4 Results and discussion

 

 

Table 1: The radio continuum sources detected in DC 303.8-14.2. Flux upper-limits are calculated assuming a point source with a peak intensity of twice the map rms. Note that source $\char93$1 is located outside the 3 cm primary beam. The positional accuracy is expected to be better than 2 $\hbox {$^{\prime \prime }$ }$. Source $\char93$4 is coincident with IRAS 13036-7644. The deconvolved angular size is the measured FWHM along the major and minor axes at 6 cm (or at 3 cm if 6 cm data not available) with uncertainty in parentheses. All the sources are unresolved. The quantity S is the intensity integrated over the source. The errors of S and $\alpha$include only the noise of the data, not the calibration uncertainty.

	Position		Angular	S(3 cm)	S(6 cm)	$\alpha$
Source number	$\alpha$(1950)	$\delta$(1950)	size [ $\hbox {$^{\prime \prime }$ }$]	[mJy]	[mJy]
1	$13^{\rm h} 01^{\rm m} 01\hbox{$.\!\!^{\rm s}$ }79$	$-76\hbox{$^\circ$ }40\hbox{$^\prime$ }10\hbox{$.\!\!^{\prime\prime}$ }6$	$5.7\times4.5$ (0.7)	<0.48	$3.22\pm0.34$	<-3.2
2	$13^{\rm h} 01^{\rm m} 36\hbox{$.\!\!^{\rm s}$ }58$	$-76\hbox{$^\circ$ }47\hbox{$^\prime$ }57\hbox{$.\!\!^{\prime\prime}$ }5$	$6.3\times4.5$ (0.9)	$0.30\pm0.13$	<0.16	>1.1
3	$13^{\rm h} 02^{\rm m} 36\hbox{$.\!\!^{\rm s}$ }46$	$-76\hbox{$^\circ$ }43\hbox{$^\prime$ }27\hbox{$.\!\!^{\prime\prime}$ }2$	$5.0\times3.1$ (0.3)	$1.00\pm0.17$	$1.55\pm0.15$	$-0.7\pm0.1$
4	$13^{\rm h} 03^{\rm m} 42\hbox{$.\!\!^{\rm s}$ }11$	$-76\hbox{$^\circ$ }44\hbox{$^\prime$ }08\hbox{$.\!\!^{\prime\prime}$ }7$	$5.3\times4.2$ (0.5)	$0.92\pm0.11$	$1.16\pm0.11$	$-0.4\pm0.1$

Figure 1: A map of the integrated intensity of 12CO (J=1-0) emission over the velocity intervals 0-1 (solid contours) and 6-7 km s-1 (dashed contours). The contours are from 0.1 to 0.8 in steps of 0.1 K km s-1. The radio continuum sources are indicated with filled circles, with numbers referring to the list in Table 1. We identify source #4 with IRAS 13036-7644. Source #1 is located outside this map. The (0,0) position is at RA $(1950)=13^{\rm h} 3^{\rm m} 41\hbox{$.\!\!^{\rm s}$ }4$, Dec. $(1950)=-76\hbox{$^\circ$ } 44\hbox{$^\prime$ }18\hbox{$.\!\!^{\prime\prime}$ }0$.

We detected seven sources in all, but only four within DC 303.8-14.2's boundaries. Of these, two were detected at both 3 and 6 cm. All the sources were unresolved, and no large scale structure was apparent in the maps. We therefore used the AIPS task IMFIT to derive fluxes and positions, along with their uncertainties, by fitting 2-dimensional Gaussians to each source. We have verified the unresolved nature of the sources by making contour plots of the sources along with the plots of the beams. The measured positions, deconvolved angular sizes, integrated fluxes and derived 3 cm-6 cm spectral indices are shown in Table 1. The errors of fluxes and spectral indices include only the statistical noise of the data, not the calibration uncertainty. Note that the source #1 is situated beyond the 3 cm primary beam, and is located in the outer parts of the optical extent of DC 303.8-14.2. When calculating flux upper-limits for sources #1 and #2, which were not detected at both wavelengths, we assumed a point source with a peak intensity twice times the map noise level.

The source #3 has a spectral index $\alpha=-0.7$, indicating non-thermal spectrum. Interestingly, it is located exactly on the axis of the blue-shifted outflow component, and raises the question whether it is related to the outflow, because non-thermal emission from Herbig-Haro objects has been detected (Reipurth 1991). In this context, however, we do not consider this source any longer.

Source #2 is detected only at 3 cm. Using a 2$\sigma$ value for the flux at 6 cm, its spectral index is $\alpha>1.1$, i.e. the emission is thermal, possibly from a source located within the cloud.

4.1 Emission diagnostics for the IRAS source

Source #4 is located within the positional error ellipsoid of IRAS 13036-7644, thus we have identified these sources to be the same. The observed flux is typical for YSOs located in Bok globules (Yun et al. 1996). The spectral index $\alpha$ of this source is $-0.4\pm0.1$, which indicates non-thermal emission, but is close to the value of $\alpha$ for optically thin thermal emission, -0.1. A calibration error of 10% at both wavelengths is enough to bring the spectral index into the regime of optically thin thermal emission. We note that if we simply take the peak value of the observed surface brightness, which for point sources is the same as the total flux, we derive a spectral index of -0.1. We conclude that within uncertainties, the spectral index of the source #4 is consistent with optically thin thermal emission.

In the case that a source is resolved, the most secure way to discriminate between thermal and non-thermal emission would be the brightness temperature: thermal emission is charecterized by low brightness temperature, $T_{\rm b} \leq 10^{4}$ K, while non-thermal emission is charecterized by high brightness temperature, $T_{\rm b} \geq 10^{7.5}$ K. However, none of the measured FWHM for our sources are sufficiently larger than the synthesized beams to qualify as being resolved.

The value of $\alpha$ rules out the Bertout's (1983) accretion model, and the "standard'' spherical wind model, which both have $\alpha \approx 0.6$. We do not consider in detail the nonspherical stellar wind models (Schmid-Burgk 1982; Reynolds 1986) due to complexities of these models compared with our sparse data. We just note that for a fully ionized collimated flow, in the case of constant temperature, velocity and ionization fraction, i.e. the "standard'' case except that the flow is confined, the spectral index has the form

$$\alpha = 1.3 - 0.7 / \epsilon$$ (3)

where $\epsilon$ is the power-law index which determines the dependence of the width of the flow (w) on the distance from the central object (r), $w \propto r^{\epsilon}$ (Reynolds 1986). If we use the value $\alpha=-0.1$, we derive $\epsilon = 0.5$.

In the following we discuss in more detail the models which predict a spectral index $\alpha$ compatible with our observations. A lower limit to the flux from accretion (Eq. (1)) can be predicted by setting the logarithmic term to unity. In reality we expect the condition $r_{\infty} \gg r_{\rm c}$ to be prevalent, i.e. the logarithmic term is much greater than unity. Assuming that T=10⁴ K, $M=1~M_{\odot}$, and d=200 pc, the mass accretion rate required to produce the observed flux is $\sim$ $4\times10^{-10}~{M}_{\odot}~\mbox{yr}^{-1}$. This rate is about four to five orders of magnitudes smaller than the typical rate expected for low-mass YSOs, $\sim$ $10^{-6}{-}10^{-5}~{M}_{\odot}~\mbox{yr}^{-1}$. Clearly, the flux predicted by this spherical, free-fall accretion model has to be reduced somehow, if it is to be the emission mechanism for a YSO with a typical mass accretion rate. The flux can be reduced if e.g. (1) the accretion is not spherical due to a disk and/or magnetic fields, as evidenced by observations in other systems and models, (2) the accretion is not described by free-fall motions due to e.g. magnetic fields, turbulence, outflow or winds.

The flux predicted by the shock emission model (Eq. (2)) has been calculated by assuming T=10⁴ K, $\dot{M}$ between $10^{-8}{-}10^{-7}~{M}_{\odot}~\mbox{yr}^{-1}$, $v_{\infty}=200~\mbox{km~s}^{-1}$, and d=200 pc;
$S_{\nu}\mbox{(shock)} \approx (\Omega/4~\pi) (0.2-2)$ mJy. Due to the many unknown parameters this estimate is only suggestive, but it shows that with typical parameters the shock emission model can produce the observed flux.

The predicted flux of thermal dust emission is about two orders of magnitudes less than the rms noise level in the maps. Furthermore the flatness of the spectrum between 3 and 6 cm is not consistent with thermal emission from dust. The 1.3 mm continuum emission around IRAS 13036-7644 is extended on a map made of 3$\times$3 grid points with a grid distance of 20$\hbox {$^{\prime \prime }$ }$, measured with the SEST (Swedish ESO Submillimeter Telescope) bolometer (Lehtinen et al.  unpublished data). Thus thermal emission from dust grains at cm-wavelengths would also be extended. We can exclude the dust emission as mechanism for source #4.