Extended sources

G003.6-0.11

This source, seen just west of G003.7$-$0.2 in the 1428 MHz image by Gaensler (1999), is also detected as a resolved source at 327 MHz. The GMRT 327-MHz image and the VLA 1428-MHz images of this sources are shown in Fig. 6.1. The resolution in the 327- and 1428-MHz images is $\approx 20\times11 {\mathrm{arcsec^2}}$ and $15\times9 {\mathrm{arcsec^2}}$ respectively. This source was also observed by Yusef-Zadeh (private communications) using the VLA in the D-array configuration at 4.9 GHz. We therefore mapped this source at 4.9 GHz using the data acquired from the VLA achieves (Fig. 6.2). The size of this source is $\sim1{^\prime}$ with the centre located at ${\mathrm{RA}_{J2000}}=17^h54^m32^s$, ${\mathrm{Dec}_{J2000}}=-25{^\circ}51{^\prime}30 {^{\prime\prime}}$. There is significant emission in the IRAS $60\mu$m image peaking at ${\mathrm{RA}_{J2000}}=17^h54^m32^s$, ${\mathrm{Dec}_{J2000}}=-25{^\circ}50{^\prime}33{^{\prime\prime}}$ (Fig. 6.2).

Figure 6.1: The left panel shows the GMRT 327-MHz image of G003.6$-$0.1. The resolution in this image is $\sim15\times20 {\mathrm{arcsec^2}}$ and the RMS noise of $\sim 5$ mJy. The VLA image at 1428 MHz (Gaensler1999) is shown in the right panel. Resolution in this image is about $15 \times 9 {\mathrm{arcmin^2}}$. The shell type structure of this object is clearly seen in this image. The general morphology is same as that seen in the 327-MHz GMRT image and the source in the north-west is also resolved as a separate compact source.

Figure 6.2: The VLA 4.8-GHz image of G003.6$-$0.1 using the data acquired by Yusef-Zadeh in Feb. 1987 is shown in the left panel. The resolution in this image is $\sim 11\times 10{\mathrm{arcmin^2}}$, P.A.= $-70{^\circ}$. The general morphology of a compact core at the south-eastern edge of the lower surface brightness shell is clearly visible. The right panel shows the IRAS 60 $\mu$m image of this region. The peak in this image occurs at ${\mathrm{RA}_{J2000}}=17^h54^m32^s$, ${\mathrm{Dec}_{J2000}}=-25{^\circ}50{^\prime}33{^{\prime\prime}}$. With the coarse resolution of IRAS, this peak covers the entire radio source.

Radio Recombination Lines (RRLs) have been detected towards ${\mathrm{RA}_{J2000}}=17^h54^m30^s$, ${\mathrm{Dec}_{J2000}}=-25{^\circ}51{^\prime}23{^{\prime\prime}}$ near 5 GHz (H109$\alpha$ and H110$\alpha$ lines) by Caswell & Haynes (1987) using the Parkes $64$-m single dish with a HPBW of $\approx 4{^\prime}$ and by Lockman (1989) at $3$ cm with a HPBW of $\approx 3{^\prime}$. The RRL parameters reported by Caswell & Haynes (1987) are: $T_l/T_c=0.05/0.7$, $\theta=2{^\prime}$, $\Delta V=28$ km sec$^{-1}$, $V=3$ km sec$^{-1}$, $T_e=5600$ K, $D=1.5$ or 15.5 kpc (assuming the standard IAU parameters (Kerr & Lynden-Bell1986) for solar orbital velocity $\theta_\circ=220$ km sec$^{-1}$ and distance to the Galactic Centre $R_G=8.5$ kpc). However, since the Galactic longitude is small, the measured radial velocity could entirely come from non-circular motion which makes the distance estimates unreliable and must be treated as a nominal distance. No optical counterpart is detected implying that a distance of 15.5 kpc is more likely. The surveys by Downes et al. (1980) and Wink et al. (1983) for H76$\alpha$ and H110$\alpha$ transitions respectively do not detect any RRL towards G003.7$-$0.1 due to the sensitivity limits of the surveys ( $S_{4.9GHz} > 1$ Jy and $S_{14.7GHz} > 2$ Jy respectively). This source is listed as a HII region in the continuum survey done by Wink et al. (1982) at 14.8 and $4.9$ GHz (HPBW= $2{^\prime}.6$) and is also listed in the list of compact sources from the 11 cm Galactic plane survey by Fürst et al. (1990). The PMN source PMN J1754-2551 at ${\mathrm{RA}_{J2000}}=17^h54^m31^s$, ${\mathrm{Dec}_{J2000}}=-25{^\circ}51{^\prime}02{^{\prime\prime}}$ covers all the components of this source (Griffith et al.1994).

A circular shell of emssion with a peak of emssion along the southern edge of the shell along with two compact sources in the north is visibile in the 327-, 1428- and 4850-MHz images. For further discussion of the nature of this source, we label the peak of emission in the south-eastern part of the shell as A, the shell itself as B, the compact source on the northern rim of the shell as C and the moderately resolved northern most source as D. There is an indication of a connecting bridge of emission between C and D in all the three images. The disctinct semi circular gap around C, followed by a concentric arc seen in both the 5 and 1.4 GHz images is suggestive of an interaction between A and B/C. HI absorption spectra was measured using the GMRT towards A, C and D (Fig. 6.3). The HI absorption spectra for A and C is very similar, implying that they are at a similar kinematic distance. The extra $-27$ km sec$^{-1}$ feature in the spectra of D implies that this component is farther away compared to A and C. Kinematic distance of D corresponds to $>20$ kpc and could also be extragalactic. The $V_{LSR}$ of $3$ km sec$^{-1}$ for the RRL and the $+30$ km sec$^{-1}$ feature seen in the spectra of all the three components implies that the line of sight absorber at a velocity of $+30$ km sec$^{-1}$ is not associated with the RRL emitter. This absorber corresponds to a kinematic distance between 6 and 12 kpc. Assuming that the RRL is associated with extended nebula and the component A, a distance of 15.5 kpc is then consistent with the RRL and HI absorption velocities. The absence of the $-27$ km sec$^{-1}$ feature in the spectra of A and C places these components in front of the component D and consequently the component D at $>20$ kpc.

Figure 6.3: HI absorption profiles against the components A, C and D of G003.6$-$0.1 respectively, measured using the GMRT.

However, HI absorption at negative velocities are observed in the first quadrant of the inner Galaxy, possibly from clouds in the Galactic Centre region. If the observed RRL is not associated with the nebula or any of the components, and the $-27$ km sec$^{-1}$ feature is due to line of sight clouds with anomalous chaotic motions (Shaver et al.1982; Belfort & Crovisier1984), the HI absorption profiles are consistent with components A and C at a distance between $R_G$ and 6 kpc, while D at a distance $>R_G$. This ambiguity can be resolved with a high resolution RRL observation towards this direction to identify the RRL emitter.

Table 6.6: Peak flux densities for the components A, C and D for G003.6$-$0.1 at 0.327, 1.4 and 4.8 GHz.

Component	$S_{327MHz}$	$S_{1.4GHz}$	$S_{4.8GHz}$
	(mJy)	(mJy)	(mJy)
A	48.2	32.5	33.5
C	38.6	24.8	46.1
D	42.5	22.4	26.2

Figure 6.4: Continuum spectra of the components A, C and D of G003.6$-$0.1. Top panel shows the peak flux density spectra. The spectral index of component C between 1.4 and 4.8 GHz is $\sim 0.5$ while that of components A and D are 0.0 and 0.1 respectively. Bottom panel shows the integrated flux density spectra for the three components with the average 327-MHz flux density in the vicinity subtracted for the components D and C which are not distinctly detected as separate sources at 327 MHz.

The peak flux densities of the components A,C and D were measured from the 0.327, 1.4 and 4.8 GHz maps after convolving them to same resolution. The measured peak flux densities are listed in Table 6.3.1 and plotted in Fig. 6.4. Emission from the nebula itself is indicative of non-thermal emission. At 327 MHz, C does not appear to be distinct from the shell and the flux density at this frequency may be contaminated by that due to the shell itself. The 327-MHz flux density of D too may thus be contaminated. Subtracting the average flux density measured around these components may give an estimate of the 327-MHz flux density of these components. The bottom panel of Fig. 6.4 shows the integrated flux density spectra of the three components, with the average 327-MHz flux density in the vicinity of components C and D subtracted. A and D in this plot show non-thermal spectra. The spectrum of C is consistent with spectral index of $\sim 0.5$ between 327 MHz and 4.8 GHz. The spectral indices thus determined, however, must be treated as only tentative. Observations at other frequencies between 327 and 1400 MHz (e.g. 610 MHz) will help in clarifying the spectra and the nature of these components. The peak flux spectral index between 4.8 and 1.4 GHz ( $\alpha^{4.8}_{1.4}$) for the components A, C and D was measured to be 0.0, 0.5 and 0.1 respectively while the spectral index between 1.4 and 0.327 GHz ( $\alpha^{1.4}_{0.327}$) was measured to be $-0.3$, $-0.3$ and $-0.5$ respectively.

Could it be that the 327-MHz flux density is over-estimated and the true value is less than the value at 1428-MHz, as expected from a source of thermal emission? For this to happen, the 327-MHz flux density must be over estimated by a factor of about $\sim 2-3$, which is very unlikely. Another way the flux density can be over estimated is due to a slowly varying background emission. The average flux density in an approximately $1{^\prime}$ box around this source is $\approx 0.15$ Jy, comparable to the error bar of 0.1 Jy for the measured flux density. Hence, this also cannot account for an over estimated flux density. To further eliminate the possibility of a systematic flux density calibration error, we compared the flux densities of the VLA calibrator in the field located at $Dec_{J2000}=17^h51^m52^s$, $RA_{J2000}=-27^\circ24{^\prime}01.33{^{\prime\prime}}$ and the SNR G003.7$-$0.2. Unfortunately, this VLA calibrator is not a good P-band calibrator and hence the VLA 327-MHz flux density is not known. However, it is listed in the Texas catalogue, and at 365 MHz, its flux density is $1.41\pm0.09$ Jy corresponding to $1.14 \pm 0.09$ Jy at 327 MHz (spectral index of $1.8\pm0.7$). The measured flux density from the GMRT primary beam corrected image is $1.3\pm0.2$ Jy. The flux density measured for G003.7$-$0.2 (see section 5.3.1) was also found to be in good agreement with the extrapolated value at 327 MHz.

We now suggest a model for the compact source C. The class of symbiotic stars is defined by the basic characteristic of an optical spectrum containing both high excitation emission lines and absorption features of a cool, late-type star (Seaquist & Taylor1990). While most of these stars emit radiation at IR wave bands, few of these stars are also detected at radio frequencies. High frequency (GHz range) spectral index is invariably positive, ranging from 0 to 1.2 and the emission mechanism is thermal bremsstrahlung. This range of spectral indices can be explained by a simple binary model, where, the wind in the form of uniform mass loss from a cool star is ionized by a hotter companion star (Taylor & Seaquist1984). This model predicts a lower limit of 0.6 and an upper limit of 1.3 for the spectral index, depending upon the viewing angle of the binary system, the mass loss rate from the cooler star and the ionizing photon flux from the hotter star. Some fraction of the emission, however, might be optically thin and an observed optically thick spectral index lower than 0.6 may be consistent with the binary model. The radio spectra of most of these stars, above a few GHz, turn over to a relatively flat spectral index of $\sim 0.1$. A radio survey of such stars in the Galaxy done by Seaquist et al. (1984) found a mean spectral index of $+0.6$ and a cut off at 1.2, in excellent agreement with the binary model.

The high frequency spectral index ( $\alpha^{4.8}_{1.4}$) of component C is therefore consistent with it being such a radio loud symbiotic star. The binary-model for such objects relates the turn-over frequency and the optically thick spectral index to the physical properties, namely the mass-loss rate and the hydrogen ionizing photons flux. The present high frequency data on this source probably samples only the optically thick part of the spectrum. Continuum observations a few higher frequencies will be required to determine the true nature of this source.

About 20% of the UC H II regions mapped in the incomplete survey of Wood & Churchwell (1989b) and Kurtz et al. (1994) where of the cometary morphology. A typical example of this morphology is G034.2$-$0.2. van Buren et al. (1990) proposed a model of a bow shock created by a wind-blowing massive star moving supersonically through a molecular cloud. The required velocities of less than $\sim10$ km sec$^{-1}$ are comparable to the observed velocity dispersion of stars in OB associations. The gross structure seen in the radio continuum and the velocity structure in the hydrogen recombination and molecular lines is well explained by this model. Their model also make specific predictions about the OH maser spots in the leading edge of the shock as well as detectable proper motion of maser sources over a time scale of few years for the nearby UC H II region.

The higher frequency spectra of component A is flatter, typical of HII regions. The morphology of the nebula associated with the component A, as seen in the images at 1.4 and 4.8 GHz is suggestive of a cometary UC H II region. A typical UC H II region has a size of $<0.1$ pc, $n_e> 10^4\mathrm{cm^{-3}}$, and EM $>10^7\mathrm{pc cm^{-6}}$. The linear sizes of the nebula corresponding to the distances of 18.5 and 7 kpc is $\sim 1.5$ and $\sim 5$ pc respectively. Clearly, the size of the nebula in this field is much greater than the size of typical UC H II region. However, the extended emission seen towards this source could then have the same origin as the extended emission seen associated with UC H II regions in recent observations at 1.4 and 5 GHz (Kim & Koo2001; Kurtz et al.1999). EM for this sources, assuming $T_e=5600$K as derived from the RRL observations, is also smaller than the typical value for UC H II region. However, the $T_e$ estimates from the existing low resolution RRL observations may be underestimated.

The morphology in the 327-MHz image of this sources is markedly different from that at 1.4 and 4.8 GHz. The brightness contrast between the component C and the extended emission associated with the nebula is negligible. The structure of the nebula itself is replaced by two arcs of emission of comparable brightness. The flux density of all the three components at 327 MHz are higher than the values at higher frequencies, which is inconsistent with a purely thermal emission. The 327-MHz data is therefore suggestive of a foreground source of non-thermal emission. The spectra of A (Fig. 6.4) between 327 MHz and 4.8 GHz is also non-thermal with a spectral index of $\sim -0.4$. The spectral index, which is typical of shell-type SNRs, and the morphology seen in the 327-MHz image are consistent with the extended emission being a new SNR. High resolution observations at 233 and 610 MHz with the GMRT, along with higher resolution imaging with the VLA at higher frequencies will help in resolving the nature of the shell. A conclusive evidence of non-thermal emission could come from the detection of polarized radio emission.

Linear structure in the field of G356.3$-$1.5

Figure 6.5: Sub-image of the linear structure seen in the full primary beam image of the region containing G356.3$-$1.5. This structure is seen clearly in the low resolution image as well (Fig. 4.8) and appears to be a real feature and not an artifact.

The linear structure seen $\sim30{^\prime}$ north-east of the barrel shaped SNR G356.3$-$1.5 in the low resolution image is also clearly visible in the the high resolution image. The sub-image of this structure is shown in Fig. 6.5. The compact unresolved sources (at the ends of the linear structures) are also present in the NVSS image of this region. However, the linear structure is not detected in the NVSS image at 1420 MHz, probably due to the sensitivity and dynamic range limits (due to snapshot uv-coverage) of the NVSS in this region. Its detection in the GMRT 327-MHz image at the level of $5-7$ mJy/beam, compared to its non detection in the NVSS image, is indicative of non-thermal emission. To eliminate the possibility of this feature being an artifact of data processing or due to the presence of bad data, this data was mapped at low and high resolution. Maps were also made using a single frequency channels as well as using a number of RFI free frequency channels. The low resolution map was made using only the GMRT Central Square antennas at a resolution of $3-4$ arcmin which does not require 3D imaging. The high resolution map used all the available arm antennas and required 3D imaging. Multiple frequency channels were used in two ways. Five adjacent frequency channels were averaged and four of these averaged channels (corresponding to twenty frequency channels at the original frequency resolution) were used for gridding the visibilities. In the second method, all twenty frequency channels were used directly for gridding. This linear structure was seen in all these images.

Marginally extended linear features similar to this structure (see also the lower resolution image in Fig. 4.8) have been detected earlier in the 843 MHz survey of the inner Galaxy (Gray1996). One such feature, namely G357.1$-$0.2, was later imaged at higher resolution using the VLA at 5 GHz revealing a bizarre source with 'tubes' of highly confined emission in an even more bizarre morphology. The nature of this source, which is fairly close to the Galactic plane, is not known. The extended structure seen in this GMRT image is also suggestive of a similar source. Higher resolution continuum imaging, polarimetry and HI absorption observation towards this source will be required to get some handle on its nature.