3. Color - color diagram
From data in Tables 3 and 4, after the IRAS fluxes are transformed into magnitudes without color corrections according to Cheeseman et al. (1989), different infrared color-color diagrams can be plotted: () - () in Fig. 1; () - () in Fig. 2; () - () in Fig. 3 and () - () in Fig. 4. In those figures, the open and filled circles indicate S stars with and without Tc, respectively, and the straight line corresponds to the blackbody distribution.
It is obvious from Fig. 1 that although samples with Tc distribute in a rather wider area than the ones without Tc, both species are basically located in almost the same region. These distributions imply that, statistically, both categories of S stars have almost the same colors and temperatures in the near infrared, and they can not be clearly distinguished on the basis of their near infrared color.
It is shown from () - () diagram in Fig. 2 that a rather clear segregation between Tc-rich and Tc-deficient S stars is visible by different but not by except for a peculiar Tc-deficient star DY Gem (it will be discussed in more detail later). This implies that the color is a good probe for distinguishing these two kinds of S stars, and Tc-rich S stars really exhibit on infrared excess in the 12 µm and 25 µm bands. They have cooler temperature and hence are probably more evolved. On the other hand, Tc-deficient stars have no or little excess in the 12 µm and 25 µm bands which indicates a photospheric origin of the mid-infrared flux. It is also noted that a few Tc-rich S stars including NQ Pup, Ori, Cyg and HR Peg populate the same region as Tc-deficient S stars as Jorissen et al. (1993) already pointed out. In addition, it is seen from Table 4 or Fig. 2 that Tc-deficient S stars with good quality 60 µm flux are not numerous (only 7, out of 20 in Table 4). Therefore, other colors not involving 60 µm flux must be used to show their infrared property.
In order to include more samples than that in Fig. 2, the () - () diagram is plotted in Fig. 3 from which it is obviously seen that except for DY Gem and V Cnc (they will be discussed later), Tc-deficient S stars are concentrated on a very small region of and , and very close to the blackbody line, whereas almost all Tc-rich S stars are out of this region, far away from the blackbody line, and spread over a much wider area that corresponds to a much larger infrared excess either in or in colors, hence with much lower temperature and much higher mass loss. Only four stars: Ori, NQ Pup, HD 170970 and V679 Oph are located in the Tc-deficient sample region. This suggests that both and are good probes to distinguish Tc-deficient S stars from Tc-rich ones.
The () - () diagram is also plotted for the two kinds of S stars in Fig. 4. Again except for DY Gem and V Cnc, most Tc-deficient S stars can be separated from Tc-rich ones with color, but not with color. It should be noted that the () color has been used in Figs. 3 and 4, although the observational results in K and 25 µm were not obtained at the same epoch. From the previous observations in K (Gezari et al. 1993) it is found that, on the average, the difference between the previous data and the new ones is 0.22. This has no serious influence on the result shown in Fig. 3. However, the similar comparison is not possible for samples in Fig. 4 for lack of earlier data.
The conclusion that can be drawn out from the analysis of infrared color-color diagrams mentioned above is that the more sensitive colors for segregating Tc-rich and Tc-deficient S stars are and , hence the more appropriate color-color diagram for distinguishing them is the () - () diagram. Chen et al. (1995) gave more than 700 IRAS associations of S stars in their Table 1, but more than 2/3 of the samples have no good 60 µm flux. If one wants to extract candidates of Tc-deficient S stars according to their infrared colors, only color is not enough. Fortunately, if K magnitudes have been measured, by using the () - () diagram the candidates of Tc-deficient S stars can be well selected despite the absence of 60 µm flux for almost all samples.
© European Southern Observatory (ESO) 1998
Online publication: April 20, 1998