65$.Black filled circles show the colors of the 43 sources in our surveyed region (one source detected only at 100 $\mu$m is not present in the first two plots). Many of the sources only have upper limits at one or more IRAS bands, the colors for these sources are either upper, lower limits, or are undetermined. We have not marked these sources with special symbols as we have not corrected the color-color surface density contours for upper limits. The five IRAS sources with a radio continuum counterpart are marked with plus symbols

We note that the 43 IRAS sources in our field tend to populate the color-color planes in the same fashion as the entire inner galactic plane sample (contour plots), which, of course, is what we expected. It is remarkable, however, that all, and only, the IRAS sources detected in radio continuum (marked with plus symbols in the figure) lie in well-defined, low-density parts of the planes. This is the part of the color-color planes where ultra compact HII (UCHII) regions are expected to be found (WC89; Wood & Churchwell 1989b; Kurtz, Churchwell & Wood 1994; White et al. 1991; Becker et al. 1994).