Online material

Appendix A: Notes on individual RR Lyrae stars

V19 and V21: in the catalog of C13, these variables are listed as possible non-members because their Fourier parameters are anomalous compared with other variables with similar periods (Clement & Rowe 2000). However, and as indicated in Table 1, the proper motion studies of vL00 and B09 found a membership probability of (100%,  100%) and (100%, 99%) for V19 and V21, respectively. The membership status for these two stars is further supported by their position in our NIR CMD, which locates them in the middle of the RRL IS (see the bottom panel of Fig. 1).

V52: this star appears to be much brighter than other RRab stars in ω Cen, as can be noted in Fig. 1. Its light curve has high photometric errors despite its high luminosity, thus suggesting contamination due to blending. Indeed, vL00 reported that V52 may have an unresolved companion, which would account for the fact that it is brighter and has a lower amplitude than other RRL variables with the same period. According to vL00 and B09, the membership probability of this star is 42% and 53%, respectively. Based on its position in our NIR CMD, the projected distance to the cluster center, d ~ 2′, and the high amount of noise in the light curve, V52 is most likely a member of the cluster in a very marked blend. Inspection of Hubble Space Telescope (HST) images (Anderson & van der Marel 2010) confirms the presence of a companion that is unresolved in our images, located just about 0.5′′ from the RRL star, but brighter by about 0.5 mag in V, and bluer by about 0.25 mag in B − V.

V56 and V168: the field status of these two stars was based on a radial velocity study by Liller & Tokarz (1981). Subsequently, V56 was again considered a member of the cluster based on its high membership probability of 98% and 100% found by vL00 and B09, respectively – a status in excellent agreement with its position in our NIR CMD. These inconsistent conclusions can probably be explained as follows: When assessing membership status of an RRL variable from its radial velocity, it is necessary to take into account the phase in its light and velocity cycle. Liller & Tokarz obtained this information from Dickens (1980, priv. comm.), based on an epoch established in 1974. The observations of Liller & Tokarz were made in 1978. A period change study by Jurcsik et al. (2001) indicates that V56 undergoes random period changes. Thus it is highly likely that the light-curve phase assumed by Liller & Tokarz was incorrect. In contrast, all proper motion studies agree in classifying V168 as a non-cluster member, which also agrees with our photometric classification as a field star: Fig. 1 clearly shows that this RRL star is a background interloper.

V68 and V84: these two stars have controversial mode classification in the literature. First thought to be RRLs by Bailey (1902), they were later suggested to belong to the ACEP group by Nemec et al. (1994) because their optical magnitudes are brighter than those of the other RRLs of the cluster. ACEP stars are Population II stars with periods between 0.4 and 2.4 days that do not follow either the RRL or the Type II Cepheid PL relations. Most of them are found in dwarf spheroidal galaxies (Nemec et al. 1994) and the Large Magellanic Cloud (Soszyński et al. 2008), a few ACEPs are known in Galactic globular clusters (see Zinn & Dahn 1976; Sollima et al. 2010; Osborn et al. 2012; Arellano Ferro et al. 2013) and only one in the Galactic field (XZ Ceti; Teays & Simon 1985; Szabados et al. 2007).

V68 appears to be an RRc cluster member star according to its proper motions and its position in the CMD (see Table 1 and the bottom panel of Fig. 1). Although it has the longest period among the RRc stars of the cluster (0.56 days), which is longer than even the short-period end of the RRab period distribution, the amplitude of its light curve is small enough to place it within the RRc stars group (see Fig. 2). V68 also follows the PL relations for RRc stars in J and K_S (Navarrete et al., in prep.). This star, on the other hand, also appears to follow the optical (B, V) PL relations for ACEP stars, as derived by Nemec et al. (1994).

On the other hand, V84 was discarded as a cluster member by vL00, with a membership probability . Unfortunately, we were unable to confirm this result using proper motions from B09, since V84 is outside their FoV. Our NIR CMD shows that V84 is clearly located in the RRL instability strip, opening the possibility that it is a cluster member. However, as Del Principe et al. (2006) pointed out, V84 does not follow the NIR PL relation for RRab stars. This result is confirmed using our PL relations (Navarrete et al., in prep.), with V84 appearing brighter than other stars with the same period. Conversely, Nemec et al. (1994) found that V84 was placed slightly below the B- and V-band PL relations for ACEPs, which could be explained if the star was not an ACEP but instead an RRL star leaving the HB.

Using our NIR observations, we cannot adopt a definitive classification for either of these stars. If they are RRLs, the most probable scenario is that V68 is an RRc member and V84 is an RRab star from the field. However, the ACEP classification cannot be completely discarded. Indeed, V68 and V84 follow the K_S-band PL relation for ACEPs in the Large Magellanic Cloud derived by Ripepi et al. (2014), assuming a distance modulus for ω Cen of (m − M)₀ = 13.63 ± 0.1 mag – which agrees well with the one derived by Matsunaga et al. (2006) using Type II Cepheids. If these two stars are indeed ACEPs, they probably are cluster members.

V80: this star was detected in the first variability study on ω Cen carried out by Bailey (1902). In that work, V80 was reported as difficult to measure because it is distant from comparison stars, and accordingly only sparse measurements of differential magnitudes were published, without an accompanying light curve. Bailey suggested a period of 0.45 or 0.31 days for this star, wherewith it was tentatively classified as an RRL-type star. Based on our wide-field VISTA observations, we determined its period, P = 0.37718 d (Table 1), and derived its complete, phase-folded light curve (Fig. A.1).

According to its coordinates, ID-53, P = 0.377 d from Weldrake et al. (2007), appears as the nearest star to V80 among the list of new variables discovered by these authors. However, Weldrake et al. did not associate ID-53 with V80, probably because a period value had never previously been published for V80 – and accordingly, they claimed that their ID-53 was a newly discovered variable star. Instead, our independently derived period and coordinates alike indicate that V80 and ID-53 are one and the same star.

V80 is not present in any proper motion study. Its derived period and position in the CMD indicates that V80 is an RRc star that is also a cluster member.

V118, V135, and V139: these three stars appear to be clearly located within the RRab region in the IS (see Fig. 1). Based on the vL00 and B09 proper motion studies, V118 has a membership probability of (100%, 96%), respectively, whereas V139 has (100%, 90%). V135 is only present in B09, with a membership probability of 97%. However, all appear to be brighter than the other RRab stars with similar periods in the J and K_S PL relations (Navarrete et al., in prep.).

V118 is located in a very crowded region. According to the HST catalog of Anderson & van der Marel (2010), there are more than ten stars inside a 1″ radius around its coordinates. A similar situation occurs with V139, having a neighbor located at only 0.432′′, according to the catalog of Anderson & van der Marel – which is a difference of less than two pixels, given the VIRCAM pixel scale. Hence, this star is almost certainly brighter in our photometry than it would be if it were isolated.

V135 is suffering the influence of a disturbing neighbor, located 2.5″ away, which is saturated. This situation was previously reported by Samus et al. (2009), with both stars being labeled in that study as visual binaries.

V143: it is certainly a member of the cluster according to vL00, with 99% membership probability. B09 give instead for this star. However, it appears a few tenths of a magnitude brighter than other RRab stars with the same color. A visual inspection of this star in some of our images suggested to us that its anomalous position in the CMD is not due to a blend effect of an unresolved system, but rather to a saturated companion located close by (i.e., at a distance of ≈3.4″) that is somehow disturbing the photometric quality in its neighborhood. Indeed, this situation was explicitly reported by Shokin & Samus’ (1996), where the presence of a bright star located southwest from the position of V143 is remarked upon. Therefore, despite its anomalous position in the CMD, we suggest that V143 is indeed a cluster member.

V151: listed as an RRab star by Kaluzny et al. (2004) with a period of 0.4078 days, this star would appear to be the shortest-period fundamental-mode RRL in ω Cen. However, despite the fact that our estimated period (P = 0.407756 d) excellently agrees with the one derived by Martin (1938), we note that its light curve shows a very small amplitude, which is among the lowest for RRab stars in our study. Small amplitudes are unusual for short-period RRab stars. Figure 2 shows the position of V151 in the Bailey (period-amplitude) diagram, as compared to other ω Cen RRL stars. This plot clearly shows that V151 is well located among the RRc-type variables. In addition, its position in the PL relation (Navarrete et al., in prep.) is also consistent with first-overtone pulsation. Thus, we conclude that this star should be classified as an RRc. We note that Martin (1938) also classified V151 as an RRc star, hence the erroneous RRab type adopted by Kaluzny et al. is probably a typographical error, which was later inadvertently reproduced in the online catalog of C13. V151 was not included in the proper motion studies of vL00 and B09; however, our NIR CMD suggests that this variable probably belongs to the cluster.

Fig. A.1 VISTA NIR light curves for all the RRLs that lacked a derived light curve or period in the literature. For each star, upper panel: J-band curve, whereas the lower one displays the KS band instead. Our derived period for each star is provided in each panel. Note that the V180 light curve is consistent with a W UMa classification. Variables ID-91 and (especially) ID-99, from Weldrake et al. (2007), have light-curve shapes that are different from all the other RRab in ω Cen, and are suspected to be field stars.

Open with DEXTER

V172, V173, V175, V177, and V180: Wilkens (1965) announced the discovery of V171-V180. Of these ten stars, V176 was subsequently discarded as a non-variable star, V174 classified as an EB, and the remaining stars classified as RRab’s, except for V180, which was classified as an RRc instead. To the best of our knowledge, no further variability studies of these stars have been carried out, except for V172 and V177, which were included in the observations by Weldrake et al. (2007) as ID-18 and ID-49, respectively. None of them has had a period assigned in the catalog of C13.

Four of Wilkens’s stars, namely, V171, V175 (without a previous type assigned), V178, and V179 are all very far from the cluster center and thus were not observed by us, despite VIRCAM’s large FoV. Consequently, only light curves for V172, V173, V177, and V180 were derived, and they are shown in Fig. A.1. Only V180 has membership probability derived based on proper motions studies, being considered a non-member of the cluster with a null membership probability by vL00. According to their positions in the CMD (bottom panel of Fig. 1), we suggest that V172, V173, and V177 are cluster members. We confirm the ab-type classification for V172, but contrary to Wilkens, we did not find an RRab behavior for either V173 or V177, and we suggest instead an RRc classification for both these stars. This is consistent with their positions in the Bailey diagram, as can be seen from Fig. 2. V175, according to Vivas (2014, priv. comm.; see also Fernández Trincado et al. 2013) is a foreground c-type RRL. As can be seen in Fig. 1, V180 is located in the RRab region, redder than any other RRc star. Moreover, it has a larger amplitude than other RRc stars (Fig. 2). Our estimated period produces a light curve clearly showing two different minima. Accordingly, we propose to classify V180 as a contact EB of the W Ursae Majoris type, rather than as an RRc, which is consistent with its fairly red color and the fact that this star is not a cluster member.

V181 and V183: V181, V182, and V183 were discovered by Wesselink (1969, unpublished, priv. comm. to Sawyer-Hogg), who indicated the stars’ coordinates and periods, but without providing light curves or variability types for these stars. Clement et al. (2001), based on Wesselink’s periods, classified V181 and V182 as RRab’s, and V183 as an RRc. These stars were outside the FoV in subsequent investigations by other authors. Unfortunately, V182, at 59.7 arcmin from the center of the cluster, is also outside the FoV of VISTA.

Figure A.1 shows the first complete light curves for V181 and V183. They are ab- and c-type RRL stars, respectively, in agreement with the periods originally derived by Wesselink. These two stars are very distant from the cluster center (though still inside its tidal radius), and thus they are absent from proper motion studies of the cluster. However, based on their positions in the CMD (see the top panel of Fig. 1), they both appear to be field RRL stars.

V268: vL00 claimed that V268 is a field star with a membership probability of 0%. However, we found some inconsistencies in this membership status using our coordinates and those from the literature: in fact, the nearest star to V268 in the catalog of vL00 has a membership probability of 70%. Moreover, according to B09, this star has a 99% membership probability. Using the NIR CMD to determine the membership status of this star, it is found that V268 is located well inside the RRL instability strip, as the other RRab members of the cluster (Fig. 1, bottom panel). We thus conclude that V268 is indeed a cluster member.

V283: this star was neither studied by vL00 nor by B09. Based on its faint visible magnitude, Kaluzny et al. (1997) claimed that it is a background halo RRL. This field status is confirmed by our photometry, which locates the star a few magnitudes below the HB, even fainter than the turn-off point level (see CMD in the top panel of Fig. 1).

V349 and V351: these two RRc variables were discovered by Kaluzny et al. (2004) and are located ~1.5 arcsec away from the cluster center. The huge amount of crowding at their positions did not allow those authors to present properly calibrated light curves. Interestingly, Kaluzny et al. noted that V351 is a multiperiodic variable, and suggested that it may show non-radial pulsation modes, as previously proposed by Clement & Rowe (2000) for several other RRL stars in ω Cen. Using the coordinates and periods provided by Kaluzny et al. (2004), these two variables were not found either in our ALLFRAME or in our DoPhot photometry. The matching radius used was 1 arcsec. Increasing the matching radius to 10 arcsec still did not return any variable stars. The last check step was to consider all the stars within 30 arcsec around the listed coordinates, and check for the presence of periodicity in this enlarged sample using the analysis of variance statistic (ANOVA; Schwarzenberg-Czerny 1989). Again, no periodic variables were detected in this way. Therefore, V349 and V351 could not be recovered using PSF photometry in our VISTA images.

Appendix B: RR Lyrae stars from Weldrake et al. (2007)

Weldrake et al. (2007) reported the discovery of five new RRL stars in ω Cen: ID-53, ID-91, ID-99, ID-144, and ID-145. By carefully checking the positions and periods of these stars, we have been able to establish that three out of these five variables were not really new discoveries. More specifically, according to our analysis, we were able to establish that ID-53 = V80, ID-144 = V272, and ID-145 = V15.

ID-91 and ID-99, on the other hand, do appear to be bona fide new variables. Weldrake et al. (2007) listed optical magnitudes of V ~ 17.1 and 17.45 mag for these two stars, respectively, which is consistent with their being field interlopers. Unfortunately, neither star appears in the proper motion catalogs of vL00 and B09. For their variability status, Weldrake et al. suggested that ID-99 could be either a long-period RRc with a 0.671-d period or an EB with a period twice as long. For ID-91, in turn, they assigned an RRab type. Both stars’ light curves are shown in Fig. A.1, according to which the RRab status of ID-91 seems reasonable, whereas the RRL nature of ID-99 remains a possibility. However, if indeed an RRL, ID-99’s long period might more naturally suggest an RRab classification, as opposed to the RRc type suggested by Weldrake et al. To further check their classification, ID-91 and ID-99 were plotted in the Bailey diagram (see Fig. 2). Their location suggests that both stars are either extreme examples of OoI RRab stars and/or show a modulation in their light curves (e.g., the Blazhko effect), and/or are unresolved blends. The latter possibility receives some support from the positions of the stars in the CMD (Fig. 1), which show that both stars are slightly redder than expected for RRL stars. However,

these stars are also quite far away (~25′) from the cluster center, and inspection of our images reveals that they are in relatively unpopulated areas, which makes blends seem unlikely. Unfortunately, HST images of these fields are not available, therefore we cannot conclusively establish that unresolved companions to these stars are indeed absent. Similarly, no Blazhko effect is supported by either our photometry (Fig. A.1) or the photometry of Weldrake et al. (2007).

Based on the astrometry that is provided along with the VISTA images (which is tied to the 2MASS system), we found some erroneous matches between some of the known variables in the catalogs of Weldrake et al. (2007) and Kaluzny et al. (2004): ID-115 = V264 (instead of V48) and ID-135 = V266 (not V356). Moreover, Weldrake et al. erroneously claimed that ID-133 was the same star as V144 from the catalog of Clement et al. (2001). However, as already reported by Navarrete et al. (2013), ID-133 is indeed a different variable star. C13 has already added this new RRab star in her online catalog, where it is now officially listed as V411.

C13 has included ID-74 of Weldrake et al. (2007), a possible RRL based on its period (0.6671 d, according to Weldrake et al.), as V433 in her catalog. This star has V ≈ 16.5 and (V − I) ≈ 1.0 mag, which suggests that it is a field star. Indeed, according to our NIR magnitudes, V433 is placed above the cluster’s turn-off point, above the main sequence and close to (but redder than) the base of the red giant branch (see the top panel of Fig.1). Still, the star does appear to show some short-period variability, with a period of (0.6681 ± 0.0004) d, and an amplitude that probably does not exceed 0.05 mag in K_S, again preventing us from classifying it as an RRL. The star’s red color is probably not caused by a blend with a red star, since it is ~30′ away from the cluster center, and without any obvious close companions in our images. According to Da Costa & Coleman (2008), the star is a radial velocity non-member. In conclusion, we follow Weldrake et al. (2007), and consider V433 as a variable star with an uncertain variability type.

© ESO, 2015