Red giant pulsations from the suspected symbiotic star StHA 169 detected in Kepler data
aa r X i v : . [ a s t r o - ph . S R ] A p r Mon. Not. R. Astron. Soc. , 000–000 (0000) Printed 21 October 2018 (MN L A TEX style file v2.2)
Red giant pulsations from the suspected symbiotic starStHA 169 detected in
Kepler data
Gavin Ramsay , Pasi Hakala , Steve B. Howell Armagh Observatory, College Hill, Armagh, BT61 9DG, UK Finnish Centre for Astronomy with ESO (FINCA), University of Turku, V¨ais¨al¨antie 20, FI-21500 PIIKKI ¨O, Finland NASA Ames Research Center, Moffett Field, CA 94095, USA
Accepted 2014 April 23. Received 2014 April 23; in original form 2014 February 13
ABSTRACT
We present
Kepler and
Swift observations of StHa 169 which is currently classifiedas a symbiotic binary. The
Kepler light curve shows quasi periodic behaviour with amean period of 34 d and an amplitude of a few percent. Using
Swift data we find a rela-tively strong UV source at the position of StHa 169 but no X-ray counterpart. Using asimple two component blackbody fit to model the combined
Swift and 2MASS spectralenergy distribution and an assessment of the previously published optical spectrum,we find that the source has a hot ( ∼ ∼ Kepler light is dominated by the cool component and we attribute thevariability to pulsations in a red giant star. If we remove this approximate month longmodulation from the light curve, we find no evidence for additional variability in thelight curve. The hotter source is assigned to a late B or early A main sequence star.We briefly discuss the implications of these findings and conclude that StHA 169 is ared giant plus main sequence binary.
Key words:
Stars: individual: – StHa 169 – Stars: binaries – Stars: symbiotic stars
Symbiotic stars are interacting binary systems containinga red giant star and a hotter component, which can be awhite dwarf, a main sequence star or even a neutron star (seeMikolajewska 2007 for a review). A relatively small fractionof these binaries show evidence for accretion onto the hotcomponent via a disc, while the remainder show evidenceof accretion via the wind from the giant star and, in somesystems, nuclear burning occurs on the surface of the hotcomponent (see Kenyon & Webbink 1984). Some systemssuch as CH Cyg, have produced jets (e.g. Taylor, Seaquist& Mattei 1986, Crocker et al. 2001, Galloway & Sokolski2004) and large variations ( ∼ Kepler misson (Borucki et al. 2010) provides aunique opportunity to study objects such as symbiotic bi-naries on short timescales (1 min) and also much longertimescales (the initial
Kepler pointing lasted approximately4 years). There were two objects classified as symbiotic bi-naries in the
Kepler field – the previously mentioned systemCH Cyg – and StHA 169. This paper presents an analysisof
Kepler and
Swift observations of StHA 169.
StHA 169 (also known as NSV 12466 and S169) was classi-fied as a symbiotic binary by Downes & Keynes (1988), andis in the catalogue of symbiotic binaries of Belczy´nski et al.(2000), as a result of its optical spectrum which ‘resembles c (cid:13) Figure 1.
The upper light curve shows the normalised long cadence data of StHa 169. The lower light curve has been de-trended andnormalised and corrected so that there are no step changes in flux between different quarters. that of the quiescent phase of the symbiotic recurrent novaRS Oph’. StHA 169 has the identifier KIC 9603833 in the
Kepler
Input Catalogue (Brown et al. 2011) and has a mag-nitude of g =14.37 and colour g − r =1.60 in the Kepler
INTSurvey (Greiss et al. 2012) and g =14.12 and g − r =1.42 inthe RATS-Kepler
Survey (Ramsay et al. 2014). StHa 169 isrecorded in the ASAS survey of the
Kepler field (Pigulskiet al. 2009) as a source with ‘no well-defined periodicity inlight variations’.
The detector on board
Kepler is a shutterless photometerusing 6 sec integrations and a 0.5 s readout. There are twomodes of observation: long cadence (LC), where 270 integra-tions are summed for an effective 29.4 min exposure (thisincludes deadtime), and short cadence (SC), where 9 inte-grations are summed for an effective 58.8 s exposure. Gaps inthe
Kepler data streams result from, for example, 90 ◦ space-craft rolls every 3 months (called Quarters), and monthlydata downloads using the high gain antenna. Kepler data are available in the form of FITS files whichare distributed by the Mikulski Archive for Space Telescope
Long Cadence Short Cadence(Quarter) (Quarter/Month)0,1,2,3,4,5,6,8,9,10,12,13,14,16,17 6/2, 14/2
Table 1.
Journal of
Kepler observations. Each quarter nominallylasts 3 months with a short gap between months. Short Cadenceobservations are made on a monthly basis. Quarter 17 was trun-cated to approximately 1 month. (MAST) . For LC data each file contains one observing quar-ter worth of data whereas for SC data one file is created permonth. After the raw data are corrected for bias, shutter-less readout smear, and sky background, time series are ex-tracted using simple aperture photometry (SAP). We notein Table 1 the Kepler
Quarters in which LC data was ob-tained and also the
Kepler months in which SC mode datawas obtained. The first data to be taken (Q0) started in May2009 and the final data (Q17) finished in May 2013. http://archive.stsci.edu/keplerc (cid:13) , 000–000 ed giant pulsations detected from StHA 169 Using the data downloaded from MAST we used the ‘SimpleAperture Photometry’ (SAP) data and removed data whichdo not conform to ‘SAP QUALITY=0’ (for instance, timeintervals of enhanced solar activity) and then normalisedthis light curve so that the mean count rate was unity (Fig-ure 1). There are clear flux variations on a timescale of tensof days and a semi-amplitude of several percent.To remove systematic trends in the data (e.g. Kine-muchi et al. 2012) we used the task kepcotrend which ispart of the
PyKE software (Still & Barclay 2012) . We thenapplied a small offset so that there are no discrete jumps influx between the different quarters of data. This light curveis also shown in Figure 1 and shows similar features but withfewer large flux variations at the start and end of quarters.We show the Lomb Scargle power spectrum (Lomb1976, Scargle 1982) of the corrected light curve in Figure2. The peaks correspond to periods of 40.0, 38.5 and 36.3days. There are also peaks in the power spectrum at ∼ Kepler year(372.5 d) in the data of giant stars (Banyai et al. 2013),some caution is necessary in interpreting long period signalsin power spectra such as these.To further investigate the nature of the light curve, wedetermined the time of maximum for every peak in the lightcurve by eye (the error on the time of maximum was gener-ally 0.3 days which is very much smaller than the range ofduration of each cycle) and then calculated the time differ-ence between successive peaks. The duration of each cycleis shown in Figure 3 and shows a considerable range in theduration of each cycle, ranging from 23 to 52 days, with amean of 34.2 d and σ =8.4 d. StHa 169 was observed in two quarters (6 and 14) usingShort Cadence mode, which provides photometry with ef-fective exposure times of 58 sec and allows the short termphotometric behaviour to be studied in more detail. To re-move the effects of systematic trends and also the 25–50 daypulsation period we used the
PyKE task kepflatten . Afternormalising the data by dividing the light curve by the meanflux we found the rms was 0.00048 and 0.00049 for data inquarters/months 6.2 and 14.2 respectively. This result showsvery little evidence of short period variability in the
Kepler light curve. SWIFT
OBSERVATIONS
StHa 169 lies ∼ γ -ray burstGRB060105 and therefore within the field of view of theX-ray (XRT) and optical/UV (UVOT) instruments on-board Swift . Observations were made between 2006 Jan 5thand 11th. However, only in the dataset comprising ObsId00175942000 (Jan 5–6th) were observations made in the UVfilters. We therefore restrict ourselves to this dataset. http://keplergo.arc.nasa.gov/PyKE.shtml P o w e r Cycles/Day
Figure 2.
The Lomb Scargle power spectrum of the light curvecorrected for systematic trends and jumps between different quar-ters.
Figure 3.
The length of the ’month’ long pulsation period ofStHa 169 as a function of time. The length is defined as the timedifference between a given peak in the light curve and the previouspeak.
We show in Table 2 the filters in which observationswere obtained. We used the ftool task uvotsource to de-termine the mean count rate for StHa 169 and the corre-sponding flux. The exposure time of each UVOT image isgenerally 500–700 sec in duration and there were 7 images inthe UVW1 filter and 13 images in the UVM2 filter (other fil-ters had values between these numbers). Based on the countrates derived from these images we determined the rms andexpected rms (assuming Poisson statistics) and the corre-sponding value sfrac (=rms/count rate). These numbers aregiven in Table 2. In no case was the rms variation greaterthan twice the expected rms. We also did the same anal-ysis using event rate data which was obtained using theUVM2 filter. We binned the data using different binsizesand show the results for a 4 sec and a 60 sec binsize in Ta-ble 2. Again the rms/rms exp ratio is less than 2 (and was http://heasarc.gsfc.nasa.gov/lheasoft/ftools/headas/uvotsource.htmlc (cid:13) , 000–000 Filter Wavelength Rate Flux rms rms exp rms/ sfrac Points Bin Size sfrac(˚A) (Ct/s) (ergs s − cm − ˚A) (Ct/s) (Ct/s) rms exp (sec)V 5468 62.9 ± ± . × − ± ± . × − ± ± . × − ± ± . × − ± ± . × − ± ± . × − Table 2.
Details of the Swift UVOT observations made in 2006 Jan 5–6 where we indicate: the filter and its central wavelength; themean count rate and flux; the rms derived from the given number of points. We also indicate the expected rms exp which is determinedfrom the mean count rate and sfrac = rms/Rate. We also determine sfrac for two different bin sizes and were derived from the eventdata. for all attempted binsizes). This finding of very low shortterm variability is consistent with that found in the ShortCadence
Kepler observations.In their study of symbiotic binaries using
Swift
Luna etal. (2013) found that in the UV 33 non-saturated sourcesshows rms/rms exp >
2. All the sources which did not showsignificant rms variability were fainter in that band com-pared to StHa 169. StHa 169 therefore displays an unusu-ally low degree of variability compared to the majority ofsymbiotic binaries.In the dataset comprising ObsID 00175942000, the com-bined X-ray exposure was 57.8 ksec. However, there is nodetection of a source at the position of StHa 169. The countrate for the location of StHa 169 is 0.000011 ± ∼ . × − ergs s − cm − for this observation. With the available data of StHa 169 ranging from the near-UV and optical (
Swift ), the near-IR (2MASS) and far-IR(WISE) we can obtain the broad spectral energy distributionof StHa 169 and from that assess which binary componentis dominant in the
Swift and
Kepler band-passes. However,as symbiotic stars are known to show flux variations overthe long term some assessment needs to be made as to thelong term variability of StHa 169. Perhaps the most spec-tacular long term variability seen in a symbiotic star is CHCyg which has shown variations of 6 mag in the U band, al-though the variation is much reduced ( < Kepler data of StHa 169 itself which spansnearly four years. This shows a standard deviation of 3.6percent in the unnormalised light curve and 2.7 percent inthe unnormalised detrended light curve. Other data, for in-stance, the g mag from the
KIS and
RATS-Kepler surveys differ by 0.25 mag, while the V mag from the Swift missionand the
Kepler
UBV survey (Everett, Howell & Kinemuchi2012) differ by 0.28 mag.
ASAS had (incomplete) coverageof StHa 169 over a 600 day interval and found a standarddeviation of 0.13 mag in the light curve (Pigulski et al 2009).We therefore use
Swift data from the dataset ObsId00175942000 (as we used in the previous section) since pho-tometric data using six UVOT filters were made within ashort (2 days) time interval. We have also extracted fromthe archives the 2MASS and WISE flux measurements. InTable 3 and Figure 4 we show the observed fluxes from theseinstruments in mJy.Although cool stars are poorly fit by blackbody models(which we discuss in the next section) they are simple touse and allow us to easily determine which stellar compo-nent is dominant in the
Kepler and
Swift band passes. Thefractional residuals to a two blackbody model were mea-sured to have a scatter of 18 percent about the fit (Figure4). To account for any changes in the brightness at differentepochs, we therefore assigned the error in each photometricband to be 18 percent and carried out a set of 200 MonteCarlo simulations to derive the errors for the parameters.We find T1=10200 ± ± ± B − V = 0.30 ± Kepler band-pass (4200–9000˚A) is dominated by the flux from acool star, while the UV and blue flux in the
Swift
UVOT isdominated by a hotter component. In the V band the coolcomponent contributes 70 percent of the observed flux. Although the above modelling gives approximate temper-atures for the two components, it is well known that coolstars (of any spectral class) are very poorly fit by a black-body model. To obtain a better estimate of the temperature c (cid:13)000
UVOT isdominated by a hotter component. In the V band the coolcomponent contributes 70 percent of the observed flux. Although the above modelling gives approximate temper-atures for the two components, it is well known that coolstars (of any spectral class) are very poorly fit by a black-body model. To obtain a better estimate of the temperature c (cid:13)000 , 000–000 ed giant pulsations detected from StHA 169 of the cool star in StHa 169 we used the spectral atlas ofPickles (1998) and the published spectrum of StHa 169 (Fig.4 of Downes & Keynes 1988). Given that the Kepler datastrongly supports the view that the cool star is a pulsatinggiant, we compared red giant spectra given in Pickles (1998)with StHa 169. Based on the general shape of the spectraof StHa 169 and the dip near 5452 ˚A, we assign a spectraltype of M2III–M3III to StHa 169. Assuming this range ofspectral type we can then assign a temperature of 3650–3750K (Straizys & Kuriliene 1981) which is considerably hotterthan the fit derived using two blackbody components (2100K) but physically more realistic for a pulsating red giantstar.Using the information to hand we can make estimateson the distance to StHa 169 using different assumptions andassess the nature of the hot component. For an observed ( J − K )=1.17 and ( J − K ) o =0.90, Tabur et al. (2009) indicatethat this colour implies M K ∼ − . ∼ M K =–6.85implying a distance of 13.7 kpc if StHa 169 was on at thisstage in its evolution. We can also make an approximateestimate of the distance to the system by taking the observedrelationship between the pulsation period of red giants and M K (Tabur et al. 2010). For a pulsation period of 34 d weestimate M K ∼ − . ± . ± K bandmay have residual dust contamination, see § Gaia should be ableto determine the distance to StHa 169 with an accuracy of ∼
20 percent.Our blackbody fits ( §
5) suggest that the temperature ofthe hotter component is ∼ V band, and that an AOV star would lie at a distanceof 5.3kpc. On the other hand if the hotter component wasa pure hydrogen white dwarf with a temperature of 10000K, it would have an absolute magnitude of M V =12.1 (Berg-eron, Wesemael, & Beauchamp 1995), implying a distanceof 26 pc. Even if there was an accretion disk around a whitedwarf (which could brighten the absolute magnitude by 2mag), the distance would be very much closer than that in-ferred using the cooler component. We conclude that thehot component in StHa 169 is not a white dwarf. However,the hotter star must be less massive than the red giant star(estimated to be 3.3–3.5 M ⊙ ). An AO V star has a mass of2.2 M ⊙ while a B8 V star has a mass of 3.0 M ⊙ (Straizys &Kuriliene 1981). The
Swift -2MASS spectral energy distribution of StHa 169indicates there are two components, one relatively hot andone relatively cool. Using additional information such as theexisting optical spectrum of Downes & Keyes (1988) ourbest estimate for the temperature of the two components is10000K and 3700K. The
Kepler observations therefore sam-ple the cooler (and physically much larger) component, whilethe
Swift
UV and blue filters sample the hotter component.
Wavelength Filter Flux(˚A) mJy1880 Swift UVW2 1.6 ± ± ± ± ± ± ± ± ± ± ± ± Table 3.
The observed flux in mJy of StHa 169. The errors for
Swift fluxes come from the standard deviation of the observedfluxes in
Swift data obtain during 2006 Jan 5–6. The error on the2MASS and WISE fluxes come from the error on the magnitudein the 2MASS and WISE catalogues.
Figure 4.
The spectral energy distribution of StHa 169 from thenear-UV to near IR. The absorbed flux unit is milli-Jy and weshow the joint fit (solid line) from two blackbody models withtemperature ∼ ∼ E B − V =0.26.We have not used the WISE points (the three right most points)in the fit. The
Kepler light curve shows quasi periodic behaviour witha mean period of 34 d. Given that this period is not sta-ble, it is clearly not the signature of a binary orbital period.Rather it indicates that the cool component is a pulsatingred giant star. This is consistent with the suggestion madeby Downes & Keyes (1988) that the cool component in StHa169 has a M spectral type and resembles the red giant in therecurrent nova RS Oph (P orb =460 d; Dobrzycka & Kenyon1994; M0–M2 III, Dobrzycka et al. 1996).Further confirmation of the evolutionary status of thered giant in StHA 169 is provided through its frequencyspectrum.
Kepler photometry has been used extensively tocharacterize red giants as to their membership on the RGBor the AGB (Chaplin et al. 2013). The
Kepler light curve ofStHa 169 is very similar in character to, say, the red giantKIC 2986893 (B´anyai et al. 2013) which has a mean period of c (cid:13) , 000–000 ± =2.2 d. However, given that StHa 169 has a wide rangeof pulsation period, it is possible that it is a Semi RegularVariable (cf Soszy´nski, Wood & Udalski 2013). B´anyai etal. (2013) showed that M giants separate into three distinctgroups according to their period structure. StHA 169 andKIC 2986893 belong to group 1 – red giants with periodsbetween 10–100 days. Group 1 stars with a period similarto that found in StHa 169 tend to lie the upper Red GiantBranch (see Kiss & Bedding 2003, 2004) and are pulsatingdue to first and second overtone modes. (Given the mainsequence lifetime of a 3.3 M ⊙ star is 500 Myr, the system isat least this old).The identification of StHa 169 as a symbiotic star liessolely with the optical spectrum presented in Downes &Keyes (1988). The spectral energy distribution as derivedfrom Swift and 2MASS photometry and presented in § Kepler data clearly demon-strates that the cool star is a red giant. However, determin-ing the nature of the hot component in symbiotic stars isnot a trivial task (see, for instance, Sokoloski & Bildsten2010 who recount the quarter of a century debate on thenature of the hot star in the Mira AB system). Our spectralenergy distribution shows that an isolated white dwarf ora white dwarf with an accretion disk would not lie at thesame infered distances for the red giant component. Insteadour results favour that the hot star is more likely to be alate B or early A main sequence star. The absence of shortperiod variability in the UV and the non detection in X-rayssuggest that accretion was not taking place at the time ofthese
Swift observations.There are at least two other sources which bear somesimilarity to StHa 169: XX Oph and AS 325 which arethought to consist of a Be star and a red giant secondary(Howell, Johnson & Adamson 2009). Indeed, AS 325 wasoriginally taken to be a symbiotic system. The fact that theoptical spectrum of StHa 169 (Downes & Keyes 1988) showsthe Balmer lines (and He II 4686˚A) in emission may indi-cate that the B/A star is an emission star (either through awind or accretion). Stars like these are interesting from a bi-nary evolution point of view. Determining the binary orbitalperiod is a key step but will be difficult to disentangle thesignature of the binary period from the red giant pulsations.
Kepler was selected as the 10th mission of the DiscoveryProgram. Funding for this mission is provided by NASA,Science Mission Directorate. The
Kepler data presented inthis paper were obtained from the Multimission Archive atthe Space Telescope Science Institute (MAST). STScI is op-erated by the Association of Universities for Research in As-tronomy, Inc., under NASA contract NAS5 26555. Supportfor MAST for non HST data is provided by the NASA Officeof Space Science via grant NAG5 7584 and by other grantsand contracts. This work made use of PyKE, a softwarepackage for the reduction and analysis of Kepler data. Thisopen source software project is developed and distributedby the NASA Kepler Guest Observer Office. Armagh Ob-servatory is supported by the Northern Ireland Government through the Dept of Culture, Arts and Leisure. We thankthe anonymous referee for a useful report.
REFERENCES
Angeloni, R. et al., 2014, MNRAS, 438, 35B´anyai E., et al., 2013, MNRAS, 436, 1576Belczy´nski, K., Mikolajewska, J., Munari, U., Ivison, R. J., Fried-jung, M., 2000, A&AS, 146, 407Bergeron, P., Wesemael, F., Beauchamp, A, 1995, PASP, 107,1047Borucki, W. J., et al., 2010, Sci, 327, 977Brown, T. M., Latham D. W., Everett M. E., Esquerdo G. A.,2011, AJ, 142, 112Chaplin, W. J., Miglio, A., 2013, ARAA, 51, 353Crocker, M. M., Davis, R. J., Eyres, S. P. S., Bode, M. F., Taylor,A. R., Skopal, A., Kenny, H. T., 2001, MNRAS, 326, 781Dilday, B., et al., 2012, Sci, 337, 942Dobrzycka, D., Kenyon, S. J., 1994, AJ, 108, 2259Dobrzycka, D., Kenyon, S. J., Proga, D., Mikolajewska, J., Wade,R. A., 1996, AJ, 111, 2090Downes, R. A., Keyes, C. D., 1988, AJ, 96, 777Everett, M. E., Howell, S. B., Kinemuchi, K., 2012, PASP, 124,316Galloway, D. K., Sokoloski, J. L., 2004, ApJ, 613, L61Greiss S., et al., 2012, AJ, 144, 24Howell, S. B., Johnson, K. J., Adamson, A. J., 2009, PASP, 121,16Kenyon, S. J., Webbink, R. F., 1984, ApJ, 279, 252Kinemuchi, K., Barclay, T., Fanelli, M., Pepper, J., Still, M., How-ell, S. B., 2012, PASP, 124, 963Kiss, L. L., Bedding, T. R., 2003, MNRAS, 343, L79Lomb, N.R. 1976, A&SS, 39, 447Luna, G. J. M., Sokoloski, J. L., Mukai, K., Nelson, T., 2013,A&A, 559, 6Mikolajewska, J., 2007, Baltic Astron., 16 ,1Mikolajewski, M., Mikolajewska, J., Khudiakova, T. N., 1990,A&A, 235, 219Munari, U., Yudin, B. F., Kolotilov, E. A., Tomov, T. V., 1996,A&A, 311, 484Pickles, A. J., 1998, PASP, 110, 863Pigulski A., Pojma´nski G., Pilecki B., Szczygie l D. M., 2009, AcA,59, 33Ramsay, G., Brooks, A., Barclay, T., Hakala, P., Still, M., Greiss,S., G¨ansicke, B., Steeghs, D., Reynolds, M., 2014, MNRAS,437, 132Scargle, J.D. 1982, ApJ, 263, 835Sokolski, J. L., Bildstein, L., & Ho, W. C. G., 2001, MNRAS, 326,553Sokolski, J. L., Kenyon, S. J., 2003, ApJ, 584, 1027Sokolski, J. L., Bildstein, L., 2010, ApJ, 723, 1188Soszy´nski, I., Wood, P. R., Udalski, A., 2013, ApJ, 779, 167Still, M., Barclay, T., 2012, AstrophysicsSource Code Library, record ascl:1208.004,http://adsabs.harvard.edu/abs/2012ascl.soft08004SStraizys, V., Kuriliene, G., 1981, Ap&SS, 80, 353Tabur, V., Kiss L. L., Bedding, T. R., 2009, ApJ, 703, L72Tabur, V., Bedding, T. R., Kiss, L. L., Giles, T., Derekas, A.,Moon, T. T., 2010, MNRAS, 409, 777Taylor, A. R., Seaquist, E. R., Mattei, J. A., 1986, Nature, 319,38 c (cid:13)000