Appendix A: new near-IR counterparts of IRAS sources in the LMC
Periods of weather conditions that were too poor for long-term photometric monitoring at the South African Astronomical Observatory (SAAO) at Sutherland, South Africa, in December 1997 were used to search for near-IR counterparts of IRAS point sources in the direction of the LMC. This was done on the 1.9 m telescope with the Mk III scanning photometer in the K-band. An aperture of was used, chopping and nodding with a throw of . The search was limited to objects brighter than mag. The areas around five IRAS point sources suspected to be obscured AGB stars (Paper I) were searched. The candidate near-IR counterparts found are listed in Table A1, where the photometry is in the SAAO system (Carter 1990), i.e. the J-band magnitude is transformed to the 0.75 m telescope system. One object was re-observed under good photometric conditions (LI-LMC1284), together with the star HR2015 ( Dor) for photometric calibration. Positions have been estimated by comparing the position of the diaphragm in the (red) acquisition video images with the second generation Digital Sky Survey, and are accurate to .
Table A1. Near-IR stars near IRAS point sources in the direction of the LMC (LI=LI-LMC: Schwering & Israel 1990) that are candidate obscured AGB stars. Listed are IRAS flux densities (in Jy), near-IR position, distance to the IRAS source (in arcsec), near-IR magnitudes, and bolometric magnitude assuming association (see text). Values between parentheses are 1- errors.
I retrieved 12, 25 and 60 µm data from the IRAS data base server in Groningen 1 (Assendorp et al. 1995). Point sources were recovered by means of square degree maps with pixels. The flux density was measured from a trace through the position of the source using the command SCANAID in the Groningen GIPSY data analysis software. LI-LMC203 shows a hint of duplicity: two similarly bright sources separated by one arcminute. LI-LMC987 looks slightly extended, and LI-LMC1284 is on top of brighter emission. LI-LMC1522 and especially LI-LMC1795 are isolated. Assuming identification of near-IR and IRAS sources, the spectral energy distribution were integrated graphically to yield bolometric magnitudes.
LI-LMC203 is not identified with certainty. The best spatial coincidence is for the first listed in Table A1, that has blue near-IR colours incompatible with mass-losing AGB stars. There are two much brighter near-IR sources with moderately red nearby, of which the third listed in Table A1 is a cluster of stars within a diameter of . The proposed near-IR counterparts of LI-LMC987 and 1795 have near-IR colours consistent with red giants without mass loss and are probably not associated with the IRAS sources. The near-IR counterpart of LI-LMC1284 is a heavily obscured AGB star, with IR colours compatible with either carbon or oxygen-rich dust. LI-LMC1522 is also identified as a dust-enshrouded star, with IR colours suggesting oxygen-rich dust.
Appendix B: expansion velocities
Here the expansion velocities of AGB stars are briefly discussed in order to arrive at a justified calibration of the scaling relation (Eq. (3)) for LMC stars. I consider derived from the separation of the peaks of OH maser line profiles, and from the width of CO(1-0) emission for Milky Way stars without OH measurements. The latter are divided by 1.12, following Groenewegen et al. (1998; see also Lewis 1991).
The is plotted against P in Fig. B1. Periods for the LMC stars are from Wood et al. (1992) and Wood (1998), and their are from Wood et al. (1992) and van Loon et al. (1998b). Carbon stars (filled symbols) are distinguished from oxygen-rich, M-type stars (open symbols). As an AGB star evolves, P increases. Mass loss becomes important for d and reaches a maximum for d (cf. Jura 1986). This evolution is also reflected in , which increases when d and levels off at a value km s-1 for d (dotted line in Fig. B1, see also Lewis 1991). The individual data sets (and Lewis 1991) suggest that AGB stars with d evolve at constant typically 3 to 4 km s-1. This is also seen in (intrinsic) S-type stars with semi-regular variability (Fig. 18 in Jorissen & Knapp 1998). This may be a slow wind supported by another mechanism, such as radiation pressure on molecules (cf. Sahai & Liechti 1995; Steffen et al. 1997, 1998). It agrees with the observation that SiO masers, that depend on shocks by stellar pulsation (Alcolea et al. 1990), may be present for d and become ubiquitous when d (Izumiura et al. 1994).
Carbon stars appear to have somewhat larger than M-type stars at the same P, although the (few) SGC carbon stars do not obey this trend. The data is also consistent with smaller P for carbon stars at the same .
SGC stars within 1 kpc from the galactic plane (squares) are distinguished from SGC stars beyond that (large circles). The latter are presumably of sub-solar initial metallicity and have smaller than the former.
Wood et al. (1992) showed that is smaller at LMC metallicity than at solar metallicity, providing supportive evidence for Eq. (3). The LMC star with km s-1 is IRAS04553-6825, a very luminous RSG (van Loon et al. 1998b, and references therein). The other LMC stars are AGB stars with km s-1. The OH/IR stars in the Groenewegen sample with have to 20 km s-1, suggesting that initial metallicities of these Milky Way stars are higher than the LMC stars. The expansion velocities of the obscured AGB stars in the Galactic Centre range up to km s-1, and initial metallicities of two to three times solar have been suggested by Wood et al. (1998). Considering all this, Eq. (3) is calibrated by demanding a star with LMC metallicity and ( mag) to have km s-1.
© European Southern Observatory (ESO) 2000
Online publication: January 31, 2000