Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.sao.ru/hq/lizm/conferences/pdf/2006/2006_p025.pdf Äàòà èçìåíåíèÿ: Wed Apr 21 16:48:56 2010 Äàòà èíäåêñèðîâàíèÿ: Tue Oct 2 16:54:33 2012 Êîäèðîâêà:

Physics of Magnetic Stars, 2007, pp. 25­31

The magnetic fields of peculiar A and B stars in open clusters
John D. Landstreet1 , Vincenzo Andretta2 , Stefano Bagnulo3 , Elena Mason3 , Jessie Sila j1 , Gregg A. Wade4
Department of Physics & Astronomy, UWO, London, Canada INAF - Osservatorio Astronomico di Capodimonte, salita Moiariello 16, 80131 Napoli, Italy 3 European Southern Observatory, Casilla 19001, Santiago 19, Chile 4 Department of Physics, Royal Military College of Canada, P.O. Box 17000, Station `Forces', Kingston, Ontario, Canada K7K 7B4
2 1

Abstract. This pap er discusses our recent efforts to observe magnetic fields in a large sample of A and B stars in a numb er of op en clusters. The aim of this pro ject is to obtain a statistically significant sample of magnetic Ap and Bp stars, for which we can characterise the magnetic field structures and chemical abundances, and for which we have reasonably well-determined masses and ages. We exp ect that this sample will provide valuable constraints on the evolution of magnetic fields and chemical p eculiarities in these stars. Key words: stellar magnetic fields ­ upp er main sequence stars ­ op en clusters

1

Intro duction

Ab out 10% of A and B main sequence stars show p eculiar atmospheric chemistry, sufficiently p eculiar that the anomalies relative to stars of approximately solar chemical comp osition are readily seen in classification sp ectra. The chemically anomalous stars form several roughly homogeneous classification groups: Am stars, which range from ab out F0 to A0 and show a mo dest excess of iron p eak elements; Ap and Bp stars, which range b etween ab out F0 and B3, and show very substantial excesses of Cr and rare earths in co oler stars, changing gradually to excess Si and deficient He in hotter stars; HgMn stars, in the range b etween A0 and B6, which show large excesses of a small numb er of elements such as P, Mn, Ga, and Hg; He-weak stars, in the range b etween ab out B8 and B3, with clearly deficient He for their effective temp eratures; and He-strong stars, near B2, whose principal anomaly is strongly overabundant He. The two subgroups of Ap-Bp stars and He-strong stars, and some of the He-weak stars, are found to host readily detectable magnetic fields, usually in the range b etween 3 · 10 2 and 3 · 104 G. These field app ear to have a relatively simple global structure. Usually the fields are top ologically dip olar, although the detailed structure can vary considerably. These are the stars that will b e discussed in this pap er. These magnetic Ap and Bp stars have a numb er of other characteristics by which they differ from other p eculiarity groups, and from normal A and B stars. Most are p erio dically variable in one or more of photometric brightness (in typical photometry bands such as U, B , V ), sp ectrum, and magnetic field strength. When more than one characteristic varies, all vary with exactly the same p erio d, and with a fixed phase relation among variables. The p erio ds are in the range of 0.5 d up to some decades. Furthermore, the observed pro jected rotation velo city v e sin i is closely related to 25


26

LANDSTREET ET AL.

the p erio d; the smaller ve sin i is, the longer the p erio d. It was realized long ago that the p erio d of variability is the rotation p erio d, and this realization led to the phenomenological "oblique rotator mo del", in which the variations are due to an inhomogeneous distribution of elements and magnetic field vector over the stellar surface, which lead to variations in the observed quantities as the star rotates. Usually the dip olar overall structure of the field is not aligned with the rotation axis of the star; this is the origin of the "oblique" in "oblique dip ole rotator". The rotation p erio ds of magnetic Ap and Bp stars are several times longer than those of typical A and B stars; these stars typically have only ab out 10 or 20% of the sp ecific angular momentum of normal main sequence stars of similar mass. The very longest p erio d stars have less than 0.1% of the normal sp ecific angular momentum of normal stars. For recent reviews concerning many of the p oints summarized ab ove, consult the volume from the recent Poprad meeting on A stars (Zverko et al. 2004). The existence of rather large, and globally fairly simple, magnetic fields in some main sequence stars but apparently not in others (field measurements of some of the brightest normal A and B stars give upp er limits of some tens of G only: see for example Shorlin et al. 2002) raises a numb er of fundamental questions, to which some provisional answers have b een prop osed, but which are still not definitively settled. · What is the nature of magnetic fields found in Ap and Bp stars? How are they pro duced? The long-term stability, simple structure, lack of activity, and the lack of correlation b etween the field strength B and rotation rate, all suggest that the field is a fossil left from (at least) the pre-main sequence phase. · Why do Ap stars have fields while other A and B stars do not, or at most have fields some orders of magnitude weaker than those of Ap and Bp stars? There is no very satisfactory answer to this question yet. · How do es the magnetic field evolve as the star evolves (more sp ecifically, how do es the field evolve during the pre-main sequence and main sequence stages of the star's life)? If the field is a fossil, there should b e ohmic decay plus distortion and amplification due to stellar structure changes. Can this idea b e tested in some way?

2
2.1

Observational study of Ap evolution
Field magnetic stars

A large numb er of field (main sequence) Ap and Bp stars have b een studied in some detail. For many we have rotation p erio ds, overall magnetic field structure and strength, abundances of a numb er of elements and even simple mo dels of the patchy distribution of some elements. However, in general we either do know the ages of these stars, or do not know them with useful precision. For many, we also do not know the mass particularly accurately either. For any particular star in this well-studied sample, what we frequently know is that at some unknown time in the main sequence life of a star of uncertain mass, a magnetic Ap star can have an observed field structure and surface chemistry. If we were able to determine the mass and age (b oth absolute and "fractional" ­ the fraction already elapsed of the main sequence lifetime of that star) of a large numb er of Ap stars with known magnetic fields and chemistry, these stars would provide new clues ab out the evolution of magnetic fields and abundance patterns, and could b e used to test evolution theories of Ap stars. Is this practical?


THE MAGNETIC FIELDS OF PECULIAR A AND B STARS IN OPEN CLUSTERS

27

We need to determine masses, ages, and fractional ages for a substantial sample of Ap stars, since each star is only a snapshot of evolution, and these stars, even at a single mass, clearly have a distribution of such prop erties as field strength. A substantial sample might b e 50 ­ 100 stars, b ecause we would like to divide the sample into at least three or four mass bins as well as several age bins, with at least a few stars in each bin. There are enough well studied stars in the field to provide a sample of adequate size ­ but can we get masses and ages for these stars? This seems at first glance to b e a fairly straight-forward problem, esp ecially since the tremendous success of the Hipparcos satellite, which has provided hundreds of parallaxes of Ap and Bp stars that are accurate to ±10 or 20%. The obvious metho d to use is to determine T e from photometry, and log(L/L ) from apparent magnitude, b olometric correction, and distance, for a large sample of field stars. One then uses the results of standard evolution mo dels such as those of the Geneva group to determine mass and absolute and fractional age from the observed p ositions in the theoretical HR Diagramme. This would furnish a sizable sample of stars with known magnetic and abundance prop erties which can b e binned by mass and age. With the resulting sample, we could lo ok for statistical trends in field strength, chemical abundances, rotation p erio ds, etc. We would exp ect that such a sample would provide a numb er of really valuable hints ab out how b oth stellar magnetism and the asso ciated abundance anomalies evolve through the main sequence lifetime of a star. In turn, this should also provide valuable constraints on the mechanisms prop osed for field origin and evolution, for the development of chemical p eculiarities, and so on. In fact, this kind of study has already b een done by two groups. A first study was carried out by Hubrig et al. (2000), using a rather small sample of field magnetic Ap stars, essentially only ones for which the ratio of field strength to v e sin i is large enough that they have visibly split lines. More recently, this study has b een rep eated by Ko chukhov & Bagnulo (2006), for a considerably larger sample, including most of the field stars for which b oth accurate Hipparcos parallaxes and useful field strength estimates are available. In b oth cases the authors have placed the stars of their samples in the HR diagramme as accurately as p ossible, and then tried to discern evolutionary trends in field strength through the main sequence lifetime of the stars studied. Hubrig et al. (2000 ) have argued that their data supp ort a very surprising result, namely that in lower mass (M 3M ) magnetic Ap stars, fields do not appear at the surface of the stars until about 30% of the main sequence lifetime has elapsed. This result is disputed by Ko chukhov & Bagnulo (2006) on the basis of their larger sample of Ap stars, but as the Hubrig et al. conclusion raises fundamental questions (see b elow), it would b e valuable to test it farther. Is the conclusion of Hubrig et al. consistent with other things we b elieve that we understand ab out magnetic Ap stars? One ma jor problem raised by their result concerns rotation of Ap stars. As mentioned ab ove, the rotation p erio ds of these stars are typically a factor of ten longer than the rotation p erio ds of normal A and B stars, and in some cases hundreds or thousands of times longer. These stars are in general not memb ers of close binary systems, so that is not the origin of their slow rotation. Instead, it is generally b elieved (see St¸ n 2000) that this angular momentum is lost epie ´ during the pre-main sequence phase of the star's formation, when strong magnetic coupling with the accretion disk and outflowing wind allow the star to brake more efficiently than most stars do. This mo del of slowing not only explains the generally slow rotation of Ap stars, but also accounts for the fact that the slowest rotators are virtually always relatively low mass stars: the length of time sp ent as a pre-main sequence star rises rapidly as the mass decreases, so the lowest mass stars have the longest time in which to shed angular momentum. If the magnetic field only emerges for the first time well into the main sequence phase, then this picture makes no sense, and we are left with no plausible explanation of the slow rotation of magnetic Ap stars. Hence we need to carefully test the conclusion of Hubrig et al. There are imp ortant uncertainties asso ciated with efforts to place magnetic Ap stars in the HR Diagramme. A first problem concerns the effective temp erature calibrations. There are no


28

LANDSTREET ET AL.

magnetic Ap stars with fundamentally determined temp eratures (i.e. based on measured angular diameters and fluxes). Instead, we must adapt calibrations obtained for normal A and B stars to Ap stars. However, it is well known that, compared to a normal star of similar Paschen continuum slop e, Ap stars are usually deficient in flux in the Balmer continuum. Furthermore, in general, even mo del atmospheres with tuned abundances have not b een successful at repro ducing these p eculiar energy distributions, although such mo dels are often extremely go o d for normal stars (see for example Fitzpatrick & Massa 1999). In this circumstance, a numb er of empirical corrections have b een develop ed (e.g. Lanz 1984; St¸ n & Dominiczak 1989; Hauck & Kuenzli 1996) which epie ´ have generally concluded that effective temp eratures of Ap stars are a few hundred K lower than those of normal stars of similar photometric colours. This generally shared opinion has led p eople to guess that Ap Te 's may b e deduced from colours with uncertainties similar to those of normal stars, of the order of 2­300 K for A and B stars. However, this calibration has b een questioned recently by Khan & Shulyak (2006), who find, using the b est available mo dels of magnetic Ap stars (computed with LLMo dels, their own co de) that the Paschen continuum slop e of p eculiar magnetic stars is very similar to that of normal stars of the same Te . Thus, at present, we really ought to consider that effective temp eratures of magnetic Ap stars are uncertain by p erhaps as much as 500 K. This is quite a lot larger than the uncertainty assumed by Hubrig et al (2000), and it could b e a systematic effect, raising most or all T e values. Similarly, the value of log(L/L ) is uncertain by more than simply the distance uncertainty. This value is deduced with the aid of a b olometric correction (BC), usually the one for normal stars (see Co de et al. 1976; Malagnini et al. 1986). Sometimes even obsolete b olometric corrections such as that of Schmidt-Kaler (repro duced in Lang 1992) are used; Hubrig et al. used the Schmidt-Kaler BC, again intro ducing a systematic error. We have recently derived a new BC for Ap and Bp stars, which is systematically smaller than the one for normal stars. Even with this new result, we estimate that values of log(L/L ) with go o d parallaxes are uncertain by ab out ±0.1 dex. These uncertainties in turn lead to significant age uncertainties when the stars are placed in the HR Diagramme. The problem is particularly acute for stars near the b eginning of their main sequence life, as the iso chrones are quite close together, and a typical error b ox can easily result in an age uncertainty of the order of one-quarter of the main sequence lifetime, although the uncertainty in mass is relatively small, only p erhaps ±10%. A further imp ortant source of age uncertainty is the fact that we do not know the bulk comp osition (particularly the metallicity, Z ) of any particular Ap star. Since evolution tracks and iso chrones of mo dels of various Z values are displaced with resp ect to one another in the HR Diagramme (see Schaller et al. 1992; Schaerer et al. 1993), this leads to further age uncertainty; in fact, we find that, for field Ap stars, ages determined from HR Diagramme p ositions are usually only accurate enough to decide on whether a given star is in the first or second half of its main sequence life.

2.2

Cluster magnetic stars

It is certainly desirable to obtain more accurate ages than this. This can b e accomplished, esp ecially for stars early in their main sequence lifetimes, by studying magnetic Ap stars in op en clusters, for which age uncertainties are typically of the order of 0.2 dex, or less than a factor of two. This means that for very young stars (say 107 years old), the uncertainty relative to the full main sequence lifetime (which for A0 stars is of the order of 3 · 10 8 yr) is only a few p ercent, rather than roughly 50%. Note that this advantage diminishes as one lo oks at stars which are near the cluster turnoff, so that they have ages similar to that of the cluster. In this limit, the age uncertainty is similar for cluster memb ers and for field stars with go o d parallaxes. Two imp ortant recent advances have made cluster Ap stars accessible in interesting numb ers. The first is the new prop er motions from the Hipparcos pro ject, which has generated prop er motions with mas accuracy not only for the Hipparcos Input Catalogue stars, but also for the roughly


THE MAGNETIC FIELDS OF PECULIAR A AND B STARS IN OPEN CLUSTERS

29

2 million stars detected with the guide system, now publicly available as the Tycho-2 catalogue (HÜg et al. 2000). These new prop er motions are fairly complete to fainter than V 10, making them p owerful discriminants of cluster memb ership for A stars out to distances of several hundred parsecs. This fact makes it p ossible to confirm or reject memb ership of magnetic Ap stars in dozens of clusters, so that a usefully large sample of magnetic stars may b e gathered. In parallel, observations with Geneva and esp ecially a photometry (see Maitzen 1993) have made identification of probable magnetic Ap stars much more efficient by allowing easy selection of go o d candidate stars. The second advance is the presence of high-efficiency sp ectrop olarimeters on large telescop es. Two imp ortant recent additions to the previously available instruments (such as the Main Stellar Sp ectrograph on the SAO 6-m telescop e) are FORS1 on the ESO VLT, and ESPaDOnS at the Canada-France-Hawaii telescop e. FORS1 is a low-disp ersion multi-ob ject sp ectrograph with optional p olarisation optics, which has b een found to b e very efficient for magnetic measurements (see Bagnulo et al. 2002). It primarily relies on detecting fields through the Balmer lines, but has a resolving p ower which is just high enough that it can also detect the Zeeman p olarization in the metallic sp ectrum, although with substantially reduced efficiency. In contrast, ESPaDOnS is a single-ob ject high-resolution sp ectrograph sp ecifically designed for sp ectrop olarimetry, which has a very wide wavelength coverage. For field measurements of sharp-lined stars, ESPaDOnS is able to fully exploit the information content in the metallic sp ectrum, and for such stars, it is even more efficient at detecting fields than FORS1, in spite of the fact that the CFHT has only 20% of the surface area of an 8-m telescop e.

3
3.1

Magnetic fields in cluster stars
New observations

A few previous surveys have observed a small numb er of cluster stars, and also a substantial sample of stars in the Ori OB1 and Sco OB2 asso ciations (Borra 1981; Thompson et al. 1987). However, there are certainly not enough observations of cluster and asso ciation Ap stars available in the literature to study the evolution of fields using such stars. A new survey was required. Bagnulo et al. (2006) have used FORS1 to carry out a ma jor survey of probable Ap stars in more than 30 op en clusters. The goal of this survey is to obtain a significant sample of detected magnetic stars of relatively well-known age. Almost 100 candidate Ap stars (stars identified as probable Ap's on the basis of photometric indices or classification sp ectra, and probable cluster memb ers) have b een observed with a median uncertainty of ab out 80 G. Fields have b een detected in 41 of the observed stars; for 36 of these stars this is the first rep orted detection. This survey has required a lot of work (particularly on the part of Bagnulo and Mason) to develop robust and reliable reduction techniques for the data. The success of this effort is shown by the fact that no field was detected in any of the roughly 160 non-Ap stars observed during the survey, which shows clearly that this metho d of field measurement is not prone to spurious field detections. Furthermore, Bagnulo et al. have shown that for stars in which a field is detected from the Zeeman signature in Balmer lines, it is often p ossible (when the star observed has a rich sp ectrum of strong lines) to detect the Zeeman signature in the low-resolution metallic sp ectrum. The field measurement obtained from the metallic sp ectrum is generally in go o d agreement with that from the Balmer lines except for large fields (ab ove ab out 1 kG), for which the weak-field approximation used in data analysis breaks down for metal lines. Thus the metallic sp ectrum can often b e used to confirm or reject a marginal detection in Balmer lines. It is clear from our work that FORS1 is capable of obtaining field measurements with a standard error of the order of 30 or 40 G if enough exp osures are made, although achieving this error level requires very careful reduction. It is not known at present if the instrument is capable of achieving


30

LANDSTREET ET AL.

still lower uncertainties, or if this flo or is set by instrumental instabilities or remaining unidentified reduction difficulties. One particularly interesting star that was found during the course of this survey is star numb er 334 in the very young (2­3)·106 yr) cluster NGC 2244 in the Rosette Nebula, which had a measured Bz value of ab out 9 kG, the second highest value ever found in a non-degenerate star. This star has V 13, and is the faintest main sequence star in which a field has ever b een detected (Bagnulo et al. 2004). It app ears otherwise to b e typical Si-rich, He-weak Bp star.

3.2

Discussion of results

For the stars observed in the survey of Bagnulo et al. (2006), and the stars available from the literature, we have re-examined cluster memb ership on the basis of the b est available parallaxes, prop er motions, radial velo cities, and photometry. The data now available are usually sufficient to decide whether a star is a probable cluster memb er or not. For cluster memb ers, we then determine effective temp eratures from available uv by and Geneva photometry, using the calibrations of St¸ n epie ´ & Dominiczak (1989) and of Hauck & Kunzli (1996). Luminosities are found from cluster distance ¨ mo duli and apparent magnitudes, together with a new set of b olometric corrections sp ecifically for Ap and Bp stars. We are then able to place the stars in the HR diagramme, and compare their p ositions to standard evolution tracks and iso chrones (e.g. Schaller et al. 1992), using the cluster ages to constrain the range of allowed absolute and fractional ages (Landstreet et al. 2007). The analysis of this data set is not yet completed, but we do have a numb er of preliminary results. · The sample of known magnetic stars which are probable clusters is now sufficient for a useful statistical analysis. The present sample is rich in relatively young stars, reflecting the fact that typical cluster and asso ciation lifetimes are smaller than the main sequence lifetimes of A stars. The sample is also rich in relatively massive Bp stars, reflecting the fact that young clusters still have many of their more massive memb ers. · The median field strength found is somewhat larger for stars of more than 3 M than for lower mass stars. This is consistent with the results of Thompson et al (1987 ) for Sco OB2 and with Ko chukhov & Bagnulo (2006) for field stars. · The mean field strength Bz seems to decline somewhat with increasing star age in each mass interval. This is consistent with the results of Ko chukhov & Bagnulo for field stars. The decline is also consistent with flux conservation as stellar radius expands during main sequence evolution. · Contrary to the prop osal of Hubrig et al. (2000), a substantial numb er of stars which have gone through less than 30% of their main sequence lifetimes are found to exhibit magnetic fields. This was already demonstrated by an early result of our survey, in which a strong field was detected in the star HD 66318, a star which has completed only ab out 16 ± 5% of its main sequence lifetime (Bagnulo et al. 2003). This result is also consistent with the results of Po ehnl et al. (2003) and of Ko chukhov & Bagnulo. · However, hardly any stars in of the fact that magnetic Ap Ko chukhov & Bagnulo 2006); the cluster sample is deficient our new sample have masses of less than ab out 2 M , in spite stars o ccur in the field with masses down to ab out 1.6 M (see the roAp stars are all such stars. It is not known at present why in such stars.

The next step in this programme will b e to study the evolution of atmospheric chemical patterns with age thorough the main sequence life of the stars in our cluster-asso ciation sample. We


THE MAGNETIC FIELDS OF PECULIAR A AND B STARS IN OPEN CLUSTERS

31

exp ect that this will provide valuable clues ab out the long-term op eration of the sorting and mixing mechanisms that op erate inside such stars and that lead to the remarkable variety of Ap abundance patterns observed.
Acknowledgements. JDL, JS and GAW acknowledge support by the Natural Sciences and Engineering Research Council of Canada.

References
Bagnulo S., Hensberge H., Landstreet J.D., Szeifert T., & Wade G.A., 2004, A&A, 416, 1149 Bagnulo S., Landstreet J.D., Lo Curto G., Szeifert T., Wade G. A., 2003, Astron. Astrophys., 403, 645 Bagnulo S., Landstreet J.D., Mason E., Andretta V., Sila j J., Wade G. A., 2006, Astron. Astrophys., 450, 777 Bagnulo S., Szeifert T., Wade G.A., Landstreet J.D ., & Mathys G. 2002, Astron. Astrophys., 389, 191 Borra E.F., 1981, ApJ, 249, L39 Code A.D., Davis J., Bless R.C., Hanbury Brown R., 1976, ApJ, 203, 417 Fitzpatrick E.L., Massa D., 1999, ApJ, 525, 1011 Hauck B., Kunzli M., 1996, Baltic Astr., 5, 303 ¨ HÜg E., Fabricius C., Makarov V. V., Urban S., Corbin T., et al., 2000a, Astron. Astrophys., 355, L27 Hubrig S., North P., Mathys G., 2000, ApJ, 539, 352 Khan S., Shulyak D., 2006, Astron. Astrophys., 454, 933 Kochukhov O., Bagnulo S., 2006, Astron. Astrophys., 450, 763 Landstreet J. D., Bagnulo S., Andretta V., Mason E., Sila j J., Wade G. A., 2007, in preparation Lang K. R., 1992, Astrophysical Data: Planets and Stars, Heidelberg, Springer-Verlag, p. 137 Lanz T., 1984, Astron. Astrophys., 139, 161 Maitzen H.M., 1993, Astron. Astrophys., 102, 1 Malagnini M. L., Morossi C., Rossi L., Kurucz R. L., 1986, Astron. Astrophys., 162, 140 P¨ ohnl H., Maitzen H. M., Paunzen E., 2003, Astron. Astrophys., 402, 247 Shorlin S. L. S., Wade G. A., Donati J.-F., Landstreet J. D., Petit P., Sigut T. A. A., Strasser S., 2002, Astron. Astrophys., 392, 637 Schaerer D., Meynet G., Maeder A., & Schaller G., 1993, Astron. Astrophys., 98, 523 Schaller G., Schaerer D., Meynet G., & Maeder A., 1992, A&ASS, 96, 269 St¸ n K., 2000, Astron. Astrophys., 353, 227 epie´ St¸ n K., & Dominiczak R., 1989, Astron. Astrophys., 219, 197 epie´ Thompson I.B., Brown D.N., & Landstreet J.D., 1987, ApJS, 64, 219 z Zverko J., Zinovsky J., Adelman S. J., Weiss W. W. (eds.), 2004, "The A-Star Puzzle" (Cambridge University ´ Press, IAU Symposium No. 224) p. 633