4 Classification and interpretation of the objects towards the Magellanic Clouds

In previously published work, 192 radio sources towards the LMC and 93 towards the SMC have been classified as either Hii regions, SNRs or background sources.

Of the 192 sources towards the LMC, 188 were detected in recent Parkes surveys (see Table 5). Of these 188 sources, 53 appear to be Hii regions, 34 are confirmed SNRs, 24 are possible SNRs and 71 are listed as background sources. Six sources are listed as ambiguous in Paper VI.

Four confirmed LMC SNRs listed in Mathewson et al. (1983a), Mills et al. (1984b), Tuohy et al. (1982) and Chu et al. (1997) could not be detected (B0505-679, B0509-675, B0543-689 and J0540-697) in any of our six surveys. We have discussed these four SNRs in Paper VI. Gurwell & Hodge (1990), mainly using optical identifications, have reported a number of galaxies behind the LMC, but only a few of these have detectable radio emission.

The remaining 295 (483-188) sources listed in Table 5 provide new candidates for Hii regions, SNRs and background galaxies or quasars. For 229 of these there is a fitted spectral index.

In the SMC, 102 sources were known and classified from previous radio investigations. Out of these, 92 were detected in the Parkes surveys (Table 6). Many of the SMC sources were discussed in Ye (1988). Of these 92 sources, 14 appear to be Hii regions, 10 are confirmed SNRs, 2 are possible SNRs and 66 are listed as background sources.

Six confirmed SMC SNRs (SNR B0044-7324, B0045-7322, B0046-7335, B0056-7227, B0056-7223 and B0104-7221) and three SNR candidates (B0051-7259, B0056-7250 and B0057-7308) in the SMC could not be detected in any Parkes surveys because of their small angular size, weak radio emission and confusion. For the same reasons a number of background galaxies detected by Ye (1988) were missed.

The remaining 131 (224-92) unclassified sources (of which 73 have a fitted spectral index) towards the SMC are investigated to determine their nature.

4.1 Selection criteria

We have used the following criteria to classify the discrete radio sources listed in Tables 5 and 6:

1.

For Background source candidates -
background sources are either radio galaxies with generally steep radio spectra ($\alpha \le -0.8$) or quasars which have flat spectra ($\alpha \sim 0.0$). Objects well removed from the stellar/nebular component of the MC are expected to be clear of the MCs objects and, by default, to be populated by background objects. Background objects are defined using following criteria.

$\bullet$

All radio sources outside the defined area of the LMC or the SMC are background sources. The area of the LMC is defined between RA(B1950)=04$\rm ^h$45$\rm ^m$ to RA(B1950)=06$\rm ^h$00$\rm ^m$ and Dec(B1950)=-65$^\circ$00$^\prime$ to Dec(B1950)=-72$^\circ$30$^\prime$:  the area of the SMC as defined between RA(B1950)=00$\rm ^h$30$\rm ^m$ to RA(B1950)=01$\rm ^h$30$\rm ^m$ and Dec(B1950)=-71$^\circ$40$^\prime$ to Dec(B1950)=-74$^\circ$00$^\prime$.

$\bullet$

Radio sources with very steep spectra ($\alpha \le$ -0.8) are more likely to be background objects. However, there is a wide range of radio spectral index for background sources.

$\bullet$

There is no H$\alpha$ or IR detection (however some background sources can be detected in one of these frequencies but this is not likely) (Paper VIII).

2.

For SNR and SNR candidates -
supernova remnants (SNRs) in the MCs have been the subject of numerous studies, both as individual SNRs and as a statistical sample. A series of criteria in several spectral ranges have been defined to identify SNRs in the MCs. Following Mathewson et al. (1983a and reference therein) and Rosado (1993b) the candidate must satisfy at least two of the following conditions. It should:

$\bullet$

be an extended soft X-ray source appearing either on Einstein surveys (Long et al. 1981; Wang et al. 1991) or the ROSAT all-sky survey (Pietsch & Kahabka 1993; Pietsch et al. in preparation). A negative Hardness Ratio 2(HR2) indicates that the source is a SNR (Paper VI);

$\bullet$

be a non-thermal radio source ($\alpha$=-0.4 on average but there is a wide range and so spectral index is not itself conclusive) (Sect. 4);

$\bullet$

have optical emission with the following characteristics:

a)

[SII]/( H$\alpha$+[NII]) $\ge$ 0.4
Possibly the most sensitive diagnostic for identifying SNRs is narrow-band optical CCD observations in the light of H$\alpha$ and [SII]. The ratio of the [SII] $\lambda$6717, 6713 to H$\alpha$ emission in nebulae provides a good discriminator between Hii regions, which typically show [SII] H$\alpha$ of $\sim$0.1, and SNRs, which usually have ratios >0.4. This is because sulphur is ionized almost completely to the S⁺⁺ state in Hii regions, while a wide range of temperatures and ionization states in the cooling regions behind SNR shocks gives rise to relatively strong [SII] emission;

b)

have H$\alpha$ with looped or filamentary structure and

c)

an H$\alpha$ flux density different from that expected for the radio counterpart in the case of thermal source (in the case of finding an Hii region coincident with radio sources) (Ye 1988).

*

have weak or no IR emission. SNRs are not likely to appear in IRAS surveys unless they are embedded in Hii regions (Paper VIII).

These criteria are biased toward radio properties and further kinematical information is probably required. Furthermore, one must keep in mind that several well-known SNRs in the MCs could not fulfil 2 of the 3 criteria given above because of the following:

-

some radio SNRs could be missed in X-ray surveys if their X-ray emission is weak (or if it is attenuated by absorbing matter) (Paper VI);

-

limitations in sensitivity of radio surveys. Some known MCs SNRs are weak radio emitters and near the sensitivity limit of the Parkes telescope, seen in only one radio frequency, and therefore have no radio spectral information. One example is DEM L298;

-

confusion with the surrounding Hii regions. Most of the LMC SNRs are embedded in Hii regions.

3.

For Hii regions candidates -
Hii regions are expected to be within the different nebular regions of the MCs. They are also expected:

*

to be thermal radio sources (with $\alpha$=-0.15 on average) but because of wide range of spectral indices this criterion is not itself conclusive;

*

Not to appear at X-ray frequencies - Hii regions are not likely to be seen at X-ray frequencies (Paper VI);

*

to have H$\alpha$, UV and/or IR positive detections (Paper VIII).

4.2 Background sources in the field of the MCs

Because the MCs are transparent to radio continuum, background radio galaxies and quasars are seen throughout the MCs. It is very difficult to distinguish radio galaxies and quasars just on a basis of radio spectra and therefore, for this work, these sources will be treated together.

4.2.1 Background sources in the field of the LMC

Lists of background sources towards the LMC have been published in Gurwell & Hodge (1990), Dickey et al. (1994) and Ye (1988). From their lists, 65 sources (marked in Table 5, Col. 17 as BG) are detected in our radio surveys as background, and an additional six as very probable background sources (marked in Table 5, Col. 17 as BG?). All of these 71 sources obey the criteria from Sect. 4.1. and will be treated as definite background sources.

Outside the defined area of the LMC, 97 sources (marked in Table 5 Col. 17 as bg) have been found. These sources are very strong candidates for background sources. Also, all of these sources obey the criteria established in Sect. 4.1. Another group of 34 sources (marked in Table 5, Col. 17 as bg?) satisfy some criteria of background sources and so are treated as background-source candidates.

In total, at least 202 sources towards the LMC are classified as background or background candidates, which is 42% of all sources found in the field of the LMC. Most of the 75 unclassified sources are likely to be background objects according to our source-density study (Sect. 4.2.3).

4.2.2 Background sources in the field of the SMC

The most complete list of background sources towards the SMC can be found in Ye (1988). From that list, 63 sources (marked in Table 6, Col. 14 as BG) are detected in our radio surveys as background and an additional three as very probable background sources (marked in Table 6, Col. 14 as BG?). All of these 66 sources obey the criteria from Sect. 4.2. and should be treated as definite background sources.

There are 108 sources outside the SMC area defined in Sect. 4.1. which have not been previously classified (marked in Table 6, Col. 14 as bg). These are very strong candidates for background sources as all of them satisfy criteria established in Sect. 4.1.

In total, at least 174 sources towards the SMC are classified as background or background candidates, which is 78% of all sources found in the field of the SMC. This number could increase slightly as the 12 unclassified sources are likely to be background according to the source-density study (Sect. 4.2.3).

4.2.3 Statistics of the background sources

The expected background integral source counts at our six survey frequencies were obtained by interpolating the polynomial fits (Log N - Log S study) of Wall (1994). These are the predicted number of background sources in the observed fields of the LMC and the SMC at each observing radio frequency (Tables 10 and 11).

Column 1 (Tables 10 and 11) lists the observing radio frequency and Col. 2 gives the 5$\sigma$ cut-off level in mJy. Column 3 lists the total number of sources found at each observing frequency and the area covered by each survey is given in Col. 4. The number of sources found in our surveys to be stronger then 5$\sigma$ at each frequency is given in Col. 5 and the total number of background sources stronger then 5$\sigma$as given in Col. 6. The number of sources estimated from Wall (1994) for the given cut-off level (Col. 2) is given in Col. 7.

  

Table 10: Log N - Log S estimates for the LMC

  

Table 11: Log N - Log S estimates for the SMC

Comparing these results (Cols. 6 and 7), the number of background sources classified in this work (Sects. 4.2.1. and 4.2.2.) are very close to the expected range of the Log N - Log S study. The agreement in numbers, at the 10% level, is as close as can be expected, given the sensitivity of integral counts to the flux-density calibration and flux cut-off. These findings strengthen our confidence in the selection criteria and source classification.

Figures 5 and 6 show the distribution of all classified background sources towards the LMC and the SMC. These source distributions are uniform, as expected.

  

Figure 5: Distribution of background sources towards the LMC

  

Figure 6: Distribution of background sources towards the SMC

4.3 The Magellanic Clouds sources

Radio sources intrinsic to the MCs are considered to be either SNRs or HII regions. SNRs are formed by the high-velocity and high-temperature material ejected when a star ends its life with a violent explosion. The notation "Hii region'' is usually reserved for relatively dense regions of the interstellar gas which are nearly completely ionized by the Lyman continuum radiation from massive and hot stars (Mezger 1972). Hii regions in the MCs provide the opportunity to study the ionized gas component of the interstellar medium and star formation on a galactic scale, with a sensitivity and spatial resolution which are comparable to most of the Galactic surveys.

4.3.1 Supernova remnants in the LMC

Out of 62 SNRs (excluding the remnant of SN 1987A) and SNR candidates, 34 (marked in Table 5, Col. 17 as SNR) were confirmed to obey criteria from Sect. 4.1. and therefore will be treated as certain SNRs. Four well-known SNRs (see Sect. 4) could not be detected in any of our Parkes radio surveys. The other 24 sources (marked in Table 5, Col. 17 as SNR?), remain SNR candidates because of lack of confirmation. Thus only 12% (58 out of 483) are SNRs or SNR candidates. However, it is likely that more SNRs may exist which are confused with Hii regions.

4.3.2 Hii regions in the LMC

From previous studies, 53 radio sources (marked in Table 5, Col. 17 as HII) are confirmed as Hii regions. They all obey the criteria established in Sect. 4.1. and therefore they are classified as certain HII regions. A further 41 previously unclassified sources (marked in Table 5, Col. 17 as hII) have characteristics typical for Hii regions and hence they will be treated as very good candidates for Hii regions. Another 20 sources (marked in Table 5, Col. 17 as hII?) obey at most two different criteria typical for Hii regions and therefore they are suggested Hii region candidates. Until further analysis one cannot be sure about the nature of these sources.

Of the 483 sources towards the LMC, 114 (24%) are thus listed as Hii regions. This number could increase as 34 sources (marked in Table 5, Col. 17 as ?) have some characteristics of MCs sources but one cannot be sure of their intrinsic nature.

4.3.3 Supernova remnants in the SMC

From Ye's (1988) list of 19 SMC SNRs and SNR candidates, 10 (marked in Table 6, Col. 14 as SNR) were confirmed by the criteria from Sect. 4.1. and therefore they are certain SNRs. Six well-known SNRs and three SNR candidates (see Sect. 4) could not be detected in any of our Parkes radio surveys due to the flux density limits. The remaining two SNR candidates (marked in Table 6, Col. 14 as SNR?) are still under study and, at present, they will remain as SNR candidates.

Only 6% (12 out of 224) of sources towards the SMC shows characteristics of SNRs. However, as for the LMC, it is likely that this number of SNRs could be higher because of their confusion with Hii regions.

4.3.4 Hii regions in the SMC

From previous studies, 14 radio sources listed in Paper V (and also marked in Table 6, Col. 14 as Hii) are confirmed to be Hii regions. They obey all criteria for Hii regions (Sect. 4.1.) and therefore they are classified as certain Hii regions. Another ten previously unclassified sources (marked in Table 6, Col. 14 as hII) have all characteristics typical for Hii regions and hence they will be treated as very good candidates for Hii regions.

Of the 224 sources towards the SMC, only 24 (11%) are listed as Hii regions or Hii region candidates. Two other sources (marked in Table 6 Col. 14 as ?) also have some characteristics of Hii regions but their status also remains unclear.