4.1. Cold dust in molecular clouds
We zoomed to the Chamaeleon part of the Fig. 1 b for a detailed view on the coldest regions and present the colour temperature distribution as derived from the background subtracted and intensities.
The average intercloud dust colour temperature in Chamaeleon is 16 K (Fig. 1, Fig. 3), in between the 17.5 K of galactic cirrus and the 15 K cold dust component derived by (based on DIRBE (Lagache et al. (1998))). The dust of around and below 14 K is strictly concentrated inside the well known molecular clouds and is not found in the diffuse intercloud medium. The average intracloud temperature of Cha I, Cha II and Cha III of 14 K is close to the DIRBE based values of 14.2 K-14.9 K (Boulanger et al. (1998)).
Fig. 3 can be used to locate molecular clouds since the regions which show colour temperatures 15 K are located inside the large CO clouds of Dame et al. (1987) and are associated with 13CO (J=10) line intensities (see also contour lines in our Fig. 3 ; Mizuno et al. (1998)). Furthermore, all the cold regions are opaque on the Digitized Sky Survey, with an average extinction of 2 mag which is characteristic for molecular clouds.
We have investigated the functional relationship between and the visual extinction, using the extinction maps of Cambresy et al. (1997) and Cambresy (1999b), which are based on J band star counts. We have found that the linear dependence between and holds for extinction values up to 7 mag, i.e. is an excellent tracer of dust column density as seen in Fig. 7.
Boulanger et al. (1998) found in the Chamaeleon main clouds a linear correlation between and blue starcount based , which, however, holds for 2 mag only. The better correspondence of with as compared to is a result of two effects: (i) 170 µm emission comes from the large grain population only, which is also responsible for the optical extinction, whereas at 100 µm smaller grains may still contribute. (ii) In high density regions the effective shielding of the ISRF leads to colder dust, which is seen better at the longer wavelength.
As measures a single dust component, we have correlated it with the quantity in order to remove contributions from other grain components to the 100 µm sky brightness. was introduced by Boulanger et al. (1998) based on Laureijs et al. (1991). It quantifies the excess of over and was meant to trace the pure cold dust 100 µm emission on the basis of the IRAS bands. For this purpose, we have extracted 60 µm ISSA intensities at the same slew positions used already for the vs. comparison (and scaled similar as before: ).
The scatter plot (Fig. 8) reveals a linear correlation between and . The 30 offset at the -axis is due to the large scale galactic background. The scatter is driven by the temperature variations inside the Chamaeleon clouds due to different degrees of attenuation of the ISRF (e.g. see the VCCs). There is no indication that either of the two quantities is affected by other than the classical dust grain component. This finding justifies the widely used way of locating cold regions by their IRAS based 100 µm excess. This is particularly important, as the area filling factor of cold clouds is usually low - in the investigated field only 3 %. Therefore one needs the resolution of IRAS to detect this kind of objects.
The slope of the fitted correlation line can be converted to dust colour temperature as in Fig. 1. We have found = 13.1 K, which is slightly colder than the 14 K derived in Sect. 3. The small difference may come from different calibration procedures of Boulanger et al. (1998), affecting the conversion factors entering the definition of . If the difference is a trace of an contamination by other grain components, this effect would in principle also apply to the temperatures of the VCCs (Sect. 4.2). But as the VCCs are not visible in the ISSA 60 µm maps, the correction is negligible.
4.2. Very cold cores
Two of the very cold cores (VCCs) are included in the ISOPHOT P22 maps of Lehtinen et al. (2000). See their paper for details on the measurement and calibration. The VCCs no. 3 and no. 4 are both seen on their sized 100 µm 150 µm and 200 µm maps (see Fig. 9). The background FIR radiation was subtracted using the average brightness of pixels surrounding the ellipse shaped cloud cores. The FIR spectral energy distributions of these cores indicate an absence of warm dust contribution, as it is shown in Fig. 10. The excess surface brightnesses associated with the cores are significantly different but the dust temperatures of the two cores are both K. This PHT22 based dust temperature value is similar to the ISOSS results, within the uncertainty limits. It confirms the classification of these cores as cold sources. The position and extent of these VCCs are approved, as well as the fact that they are colder than their surrounding. The difference in the PHT22 and ISOSS based temperature estimates may rise both from the methods deriving the temperature, and the difference in the ways of eliminating the FIR background.
The positions of the VCCs have been compared with visual extinction and C18O data (see Fig. 6). The map resolves the internal structure of the cloud, as the near-infrared star count method can trace the visual extinction up to 11 mag. All but one of the VCCs are associated with high extinction peaks ( 4 mag, see also Table 2). Additionally, all the VCCs lie inside the CO K km s-1 contour of the 13CO (J=10) line intensity (see Fig. 3 and Mizuno et al. (1998)), and associated with C18OKkms-1 C18O line intensity peaks. Hence, our VCCs are indeed opaque, high column density molecular cloud cores. We may assume their line of sight diameters is 0.2pc just as their mean half power diameter is (see Table 1 and the distance), and that half of the observed C18O column density towards the VCCs is actually by the VCCs itself. The VCC gas densities would then be around Hcm-3.
Table 2. Physical parameters of the very cold cores. based on B starcounts (Toriseva & Mattila (1985))
The distribution of dense gas is shown best by Mizuno et al. (1999), who presented full coverage mapping observations of the three Chamaeleon main clouds. We have estimated the physical parameters of the gas associated with the VCCs from their 13CO and C18O spectra by Mizuno (1999), using the measurements nearest to the VCC centres and assuming LTE, thermalised CO, terrestrial isotopic ratios, and .
The dust and gas parameters are compared in Table 2, where the columns are: (1) Running number as in Table 1, (2) ISOSS name, (3-4) average and peak visual extinction (Cambresy et al. (1997); Cambresy (1999b)), (5) dust colour temperature, derived from column 8 of Table 1, errors in are 1 - 2 K (6) C18O excitation temperature, (7) optical depth of C18O (at line centre), (8) H2 column density, derived from C18O.
Seven of the nine VCCs have been found cold with . However, can provide only a first order estimate of the kinetic temperature, since C18O is expected to be sub-thermal in dense cores. NH3 is a better indicator of gas temperatures in dense cores. An ammonia survey of the VCCs would thus be necessary to confirm the low gas kinetic temperatures.
A general correlation is not expected between gas temperature and colour temperature of large dust grains. The heating and cooling mechanisms are different. Gas - dust interactions play a dominant role in the thermal balance only at high densities, and gas - dust encounters are most frequent between molecules and the more abundant smaller dust grains. Both gas and dust might, however, cool down below 15 K already at moderately high densities of n(H2) 104cm-3 when the incident flux is low (Krügel & Walmsley (1984)). We claim that this is the case in the dense parts of the Chamaeleon clouds, where the shielding from the outside radiation field is effective and the star forming regions do not form massive stars. The radiation from the low mass stars heats the ISM only in their vicinity.
The very cold cores have very low gas temperatures as it is seen from molecular line measurements (Table 2). To compare with the calculations of Krügel & Walmsley (1984) we assume that the mean intensity of the ISRF impinging at the Chamaeleon VCCs is one third of the mean ISRF in the solar neighbourhood i.e. ergcm-2s-1 (Metzger (1990)), and that the UV radiation of hot stars and HII regions is effectively shielded by the ISM into which the VCCs are embedded. Thus we assume a soft radiation field ( K) with low incident flux of ergcm-2s-1. For such a radiation field, a grain size distribution (Mathis et al. (1977)), and a gas density of the order Kr"ugel & Walmsley (1984) calculated gas and dust temperatures of K and K respectively which are in the range we found for our VCCs. Hence, we regard our method for finding very cold FIR sources as an effective way to locate the coldest cloud cores in terms of gas temperature, too.
4.3. Completeness and reliability of the cold source search
The C18O survey of Mizuno et al. (1999) fully covers the three Chamaeleon main clouds with an angular resolution (2.7´) similar to ours. We have therefore used their list of 23 C18O cores to investigate the completeness of our method to locate cold cloud cores: 15 of the 23 cores have been crossed by ISOSS slews. All these 15 cores have been detected. 11 (73 %) are cold with , 6 (40 %) are associated with VCCs.
The characteristics of the C18O cores differ in the three clouds. We therefore compare the gas and dust properties in the most densely covered Chamaeleon I region, where all six C18O cores have been crossed by ISOSS slews: The C18O cores no. 1, 2 and 4 have been identified as VCCs no. 1, 2 and 6, respectively. C18O core no. 5 contains the two separated VCCs no. 3 and 4 (Fig. 6).
While the four C18O cores of Chamaeleon I associated with VCCs have colour temperatures 13 K (Table 2) and C18O core no. 6 is associated with sources of , C18O core no. 3 has (warm sources are not indicated in the Fig. 5 and Fig. 6). This core, which contains much warmer dust than all others, is seen towards a peak in surface density of young stellar objects.
4.4. Two sources of Cha I with warm dust
In this paper we have not intended to test the capabilities of the ISOSS/IRAS analysis studying the dust in warm regions. The two warm sources seen in Cha I are both heated internally by embedded young stellar objects. CED 112 is at the southern edge of a group of pre-main-sequence stars (North et al. (1996)) with mid-infrared excess (Persi et al. (2000)). The bipolar nebula Cha IRN was classified by Ageorges et al. (1996) as a free-falling cloud molecular core with bipolar outflow cavities and embedded young star-disk system. Cohen & Schwartz (1984), using KAO FIR measurements with effective FWHM beamsizes of about 45" derived 65 K as the temperature of the coldest dust component seen there. The fluxes of these warm spots are mainly caused by point sources. The colour temperatures derived from our 5´ angular resolution ISOSS/IRAS data (24 K and 19 K respectively) are due to a mixture of radiation of circumstellar and interstellar dust seen in the beam with various temperatures. The physics of large dust particles in heated environment can be studied by ISOSS at regions where external heating is dominant, i.e. the heating source can be separated with our resolution.
© European Southern Observatory (ESO) 2000
Online publication: January 29, 2001