Astron. Astrophys. 353, L21-L24 (2000)

3. Discussion

The optical counterpart of MIR1 appears to be the variable star 224 discovered by Balona (1992). During the observing run of six nights when it was monitored by Balona, it faded by 0.2 mag and was distinctly variable within a night. Although periods around the 1 day expected for a Cepheid were indicated, no period gave a satisfactory fit to the data so the observed scatter and red color led Balona to suspect that it may be a double mode Cepheid on the red edge of the instability strip. Independent observations of NGC 330 (Sebo & Wood 1994) made over a 4 year period verified the variability of MIR1 (their star 515V)with a and , but again no regular period was evident. Strikingly, the average V magnitude over six days (17.12; Balona, 1992) is very similar to the average V magnitude over 4 years (17.17; Sebo & Wood, 1994).

The optical counterpart of MIR1 was found to be a strong H source. Observations in the narrow-band (nm) H filter showed that this object (star 485, Keller et al. 1999c) was the second strongest H emitter in the field of NGC 330 after the planetary nebula L305. This object is also listed in the SMC H source catalog of Meyssonnier & Azzopardi (1993) as object 906.

The strong H emission and the prominent mid-IR excess are difficult to assess within the evolutionary scenario of a classical Cepheid. Indeed it is possible that this object is a binary system, however the discussion of this possibility in the view of the scarce observational facts seems rather premature. The color index is much larger in MIR1 than in any classical Be star in NGC 330 (Keller et al. 1999c), which, together with a strong mid-IR excess, indicates that MIR1 is unlikely to be a classical Be star. Therefore, we will further concentrate on the Be supergiant, Herbig Ae/Be and post-AGB star scenarios instead.

3.1. Be supergiant and Herbig Ae/Be star scenarios

One of the possible alternatives for constraining the evolutionary status of MIR1 is a Be supergiant scenario. This is indeed supported by the existence of H emission, which is typical to all types of Be stars. Spectral observations of MIR1 obtained by Keller (1999a) confirm that this object is a very strong H emitter; the spectrum clearly shows H line though no H or higher.

Although the observed optical color indices of MIR1 are distinctively different from those of Be supergiants in the Magellanic Clouds (Zickgraf et al. 1992), this may be a consequence of the interstellar or circumstellar reddening. A dereddening procedure employing the reddening-free Q parameter yields (calculated assuming the standard excess ratio) which indicates that the spectral type of this object (depending on the luminosity class) should be O8-B2. Taking B0 as a representative of these values, one obtains , and , and, using the SMC distance modulus of 18.9, . Assuming that the bolometric correction for the spectral type B0 is , we derive . Taking into account the errors of the spectral type determination (which set a range of possible between K), the obtained and are indeed comparable with those of Be supergiants in the MCs (cf. Zickgraf et al. 1992). Keller et al. (1999b) show a HR diagram of the cluster from the HST data and the Be stars at the cluster turnoff have and . They also have one Be star (B13) like a blue straggler with and . These temperatures and luminosities are similar to the ones obtained for MIR1. The derived , however is much higher than the average in the field of NGC 330 (which measures the range from derived by Carney et al. (1985) to obtained by Bessell, 1991), and therefore indicates a significant circumstellar extinction.

Indeed, the spectral energy distribution of MIR1 shows a strong mid-IR excess (Fig. 2). The estimate of the ratio of ISO LW10 band flux over the V band flux in MIR1 yields . This is comparable with the observed in a `representative' Be supergiant GG Car (Waters et al. 1998) and thus could be viewed as an additional argument supporting the Be supergiant scenario.

Fig. 2. Spectral energy distribution of MIR1, constructed from optical photometry and ISO data (Table 1). Error bars of the mid-IR data are formal IRAF/APPHOT errors. Solid line shows two-blackbody fit to the optical and mid-IR data ( K and K); dashed line indicates K fit to the mid-IR ISOCAM data used to estimate the infrared luminosity (see text for details).

Employing theoretical evolutionary tracks of Fagotto et al. (1994) and making use of the derived and we obtain a stellar mass of and the age of 8-14 Myr. The derived age of MIR1 is comparable with the cluster's age (10-20 Myr, Cassatella et al. 1996), suggesting that the candidate Be supergiant could be a cluster member.

Prominent mid-IR excesses are also common in Galactic Herbig Ae/Be stars with cool circumstellar shells (group II objects, see Hillenbrand et al. 1992). However, Herbig Ae/Be scenario seems rather unlikely for the case of MIR1. First, the available observations of NGC 330 do not show any evidence for the ongoing star formation in the field of NGC 330. Second, although some Galactic Herbig Ae/Be stars are observed as isolated objects, they are usually low-mass stars (cf. Hillenbrand et al. 1995) and therefore the high mass of the possible Herbig Ae/Be candidate ( ) inferred from the dereddened photometry of MIR1 and the SMC distance modulus rules out this possibility too.

3.2. Post-AGB star scenario

Post-AGB stars have been long recognized as one of the evolutionary groups showing the Be phenomenon. Indeed, strong H emission is typical for most post-AGB objects and thus the existence of H emission in MIR1 works in favor of this scenario too.

Most of the post-AGB objects show a double-peaked spectral energy distributions (e.g., Kwok, 1993; Zhang & Kwok, 1991), similar to the one observed in MIR1 (Fig. 2). A simple estimate of the infrared luminosity obtained from the blackbody fit to the ISOCAM data yields with a blackbody dust temperature K. The estimate of the dust mass in the circumstellar shell, , can be made then using the following expression (Gurzadyan, 1997):

where and are the dust temperature and the infrared luminosity, respectively. Taking the and values derived above, one obtains , which is comparable with the dust masses typical for the post-AGB objects (e.g., Pottasch & Parthasarathy, 1988). Two facts should be noted, however. Firstly, the obtained blackbody dust temperature ( K) can be considerably overestimated, since its derivation relies on the mid-IR data only and does not take into account any information about the dust radiation at longer wavelengths. Secondly, at the dust temperatures typical for the post-AGB objects, a large fraction of infrared flux is emitted at wavelengths longer than m and thus can be considerably higher than the presently derived value. Therefore, the obtained estimate of indicates only a lower limit for the dust mass in MIR1.

The upper limit for the effective temperature of the central star of the possible post-AGB object can be inferred from the following considerations. If MIR1 is assumed to be a normal planetary nebula (i.e., past the PPN stage), the effective temperature of the central star should be at least K and the observed would indicate a considerable circumstellar extinction. Indeed, the central star with K should have , and hence the , that is, and . Assuming that the bolometric correction is one obtains , which is very close to the classical luminosity limit for the post-AGB stars (, e.g. Shaw & Kaler, 1989). Thus we conclude, that the classical luminosity limit for the post-AGB objects sets the upper limit for the effective temperature of the central star at about 30 000 K.

The lower limit for the effective temperature of MIR1 can be constrained from the observed SED. The two-blackbody fit to the optical and mid-IR data (see Fig. 2) gives a lower limit estimate of the total luminosity of MIR1, . Using a simple iteration procedure one can obtain the , and therefore , which would produce the observed with the observed . Such procedure yields , , and K, seting this as a lower limit for the effective temperature of the central star.

The obtained temperature range suggests that MIR1 can be a good proto-planetary nebula (PPN) candidate. This is reinforced by the fact, that the infrared to the total luminosity ratio in MIR1 is , which is considerably higher than the value typical for the planetary nebulae (, see e.g. Pottasch, 1997). Since the presently estimated total luminosity of MIR1 is only , it is rather unlikely that this object could be a high mass post-AGB star belonging to NGC 330; instead, it is probably a low mass field star. However, the mass and thus the evolutionary status of the possible PPN can not be constrained precisely yet. Therefore the tighter constraints on this scenario should be set by future optical spectroscopy of MIR1, which would provide additional information both about the central star and the nebula.

© European Southern Observatory (ESO) 2000

Online publication: January 18, 2000
helpdesk@link.springer.de