3 Results

3.1 Identifying the NIR counterparts  

In order to identify the NIR counterparts of the IRAS sources, we need to determine where the IRAS uncertainty ellipses (which express the position error at a 95% confidence level) lie within each frame. After finding $\alpha$ and $\delta$ coordinates as a function of pixel rows and columns for each K image (see Sect. 2.3), the IRAS ellipses could be properly placed on the observed fields. All K sources within an area of $\sim 60\times 60$ arcsec² centred on the IRAS ellipse were then examined for finding the best candidates. The NIR counterpart, i.e., the object which effectively contributes the most to the FIR flux, is tentatively selected in each field according to the following criteria:

1.

spectral index ;

2.

maximum luminosity in the K band;

3.

intrinsic excess in (H-K, J-H) diagrams (hereafter, we will refer to these simply as colour-colour diagrams);

4.

the spectral index s should be the largest one;

5.

closeness to the IRAS ellipse centre.

We noted that sources with spectral indices s<0 are generally visible in the DSS optical plates, whereas sources with are not, so item 1 cut off at least bright visible stars. Item 3 should allow to select YSO's, though this criterion can be relaxed since low luminosity Class I sources sometimes do not exhibit IR excesses in colour-colour diagrams and may appear as very reddened stars (see, e.g., Lada & Adams 1992; in the following, we will refer to this paper in analyzing colour-colour diagrams). If the FIR flux was due to a cluster of sources, rather than a single one, we anyway would expect the IRAS uncertainty ellipse to include the location of the cluster (brightness-weighted) centre. To accurately assess the above requirements, for each field we put all K sources detected within the $\sim 60\times 60$ arcsec² area in a (H-K, J-H) diagram and plotted their SED's in the form of a $\log(\lambda)-\log(\lambda F_{\lambda})$ diagram (see Figs. 3-14, which however report just the SED's of candidate counterparts). This allowed us to rapidly decide which is the NIR counterpart in almost all cases. In the following, we discuss in detail all the observed fields; NIR source numbers refer to those given in Table 4.

3.1.1 IRS 13

This field is, at the moment, the only one for which we could not obtain a reliable astrometry; in fact, the DSS plate of the region contains too few stars with a NIR counterpart in the K frame, confirming that the extinction is large in this area (A_V> 30 mag, as reported in appendix to Paper I). However, the most prominent structure in the field is a reflection nebula (about 15$\hbox{$^{\prime\prime}$}$ $\times$ 15$\hbox{$^{\prime\prime}$}$) hosting at least 4 point sources (# 22, 25, 27 and 29; see Fig. 3c). Out of these, source # 29 dominates the emission in the K and L' bands, and fits the above criteria (see Figs. 3a and c). In the colour-colour diagram (Fig. 3a) it lies in a region typically occupied by luminous Class I sources. Its J, H and K magnitudes (see Table 4) are equal, within the errors, to those obtained by single channel photometry (given in Paper I). Though this one appears as the main source contributing to the FIR flux, since its K brightness exceeds that of other sources of about a factor 5 at least, also # 25 has typical colours of Class I sources and # 22 and 27, which were not detected in the J band, could be protostellar objects of the same type as well. Although, as we will show later, we cannot exclude that such red objects with lower limits at J may be heavily obscured background stars, it is quite reasonable that all these sources belong to a very young embedded star cluster.

  

Figure 3: a) Colour-colour diagram, b) SED's of some of the possible NIR counterparts of IRS 13 (along with fluxes in the IRAS bands) and c) contour plot of the K flux around the IRAS uncertainty ellipse (dotted line). Contours are in steps of $\sim 1\sigma$ from $\sim 1\sigma$. The solid line in the colour-colour diagram marks the locus of main sequence stars (Koornneef 1983), whereas dashed lines are the reddening law according to Rieke & Lebofsky (1985); 10 magnitudes intervals of AV are indicated by crosses. Data point on the right of the two dashed lines represent sources with an intrinsic NIR excess. Upward arrows indicate lower limits in (J-H) and vertical segments with rightward arrows indicate sources with only upper limits at J and H

3.1.2 IRS 14

The IRAS uncertainty ellipse is roughly centred on a possible reflection nebula hosting a small cluster which contains more stars than reported by West (1980) from optical plates (his designation: ESO 313-N*10). A single point source, # 37, is by far the brightest object in the K and L' bands and, as shown in Fig. 4, is the best candidate to be the NIR counterpart of IRS 14. It clearly exhibits a NIR excess in the colour-colour diagram of Fig. 4a and its J, H and K magnitudes coincide with those given in Paper I. The colour-colour diagram indicates the presence of other sources with a possible NIR excess in a region that is typically occupied by Classical T-Tauri or Herbig Ae/Be stars; indeed, this means that a more plausible identification for IRS 14 is that it is a Herbig Ae/Be star. Furthermore, both the low mm-continuum flux (see Table 1) and the failing to detect CS(2-1) emission (see Paper I) suggest that this region may be more evolved with respect to IRS 13.

  

Figure 4: Same as Fig. 3, but for IRS 14

3.1.3 IRS 17

Again, the IRAS uncertainty ellipse coincides with an extended NIR source, probably a reflection nebula, and is roughly centred on the brightest object (# 57) in the K and L' bands, which lies inside the nebula itself. Source # 57 has the "right'' spectral index (see Fig. 5b) and fulfils the counterpart requirements. Its colours (Fig. 5a) are typical of Class I sources. Our photometry coincides reasonably, within the errors, with the single channel photometry (Paper I). Even if this object is surrounded by other sources satisfying some of the counterpart requirements (# 40, 49, 67, 25; see Fig. 5b), these are fainter, with the brightest ones essentially located to the south and south-west of the ellipse and not symmetrically arranged around it; thus # 57 appears as the main source of FIR flux. However, we note that # 40, an object with a NIR excess and also detected in the L' band, has a K flux which is only a factor of 2 less than that of #57. Its protostellar nature is confirmed by NIR images in a narrow band centred at the H₂ v=1-0 S(1) line (2.12 $\mu$m) showing a jet which appears to be driven by the source (Massi et al. 1997). The colour-colour diagram indicates the presence of other sources with NIR excesses and colours typical of classical T-Tauri stars, Herbig Ae/Be stars and Class I sources, suggesting this field is an extremely young embedded star cluster.

  

Figure 5: Same as Fig. 3, but for IRS 17

3.1.4 IRS 18

The best NIR counterpart candidate, as shown in Fig. 6, is source # 119, which lies at the margin of the uncertainty ellipse and has the greatest spectral index. Our photometry indicates that the previously found NIR source (Paper I), corresponding to our # 137, must be discarded because of both its negative spectral index (see Fig. 6b) and its lack of a NIR intrinsic excess (see Fig. 6a). In fact, according to Fig. 6a its colours are more or less compatible with those of a reddened A0 V star. Source # 138, which lies towards the north-eastern edge of the ellipse, and source # 121, located within the ellipse, also fulfil some of the counterpart criteria and could contribute to the FIR emission (as suggested by the location of the IRAS ellipse, straddling # 119, # 121 and # 138), though their K fluxes are 3.5 and 5.5 times less than that of # 119, respectively. We note the presence of another object (# 176) with possible colours of a luminous Class I source (its J magnitude is a lower limit), even brighter than # 119 in K, but lying $\sim 30 \hbox{$^{\prime\prime}$}$ north-east of the IRAS ellipse centre, then probably unrelated to the bulk of FIR flux. No images in the L' band are available.

It is interesting to note that the presence of main sequence stars just towards IRS 18 (source # 137) may serve to constrain the distance to the region. As said, JHK colours indicate that # 137 may be an A0 V star with negligible extinction; since the GSC reports a star with V=12.38 mag in the same position, this confirms that the identification is roughly correct and that an A0 V star with intrinsic brightness $V \sim 10.8$ mag and A_V $\sim 1.5$ mag could account for the VJHK magnitudes. Allen (1976) quotes an absolute magnitude V=0.7 for an A0 V star, hence source # 137 results to be located at a distance of $\sim 1000$ pc. We can exclude a supergiant star since a B7 I, which has roughly the same VJHK colours as an A0 V (Koornneef 1983), would lie at a distance >16000 pc (which is much more than we can reasonably assume for the VMR; see Paper I). Considering a mean extinction gradient of 1.9 mag kpc^-1 in the Sun neighbourhood (Allen 1976), this means that the star has to be foreground; however, the DSS plate shows an optical cluster just towards IRS 18 which may represent the front end of a larger star aggregate belonging to VMR-D, so the given value is not just a lower limit for the cloud distance. Then, this estimate is in agreement, within the uncertainties and the simplifying assumption of an A0 V star, with the distance to the VMR-D of $700 \pm 200$ pc given in Paper I.

Throughout the examined fields we have found objects with only a lower limit at J, hence with either an intrinsic NIR excess and typical colours of Class I sources, or very large extinctions (A_V $\sim 30-40$ mag) if falling within the main sequence reddening band in colour-colour diagrams. Using source # 176 as a test case, we cannot exclude, only on the basis of the K fluxes, that such objects are heavily obscured background stars. In fact, assuming it is a reddened star, we can estimate an extinction A_V $\sim 45$ mag from the colour-colour diagram and derive its intrinsic V magnitude (using the colours given by Koornneef 1983). Then, comparing the obtained value with absolute visual magnitudes of a wide range of spectral types and luminosity classes (Allen 1976), we checked that supergiant stars yielding the same obscured K flux should be located at distances in excess of $\sim 2000$ pc, well behind the VMR-D cloud. Only small differences arise in terms of extinction from the fact that the locus of supergiant stars does not coincide with the main sequence.

  

Figure 6: Same as Fig. 3, but for IRS 18

3.1.5 IRS 19

The northern part of an extended source, again a possible reflection nebula, is located inside the eastern half of the IRAS uncertainty ellipse. This nebula hosts the brightest point source in the field, # 49, which, as clearly shown in Fig. 7, largely dominates the K flux and obeys the counterpart criteria. In the L' image it appears as the most prominent source. Its JHK magnitudes are also equal, within the errors, to those found through single channel photometry (Paper I), meaning the previous identification is correct. All other sources within or around the ellipse are at least one order of magnitude fainter in the K band and have spectral indices less than that of source # 49. Thus, they are unlikely to substantially contribute to the FIR flux. The identification of # 43 (towards the southern part of the nebula; see Fig. 7) as a point source may be questioned and it could just represent the brightest part of the reflection nebula. The colour-colour diagram shows a large number of objects with possible NIR intrinsic excesses, suggesting, also in this field, the presence of a young embedded cluster.

  

Figure 7: Same as Fig. 3, but for IRS 19. An asterisk in a) indicates the colours of the observed nebula, whereas the curved line represents a $\lambda ^{-4}$ scattering law drawn by varying $\tau_{0}$ (see Sect. 4.2)

3.1.6 IRS 20

Here, the IRAS uncertainty ellipse falls roughly among 3 extended nebulosities (see Fig. 8c), two of them lying within the ellipse, at least partially. The southernmost nebula, whose photometry coincides with the single channel photometry, was previously identified as the NIR counterpart of IRS 20 (Paper I). Yet, as shown in Fig. 8c, this nebulosity is located at the edge of the IRAS ellipse and hosts no star-like objects. In fact, we doubt that sources # 82, # 96 and # 109 are real stellar objects and suspect they correspond to the densest regions in the nebulae. Other sources (# 73 and 74; see Table 4) could be blobs of nebular emission, as well. We think that source # 98, which has the steepest spectral index and is near to the ellipse centre represents the best counterpart candidate. No images in the L' band are available. We note that # 93, which however lies $\sim 46 \hbox{$^{\prime\prime}$}$ west of the ellipse centre (hence, out of the field depicted in Fig. 8c), also may have the colours of a Class I source (depending on the real value of its J brightness, which is a lower limit in magnitudes). Similar argumentations as for IRS 18 suggest that a heavy reddened background star cannot be rejected in this case.

The colour-colour diagram of Fig. 8 shows a high dispersion around the reddening band, but most of the sources located on the left of the band have large error bars, such as a few of those on the right. On the contrary, the sources within the reddening band have well defined colours and display a high degree of extinction. We note also that a number of sources are present with only lower limits at J and H. Hence, it is quite likely the existence of a young embedded star cluster.

  

Figure 8: Same as Fig. 3, but for IRS 20. The asterisks in a) mark the colours of the observed nebulae, whereas the curved line represents a $\lambda ^{-4}$scattering law drawn by varying $\tau_{0}$ (see Sect. 4.2)

3.1.7 IRS 21

The IRAS uncertainty ellipse overlays a group of point sources, and is off-centred towards the south with respect to the latter. The brightest objects of the group (in the K band) are located within the ellipse or close to its southern edge (see Fig. 9c), meaning they contribute to the most of the FIR flux. Source # 50, which has a lower limit at J, anyway fits the counterpart criteria, although the source previously identified through single channel photometry (see Paper I) is probably # 32. The latter, however, is outside the ellipse, has the colours of a reddened main sequence stars (see Fig. 9a) and has a smaller spectral index with respect to source # 50 (see Fig. 9b). Unfortunately, due to the same arguments as used for IRS 18, we cannot exclude the possibility that # 50 is a very reddened background star.

A contribution from source # 27 must have affected the single channel photometry in the K band, since this object is only a few arcsec from # 32. Because of its spectral shape and typical colours of a luminous Class I source, this appears to be a counterpart candidate (see Figs. 9a,b); nevertheless, its location with respect to the ellipse may indicate that source # 50 is prominent at FIR wavelengths. Similarly, sources # 47 and # 35, which have the colours of Class I sources, lie to the south or to the south-east of the IRAS uncertainty ellipse. However, they witness again the existence of a young embedded star cluster. No images in the L' band are available.

  

Figure 9: Same as Fig. 3, but for IRS 21. An asterisk in a) marks the colours of the observed nebula, whereas the curved line represents a $\lambda ^{-4}$scattering law drawn by varying $\tau_{0}$ (see Sect. 4.2)

3.1.8 IRS 62

Figure 10b suggests that source # 27, coinciding with the one already found through single channel photometry (as indicated by its J, H and K magnitudes; see Paper II), is a suitable candidate. However, as shown in Fig. 10c, it is clearly outside the IRAS ellipse, so # 26, which is very close to the ellipse centre and has a similar spectral index, may be the real counterpart. Both objects have colours which are typical of luminous Class I sources, yet # 27 displays equal fluxes at K and 12 $\mu$m, so we cannot exclude that it contributes a not negligible part to the FIR flux. Indeed, the IRAS compactness parameter is E at 12 $\mu$m (see Paper II), suggesting a slight deviation from point-likeness at this wavelength. No images in the L' band are available.

  

Figure 10: Same as Fig. 3, but for IRS 62

3.1.9 IRS 63

In this case, the IRAS uncertainty ellipse does not include point sources, and clustering, if any, is small. There are only a few objects lying at the edge of the ellipse (see Fig. 11c) and, furthermore, source # 27, which dominates in the K band and whose J, H and K magnitudes roughly coincide with those found through single channel photometry, clearly has a negative spectral index (it is also plainly visible in the DSS plate). The reddest objects are # 33, 11 and 21, but these are outside the ellipse. Source # 33 lies about $20 \hbox{$^{\prime\prime}$}$ east of the ellipse centre and appears embedded in a faint nebulosity (see Fig. 11c), source # 21 is the nearest to the ellipse whereas source # 11 is about $40 \hbox{$^{\prime\prime}$}$ south-east of the ellipse centre. The protostellar nature of # 22, an object undetected at J and H, is suggested by narrow band NIR images which show H₂ v=1-0 S(1) compact emission coinciding with this source (Massi et al. 1997). Although similar argumentations as for IRS 18 can be used in order to show that # 21 could be a heavily reddened background star, given its closeness to the IRAS ellipse and its slightly steeper SED in the NIR, this source appears, at the moment, the best counterpart candidate; an important contribution to the FIR flux may come from sources # 33 and 11, as well. No images in the L' band are available.

  

Figure 11: Same as Fig. 3, but for IRS 63

3.1.10 IRS 66

As shown in Fig. 12a, source # 31 has typical colours of Class I sources; however, it lies $\sim40 \hbox{$^{\prime\prime}$}$ from the ellipse centre. No other sources fulfil the requirements to be considered as NIR counterparts and, conversely, their colours are compatible with those of reddened main sequence stars. In Paper I (see Appendix) it was speculated that the FIR flux may in part arise from the HII region RCW 32; this could also explain the larger positional error with respect to the IRAS ellipse. Otherwise, the real counterpart should have . No images in the L' band are available.

  

Figure 12: Same as Fig. 3, but for IRS 66

3.1.11 IRS 67

Source # 61 fulfils the counterpart criteria, and has the colours of a Class I source (as judging from Fig. 13a). Previously, no counterparts had been found within K=12.5 mag (see Paper II). All other sources surrounding the IRAS uncertainty ellipse are very faint, in fact most of the points on the left of the main sequence in the colour-colour diagram have large error bars, as a number of point on the right. Nevertheless, some of the NIR excesses on the right appear to be better established, suggesting there may be a few classical T-Tauri stars. No images are available in the L' band.

  

Figure 13: Same as Fig. 3, but for IRS 67

3.1.12 IRS 71

Source # 66 is located near the center of the uncertainty ellipse, but probably it is # 85, north of the ellipse, that was found through single channel photometry (see Paper II). However, a few objects either show a possible IR colour excess, or appear as very reddened stars (see Fig. 14a), or have a greater spectral index (see Fig. 14b), namely # 44, 55 and 70. All these sources lie in proximity of the eastern edge of the ellipse, and may contribute to the FIR flux. We note the presence of sources # 59 and 61 at the eastern edge of the ellipse, which are very reddened, remaining undetected in the J and H bands: they probably have the greatest spectral index, as shown in Fig. 14b. Thus, # 59, the brightest one, may be the main contributing source to the FIR flux. No images in the L' band are available.

  

Figure 14: Same as Fig. 3, but for IRS 71

3.2 Comparison with the results of Papers I and II

As seen in the previous section, 4 (out of 10) of the NIR objects indicated as counterparts of IRAS sources in Papers I and II have been confirmed by our NIR data, whereas in 2 fields no suitable candidates had been found. The 4 sources show mean differences between aperture photometry (this paper) and single channel photometry (Papers I and II) of K=0.1, H=0.7 and J=1.2 mag, respectively. The greatest differences are for IRS 17 in K (0.8 mag), H (1.6 mag), and J (2.8 mag). These discrepancies are due to both confusion in the 15$\hbox{$^{\prime\prime}$}$ beam used in single channel photometry and differences in the sky areas used to estimate the background values. In this sense, single channel photometry and aperture photometry coincide within errors for all these sources.

As for the remaining 6 fields, the identifications given in Papers I and II are wrong and the (possible) NIR counterparts we have found are significantly fainter in all 3 bands with respect to the previous ones. Almost all fields contain one or more objects which, in a colour-colour diagram, fall on the upper right corner, a region typically occupied by Class I sources (Lada & Adams 1992), and this cannot be purely coincidental. Whereas usually one of these red sources dominates in the K band, excepted towards IRS 63 and IRS 71, there are 3 fields (namely, IRS 21, IRS 62 and IRS 66) in which ambiguities arise since the brightest of the red objects do not lie within the IRAS uncertainty ellipse, and in 2 of these cases fainter sources with NIR excesses do exist much closer to the ellipse centre. One of the IRAS sources, IRS 14, can be classified as a Herbig Ae/Be star rather than a Class I source (see also Appendix).

Only among IRAS sources with Jy (number codes in our internal classification) we find counterparts with K <10 mag, whereas IRAS sources with Jy (number codes > 55) tend to have K >10 mag. However, no clear correlation is evident between IRAS and NIR fluxes, as can be deduced from Table 2, which lists the VMR-D IRAS Class I sources along with their NIR counterparts, sorted according to the 60 $\mu$m flux. The coordinates of the NIR counterparts are given in the table, as well; they are hereafter assumed as positions of the newly identified Class I sources.

  

Table 2: NIR counterparts of the VMR-D IRAS Class I sources. The latter are named according to our internal classification, whereas the former are numbered following Table 4. CC indicates the IRAS compactness parameter (see Paper I). The coordinates of the NIR counterparts are also given

3.3 mm data

The 1.3 mm observations can be used to determine the masses of circumstellar envelopes, since dust emission is likely to be optically thin at this wavelengths. Adopting the formalism of Hildebrand (1983), the total gas mass, $M_{\rm env}$, is given by:

 

$$M_{\rm env}=\frac{F_{\nu}\cdot D^2} {k_{\nu}\cdot B_{\nu}(T_{\rm d})}$$ (1)

where $F_{\nu}$ is the observed flux density, D is the distance to the source, $B_{\nu}(T_{\rm d})$ is the Planck function and $k_{\nu}$ is the opacity per unit (gas) mass. Ossenkopf & Henning (1994) discuss the best choice for $k_{\nu}$ on theoretical grounds, concluding that a value $\sim 1\times(M_{\rm dust}/M_{\rm gas})$ cm² g^-1, is likely in dense circumstellar envelopes, which, for a ratio $M_{\rm dust}/M_{\rm gas} = 0.01$, is greater than the interstellar medium opacity (0.0026 cm² g^-1), because of grain coagulation and ice mantle growth over them. The envelope masses for the 12 Class I sources have been calculated assuming a dust temperature $T_{\rm d} = 30$ K and are listed in Table 3. The uncertainty on $k_{\nu}$, as quoted by Ossenkopf & Henning (1994), should amount to a factor of 2, though an upper limit on $M_{\rm env}$ can be set using the interstellar medium dust opacity (yielding masses 5 times greater than indicated in Table 3). Since the telescope was pointed towards the IRAS uncertainty ellipse centres and the beam is roughly comparable with the ellipse sizes, this could result in a flux (i.e., mass) underestimate whenever the identified NIR counterpart lies far from the IRAS uncertainty ellipse centre.

  

Table 3: Bolometric luminosities and circumstellar envelope masses derived from 1.3 mm dust emission

Finally, we can evaluate whether the values derived for the envelope masses are consistent with the central masses derivable from the observed luminosities. The bolometric luminosities obtained by integrating the observed SED's from 1.25 $\mu$m (J band) to 1.3 mm are reported in Col. 2 of Table 3, excepted for IRS 71 (because of un upper limit at 100 $\mu$m) and IRS 66 (no millimetric data). These are increased by a factor $4 \pm 1$ with respect to the values given in Papers I and II. As expected, the bolometric luminosities are dominated by the emission at IRAS and submm wavelengths. In addition, as a consequence of our selection, the luminosities are relatively high (120 $\leq$ $L/L_{\odot}$ $\leq$ 5600) and probably reflect central masses in the range 3.5 < $M/M_{\odot}$ < 10, according to the models of Palla & Stahler (1993), i.e. protostellar candidates of intermediate mass. Noting that (see Table 3), we can conclude that the envelopes seem to have masses which are not a small fraction of that condensed on the central object. This finding indicates these sources might have not yet accumulated most of their stellar mass.
In summary, the NIR images and the 1.3 mm photometry allow us to determine coordinates (given in Table 2) and SED's for 11 newly identified Class I objects and a Herbig Ae/Be star.