This site uses cookies. By continuing to use this site you agree to our use of cookies. To find out more, see our Privacy and Cookies policy. Close this notification

Variable Stars in the Field of NGC 6882/6885: The Case of V381 Vulpeculae and V382 Vulpeculae

and

Published 2005 July 22 © 2005. The Astronomical Society of the Pacific. All rights reserved. Printed in U.S.A.
, , Citation Eric G. Hintz and Michael B. Rose 2005 PASP 117 955 DOI 10.1086/432147

1538-3873/117/835/955

ABSTRACT

We present photometric and spectroscopic results for two reported δ Scuti stars in the field of NGC 6882/6885. We find that V381 Vul has a period of 0.1185 days and is a δ Scuti variable, as previously reported. The spectra of V382 Vul shows it to be a B3 star and therefore not a δ Scuti. All evidence points to V382 Vul being a β Cephei star with a period of 0.1808 days. Additionally, we report five new variables and eight suspected variable stars. Of the five new variables, two are pulsators and three are eclipsing binary systems. In our search for new variable stars, we use a "robust median statistic" that is proven to be better at finding low‐amplitude variables than the traditional error curve approach.

Export citation and abstract BibTeX RIS

1. INTRODUCTION AND MOTIVATION

During the last 3 years a set of spectral data in the region of Hβ has been collected at the Dominion Astrophysical Observatory in Victoria, Canada. This is a full survey of all northern δ Scuti stars brighter than 13th magnitude as selected from Rodriguez et al. (2000). During one observing run, two reported δ Scuti stars, V381 Vulpeculae and V382 Vulpeculae, in the open cluster NGC 6882/6885 were observed. These two stars were both reported as potential δ Scuti variables by Peña et al. (1990). Surprisingly, the spectra for one of these stars, V382 Vul, had an emission feature centered in the Hβ absorption line. This would be a very unusual result for a δ Scuti variable.

With this result still fresh on the screen, a search was made of the literature to check for previous reports on V382 Vul. In Henize (1976) the star was reported as an emission‐line star, but no other information was given. It is also listed in a catalog of Hα emission stars compiled by Kohoutek & Wehmeyer (1999). A search was then made to confirm the identification of the star. In Hoag et al. (1961) they presented UBV photometry for stars in the region of NGC 6882/6885 and provided the generally accepted numbering system. A careful search was made of the Hoag et al. (1961) maps for NGC 6882/6885 to confirm the Peña et al. (1990) identification. We are confident that the star identified as a potential δ Scuti variable is the same star reported as an emission‐line star.

The open clusters NGC 6882 and NGC 6885 were originally described by Trumpler (1930) as potentially a single cluster, with NGC 6882 being a condensation inside of NGC 6885. Johnson (1961) did not believe NGC 6885 to be an actual cluster. However, the work of Svolopoulos (1961) seemed to indicate that NGC 6885 should be considered a cluster. The recent WIYN open cluster survey (Platais et al. 2003) again brought up the question of just what constitutes a cluster in this region. Therefore, for simplicity we refer to the cluster as NGC 6882/6885 throughout this paper.

A number of important physical characteristics for NGC 6882/6885 have been reported in the literature. From high‐dispersion spectroscopy of two stars in the region of NGC 6882/6885, Luck (1994) determined [Fe/H] = -0.02. From Washington system photometry, Geisler et al. (1991) found [Fe/H] = -0.15. Finally, Strobel (1991) reported [M/H] = -0.20 and a distance of 581 pc. These values all point to NGC 6882/6885 being just slightly metal poor. The distance to NGC 6882/6885 in general has been reported as 600 pc (Johnson 1961; Svolopoulos 1961; Becker & Fenkart 1971), with one value a little closer, at 525 pc (Hoag & Applequist 1965). However, from four stars examined by the Hipparcos satellite, Robichon et al. (1999) got a parallax of 2.52 ± 0.36 mas, which corresponds to a distance of 397 ± 50 pc. Platais et al. (2003) concluded that NGC 6882 is considerably reddened [E(B - V) = 0.6] and at a much greater distance (over 1 kpc) than in previous studies. They also believed that NGC 6885 is incorrectly classified as a cluster and might be part of an OB association. Although we do not directly address the question of what constitutes a cluster, we can place the two stars of interest in proper context.

Two unusual possibilities provided the motivation for this project. The existence of an emission‐line δ Scuti would be interesting, and the contradictions presented by NGC 6882/6885 make the discovery of variables in the region useful. Therefore, follow‐up photometric observations of NGC 6882/6885 were taken. The images cover both V381 Vul and V382 Vul, along with many other stars in the region. Below, we take a detailed look at both suspected δ Scuti variables and at previously unreported variables in the field.

2. OBSERVATIONS

Spectroscopic data were secured in 2003 September at the Dominion Astrophysical Observatory on the 1.8 m Plaskett Telescope. Observations were made with the Cassegrain spectrograph, using the 21121B grating. This grating is blazed at 4100 Å and yields 15 Å mm−1. Using the SITe‐2 CCD with 15 μm pixels gives 0.23 Å pixel−1. The grating was set to give a central wavelength of 4773 Å, with coverage from 4570 to 4970 Å.

Photometric observations were secured with the 0.4 m David Derrick Telescope (DDT) of the Orson Pratt Observatory on the Brigham Young University campus. The DDT was equipped with an Apogee Ap47p CCD camera mounted at the Newtonian focus. This provided a plate scale of 1 farcs32 pixel−1 and a field of view 22 farcm5 on a side. This field of view allowed V381 Vul and V382 Vul to be observed simultaneously. However, the location of V381 Vul permitted us to observe only the northwest corner of the cluster. The field of view is shown in Figure 1, along with our numbering system.

Fig. 1.—

Fig. 1.— Portion of field near NGC 6882/6885 covered by CCD frames. The field of view shown is 30' × 30' in order to cover all telescope positions. The numbers in the figure match those from Table 3.

An Optec filter slide was used with a set of VRI filters and a set of BVR filters, both modeled after Bessell (1990). Total time coverage is 15 hr of data over five nights. In addition, two of the nights were calibrated against standard stars selected from Landolt (1992). All frames were reduced using standard IRAF methods.

3. Hβ SPECTRAL REGION OF V381 VUL AND V382 VUL

In Peña et al. (1990) they classified V381 Vul as an F5 star, based entirely on its (B - V) color. Published spectral observations have given a spectral type of A7 (Cannon & Mayall 1949). These are dramatically different classifications. Our new spectra for V381 Vul are shown as the bottom spectra in Figure 2. Using the online digital spectra detailed by Bagnulo et al. (2003), and comparing the depths of a number of features, we find that we agree with Cannon & Mayall (1949) and classify V381 Vul as an A7 IV. The discrepancy between our value and Peña et al. (1990) clearly indicates that V381 Vul has a significant amount of reddening, which we discuss further in the next section.

Fig. 2.—

Fig. 2.— Comparison of V381 Vul spectra with three δ Scuti variables of known metal content. These three are SS Psc, RV Ari, and BP Peg, which have published [Fe/H] values of −0.10, 0.00, and −0.02, respectively.

A number of additional stars in the A to F spectral range that have established values for the Hβ index (Rodriguez et al. 2000) were also observed on the same night as V381 Vul and V382 Vul. Spectrophotometric indices were calculated using the SBANDS package in IRAF and were zero‐point–corrected using the established values. This gave a value for V381 Vul of Hβ = 2.77 ± 0.02, which is in good agreement with the spectral type determined earlier. A comparison of V381 Vul to three other δ Scuti stars (BP Peg, SS Psc, and RV Ari) is shown in Figure 2. These three stars have established values for [Fe/H] of −0.02, −0.10, and 0.00, respectively (McNamara 1997). The metal content of V381 Vul appears to be similar to or slightly more metal‐poor than SS Psc. This estimate for the metal content is consistent with the value given by Geisler et al. (1991).

The case of V382 Vul is much more complex. Once again Peña et al. (1990) gave a spectral type of F6, based entirely on the (B - V) color. The spectral type was given as B8 by Cannon & Mayall (1949). Our current spectra of V382 Vul is shown in Figure 3. Because of the emission line in the center of Hβ, we do not use it for classification. We turn to two He i lines to determine the spectral type of V382 Vul. The first is the deep line at 4721.929 Å, and the second is the doublet of 4713.143 and 4713.373 Å. From a comparison of these lines with digital spectra from Bagnulo et al. (2003), we determine an earlier spectral type of B3 V for V382 Vul. However, there are additional lines that cannot be explained by a B3 V. The source of these additional lines is unknown. The combination of these spectral lines can be seen in the changing shape of the spectral features near the 4713 Å He i lines, as shown in Figure 4.

Fig. 3.—

Fig. 3.— Continuum‐corrected spectra of V382 Vul in the region of Hβ.

Fig. 4.—

Fig. 4.— Three spectra of V382 Vul centered on 4713 Å. The shape of this line is clearly changing due to the presence of a companion star. Frames are approximately 15 minutes apart in time.

The final information gathered from the spectra of V381 Vul and V382 Vul are the radial velocities for each star. The radial velocities were determined using the RVIDLINES package from IRAF. We obtain an average velocity from three observations of 31.2 ± 1.8 km s−1 for V381 Vul. The radial velocity solutions for V382 Vul are again more complex and are divided into the emission line and the absorption lines. For the Hβ emission line, we find a velocity of 14.0 km s−1. From the absorption lines, we give a best estimate for the velocity of − 4.6 ± 2.5 km s−1. The complex changing of the lines makes this very difficult. However, our value is within the stated value and error of 1.9 ± 11.2 km s−1 given by Geisler (1988). The radial velocities for individual observations are gathered in Table 1.

4. VRI PHOTOMETRY

To aid in our discussion of the photometric measures in the region of NGC 6882/6885, we collect previously published values for spectral types, radial velocities, and UBV photometry in Table 2. The first stage in the reduction of the new observations was the calibration of 95 stars in the field, using the Landolt (1992) standards. In the final reduction, star 76 was found to be saturated and therefore was removed from further consideration. Between 40 and 50 observations of each star were secured. The average values for V, (V - R), and (R - I) are reported in Table 3, along with estimates for the error of the mean. We compared our V magnitudes with those found in Hoag et al. (1961) and found them to deviate only by a small zero‐point shift of 0.01 in the V filter.

As mentioned in the previous section, there is significant reddening in this region. In Figure 5 we plot the (V - I) color of each star for which we have a published spectral type, as given in Table 2, along with standard (V - I) colors drawn from Fitzgerald (1970) and Ducati et al. (2001). From this we find two distinct groups of stars. The upper left group consists of V382 Vul and star 22. These two stars give an average E(V - I) = 0.78. Using the formulae found in Taylor (1986), we convert this to E(B - V) = 0.61, which is in excellent agreement with Platais et al. (2003). Correcting for the reddening and estimating the absolute magnitude from the spectral types, we estimate that these two stars are at approximately 1200 pc. This places V382 Vul in what Platais et al. (2003) consider NGC 6882. The second group of stars is made up of V381 Vul and stars 2, 4, 5, 25, 69, and 72. These stars have E(V - I) = 0.22 or E(B - V) = 0.17. This gives an average distance for this group of 540 pc. This group of stars does not belong to either the cluster favored by Platais et al. (2003) or the closer group reported in Robichon et al. (1999). In fact, this group seems to be located at the original distance given for the combined cluster NGC 6882/6885 (Johnson 1961; Svolopoulos 1961; Becker & Fenkart 1971).

Fig. 5.—

Fig. 5.— Plot of (V - I) color vs. spectral type. The solid line was generated from data drawn from Fitzgerald (1970) and Ducati et al. (2001). The filled circles indicate stars from this survey with known spectral types (0 is B0, 10 is A0, etc.).

The data from Table 3 are used to generate the color‐magnitude diagram shown in Figure 6 by the open circles. Data from five stars observed by the Hipparcos project are plotted as filled circles. The two filled circles that overlap with two open circles are the two stars in common between the two sets. The agreement is within the errors. Four of the Hipparcos stars have good parallax measurements (Robichon et al. 1999), with an average value of 2.52 ± 0.36 mas. This gives a distance of 397 ± 50 pc. Using the Y2 isochrones reported in Yi et al. (2003), we generated two main sequences, which are also shown in Figure 6. The first is for a distance of 400 pc, with no reddening, and is represented by the solid line. The second is for a distance of 1200 pc, E(B - V) = 0.61, and the age estimated by Platais et al. (2003), and is shown by the dashed line that is lower and to the right in Figure 6. The reddening and distance of the more distant cluster place its main sequence on the same line as the unreddened nearby stars, making interpretation difficult.

Fig. 6.—

Fig. 6.— Color‐Magnitude diagram for region of NGC 6882/6885 covered by this survey, along with stars from the Hipparcos catalog. The open circles represent data from the current survey, and filled circles show stars taken from the Hipparcos catalog. Isochrone 1 is for an unreddened clusters at 400 pc. Isochrone 2 is for a heavily reddened cluster at 1200 pc.

5. VARIABLE STARS IN THE FIELD

In addition to V381 Vul and V382 Vul, we examined 92 other stars in the field to look for variability. The differential photometric reduction techniques are detailed in Hintz et al. (1997). There are three previously known variable stars in the field of NGC 6882/6885. SU Vul, identified as star 23, is listed as a red irregular variable in the General Catalog of Variable Stars (GCVS). V381 Vul and V382 Vul, stars 1 and 28, respectively, are both listed as potential δ Scuti stars, as mentioned earlier. For each night, the ensemble was slightly different, due to centering differences, but in general each final ensemble contained 15–20 stars.

The differential magnitudes were then converted to apparent magnitudes, using the values discussed in the previous section. All magnitudes for each star were used to calculate a mean value and a standard deviation. In Figure 7 we show the plot of error per observation versus magnitude for all 94 stars. Although a few stars appear as potential variables on this graph, we find that the three known variables are not part of this group. Therefore, we use the robust median statistic (RoMS) detailed in Enoch et al. (2003). This statistic is greater than 1 for variables, and Enoch et al. (2003) estimate that stars with a robust statistic of 0.8 have about a 50% chance of being variable. In order to get an appropriate value for the error per observations used in the RoMS, we fit a line to the lower edge of the standard error curve. This method overestimates the RoMS, yielding more potential variables, but reduces the chance of missing a variable.

Fig. 7.—

Fig. 7.— Plot of error per observation vs. magnitude for all 94 stars examined in this survey.

From the RoMS, we found 28 stars with values greater than 0.8 in at least one of the three filters. All three of the previously known variables are included in this group. At this point, each of the 28 potential variables was examined visually and with Period98. Of the 28 stars, we confirm 5 new variable stars and determine new periods for V381 Vul and V382 Vul. An additional eight stars show varying degrees of variability, with SU Vul being included in this group. However, the time coverage was insufficient in all cases to classify the type of variation. Below, we examine these 15 stars in more detail. The results are summarized in Table 4, and the positions in the HR diagram are shown in Figure 8.

Fig. 8.—

Fig. 8.— HR diagram of the stars from this survey. Stable stars are represented as open circles, variable stars as filled circles, and suspected variables as filled triangles.

5.1. Established and New Variable Stars

5.1.1. Star 1—V381 Vul

V381 Vul is one of two variables reported by Peña et al. (1990) as new δ Scuti variables. Above, we find that V381 Vul is an A7 IV at a distance of approximately 540 pc. Figure 9 shows the five individual nights of data. We find a period for V381 Vul of 0.1185 days. The amplitudes for the V, R, and I filters are 0.010, 0.006, and 0.004, respectively. From very limited data, the amplitude in the B filters is found to be approximately 0.02. These values for the amplitude and period are significantly different from those of Peña et al. (1990), who found a period of 0.056 days and an amplitude of 0.034 in the V filter. In Figure 10 we show the phased data in the V and R filters. Given the spectral type, period, and amplitude measurements, we continue to classify this star as a δ Scuti.

Fig. 9.—

Fig. 9.— Five nights of V data for stars 1, 8, 13, 28, 72, and 78.

Fig. 10.—

Fig. 10.— Phased light curves for star 1, V381 Vul, in the V and R filters.

5.1.2. Star 8—New Variable Star

Star 8 is located near star 1 in the northwest corner of the cluster. Our red data is quite noisy, so the analysis concentrated on the V and I data. The five nights of V data are shown in Figure 9. We determine a period of 0.2247 days and an amplitude of 0.015 in both filters. The (V - I) color for this star is 0.567 and is right on the main sequence shown in Figure 8. With the reddening values discussed earlier, this would correspond roughly to a G dwarf. From the consistency of the amplitude and a G dwarf spectral type, we conclude that this is likely a W UMa system with a period of 0.4494 days.

5.1.3. Star 13—New δ Scuti Variable

In Figure 9 the five nights of V data are shown for star 13. We find this star to have a period of 0.0776 days and amplitudes of 0.031, 0.015, 0.010, and 0.010 in the B, V, R, and I filters, respectively. Using the (V - I) color, the data in Figure 5, and an assumption that the star is in the group at about 540 pc, we find a spectral type of F3 for this star. The magnitude difference between star 1 and star 13 is consistent with the spectral type difference. From the period, amplitudes, and color, we classify this star as a δ Scuti. Figure 11 shows the phased V light curve of star 13.

Fig. 11.—

Fig. 11.— Phased light curve for the new δ Scuti variable star 13 in the B and V filters.

5.1.4. Star 28—V382 Vul

This is the second star reported as a δ Scuti by Peña et al. (1990). However, we show above that this star has a spectral type of B3 V, plus additional unidentified lines, which would clearly preclude the star from being a δ Scuti. The spectral observations also show a large variation in radial velocity for the spectral features. However, the radial velocity of the emission line is roughly constant at 14 km s−1. This indicates that the star is inside a shell that is the source of the emission line.

From the photometry, we find a short‐period variation of 0.1808 days. The five nights of data are again shown in Figure 9. This new period is far longer than the period reported by Peña et al. (1990), and we find a nearly constant amplitude of 0.011 in the V, R, and I filters. From the spectral type of B3 V, the period of 0.1808 days, and the radial velocity variation, we classify this star as a β Cephei. The constant amplitude is a concern for a pulsating star. However, Sareyan et al. (1997) show fairly constant amplitudes in redder filters for another β Cephei star, 16 Lacertae.

5.1.5. Star 47—New Eclipsing Binary System

Although we clearly show this star to be variable star, we have no way to determine the period. In all our nights of data, we caught only one eclipse, as shown in Figure 12. In this figure, we show the variation of the star in the B, V, and R filters. The depth of the eclipse is 0.22, 0.13, and 0.05 in the B, V, and R filters, respectively. Since star 47 is a redder star, with (V - I) = 0.621, we conclude that the companion is bluer and likely a white dwarf.

Fig. 12.—

Fig. 12.— One night of data for star 47, showing the B, V, and R data.

5.1.6. Star 72—New Variable

Star 72 is one of the stars farthest from the traditional center of NGC 6882/6885 that we examined. As given in Table 2, it has a published spectral type of A2 and appears to be in the group at 540 pc. The (V - I) of 0.158 calculated from data in Table 3 is consistent with an A2 spectral type. Only four nights of data were obtained on this star, since it was close to the edge of our field. The four nights are shown in Figure 9. We find a period of 0.2220 days, with a constant amplitude across the V, R, and I filters of 0.008. With a spectral type of A2, it is unlikely that this is a W UMa system (Csizmadia & Klagyivik 2004). However, we believe that this is an eclipsing system with a period of 0.4440 days.

5.1.7. Star 78—New Variable

This star is located in the extreme southeast corner of our field of view and for some nights was off the frame. Therefore, we only have three nights of data, and on one of these nights it appears on only half the frames. The three nights of V data are shown in Figure 9. We find a period of 0.0795 days for star 78, with amplitudes of 0.056, 0.036, and 0.032 in the V, R, and I filters. Based on the (V - I) color and the data in Figure 5, we determine a spectral type of B5 if star 78 is in the more distant cluster, or F0 if it is in the group at 540 pc. In neither case does the magnitude seem to fit with the determined spectral type. The period and amplitudes indicate that this is a δ Scuti variable, but more information is needed to confirm this classification. The phased light curve for star 78 is shown in Figure 13.

Fig. 13.—

Fig. 13.— Phased light curve for star 78 in the V filter.

5.2. Suspected Variable Stars

In addition to the seven variable stars detailed above, we find eight stars that show some indication of variation. Five nights of V data are shown for seven of these stars in Figure 14. Star 29 is not included in the list, since it only shows variations in the I filter.

Fig. 14.—

Fig. 14.— Five nights of V data plotted for the suspected variables. This includes stars 23, 25, 38, 41, 84, 91, and 92.

5.2.1. Star 23—SU Vul

Star 23 is listed as a variable star in the GCVS. It is listed as a long‐period red variable. Our data is insufficient to find a period for this star, but does show a long‐period variation. If this star had not been listed as a variable, it would have been easy to overlook it as a potential variable star.

5.2.2. Star 29

This is an interesting suspect. It is the faintest star for which we suspect variability at 13.384 in the V filter. The exposures in the V filter were not long enough to get good results, and the RoMS was only 0.811 in the V filter. However, in the I filter the statistic was 1.2 where the magnitude is 10.993. Therefore, the only analysis we can perform is in the I filter. From this data we find a period of 0.1662 days with an amplitude of 0.010. This star has an extremely red color of (V - I) = 2.39. More observations are needed to determine the type of variation for this star.

5.2.3. Star 41

For star 41 we find a period of 0.2824 days, with a constant amplitude in the VRI filters of 0.009. The (V - I) color for this star is 0.581, very similar to that found for star 8. The position in Figure 8 place this star as a G dwarf. The period, amplitude, and position in the HR diagram make it likely that this is a W UMa system with a period of 0.5647 days. Although we are able to get a reasonable solution for this star, we still feel it should be on the list of suspected variables and not on the list of variables.

5.2.4. Stars 25, 38, 84, 91, and 92

The data for these stars is shown in Figure 14, but there is little else that can be reported for these stars. No definitive period was found for any star of this group, and only the trends presented in the figure give an indication of variability.

6. CONCLUSION

We examine the two δ Scuti variables reported by Peña et al. (1990). We find that V381 Vul is indeed a δ Scuti; however, we find a longer period of 0.1185 days and an amplitude 0.010. This star is located in a group of stars at approximately 540 pc and is an A7 IV. For V382 Vul we find that the classification as a δ Scuti is in error. V382 Vul has a spectral type of B3 V, plus a late‐type companion, and is located in a group at 1200 pc. We classify V382 Vul as a β Cephei star with a period of 0.1808 days and an amplitude of 0.011.

Using RoMS = 0.8 as a cutoff value, we find 28 candidate variable stars in the field of NGC 6882/6885. Seven stars are found with RoMS>1.0. Of these, we find one star with a single night of bad data and one star with nearby neighbors that cause large errors. The remaining five stars with RoMS>1.0 are all found to be variable. For 0.9<RoMS<1.0 we have six stars, with three confirmed as variables, one suspect, and two nonvariables, or about 66% variable. For the group with 0.8<RoMS<0.9, we examine 15 stars, with six suspects and nine nonvariables. All stars with RoMS<0.8 were found to be nonvariable. These statistics are consistent with the predictions of Enoch et al. (2003). We find the RoMS technique to be superior over an examination of an error curve in finding low‐amplitude variable stars.

For the region covered by our data, we find stars located at two distinct distances. There is a group that includes V381 Vul and that is at the traditional distance of about 600 pc for NGC 6882/6885. All but one of these stars is in the northern part of our field of view. We also find stars, including V382 Vul, at a distance greater than 1 kpc, as favored by Platais et al. (2003). In the end, a more detailed approach is needed in order to separate the stars in this region into their proper clusters and define NGC 6882 and NGC 6885 appropriately.

We acknowledge the Brigham Young University Department of Physics and Astronomy for their continued support of our research efforts. In particular, we wish to acknowledge an Environment for Mentoring Grant from BYU's ORCA office. We also acknowledge a grant from the Dunham Foundation and a AAS small research grant, which have been used to help equip the BYU campus observatory. We also acknowledge Ronald Hintz for his help acquiring the spectral data.

We acknowledge the use of the 1.8 m Plaskett Telescope at the Dominion Astrophysical Observatory, Herzberg Institute of Astrophysics, National Research Council of Canada. We also acknowledge the use of the online Library of High Resolution Spectra of Stars across the HR Diagram of the UVES Paranal Observatory Project (ESO DDT Program ID 266.D‐5655).

Please wait… references are loading.
10.1086/432147