Forum Springer Astron. Astrophys.
Forum Whats New Search Orders

Astron. Astrophys. 356, 517-528 (2000)

Previous Section Next Section Title Page Table of Contents

7. Discussion

We start this section with an examination of the individual cluster membership.

7.1. Cluster membership

Although our CASPEC spectrograms cannot be used for precise radial velocity measurements, we have tried to investigate the membership problem for our program stars. To avoid any instrumental errors, we have measured only the mean shifts between calculated and observed spectra and compared them with other stars in the same cluster. Since all of the stars were observed in just a few days, these relative shifts should be equal to the real radial velocity differences between the stars. We did not investigate possible measurement errors, but we do not expect the accuracy to be less than [FORMULA] 5 km s-1. Otherwise, wavelength variations in the shifts between the calculated and observed spectra become clearly visible.

7.1.1. NGC 3496

We found that the shift is practically the same for S 115 and S 117 (mean value of about +0.1 Å), while for S 43 the shift appeared to be much greater (-0.6 Å), corresponding to a velocity difference about 50 km s-1. Such a difference could imply either that S 43 is a variable star, or that there is no connection between this star and S 115 and S 117, while the latter stars are physically connected.

This explanation is supported by estimates of the distance moduli of these stars. For S 43 and S 115 we have the necessary photometric data, and using Crawford's relation for B stars (Crawford 1978), we get [FORMULA] 13.0 and [FORMULA] 9.5. The lack of [FORMULA] photometric data for S 117 prevents such an estimate for this star. In this case we attempted to find the distance modulus by supposing that it is B8-B9 main sequence object (see values of [FORMULA] and [FORMULA] are from Table 4). A star of this spectral type should have [FORMULA]. With this value we get [FORMULA] 0.2, or [FORMULA] 11.5.

For NGC 3496 Balona & Laney (1995) have estimated the distance modulus as approximately 10.7. If we accept this value, then the distance moduli of S 115 and S 117 are consistent with cluster membership whereas that of S 43 is not, in agreement with the differences in radial velocity. However, we should be careful here, since the above mentioned authors have stated that, "we strongly suspect that NGC 3496 is not a true cluster but contains at least two different group of stars of different age, very likely at different distances".

7.1.2. NGC 6475

For both blue stragglers from this cluster we obtained the same shift between the observed and calculated spectra, with an uncertainty of only a few km s-1. This is in agreement with the results of Mermilliod (1982) and Gieseking (1985), who found that the radial velocities of these two stars are close to each other and also close to the mean radial velocity for NGC 6475 cluster.

7.1.3. NGC 6633

According to the compilation of Ahumada & Lapasset (1995), the membership probabilities for the two blue stragglers HD 169959 and HD 170054 are 90% and 70%, respectively. For the main sequence stars, HD 170011 and HD 170158, only a qualitative photometric membership probability was reported (Schmidt 1976, Hiltner, Iriart & Johnson 1958). According to Levato & Abt (1977), HD 170095 is also a member of NGC 6633 (but the authors do not give any details). In the present study we determined the differences in radial velocities [FORMULA]= [FORMULA](star)-[FORMULA] (HD 169959), i.e. between each star and HD 169959, whose membership probability is highest. As a result we have got the following [FORMULA]: +19 km s-1 (HD 170011), +1 km s-1 and +6 km s-1 (HD 170054, two spectra), +1 km s-1 (HD 170095), +6 km s-1 (HD 170158). As one can see, our estimates based on the CASPEC spectra seem to support the membership of the program stars. Only one main sequence star of the sample possesses an anomalous radial velocity, but for the final conclusion about its membership in NGC 6633, additional studies will be needed.

7.1.4. IC 2602

All of the stars observed by us in the field of this cluster are physical members (Hill and Perry 1993). The relative radial velocities (with respect to the velocity of HD 93540) estimated from our spectra are the following: -9 km s-1 (HD 92385), +2 km s-1 (HD 92837), -4 km s-1 (HD 93098), +5 km s-1 (HD 93194). These results also support the membership of the investigated stars in IC 2602 (note that HD 92385 with a highest deviation is known as variable star of the [FORMULA] CVn type).

7.2. Rotational velocities

From Table 4 one can immediately conclude that all the investigated blue stragglers have small projected rotational velocities, while the main sequence stars are fast rotators. For the latter objects (from NGC 6633 and IC 2602) [FORMULA] values are close to those which are expected for their spectral types. Similar small [FORMULA] values were reported by Mathys (1991) for ten blue stragglers from M 67 and by Andrievsky for 40 Cnc (Paper I). Another well-known blue straggler of the Hyades cluster, 68 Tau, is also very likely to be a slow rotator (projected rotational velocity is [FORMULA] = 15 km s-1, Abt & Morrell 1995). A detailed study of blue stragglers from four open clusters undertaken by Schönberner & Napiwotzki (1994) also revealed that all blue stragglers investigated have small projected rotational velocities (seven stars from NGC 7789, one from M 67, one from NGC 752 and one from NGC 2632 - 40 Cnc).

7.3. Chemical peculiarities

7.3.1. Helium and carbon

In our present study, we discovered two new He-weak stars: the blue straggler HD 169959 and the main sequence star HD 92385. In Fig. 12 we show the synthetic spectrum for HD 169959 in the vicinity of the [FORMULA] 4471 Å line calculated with a tenfold increase in the [FORMULA] abundance (i.e. solar). One can see that for this case the synthesized line appears to be much too strong.

[FIGURE] Fig. 12. Fragment of observed and synthetic spectra for HD 169959 in the vicinity of the [FORMULA] 4471 Å line: a solar helium abundance was adopted for the synthetic spectrum.

The blue straggler HD 162374 is also an extremely helium deficient star. In fact, this has been known for a long time (see e.g. Norris 1971), but an accurate quantitative estimate of the helium deficiency in the atmosphere of this star has not been performed before. We show that this star possesses an atmospheric helium abundance which is about 25 times lower than solar. In Fig. 13a we demonstrate how the helium 4471 Å lines should look in the spectrum of this star for a solar abundances. The two remaining program blue stragglers do not show any significant helium anomalies.

[FIGURE] Fig. 13a and b. Fragment of observed and synthetic spectra for HD 162374 in the vicinity of the HeI 4471 Å [FORMULA] and CII 4267 Å [FORMULA] lines. Solar helium and carbon abundances were adopted for the synthetic spectrum.

All of the investigated stars (both blue stragglers and main sequence stars) with measured carbon abundances show a moderate-to-strong underabundance of this element. For example, the CII 4267 Å line of the blue straggler HD 162374 is shown in Fig. 13b and, for comparison, the profile synthesized with a solar carbon abundance is also shown.

To be sure that the reduced helium content and NLTE effects in hot B star atmospheres do not affect our conclusions about the helium and carbon deficiency detected for some stars, 1) Dr. F. Kupka calculated at our request two atmosphere models with the same parameters as those of our He-weak blue stragglers and with a reduced helium content, 2) Dr. S.A. Korotin applied the NLTE MULTI code (see for reference Korotin, Andrievsky & Kostynchuk 1999) to investigate NLTE corrections to the carbon abundances of some blue stragglers. It was found that: 1) the atmosphere models with a reduced helium content gave practically the same result as the Kurucz models with a normal helium abundance. 2) non-LTE corrections for the stars having effective temperatures in the range 12000 K-17000 K appeared to be rather small (about -0.20 dex or less with respect to LTE carbon abundance). For example, the relative LTE carbon abundance in HD 162374 is - 1.30, while the NLTE calculation gives -1.15. This indicates that the strong carbon deficiency in the atmospheres of some program stars is real.

He and CNO anomalies are known for B-A stars from clusters and the galactic field. There exists a group of so-called He-weak stars (Norris 1971) of spectral class B. The stars of that group show a relative helium abundance (He /H) [FORMULA] 0.01.

Almost all of the investigated field B stars and those from OB associations show a moderate carbon deficiency (see, e.g. the recent results concerning the NLTE carbon abundance in B stars published by Daflon, Cunha & Becker 1999 and Andrievsky et al. 1999). Similar carbon underabundances are found also for chemically peculiar stars in the temperature region from 10000 K to 17000 K. According to Roby & Lambert (1990) [C/H] varies from approximately solar to - 0.7. These authors also reported a reduction of the carbon deficit towards higher effective temperatures.

For an explanation of the He-weak star phenomenon, the mechanism of gravitational settling of neutral helium in a stable stellar atmosphere is usually invoked (Michaud 1970). Michaud et al. (1979) also state that if the envelope is stable enough, then diffusion always leads to helium underabundance in the atmosphere.

Gonzalez, Artu & Michaud (1995) considered radiative acceleration for CNO atoms in A-F star envelopes and found that, in particular, carbon should be as underabundant as [C/H] [FORMULA] for the hottest model of their sample. Our results marginally agree with the prediction of diffusion theory.

Nevertheless there is a problem that spoils this picture based on the diffusion theory predictions for He and CNO anomalies in B-F stars. The high projected rotational velocities for HD 169959 and even higher values for S 43 and HD 92385 are hardly consistent with the supposition about the stability of their atmospheres. Nevertheless, these stars show a remarkable helium (and carbon) deficiency. For S 43, one of the faintest stars of our sample, we show in Fig. 14 a fragment of the observed spectrum and the synthetic spectrum calculated with a solar carbon abundance. Note that this star has [FORMULA] = 90 km s-1. No doubt, carbon is really deficient. The blue straggler HD 169959 has [FORMULA] = 70 km s-1 which implies that its equatorial velocity could be as high as [FORMULA] 100 km s-1, which is regarded to be limiting value for the operation of gravitational diffusion. Note that most of the helium-weak stars investigated by Norris (1971) have [FORMULA] parameters smaller than 70 km s-1.

[FIGURE] Fig. 14. Fragment of observed and synthetic spectra for S 43 in the vicinity of the [FORMULA] 4267 Å line: adopted for synthetic spectrum carbon abundance is solar.

7.4. Other elements

Among the [FORMULA]-elements, silicon shows an interesting behaviour. When He is normal in the atmosphere, silicon shows an apparent deficiency. In stars with decreased helium content, silicon is enhanced. The correlation is presented in Fig. 15, which is based on data from the present work and from Paper I in the case of HD 73666.

[FIGURE] Fig. 15. Corellation between the helium and silicon abundances.

The abundances of the remaining elements (iron-group, in particular) in blue stragglers show no anomalies when compared to those in main sequence stars.

7.5. Blue straggler phenomenon

7.5.1. The present-day situation

Standard theory is not able to explain the existence of rather massive main sequence stars (having masses greater than those of turn-off stars) in stellar clusters. These stars seemed to be delayed in their evolution. Indeed, instead of evolving off the main sequence towards the red giant region, they prefer to spend on the main sequence a time interval which is longer than the main sequence lifetime for the corresponding stellar mass.

Are they single stars, binaries or merged remnants? Several hypotheses have been proposed to explain their peculiar origin. None of them was able to elucidate all the observed properties of blue stragglers from the different stellar systems: open and globular clusters, stellar associations.

The single star evolution hypothesis (Saio & Wheeler, 1980) faces some problems with localizing the source for the strong internal mixing or the non-thermal pressure which prevents stellar core contraction and thus prolongs the lifetime of the main sequence phase. Moreover, this hypothesis was criticized from the observational side by Schönberner & Napiwotzki (1994).

An examination of the blue straggler problem which supposes that they are stars formed via stellar collisions and tidal captures in open clusters was put forward by Leonard & Linnell (1992). These authors were able to account for only 10% of the observed blue stragglers in this manner.

The mass transfer hypothesis (McCrea 1964) and stellar coalescence scenario (Zinn & Searle 1976; Mateo et al. 1990) also face observational difficulties (slow rotation and a lack of the radial velocity variations for the great majority of blue stragglers).

It seems that nowadays only one definite answer concerning the origin of blue straggler star phenomenon can be given: no known cause!

To change this situation significantly, new observational efforts should be started. Up to now we still have only a few spectroscopic investigations of blue stragglers based on high resolution spectroscopic material, from which we can learn something about the properties of only about dozen stars, but the properties of the great majority blue stragglers in open clusters remain unknown. Even the primary characteristic of these stars, their membership probability, is not known for the majority with the necessary precision.

7.5.2. What is new we can learn from the present study?

Let us consider briefly how our data fit the most popular hypotheses. Mass transfer, tidal captures and stellar mergers as an origin of open cluster blue stragglers should produce the rather high rotation of the newly formed stars, which is the result of the angular momentum conservation. Otherwise, one needs to assume some special condition for these processes. A high rotational velocity should be also an important input parameter for the single star evolution hypothesis. In fact, the high projected velocities are not observed for blue stragglers.

Mass transfer and stellar mergers must also produce some specific chemical peculiarities in blue stragglers. One can expect to detect an increased helium abundance on the surfaces of blue stragglers, provided they are formed via these mechanisms, due to the appearance in the upper atmospheric layers of the CNO-processed material from the stellar interior. But the blue stragglers investigated show either normal helium abundances, or an underabundance of this element. Global mixing of the merger's remnant caused by a coalescence of two stars should result in some decreasing of the surface carbon abundance and respective nitrogen enhancement (again due to the mixing of the atmospheric gas with CNO-processed material). The carbon abundances derived for the blue stragglers seem to support this hypothesis, but similar underabundances have been also derived for the main sequence stars.

Our spectroscopic analysis of a restricted sample of blue stragglers showed that as a group, these stars demonstrate the same chemical peculiarities as main sequence stars. The only obvious distinction between the blue stragglers and main sequence stars of the same spectral type are the projected rotational velocities.

We think that any proposed hypothesis for the blue straggler stars must first be able to account for two features: 1) the low projected rotational velocities, 2) an absent of specific chemical peculiarities inherent to the class of blue straggler stars.

Previous Section Next Section Title Page Table of Contents

© European Southern Observatory (ESO) 2000

Online publication: April 10, 2000