EDP Sciences
Free Access
Issue
A&A
Volume 581, September 2015
Article Number A138
Number of page(s) 32
Section Catalogs and data
DOI https://doi.org/10.1051/0004-6361/201526424
Published online 24 September 2015

© ESO, 2015

1. Introduction

Chemically peculiar (CP) stars are upper main sequence objects (spectral types early B to early F) whose spectra are characterized by abnormally strong (or weak) absorption lines that indicate peculiar surface elemental abundances. The observed chemical peculiarities are thought to arise from the diffusion of chemical elements due to the competition between radiative pressure and gravitational settling (Richer et al. 2000; Turcotte 2003). CP stars constitute about 10% of all upper main sequence stars and are commonly subdivided into four classes (Preston 1974): metallic line (or Am) stars (CP1), magnetic Ap stars (CP2), HgMn stars (CP3), and He-weak stars (CP4).

The CP2 or Ap stars, which are the subject of the present investigation, are notorious for exhibiting strong, globally organized magnetic fields of up to several tens of kiloGauss (Kochukhov 2011). Their atmospheres are enriched in elements such as Si, Cr, Sr, or Eu and usually present surface abundance patches or spots (Michaud et al. 1981; Kochukhov & Wade 2010; Krtička et al. 2013), leading to photometric variability, which is considered to be caused by rotational modulation and explained in terms of the oblique rotator model (Stibbs 1950).

As a result, the observed photometric period is the rotational period of the star. Generally, CP1 stars are not expected to show rotational light variability; in relation to CP3 stars, this is still a matter of debate (Paunzen et al. 2013; Morel et al. 2014). Interestingly, in a recent study using ultra-precise Kepler photometry, Balona et al. (2015) have found that most of the investigated Am stars exhibit light curves indicative of rotational modulation due to star spots. With an amplitude of up to ~200 ppm, this kind of variability is reserved for high-precision (space) photometry, though. Photometrically variable CP2 stars are also designated as α2 Canum Venaticorum (ACV) variables in the General Catalogue of Variable Stars (Samus et al. 2007–2014, GCVS). Some Ap stars also exhibit photometric variability in the period range of ~520 min (high-overtone, low-degree, and non-radial pulsation modes). These stars are known as rapidly oscillating Ap (roAp) stars (Kurtz 1990) and are not the subject of the present paper.

Considerable effort has been devoted to the study of the photometric variability of the magnetic Ap stars by photometric investigations or by mining available sky survey data (Manfroid & Mathys 1986; Paunzen & Maitzen 1998; Paunzen et al. 2011; Wraight et al. 2012). The present paper investigates the photometric variability of Ap stars using the publicly available observations from the third phase of the All Sky Automated Survey (ASAS-3, Pojmański 2002) and aims at finding previously unidentified ACV variables. Observations, target selection, data analysis, and interpretation are described in Sect. 2. Results are presented and discussed in Sect. 3, and we conclude in Sect. 4.

2. Observations and analysis

The targets for the present investigation were primarily chosen among the Ap stars listed in the most recent version of the Catalogue of Ap, HgMn, and Am stars (Renson & Manfroid 2009, RM09 hereafter). At the beginning of the project, a number of Am stars from the aforementioned list were inspected, too. According to our expectations, most of them did not exhibit variability in ASAS-3 data, which led us to abandon this approach and concentrate only on the Ap stars. The three Am stars that exhibit obvious light variability might actually be Ap stars which have been spectroscopically misclassified as Am stars (Sect. 3).

The RM09 catalogue was cross-matched with the Tycho-2 catalogue (Høg et al. 2000), and all objects with VT< 11 mag that have a reasonable number of observations in the ASAS-3 database were investigated. These are mostly southern objects with δ< 0°. Additionally, other variables of spectral type late B to early A (with or without known chemical peculiarities) that have been investigated or discovered as by-products of other projects (Otero 2007) and which exhibit photometric variability typical of ACVs were included in our investigation.

The GCVS, the AAVSO International Variable Star Index, VSX (Watson 2006), the VizieR (Ochsenbein et al. 2000), and SIMBAD (Wenger et al. 2000) databases were consulted to check for an entry in variability catalogues. Objects that have already been announced as ACV variables in the literature were dropped from our sample.

2.1. Characteristics of the ASAS-3 data

The All Sky Automated Survey (ASAS) is a project that aims at continuous photometric monitoring of the whole sky, with the ultimate goal of detecting and investigating any kind of photometric variability. The first two phases of the project, ASAS-1 and ASAS-2, resulted in the discovery of about 3800 variable stars (Pojmański 1998, 2000).

The third phase of the project, ASAS-3, which has produced a catalogue of about 50 000 variable stars (Pojmański et al. 2005), lasted from 2000 until 2009 (Pojmański 2002) and has been monitoring the entire southern sky and part of the northern sky (δ< + 28°). The ASAS-3 system, which occupied the ten-inch astrograph dome of the Las Campanas Observatory (Chile), consisted of two wide-field telescopes equipped with f/2.8 200 mm Minolta lenses and 2048 × 2048 AP 10 Apogee detectors, covering a field of of sky. Data were obtained through standard Johnson I and V filters during the ASAS-2 and ASAS-3 surveys, respectively. About 107 sources brighter than about V = 14 mag were catalogued. With a CCD resolution of about 14.̋8 per pixel, the astrometric accuracy is around 35′′ for bright stars and up to 15′′ for fainter stars. Thus, photometry in crowded fields as in star clusters is rather uncertain.

The typical exposure time for ASAS-3 V-filter observations was three minutes, which resulted in reasonable photometry for stars in the magnitude range 7 ≲ V ≲ 14. The all-sky character of the survey necessitated sparse sampling. In general, a field was observed each one, two, or three days (Pigulski 2014) resulting in strong daily aliases in the corresponding Fourier spectra.

The most accurate photometry was obtained for stars in the magnitude range 8 ≲ V ≲ 10, exhibiting a typical scatter of about 0.01 mag (Pigulski 2014). However, because of the long time baseline of almost ten years, the detection limit of the ASAS-3 data for periodic signals is much lower than this value. David et al. (2014) have detected periodic variables with amplitudes of variability of 0.010.02 mag in the magnitude range of 7 ≲ V ≲ 10. Pigulski (2014) estimated that periodic signals with amplitudes as low as 45 mmag can still be detected well. Thus, the ASAS-3 data are perfectly suited to an investigation of the low-amplitude photometric variability of CP stars (Manfroid & Mathys 1986).

2.2. Data analysis and interpretation

The light curves of all target stars were downloaded from the ASAS-3 website1. No data from our targets exist in the ASAS-2 I-band catalogue. The light curves were inspected visually, and obvious outliers and data points with a quality flag of “D” (=“worst data, probably useless”) were removed. Promising candidates, i.e. stars showing a larger scatter than usually observed for apparent constant stars in the corresponding magnitude range with comparable instruments (Hartman et al. 2004), were searched for periodic signals in the frequency domain of 1 <f(c/d) < 30 using Period04 (Lenz & Breger 2005).

Objects exhibiting periodic signals well above the noise level (corresponding to a semi-amplitude of at least ~0.007 mag, as determined with Period04) were investigated. The data was folded with the resulting best-fitting frequency, and the light curve was visually inspected. ASAS-3 data sometimes suffer from systematic trends. For example, strong blending effects may induce significant additional scatter due to the inclusion of part of the neighbouring star’s flux, thus effectively rendering small-amplitude variations undetectable (Sitek & Pojmański 2014). Furthermore, instrumental long-term mean brightness trends might result in the detection of spurious periods. Consequently, objects whose light curves are indicative of strong systematic trends were rejected. Stars exhibiting convincing phase plots were kept. Borderline cases – that is to say, stars exhibiting a weak signal (semi-amplitude of 0.007 mag) that could not be attributed to systematic trends but did not produce a convincing phase plot either – were rejected as well in order to keep the sample free of possibly spurious detections that might contaminate the sample of derived rotational periods. The object was finally classified according to spectral type, colour information, period, and shape of its light curve.

The observed variability pattern of most stars is in accordance with an ACV classification. Judging from light curve characteristics, some eclipsing or rotating ellipsoidal variables might be present, too. Because ACV variables are prone to exhibiting double-wave variations in their photometric light curves (Maitzen 1980), it is sometimes difficult to distinguish between the light curves of double-waved ACVs and the variability induced by orbital motion (ellipsoidal variables/eclipsing variables), especially for objects exhibiting very small amplitudes and/or significant scatter in their light curves.

Generally, there seems to be a lack of binaries among most subgroups of Ap stars (for example, the Si, Si-Cr, and Si-Sr stars, Gerbaldi et al. 1985; Leone & Catanzaro 1999); in particular, very few double-lined spectroscopic binaries comprising an Ap star are known (Hubrig et al. 2014), and a conspicuous lack of short-period systems has been described in the literature (North & Debernardi 2004; Hubrig et al. 2005). Furthermore, the search for magnetic Ap stars in eclipsing binary systems has mostly been to no avail. Up to 2005, no such system has ever been found (Hubrig et al. 2005). RM09 list only five bona fide candidates for eclipsing Ap stars from which one, AO Velorum, has been confirmed (González et al. 2006). In their photometric study of Ap stars with the STEREO satellites, Wraight et al. (2012) did not identify any eclipsing binaries or ellipsoidal variables among their sample of 337 stars. However, some of them show distinct double-waved light curves. Furthermore, for the spectroscopic binaries with known periods among their sample, they were not able to establish a relation between the orbital period and the period of the observed light variations.

To sum up, it seems to be well established that the incidence of ellipsoidal variables or eclipsing binaries among Ap stars is very low. Accordingly, if the observed variability pattern is in general accordance with the rotational modulation caused by spots, we are inclined to interpret double-waved light curves as due to this mechanism. However, in doubtful cases, orbital motion cannot be decisively rejected as the source of the observed variability.

In fact, we have identified some stars among our sample that we consider to be promising eclipsing binary or ellipsoidal variable candidates, which – considering the size of the sample – is expected. HD 70817 is very likely an eclipsing binary, but its status as a CP star seems to be doubtful (Skiff 2014). Multicolour photometry with higher precision and/or spectroscopic studies are needed for a final conclusive classification. This will be part of a follow-up investigation.

3. Results

Most of the investigated Ap stars have never been the subject of a light variability analysis before, or, as in the case of HD 263361, they have been investigated and found constant or probably constant, and are described here as variable stars for the first time. Some of our targets have been identified as variable stars with or without a given period in the literature, but their variability types have not been determined or they have been misclassified. To the best of our knowledge, these stars are presented here as ACV variables for the first time. Interestingly, 16 of these objects already have known or suspected periods as listed in the RM09 catalogue. While this implies they are members of the class of ACV variables, these stars are not listed as known or suspected ACV variables in related papers and variability catalogues. Thus, we deemed it worthwhile to draw attention to these objects by including them in our sample.

Also included in our sample are three Am stars from the RM09 list, which exhibit light variations and periods typical of ACV variables. Two of them are marked as not likely to be CP objects in RM09. Under the assumption that these stars might possibly have been spectroscopically misclassified, we have listed them as ACV candidates (type “ACV:” in Table 2) and suggest further spectroscopic observations. The objects are flagged as Am stars in Table 4.

Some stars in our sample are listed with a twice longer or a twice shorter period in the literature. A twice longer (or shorter) rotation period cannot, in some cases, be definitely rejected for objects showing sinusoidal light variations and very small amplitudes and/or significant scatter in their light curves. However, the observed discrepancy is mostly due to classification differences. For example, some ASAS variables, which we have identified as ACV variables, have been classified as eclipsing binaries in the ASAS Catalogue of Variable Stars (ACVS) and are consequently listed there with a period that is twice as long. For very few cases, the period value listed in the literature is very different from our solution and possibly represents an alias of the real period.

We have checked the period solution of all doubtful cases (Sect. 2.2) and are confident that we have determined the period that fits ASAS-3 data best. This is also supported by the generally very good agreement of our period solutions to those from the literature. Because of the strong daily aliasing inherent to ASAS-3 data (Sect. 2.1), alias periods cannot be totally excluded.

Some stars in our sample (N = 31) have been spectroscopically classified as late B/early A-type stars, but have not been identified as (possible) chemically peculiar objects in the literature. Nevertheless, they have been included as ACV candidates (denoted as type “ACV:” in Table 2) in the present list of stars because of their typical photometric variability. Spectroscopic confirmation of these stars is needed to draw a final conclusion about their nature.

Table 1

Statistical information on the composition of the present sample.

Table 2

Essential data for the 316 stars identified as photometrically variable chemically peculiar stars or star candidates in the present paper.

Table 3

Essential data for the 7 stars identified as promising eclipsing binary or ellipsoidal variable star candidates in the present paper.

Table 4

Available information on single objects from the literature and miscellaneous remarks.

A part of our sample comprises Hipparcos unsolved variables (Koen & Eyer 2002), which have been automatically classified using random forests (Dubath et al. 2011) and a multistage methodology based on Bayesian networks (Rimoldini et al. 2012, R12 hereafter). Likewise, the probabilistic classification catalogue created by Richards et al. (2012, “Machine-learned ASAS Classification Catalog (MACC)”; RS12 hereafter) lists classifications and their probabilities for the (mis- or unclassified) ASAS variables included in our sample. Since the results of the employed automatic classification routines regarding our sample of ACV variables might potentially be of interest in estimating their performance, we have included the findings of the above-mentioned investigations in the presentation of our results. For stars having an entry in R12, we list variability types predicted by both methods, following the order of the VizieR online catalogue: predicted type random forests/predicted type Bayesian networks (probability predicted by random forests/probability predicted by Bayesian networks). It is interesting to note that some of our confirmed ACV variables are listed as doubtful cases in RM09. That these objects exhibit light variations typical of ACV variables is further evidence of their peculiar nature.

Table 1 presents statistical information on the composition of the present sample, and Tables 2 to 4 list the results of the present investigation and are presented in their entirety in the Appendix. Tables 2 and 3 present essential data for the 323 stars we have identified as photometrically variable magnetic Ap stars or star candidates. The tables are organized as follows:

Table 4 lists available information from the literature and miscellaneous remarks on individual objects. It is organized as follows:

  • Column 1: star name, HD number or other conventional identification.

  • Column 2: variable star designation from the literature.

  • Column 3: variable star type from the literature.

  • Column 4: period (d) from the literature.

  • Column 5: period (d) from this paper.

  • Column 6: reference in which – to the best of our knowledge – the object has been announced as a variable star for the first time.

  • Column 7: remarks of a miscellaneous nature: an asterisk denotes stars, whose status as chemically peculiar objects is doubtful according to RM09.

The light curves of all objects, folded with the period listed in Tables 2 and 3, are presented in the Appendix.

thumbnail Fig. 1

Distribution of the logarithmic rotational periods (upper panel) and V magnitude amplitudes (lower panel) among the photometrically variable Ap stars and Ap star candidates of the present sample (Table 2).

Open with DEXTER

thumbnail Fig. 2

Rotational periods as a function of (BV) and (JKs) of the photometrically variable Ap stars and Ap star candidates of the complete sample (upper panels) and the 99 stars of our sample for which we were able to determine the reddening (middle and lower panels; Table 5).

Open with DEXTER

3.1. Statistical analyses

Because the source of our investigation, the RM09 catalogue, is rather inhomogeneous, the present sample is not very suitable for a statistical analysis. However, to compare our results with those of other, similar investigations, we have included some informative plots. In the following, we do not include objects whose status as ACV variables remains doubtful (Table 3).

Figure 1 investigates the distribution of the logarithmic rotational periods (upper panel) and the V magnitude ranges (lower panel) of the photometrically variable Ap stars and Ap star candidates of our sample. Our results are in excellent agreement with the literature, illustrating the well-known peak around log P = 0.4 days (compare, for example, Fig. 7 of RM09) and the magnitude range of the variations (Mathys & Manfroid 1985). It is worthwhile pointing out, though, that the observed sharp decrease in numbers below a V magnitude range <0.03 mag is at least partially due to observational bias, since the ASAS-3 measurement uncertainties approach this value for the fainter objects, thus preventing the detection of very low-amplitude variables among the fainter stars. We want to stress that we might have missed some very low amplitude variations due to the selection and analysis process (Sect. 2.2). The rotational periods of Ap stars are more or less restricted to a rather narrow range. In light of the oblique rotator model, this means that a stable magnetic field configuration, which is needed to generate surface spots, can only survive for a distinct range of equatorial rotational velocities. For a detailed analysis of this important topic, one needs masses, luminosities, ages, and the inclination angles of rotating stars. This information cannot be straightforwardly determined from photometric data alone.

Figure 2 (upper panels) shows the colour indices (JKs) and (BV) versus the logarithmic rotational periods. It is a matter of debate and even some controversy in the literature whether there should be a correlation or not. While RM09 found no correlation between the periods and the (BV) colours for their extended sample (see their Fig. 8), Mikulášek et al. (2009) reached the conclusion that cooler magnetic CP stars (such as SrCrEu stars) rotate more slowly (see their Fig. 2). Of course, relying solely on colour information simplifies the problem dramatically. This approach does not take the luminosity into account, as well as the mass and thus evolutionary effects. We investigated such a possible simple colour correlation by employing a linear regression. We chose to restrict the colour index range to the most densely populated region of the diagram, which corresponds to (BV) ≤ 0.25 mag. This also allowed us a direct comparison with Fig. 2 of Mikulášek et al. (2009). In agreement with the findings of these investigators, we found a possible weak correlation between the rotational period and (BV) index (t parameter of 3.87), in the sense that cooler magnetic CP stars rotate more slowly. The slope was estimated as 0.57(19).

However, the reddening for our sample cannot be neglected. Most of our stars are located in the Galactic disk with a Galactic latitude |b| < 10° and 7 ≲ V ≲ 10 mag. For an A0 star with MV ~ 0 mag, this corresponds to distances between 250 and 1000 pc. In the Galactic disk, an absorption of AV (=3.1E(BV) = 5.636E(JKs)) of about 2 mag kpc-1 was observed for some regions (Chen et al. 1998; Dutra et al. 2002).

Table 5

Reddening values derived for 99 stars of our sample from UBV and uvbyβ photometry as well as the extinction model by Amǒres & Lépine (2005); AV( = 3.1E(BV) = 5.636E(JKs)).

The reddening for the targets was estimated using photometric calibrations in the Strömgren uvbyβ (Crawford 1978, 1979) and the Q-parameter within the Johnson UBV system (Johnson 1958). These methods are only based on photometric indices and do not take any distance estimates via parallax measurements into account. Photometric data were taken from the General Catalogue of Photometric Data (GCPD2). Where possible, averaged and weighted mean values were used throughout. For stars with Hipparcos parallax measurements (van Leeuwen 2007) with a precision better than 30%, the distance and Galactic coordinates were used to determine the reddening from the extinction model by Amǒres & Lépine (2005). If several estimates were available, they were compared and yielded excellent agreement. Table 5 lists the reddening values of the 99 stars for which we are able to determine the reddening. For the remaining stars neither photometry nor accurate parallaxes are available. The values for AV reach up to 1.5 mag.

As described above, we then employed a linear regression to the dereddened sample (Fig. 2, lower panels). No significant slope was found for this data sample. This could be interpreted as introducing a severe bias in the correlation when neglecting the reddening. To exclude the possibility of a selection effect, we also investigated the reddened colours of the sample from Table 5.

Although the slopes are slightly shallower, they are still significant (Fig. 2, middle panels). To verify our results, we investigated the behaviour of the correlation using Tycho-2 data instead of Kharchenko (2001) data, which led to the same conclusions. Furthermore, we employed the Spearman rank-order correlation coefficient (Spearman 1904), a non-parametric measure of correlation, to investigate our results. For this, we divided our sample (Table 2) into four groups using log P and (BV):

  • All stars, reddened: ρ = + 0.20, p = 0.0004, N = 316,

  • All stars with (BV) ≤ 0.25 mag, reddened: ρ = + 0.21, p = 0.0002, N = 295,

  • All stars, unreddened: ρ = − 0.05, p = 0.59, N = 99,

  • All stars, with (BV)0 0.25 mag, unreddened: ρ = − 0.03, p = 0.76, N = 98.

The two reddened samples show, with almost 100% probability, that there is a weak correlation that is not due to random sampling. In contrast, the other two samples are, with high probability, uncorrelated. Therefore, from a statistical point of view, the derived coefficients substantiate our prior conclusions. Clearly, a larger sample of CP stars with available rotational periods, along with precise stellar astrophysical parameters, is needed to investigate this topic further.

4. Conclusions

We have searched for photometric variability in confirmed or suspected Ap stars from the most recent version of the Catalogue of Ap, HgMn, and Am stars (RM09) using publicly available observations from the ASAS-3 database. We presented a sample of 323 photometrically variable magnetic Ap stars or star candidates, 246 of which are – to the best of our knowledge – described here as variable stars for the first time. A part of our sample (31 stars) is made up of late B to early A-type variables that are not included in the RM09 catalogue and that lack confirmation of their chemical peculiarity. They were included as ACV candidates because of their typical photometric variability but need spectroscopic confirmation.

We have compared our sample with the findings of similar investigations and find very good agreement. The well-known peak around P = 2 days in the distribution of rotational periods was confirmed. We found a possible weak correlation between rotational period and (BV) index, in that cooler magnetic CP stars rotate more slowly. However, this correlation seems to disappear when correcting for the interstellar reddening.

In agreement with the findings of David et al. (2014) and Pigulski (2014), we conclude that ASAS-3 data are well suited to investigating variable stars with rather low photometric amplitudes.


Acknowledgments

This project is financed by the SoMoPro II programme (3SGA5916). The research leading to these results received a financial grant from the People Programme (Marie Curie action) of the Seventh Framework Programme of the EU according to REA Grant Agreement No. 291782. The research is further co-financed by the South-Moravian Region. It was also supported by grant 7AMB14AT015 and the financial contributions of the Austrian Agency for International Cooperation in Education and Research (BG-03/2013 and CZ-09/2014). We thank the referee, Luca Fossati, for helpful comments and suggestions that helped to improve the paper. This work reflects only the author’s views so the European Union is not liable for any use that may be made of the information contained therein.

References

Appendix A: Complete Table 2

Table A.1

Essential data for the 316 stars identified as photometrically variable chemically peculiar stars or star candidates in the present paper.

Appendix B: Complete Table 4

Table B.1

Available information on single objects from the literature and miscellaneous remarks.

Appendix C: Light curves

thumbnail Fig. C.1

The light curves of all objects, folded with the period listed in Tables 2 and 3, respectively.

Open with DEXTER

All Tables

Table 1

Statistical information on the composition of the present sample.

Table 2

Essential data for the 316 stars identified as photometrically variable chemically peculiar stars or star candidates in the present paper.

Table 3

Essential data for the 7 stars identified as promising eclipsing binary or ellipsoidal variable star candidates in the present paper.

Table 4

Available information on single objects from the literature and miscellaneous remarks.

Table 5

Reddening values derived for 99 stars of our sample from UBV and uvbyβ photometry as well as the extinction model by Amǒres & Lépine (2005); AV( = 3.1E(BV) = 5.636E(JKs)).

Table A.1

Essential data for the 316 stars identified as photometrically variable chemically peculiar stars or star candidates in the present paper.

Table B.1

Available information on single objects from the literature and miscellaneous remarks.

All Figures

thumbnail Fig. 1

Distribution of the logarithmic rotational periods (upper panel) and V magnitude amplitudes (lower panel) among the photometrically variable Ap stars and Ap star candidates of the present sample (Table 2).

Open with DEXTER
In the text
thumbnail Fig. 2

Rotational periods as a function of (BV) and (JKs) of the photometrically variable Ap stars and Ap star candidates of the complete sample (upper panels) and the 99 stars of our sample for which we were able to determine the reddening (middle and lower panels; Table 5).

Open with DEXTER
In the text
thumbnail Fig. C.1

The light curves of all objects, folded with the period listed in Tables 2 and 3, respectively.

Open with DEXTER
In the text

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.