ForumSpringerAstron. Astrophys.
ForumWhats NewSearchOrders

Astron. Astrophys. 329, 161-168 (1998)

Previous Section Next Section Title Page Table of Contents

5. The stellar content

Due to the remarks above it is obvious that the present NIR study cannot further contribute to the classification of stars in terms of spectral types and/or luminosity classes. For that purpose the classical methods like UBV photometry and MK spectroscopy and/or the K -band classification system developed by Hanson & Conti (1994) are required. For main sequence stars earlier than B3, UBV photometry is still a very efficient tool as shown by the fact that all O-type stars identified by CEN as such on the basis of their UBV colours have been verified recently as ionising sources of M 17 by Hanson & Conti (1995). The purpose of the following section is to investigate whether there are additional objects similar to the "cocoon stars " found by CK2, i.e. to identify new IR excess objects on the basis of their location in various TCDs.

5.1. Normal stars

As already discussed in the previous section and in agreement with CEN there are 18 normal early type stars among the 32 "visible stars " whose spectral types must be earlier than F2 V. CEN 0 and 10 were classified as foreground giants, CEN 44 was classified as B4 V but seems to be variable. Variability is a general problem when discussing TCDs and SEDs because data taken at different epochs may simulate unusual colours and/or excesses. Thus, for the sample of "visible stars " some caution is necessary when evaluating the TCDs in Fig. 1 that contain B and V data from 1980 and NIR data from 1992 to 1995. From the fact, however, that these 18 stars do not show any excess emission in the TCDs of Fig. 2, which are based on data from a single epoch only, one may conclude that they are indeed normal. In this context we want note that the NIR variability observed for some objects in M 17 is of the order of several tenths of a magnitude (CKC) whereas an IR excess produced by circumstellar dust typically exceeds several magnitudes.

From the apparent colours [FORMULA] and [FORMULA] one may estimate the visual extinction towards the normal visible stars; the influence of the unknown spectral types is negligible: At a distance of 2.2 kpc - equivalent to a distance modulus of 11.7 mag - and with a foreground extinction of [FORMULA] mag, the faintest sources that can be reached by our limit of [FORMULA] mag have [FORMULA] mag, corresponding to spectral type A0 V. The local extinction within M 17 shifts this limit even towards earlier spectral types. The intrinsic colours [FORMULA] and [FORMULA] for main sequence stars from O6 to G3 are generally close to zero ( [FORMULA] mag) (Koornneef 1983b), i.e. they lie within our photometric accuracy. We thus may convert the apparent colours directly into a visual extinction, using the relations between [FORMULA] and the IR colours according to Table 3: [FORMULA] and/or [FORMULA]. The observed colour range of [FORMULA] translates into [FORMULA] mag for a normal extinction law with [FORMULA]. Using the appropriate R value of 4.8 as derived above [FORMULA] will increase by 50%. Similar results are obtained from the [FORMULA] colours. These estimates agree with the [FORMULA] results derived by CEN from UBV photometry.

Among the 37 "IR sources " there seem to be further 28 normal stars as witnessed by their location close to or left of the reddening lines in Fig. 2. They partly overlap with the "visible " sample but several sources are significantly redder, extending until [FORMULA] and [FORMULA]. For a normal main sequence star these colours translate into a visual extinction of about 33 mag ( [FORMULA] ) or 50 mag ( [FORMULA] ). If these sources were background giants, small corrections due to their intrinsic colours have to be applied, yielding a total extinction of [FORMULA] mag throughout the molecular cloud of M 17. In summary, one may conclude that the present NIR survey does not extend the initial mass function towards spectral types later than B but reveals a number of additional early type stars that are embedded even more deeply than the previously known cluster members.

5.2. IR excess objects

According to their location in Fig. 1 there are nine "visible stars " with clear IR excess (CEN 14, 26, 27, 32, 46, 49, 51, 52, 65); their anomalous colours can also be traced throughout Fig. 2. In addition, CEN 28 and 40 seem to be slightly below the special reddening lines in Fig. 1, indicating a faint excess. When investigating their colours in the JKL plot of Fig. 2 this excess becomes quite pronounced. Thus, we regard 11 "visible stars " as objects with an IR excess; they are marked as circles in all TCDs. Among the pure "IR sources " there are another 9 stars (IRS 1, 2, 5, 9, 10, 15, 20, 23, 37) whose location in Fig. 2 argues in favour of an IR excess; these objects are marked as open squares.

In the crowded cluster field of M 17 it is very likely that faint red companions, contained in the aperture, may affect the photometry and thus simulate an IR excess. We therefore have inspected our CCD frames at JHK (CKC) in order to search for red companions close to the 20 new IR excess objects. Down to limit of [FORMULA] mag we only find one additional source within the aperture of our photoelectric photometry. This source is located 2" SW of IRS 9 and is only marginally visible at K ; it thus cannot contribute significantly to the JHK colours of IRS 9 which are normal in Fig. 2. A source within the OFF beam of the photoelectric observations will only decrease the IR flux and thus suppress an IR excess.

The fraction of IR excess objects among the "visible " sample is 34% - among the "IR " sample [FORMULA] %; including the five cocoon stars, these numbers turn into 47% and 27%, respectively. All stars have in common that their SEDs strongly increase with increasing wavelength. Compared to the "visible stars ", the average apparent K magnitude of the "IR sample " is 0.7 mag brighter whereas the average I brightness is more than 3 mag fainter. This demonstrates that the SEDs of the pure "IR sources " decrease faster towards shorter wavelengths reflecting their larger extinction. Four IR excess objects (CEN 14, 26, 27, 40) have been observed during different epochs. They all show variability of several tenths to about 1 magnitude at NIR wavelengths. As it will be discussed in our next paper (CKC) this seems to be a common property of the class of "cocoon stars " discovered by CK2.

Previous Section Next Section Title Page Table of Contents

© European Southern Observatory (ESO) 1998

Online publication: November 24, 1997