Publications of the Astronomical Society of the Pacific Vol. 107 1995

Publications of the Astronomical Society of the Pacific

Vol. 107 1995 May No. 711

Publications of the Astronomical Society of the Pacific 107: 411-415, 1995 May

Pulsation Time Scales and Amplitudes in a Sample of Bright Semi-Regular Variable Stars

V. Cristina Cristian Harvard-Smithsonian Center for Astrophysics, 60 Garden Street, Cambridge, Massachusetts 02138, and Research Science Institute, Center for Excellence in Education, 7710 Old Springhouse Rd., Ste. 100, McLean, Virginia 22102 Robert A. Donahue Harvard-Smithsonian Center for Astrophysics, 60 Garden Street, Cambridge, Massachusetts 02138 Electronic mail: [email protected] Willie H. Soon Harvard-Smithsonian Center for Astrophysics, 60 Garden Street, Cambridge, Massachusetts 02138 Electronic mail; [email protected] Sallie L. Baliunas Harvard-Smithsonian Center for Astrophysics, 60 Garden Street, Cambridge, Massachusetts 02138, and Center of Excellence in Information Systems at Tennessee State University, 330 10th Ave. Ν., Nashville, Tennessee 37203 Electronic mail: [email protected] Gregory W. Henry Center of Excellence in Information Systems at Tennessee State University, 330 10th Ave. Ν., Nashville, Tennessee 37203 Electronic mail: [email protected] Received 1994 December 20; accepted 1995 February 13

ABSTRACT. We have analyzed the differential V-band photometric variations of records up to 8 years long in ten M III semi-regular variable stars. Periodograms constructed from each star's entire record and seasonal intervals were used to find repetitive occurrences of pulsation periods in the range of 10 to 250 days. For every star at least one locus of periods was observed, with many stars showing two distinct distributions of pulsation time scales. The observed pulsation periods and differential V-magnitude semi-amplitudes appear to be correlated such that longer periods correspond to larger semi-amplitudes. 1. INTRODUCTION baseline is required. Fortunately, aggressive monitoring of SR stars has recently begun by groups of amateur and pro- Semi-regular (SR) variable stars are typically late-type gi- fessional astronomers such as the AAVSO Photoelectric Pho- ant or supergiant stars, showing variability due to pulsation tometry Survey and at the University of Toronto (Percy and of up to a few tenths of a magnitude. The term "semi- Fleming 1992; Percy et al. 1994). Likewise, we have se- regular" describes a star which lacks a clearly observable lected a sample group of M III SR stars for long-term exami- pulsation period as seen for Cepheids or even Miras, but nation using the 0.25-m Automatic Photoelectric Telescope whose variations are not as complicated as the truly irregular (APT) at the F.L. Whipple Observatory at Mt. Hopkins, Ari- variables such as V Aql, or γ Cas. The first stars of this class zona. These stars are part of a project begun in 1986 to to be discovered were μ Cep by Sir William Herschel in monitor bright (3^y^8 mag) variable stars (Baliunas et al. 1782, and a1 Her, in 1796. The SR stars usually show vari- 1987). The list of stars appears in Table 1. ability with periods between 30-300 days superposed with We analyzed the time series obtained thus far to search for longer variation on the order of 1000 days. Additional back- pulsation time scales in the 10-250 day range. ground information is given by Percy (1985), and Querci (1986). SR stars have not enjoyed the popularity of other pulsat- 2. OBSERVATIONS ing stars such as Miras or Cepheids because of the difficulty Automated observing of semi-regular variables with the in measuring a dominant pulsation period which can subse- 0.25-m APT allows very low cost, long-term data acquisition quently be included within theoretical models. In order to from a prime photometric site. The observing sequence and determine accurately time scales of pulsation, a very long precision of the photometric data from this telescope are de-

Table 1 3 Characteristics of the M-type SRs 0.2 0.0 Name HD Spectral V Β - Comparison Check -0.2 Number Type Star Star : -0.4 * '%V A1/ Î RX Lep 33664 M6 Hie 5.68 1.46 33093 33802 -0.6 7 STUMa + f i*Jil η Gem 42995 M3 III 3.28 1.61 41116 44478 1 . . , ι . BC CMi 64052 M4 III 6.31 1.59 63799 65345 0.0 UX Lyn 77443 M4IH 6.60 1.51 76944 77912 -0.2 RS Cnc 78712 M6 Hie 5.95 1.67 78235 76572 : ·η ïlîlfj ST UMa 99592 M4 III 6.28 1.68 99606 101133 -0.4 - TU CVn Φ ? ^ f f TU CVn 112264 M5 III 5.88 1.59 112570 110834 ^ -0.6 — . . . ] . .+. I . . . 1 ' ' ι . . . — FS Com 113866 M5 Illb 5.60 1.59 113866 113848 Ii -0.2 30 Her 148783 M6 III 5.01 1.52 150997 149630 υ R Lyr 175865 M5 III 4.04 1.59 174621 177196 á -0-4 i Iii ^ -0.6 - FS Com , ^ H' aMost of the information was obtained from the SIMBAD database, oper- ated by the Centre de Données Astronomique de Strassbourg (CDS). 1.8 - 30 Her + + 1 1.6 1.4 scribed in Young et al. (1991). The brightness of each SR 1.2 : f'VH*1 w variable star was measured differentially with respect to a 1.0 .1,1.1,1.1, nearby comparison star, whose constancy was verified in turn -2.2 I-++Í ++$ ^ + by a third check star. This gives two differential magnitudes -2.4 (variable minus comparison and check minus comparison) -- ' · S é ' -2.6 !· t'\ for each of the V, R, and I filters used. A measurement of the I RLyr .+¾ + + J precision of the resulting observations is given by the sea- -2.8 1985 1987 1989 1991 1993 1995 sonal scatter of the check-minus-comparison differentials, Year which averaged approximately 0.01 mag in the V filter. All observations were corrected for differential extinction with Fig. 2—Differential V magnitudes for five of the stars listed in Table 1. seasonal mean extinction coefficients determined on occa- From top: ST UMa, TU CVn, FS Com, 30 Her, and R Lyr. sional nights of standard star observations and reduced to the Johnson system with long-term mean transformation coefficients. Night-to-night and seasonal temperature variations source of the resulting 0.01 mag scatter. This study will con- and the resulting changes in the color response of the centrate on the V (550±45 nm) measurements only. The ambient-temperature PIN photodiode detector are a major photometric differential V magnitudes are shown in Figs. 1 and 2. Each star's record is split into up to nine 200-250-day observing seasons. 1.0 0.5 Although changes in the light curves for SR stars are evi- 0.0 ΐ dent from cycle to cycle, in several cases a dominant period -0.5 ψ is clearly marked over an extended period of time. Percy et -1.0 RX Lep al. (1989) measured a persistent 63-day period for the M6 III -0.6 star EU Del. However, Percy et al. (1993) describe the light -0.8 curve of R Lyr as "sporadic" in comparison, but containing -1.0 episodes during which a pulsation period might be measured. -1.2 η Gem Likewise, Szatmáry and Vinkó (1992) note transitory periods 0.6 in the long-term variability of the M5 Ib-II star Y Lyn. 1 0.4 While no simple mechanism for pulsation exists, it may à 0.2 b 0.0 be possible to find some degree of regularity in their light < -0.2 BC CMi curves which can suggest a likely time scale of pulsation. 1.0 - UX Lyn + Thus, the semi-regular behavior can possibly be explained by 0.8 a very complex variation brought about by the combination 0.6 0.4 of two or more oscillations, each of which may possibly be 0.2 more or less regular and varying independently of the others. 1.5 In addition, the strength of these oscillations may grow and 1.0 % 4 decay with time, with the result that different modes are 0.5 more prominent at any given epoch. Therefore, the most 0.0 i *^4 -0.5 clearly recognizable time scale of pulsation may change with 1985 1987 1989 1991 1993 1995 time. Cadmus et al. (1991) noted the change in amplitude Year and period for three SR stars, and suggest mode switching between the fundamental pulsation mode to an overtone pul- Fig. 1—Differential V magnitudes for five of the stars listed in Table 1. sation as the mechanism producing the complex light curve. From top: RX Lep, η Gem, BC CMi, UX Lyn, and RS Cnc. Measurements made over longer intervals may have

© Astronomical Society of the Pacific · Provided by the NASA Astrophysics Data System PULSATION TIME SCALES AND AMPLITUDES 413 greater success at characterizing the interaction of competing Table 2 pulsation periods. For example, Loeser et al. (1986) mea- Recurring Seasonal Periods and Semi-Amplitudes from Individual Seasons sured two persistent periods for the SR star V CVn (M6 III) Star Period rms(P) Semi-amplitude rms(a) Ν (days) (days) m using 60 years of AAVSO naked-eye observations. On the ( ag) (mag) RX Lep 85 7 0.196 0.094 other hand, long-term observations of W Cyg by Percy et al. 61 4 0.048 0.030 (1993) suggest the star is beating with close but variable η Gem 220 29 0.101 0.037 periods near 130 days. The intermediate-term time series cur- BC CMi 24 5 0.051 0.010 rently available for this sample of stars indicate persistent 43 2 0.069 0.013 regularities that may eventually lead to some illumination of UX Lyn 38 4 0.073 0.036 the underlying physical processes involved in producing the 87 12 0.062 0.022 ST UMa 73 15 0.112 0.044 complex variability. 158 15 0.182 0.069 TU CVn 43 3 0.069 0.026 3. PULSATION PERIODS AND AMPLITUDES 61 4 0.062 0.024 FS Com 53 3 0.094 0.052 Periodogram analysis was used to measure pulsation time 37 2 0.061 0.017 scales for each star observed. We have concentrated our 30 Her 64 5 0.149 0.065 search for variations to time scales less than the length of the 41 1 0.055 0.009 observing season (typically 200-250 days) and have ana- R Lyr 65 3 0.094 0.048 25 1 lyzed each observing season individually. The repeated oc- 0.029 0.007 currence of a particular pulsation period (or set of periods) from season to season will be a good indication of the pri- 1969) note the presence of very long (>250 day) modula- mary pulsation periods for a particular star. Because the tions in the light curves of SR stars. While such long-term variations are semi-regular, different periods may be promi- modulation is immediately apparent in the record of stars nent in any given observing season. Therefore, each observ- such as RX Leporis, investigation and quantitative analysis ing season was analyzed individually to minimize the above of such long-term variations is difficult given as the available effects and to provide better statistics on the recurrence of time series. possible pulsation periods. This technique works best for The periods were averaged in each locus for all the sea- phenomena which occur on time scales that are shorter than sons to form (P) and the standard deviation of the distribu- the observing season, i.e., ^ 150 days. For pulsations with tion, rms(P). Table 2 contains mean periods and semi- longer periods, it may not be possible to determine the period amplitudes, and their standard deviations, obtained from a accurately from a single season of data, and therefore a season-by-season analysis, and the number of periods, N, longer baseline must be used (see below). within each locus used to derive the average period and A periodogram (Scargle 1982; Home and Baliunas 1986) semi-amplitude. The recognition of these loci was done ar- was calculated for each interval, e.g., one season and signifi- bitrarily; a statistically significant distribution of seasonal cant peaks were recorded. In a few cases, as many as three pulsation periods will require many years of data. However, peaks were observed simultaneously within the same observ- the recurrence of some periods in almost every observing ing season. To test the validity of the multiple peaks, each season suggests that we are successfully measuring signifi- was filtered from the periodogram sequentially (Baliunas et cant contributors to the semi-regular variation, within the al. 1985) and the resultant periodogram examined for the period range searched. previously recorded secondary peak. Peaks which remain af- Three stars in Table 1 (77 Gem, RS Cnc, and ST UMa) ter filtering are not artifacts due to spectral leakage but may have periods that are close to the length of an observing still be influenced by the observing window. season. Because of this, it is difficult to accurately measure the pulsation period because <2 cycles are present within 3.1 Recurring Periods from Individual Seasons the observing season. Therefore, for these stars we have re- For every star in the sample (except for RS Cnc), recur- peated the analysis using two consecutive seasons of data. ring periods were found in several seasons that lie close to This extends the baseline from 200-250 days to 550-600 one another. The semi-regular and possibly competing varia- days and thus permits a more accurate measurement of these tions could describe the scatter in observed pulsation periods longer periods. The addition of another observing season which, although not exact from season to season, may sug- does introduce complications in the power spectrum due to gest an overall time scale of pulsation. An examination of the the intra-seasonal gap which is now in the middle of the distributions of periods from each observing season revealed record. However, it is straightforward to identify and accu- clustering at one or two loci repeating from season to season. Table 3 We interpret the recurrence of these periods to suggest that Recurring Periods and Semi-Amplitudes from Consecutive Seasons they are important contributors to the overall variation. Star Period rms(P) Semi-amplitude rms(a) Ν Some periods occurred only once within one season of the (days) (days) (mag) (mag) available time series. If these periods are due to pulsation, it η Gem 231 7 0.084 0.008 5 is possible that they are more visible at other times but not RS Cnc 237 10 0.313 0.108 5 during our limited record, but might recur in longer time ST UMa 91 0.112 0.020 7 151 18 0.148 0.067 3 series. Many authors (Percy and Fleming 1992; Glasby

Table 4 Overall Pulsation Periods and Semi-Amplitudes 0,60 Star Period Semi-amplitude (days) (mag) RX Lep 0.133 ñ η Gem 238 0.083 0.40 — BC CMi 27 0.026 Xl

0.25-m APT as well as the work done by the AAVSO and Baliunas, S.L., Donahue, RA., Loeser, J.G., Quinan, E.F., Genet, elsewhere is essential to provide empirical constraints on the R.M., and Boyd, L.J. 1987, in New Generation Small Tele- intricate pulsation models needed to describe SR-star behav- scopes, ed. D.S. Hayes, R.M. Genet, and D.R. Genet (Tucson, ior. Fairbom Obs.), p. 97 Cadmus, R.R. Jr., Willson, L.A., Sneden, C., and Mattei, J.A. 1991, We thank the referee for several helpful comments and AJ, 101, 1043 suggestions. This research was supported by the Scholarly Eggen, OJ. 1975, ApJ, 195, 661 Studies Program and Langley-Abbot Program of the Smith- Glasby, J. S. 1969, Variable Stars (Cambridge, Harvard University sonian Institution, National Science Foundation Grant No. Press) HRD-9104484 to Tennessee State University, the Massachu- Home, J.H., and Baliunas, S.L. 1986, ApJ, 302, 757 setts Space Grant Consortium and NASA Grant No. NGT- Loeser, J.G., Baliunas, S.L., Guinan, E.F., Mattei, J.A., and Wacker, 40001. as well as several gracious individuals. V-C.C. was S.W. 1986, in The Fourth Cambridge Workshop on Cool Stars, supported by the Research Science Institute Program, which Stellar Systems and the Sun, ed. M. Zeilik and D.M. Gibson was sponsored at MIT during the Summer of 1994. We (New York, Springer), p. 460 greatly appreciate the tireless efforts of Louis Boyd for his Percy, J.R. 1985, IAPPP Commun., 19, 14 diligent maintenance of the 0.25-m APT at Mt. Hopkins. We Percy, J.R., and Fleming, D.E.B. 1992, PASP, 104, 96 appreciate the efforts of the Centre de Données As- Percy, J.R., Landis, H.J., and Milton, R.E. 1989, PASP, 101, 893 tronomique, Observatoire de Strassbourg in creating and Percy, J.R., Ralli, J.A., and Sen, L.V. 1993, PASP, 105, 287 maintaining the SIMBAD database which has been invalu- Percy, J.R., et al. 1994, PASP, 106, 611 able in this work. We thank LI. Shapiro for his enthusiastic Querci, F.R. 1986, in The M-Type Stars, ed. H.R. Johnson and F.R. support of the APT program. Querci (Washington, NASA), NASA SP-492 Scargle, J.D. 1982, ApJ, 263, 875 Szatmáry, Κ., and Vinkó, J. 1992, MNRAS, 256, 321 REFERENCES Young, A.T., et al. 1991, PASP, 103, 221 Baliunas, S.L., et al. 1985, ApJ, 294, 310