Cepheids and RR Lyrae stars in the K2 fields
aa r X i v : . [ a s t r o - ph . S R ] M a y COMMISSIONS 27 AND 42 OF THE IAUINFORMATION BULLETIN ON VARIABLE STARS
Number 6108 Konkoly ObservatoryBudapest28 May 2014
HU ISSN 0374 – 0676
CEPHEIDS AND RR LYRAE STARS IN THE K2 FIELDS
MOLN ´AR, L. , , PLACHY, E. , , SZAB ´O, R. Konkoly Observatory, Research Centre for Astronomy and Earth Sciences, Konkoly Thege Mik´os ´ut 15-17,H-1121 Budapest, Hungary Institute of Mathematics, Physics, and Engineering, Savaria Campus, University of West Hungary,K´arolyi G´asp´ar t´er 4, H-9700 Szombathely, Hungarye-mail: [email protected]
The
Kepler space telescope was launched into orbit in early 2009 with a task to observeabout 170 000 stars with unprecedented photometric precision and temporal coverage(Borucki et al. 2010). The original field-of-view was located in the Lyra-Cygnus regionat high ecliptic latitudes.
Kepler observed this field for almost exactly 4 years, providingrevolutionary data for several areas, but the failure of its second reaction wheel in May2013 effectively ended its original mission.The amount of classical pulsators, Cepheids and RR Lyrae stars, was somewhat limitedin the Lyra-Cygnus field. Before the start of the mission, 57 stars were proposed as RRLyrae targets but 23 of them soon turned out to be other types of stars (Kolenberg et al.2010). A few targets were discovered later either by the ASAS survey or as contaminantsin the pixel apertures of other stars, bringing back the sample size close to 50.About half of the fundamental-mode RR Lyrae stars turned out to be modulated. Theanalysis of the 4-year-long rectified data sets revealed that a very high fraction, about80% shows multiple modulation periods (Benk˝o et al. 2014). Several of the Blazhko starsshow various additional modes and/or period doubling. The sample size, however, is low:Benk˝o et al. (2014) created rectified data sets for only 15 stars after excluding the blendedtargets and the bright and heavily saturated RR Lyr, the eponym of the class. A largerpopulation of modulated stars is needed to determine the occurrence rates and especiallythe various interrelations of the dynamical effects.Other space photometric missions observed only a handful of RR Lyrae stars. MOSTmeasured AQ Leo, the archetype of the double-mode subclass but most RR Lyraes aretoo faint for the small telescope (Gruberbauer et al. 2007). Thirteen stars (10 RRab,2 RRc 1 RRd stars) were identified in the first CoRoT fields (from IR1 to LRC04) andothers are expected in the later fields but a thorough search has not been carried out yet(Szab´o et al. 2014).Before the mission, only a single classical Cepheid, V1154 Cyg was known in the
Kepler field. Despite a thorough search, no other stars were discovered (Szab´o et al. 2011). V1154
IBVS
Cyg itself turned out to be a rewarding target, revealing significant fluctuations in its lightcurve (Derekas et al. 2012). Similar effects were discovered in two other classical Cepheidswith MOST (Evans et al. 2014). A handful of stars have been classified as Cepheids in theearly CoRoT data, but those have not been investigated in detail, therefore it is unclearwhich subtypes are present in the sample (Debosscher et al. 2009).
After the failure of the second reaction wheel, the operators of
Kepler invited the scientificcommunity to submit white papers detailing the possible uses and techniques for the spacetelescope. With the help of this feedback, a new mission scenario was devised. NamedK2, after both the two-wheel operation mode and the enigmatic and challenging peak inthe Karakoram Range, the new mission will carry out shorter-duration campaigns alongthe ecliptic and will cover several fields in the sky. This setting will allow the telescope tobalance against the radiation pressure of the Sun while maintaining attitude in the othertwo directions with the remaining reaction wheels.Although the 75-day duration of the campaigns will be much shorter than the timebase of the prime mission, the availability of very diverse fields will obviously provideadequate justification for it. Moreover, the 75-day-long campaigns are still longer thanmost space-based observations. They were only exceeded by the CoRoT long runs thatlasted for about 150 days. In the foreseeable future, the regions around the ecliptic poleswill be covered by TESS for about one year and after that only PLATO, to be launchedin 2024, will carry out multi-year observations (Rauer et al. 2014).
Target selection for the first few fields, including the two-wheel engineering test run, hadto be carried out on short time scales, usually within a few weeks. Surveying the fieldswith new observations was unfeasible, hence we only proposed stars that were alreadymarked as (potential) Cepheid or RR Lyrae stars in various databases.As a start we queried the stars designated as either RR Lyrae or some subclass ofCepheids from SIMBAD. The main drawback of SIMBAD is that in most cases we cannotverify the classification with actual light curves. Therefore we searched the time-domainphotometric databases thoroughly. The All Sky Automated Survey (ASAS; Pojmanski1997) provides classifications but the photometric accuracy of the data drops considerablybelow 10-11th magnitudes. We also used the databases of the Northern Sky VariabilitySurvey (NSVS; Wo´zniak et al. 2004), SuperWASP (Paunzen et al. 2014) and the IN-TEGRAL Optical Monitoring Camera (IOMC; Mas-Hesse et al. 2003).The two most valuable sources for RR Lyrae targets were the asteroid-searching surveysthat scanned large portions of the sky repeatedly with moderate-sized telescopes to detectfaint objects. Most of these data was already mined for RR Lyrae stars to conductstatistical studies and search for stellar streams and other structures in the galactic halo.The LINEAR database contains all types of RR Lyrae stars (Sesar et al. 2013) but coversright ascensions between 8 an 18 hours only and mostly above the celestial equator . The http://skydot.lanl.gov/ http://wasp.cerit-sc.cz http://sdc.cab.inta-csic.es/omc BVS Catalina Survey data covers a larger area, most of the sky above δ ∼ − ◦ , except fora wide band along the Milky Way. A second, deeper survey (Mt. Lemmon Survey, MLS)was carried out along the ecliptic plane as well, but in both cases only the RRab starswere identified (Drake et al. 2013a,b). The LONEOS Phase I data that mostly coveredthe ecliptic was also mined for RRab stars (Miceli et al. 2008). The distribution of RRLyrae stars from the various databases are shown in Figure 1. The complete classificationof the Catalina data was released only very recently and will be used for target selectionfrom Field 4 onwards (Drake et al. 2014). SIMBAD identifications
ASAS
LINEAR (blue), LONEOS (orange)
CSS (grey), MLS (blue)
Figure 1.
Distribution of RR Lyrae stars in the five major data sets we used for target selection.
Cepheids have been classified only in SIMBAD and the ASAS and NSVS databases(Hoffmann, Harrison & McNamara 2009; Schmidt et al. 2013) so we relied upon those andtried to confirm as many of them as possible with published observations. The distributionis shown in Figure 2.
A 9-day engineering test run was carried out in February 2014 to verify the operationswith two reaction wheels.
Kepler observed some 2000 stars in the direction of Pisces.The data suffers from various pointing adjustments and drifts that make the extraction oflight curves a complicated task. We found 27 RRab, 3 RRc and a possible Cepheid staramong the observations, although the period of the latter star is about 53 days, muchlonger than the data set itself. http://catalinadata.org IBVS
SIMBAD identifications
ASAS (orange) and NSVS (blue)
Figure 2.
Distribution of various Cepheid and Cepheid-like stars in SIMBAD and the ASAS andNSVS surveys. Most stars are close the plane of the Miky Way.
The brightest RRab star in the sample was ASAS J233637-0212.7 (EPIC 60018644).We extracted the light curve of the star with the PyKE toolset to test the stability andquality of the K2 data. The telescope was repositioned after BJD = 56695.359, shiftingthe stars by a couple of pixels on the CCDs, therefore we used two pixel masks at differentpositions to extract the photometric data. The resulting background-corrected light curvecan be seen in Figure 3. A detailed summary of the RR Lyrae stars in the engineering-testdata will be published in a future paper. Figure 3.
Comparison of the ASAS, Catalina and K2 long cadence data of the RR Lyrae star ASASJ233637-0212.7. The upper panels show the light curves folded with a pulsation period of P = 0 . K2 light curve itself. After the engineering tests, a full-scale performance test was carried out. Campaign0 was a shakedown that closely resembled the proposed operations of the K2 mission. BVS Field 0 was set towards the galactic anticenter in Gemini, including the bright open clusterMessier 35. Since the asteroid searching surveys avoided this area we ended up with only10 RR Lyrae stars that fell on silicon. However, Field 0 was ideally positioned to observeCepheids: 14 stars were accepted to the long-cadence target list, including fundamental-mode and overtone stars, likely from both Type I and Type II Cepheids. Field 0 initself will double the number of Cepheids observed with space-based photometry. OneRR Lyrae (EW Gem; Schmidt & Reiswig 1993) and one first-overtone Cepheid (NSVS9770315) were selected for short cadence observations.
The first science field of the K2 mission was set towards the Leo-Virgo region and farfrom the plane of the Galaxy. With the help of the LINEAR and Catalina catalogs, weidentified and proposed no less than 133 field RR Lyrae stars. This number far exceedsthe approximately 50 stars that were observed in the original Kepler field and offered thepossibility for an unprecedented opportunity for statistical studies. Moreover, apart fromthe 118 RRab and 14 RRc stars, we found a single double-mode (RRd) star, LINEAR2122319, in the field as well. Unfortunately, very few stars made it to the final targetlist, but later fields hopefully will accumulate an expected few hundred stars to carry outthe statistical studies. On the other hand, six targets were selected for short cadenceobservations, including all three types RR Lyrae stars.In addition, we found three intriguing extragalactic stars. The dwarf spheroidal galaxyLeo IV falls into Field 1 and three RR Lyraes were identified in it by Moretti et al. (2009).These stars, along with the brightest giants and supergiants in the galaxy were includedin the target list. The brightness of the three RR Lyraes is below Kp = 21 magnitudesso they represent a considerable observational challenge. During the primary mission theestimated precision for a 21st magnitude star was about 0.15 ( P = 148700 ppm) in a singlelong-cadence observation. The precision of the K2 measurements is expected to be withina factor of 2 in fine pointing mode (Howell et al. 2014), therefore the precision of individualpoints will be around 0.2-0.3 magnitudes. Based on that, the entire data set is expectedto provide an accuracy of 5 millimagnitudes for a coherent signal. Most additional-modepeaks fall below this limit, but the strongest ones can be recovered (Benk˝o et al. 2014;Moln´ar et al. 2012). The speciality of Field 2 is the inclusion of two globular clusters, Messier 4 and 80.
Kepler ,with its 4 ′′ /px resolution was not designed to observe dense stellar fields and that leads tovarious consequences. M4 is fairly spread-out, with a half-light radius ( R h ) of 65 pixels,but its bright core would saturate the CCD. Possibly for that reason, only the northernedge falls on silicon. M80 is fainter but also more compact with a half-light radius of only4 pixels. Still, we identified many pulsating variables in the outskirts of the clusters outto about 7 R h where we expect the crowding to be acceptable. Several RR Lyrae stars,a few semiregular and SX Phe stars and a single type II Cepheid were proposed for thetwo clusters.Apart from the globular clusters, we proposed about 50 field RR Lyrae stars and a fewCepheid/W Vir candidates, as well. IBVS
The approximate positions of most of the future fields contain a large number of RR Lyraestars, including the RRc and RRd classes. Field 7 will be positioned close to the galacticplane and therefore lack asteroid survey data. Nevertheless, the GCVS contains a largeamount of otherwise unobserved RR Lyrae stars in that region (Samus et al. 2004).A comparison of the distribution of various Cepheid stars with the preliminary positionsof the K2 fields suggests that every field will contain a handful of Cepheids (barringmisidentifications in the various survey data), including all subtypes: classical Cepheids,BL Her, and W Vir stars.We identified two galaxies where it is possible to observe extragalactic Cepheids. Fields7 and 8 may include NGC 6822 and IC 1613, respectively. IC 1613 is closer and containsseveral Cepheids that are brighter than 21 magnitudes in V band (Bernard et al. 2010).NGC 6822 is somewhat farther away, but about a dozen variables are brighter than22 magnitudes and therefore we may expect reasonable photometry from those as well(Pietrzy´nski et al. 2004; Mennickent et al. 2006).The initial data products of the K2 campaigns will be the target pixel files: a timesseries of small CCD subframes containing the image of the star. We already gainedexperiences with target pixel files to create the rectified RR Lyrae data sets (Benk˝oet al. 2014), therefore the reduction of the K2 photometric data will be a relativelystraightforward task. An example (EPIC 60018657) from the 9-day engineering run isdisplayed in Fig. 4. Figure 4. K2 target pixel files from the 9-day engineering test run. Red dot marks the approximatephotocenter of the star during the first 2 days. Left panel: 1st cadence, right panel: the 150th, i.e. afterthe repositioning of the telescope. There are also four patches of reflected light in the mask, and thebrightest one contaminates the star slightly during the first part of the observations. The patches thenmove to the opposite direction compared to the stars, separating the bright blob from the target. BVS K2 mission The step-and-stare approach of the K2 mission differs significantly from the originalmission scenario. The length of the campaigns seriously limits some applications, e.g.the detection of long Blazhko periods and/or the variations in the modulation cycles.However, the availability of multiple fields opens up several new possibilities compared tothe prime mission. Continuous observations of space photometric data revealed that classical Cepheids ex-hibit light curve fluctuations (Derekas et al. 2012). These could possibly be connectedto large convective hot spots on the surface of the star (Neilson & Ignace 2014). Thereare some indications that this effect is stronger in overtone stars than in fundamental-mode pulsators. If this relation exists, it may aid the determination of pulsation modes.Confirmation, however, requires several stars with different periods to be observed.First-overtone Cepheids may turn out to be double-mode pulsators or can exhibitnon-radial modes, according to the OGLE observations (Moskalik & Ko laczkowski 2009).Some of them show strong O–C variations too, but it is not clear if that manifests in thepulsation amplitudes as well or not.Due to their lower metallicity, light variations of type II Cepheids (W Virginis and BLHerculis stars) exhibit noticeable variations. Period doubling was already detected in oneBL Her star, but hydrodynamic models predict other effects such as chaotic pulsation orlow-amplitude modulation, too (Smolec et al. 2012; Smolec & Moskalik 2014). Continu-ous, high-precision data is the best way to detect such irregularities in the pulsation ofthese stars.A few anomalous Cepheids were also identified in the fields. These stars lie between theclassical and type II Cepheids and follow a separate P-L relation in the period-luminositydiagram (Nemec, Nemec & Lutz, 1994). They have low metallicities and their origin issomewhat uncertain but may involve mass transfer in a binary system with a possible linkto blue stragglers (Szabados, Kiss & Derekas 2007). K2 could be the first space telescopeto observe anomalous Cepheids. One of the great surprises of the
Kepler mission was the detection of millimagnitude-leveladditional modes in almost all modulated RR Lyrae stars (Benk˝o et al. 2010, 2014). Theoccurrence of these modes raise serious questions about the mode selection mechanismsin RR Lyrae stars.Although hydrodynamic models can explain the occurrence of some of the additionalmodes, there is still discrepancy between the observations and the theoretical results.Based on the sample of the
Kepler field, most of the stars exhibit period doubling, re-lated to the 9th overtone (Koll´ath, Moln´ar & Szab´o 2011), and a frequency peak thatmay correspond to the second overtone ( P /P ≈ .
6; Benk˝o et al. 2014). However, inthe models, period doubling leads to the destabilization of the first overtone which wasdetected in very few stars. Finding more stars where the first overtone is excited can leadto accurate comparisons with the models, including the mode amplitudes and the signs ofmode interactions, possibly even chaos (Plachy, Koll´ath & Moln´ar 2013). Such nonlinearasteroseismic analysis was attempted only for RR Lyr itself to date (Moln´ar et al. 2012).
IBVS
Mode interactions and resonances are the best candidates so far to finally explain themysterious Blazhko effect (Buchler & Koll´ath 2011). Considering that only half of theRRab stars are modulated, a large survey is necessary to understand the relation betweenthese modes. 75 days is long enough to cover at least one modulation cycle for the majorityof the Blazhko stars.Another interesting aspect of the Blazhko effect is the apparent decline of its frequencyabove P ∼ . − . Kepler field have periods below0.69 days, therefore it could not address this issue (Nemec et al. 2013). If the K2 cam-paigns can build up a suitably large sample of modulated stars including long-period ones,the observations may shed light on the origins of this effect and the inner workings of theBlazhko effect. Modulation is present but much less common among the first-overtone (RRc) stars. Onlya few RRc stars has been observed from space so far, and none of them turned out to bemodulated yet:
Kepler could be the first space telescope to detect one.A great surprise of the original
Kepler field was that all four RRc stars turned out tobe multiperiodic. Moreover, they all shared the same properties: the additional mode wasdetected at P X /P = 0 . − .
64, and showed period doubling in all cases (Moskalik 2014).We note that similar, mysterious modes were detected not only in RRc, but in RRab andLMC Cepheid stars as well, always at the same period ratio (Moskalik 2014). A closerlook at the frequency tables of the two RRd stars where additional modes were reportedalso reveals this mode, although the authors classified them differently. Interestingly, itis connected to the first overtone in both cases (Gruberbauer et al. 2007; Chadid 2012).The origin of the P X (or P . ) modes is not yet understood. A thorough survey isrequired to find out whether all RRc and RRd stars exhibit it, or there is some connectionbetween the mode selection mechanism and the physical properties of the stars. The sameis true for the apparent period doubling of this mode. Luckily, the Catalina and LINEARsurveys will provide several RRc and RRd targets, especially from Field 4 onwards.Double-mode stars are also important on their own right. The two main modes can bemodeled accurately with the existing hydrodynamic codes, providing strong constraintson the physical parameters such as the metallicity and the mass of the star. Field RR Lyrae stars have very different metallities, between [Fe/H]= − .
05 and − . − .
54 (Nemec et al. 2013). A change in the metal content can shift themode resonance regions to different stellar parameters, as in the case of period doubling(Koll´ath et al. 2011). If mode resonances are behind the Blazhko effect, this may leadto differences in the modulation properties as well. To detect such metallicity-dependenteffects, we need to observe a large number of stars in the Galaxy.These investigations can be expanded further with the inclusion of stellar populationsthat share very similar metal content. The globular clusters M4 and M80 have distinctlydifferent metallicities: their [Fe/H] index is –1.16 and –1.75, respectively. The dwarfgalaxy Leo IV is very metal poor with [Fe/H] = − . ± .
2, so it can trace the low end ofthe metallicity sequence.
BVS Although the K2 mission was born out of the unfortunate failures of the reaction wheelson board the Kepler space telescope, the scientific potential of the new campaigns canexceed that of the original field in several areas.All types of Cepheid variables are rather rare in the Milky Way: only a handful is ex-pected in every K2 field. Therefore they do not require a large pixel budget per campaign,but the step-and-stare approach can accumulate a good sample of both classical and typeII stars. Cepheids also represent a great opportunity for extragalactic K2 observationsin the nearby dwarf galaxies. Light-curve fluctuations and detection of additional modescan provide important insights into these stars.RR Lyrae stars, like Cepheids, are an important step in the cosmic distance ladder, butthey are good tracers of the halo structures and dwarf galaxies around the Milky Way.Therefore the understanding of their pulsations is important for the galactic structureand evolution studies. Yet several open questions still remain: the Blazhko effect, themysterious P . mode, the role of mode interactions and the level of agreement betweenthe observations and the 1D pulsation models. Most of the RR Lyrae stars are faint, below14th magnitude, therefore the pixel usage would be moderate even for a high number oftargets. The capabilities of Kepler and the campaign mode of the K2 mission representan ideal opportunity to solve these questions, provided that sufficient number of stars(preferably a few hundred in total) will be observed during the mission. Acknowledgements:
The work of L. Moln´ar leading to this research was supported bythe European Union and the State of Hungary, co-financed by the European Social Fundin the framework of T ´AMOP 4.2.4. A/2-11-1-2012-0001 ‘National Excellence Program’.R.Sz. was supported by the J´anos Bolyai Research Scholarship of the Hungarian Academyof Sciences. This work has been supported by the Hungarian OTKA grant K83790,and the ‘Lend¨ulet-2009’ Young Researchers’ Programme of the Hungarian Academy ofSciences. The research leading to these results has received funding from the EuropeanCommunity’s Seventh Framework Programme (FP7/2007-2013) under grant agreementsno. 269194 (IRSES/ASK) and no. 312844 (SPACEINN). This research has made useof the SIMBAD database, operated at CDS, Strasbourg (France), NASA’s AstrophysicsData System Bibliographic Services, and PyKE (Still & Barclay 2012), an open sourcesoftware package developed and distributed by the NASA Kepler Guest Observer Office.References:Benk˝o, J. M., et al., 2010,
MNRAS , , 1585Benk˝o, J. M., Plachy, E., Szab´o, R., Moln´ar, L., Koll´ath, Z., 2014, ApJS , submittedBernard, E. J., et al., 2010,
ApJ , , 1259Borucki, W. J., et al., 2010, Science , , 977Buchler J. R., Koll´ath Z. 2011, ApJ , , 24Chadid, M., 2012, A&A , , 68Debosscher, J., et al., 2009, A&A , , 519Derekas, A., et al., 2012, MNRAS , , 1312Drake, A. J., et al. 2013a, ApJ , , 32Drake, A. J., et al. 2013b, ApJ , , 154Drake, A. J., et al. 2014, ApJS , accepted, arXiv:1405.4290 [astro-ph.SR] IBVS
Evans, N. R., Szab´o, R., Szabados, L., Derekas, A., Kiss, L., Matthews, J., Cameron, C.,2014,
IAUS , , 55Gruberbauer, M., et al., 2007, MNRAS , , 1498Hoffmann, D. I., Harrison, T. E., McNamara B. J., 2009, AJ , , 466Howell, S. B., et al., 2014, PASP , accepted, arXiv:1402.5163 [astro-ph.IM]Koll´ath, Z., Moln´ar, L., Szab´o, R., 2011,
MNRAS , 414, 1111Kolenberg, K., et al., 2010,
ApJ , , L198Mas-Hesse, J. M., et al., 2003, A&A , L261Mennickent, R. E., Gieren, W., Soszy´nski, I., Pietrzy´nski, G., 2006,
A&A , , 873Miceli, A., et al., 2008, ApJ , , 865Moln´ar, L., Koll´ath, Z., Szab´o, R., Bryson, S., Kolenberg, K., Mullally, F., Thompson, S.E., 2012, ApJ , , L13Moretti M. I., et al., 2009, ApJ , , L125Moskalik, P. A., 2014, IAUS , , 249Moskalik, P. A., Ko laczkowski, Z., 2009, MNRAS , , 1649Neilson, H., Ignace, R., 2014, A&A , , 4Nemec, J. M., Nemec, A. F. L., Lutz, T. E., 1994, AJ , , 222Nemec, J. M., Cohen, J. G., Ripepi, V., Derekas, A., Moskalik, P. A., Branimir, S.,Chadid, M., Bruntt, H., 2013, MNRAS , , 181Paunzen, E., Kuba, M., West, R. G., Zejda, M., 2014, IBVS , No. 6090Pietrzy´nski, G., Gieren, W., Udalski, A., Bresolin, F., Kudritzki, R-P., Soszy´nski, I.,Szyma´nski, M., Kubiak, M., 2004, AJ , , 2815Plachy, E., Koll´ath, Z., Moln´ar, L., 2013, MNRAS , , 3590Pojmanski, G., 1997, AcA , , 467Rauer, H., et al., 2014, Exp. Astr. , accepted, arXiv:1310.0696 [astro-ph.EP]Samus, N. N., et al., 2004, Combined General Catalogue of Variable StarsSchmidt, E. G., Reiswig, D. E., 1993, AJ , , 2429Schmidt, E. G., et al., 2013, AJ , , 61Sesar, B., et al., 2013, AJ , , 21Smolec, R., 2005, AcA , , 59Smolec, R., Moskalik, P. A., 2014, MNRAS , , 101Smolec, R., et al., 2012, MNRAS , , 2407Still, M., Barclay, T., 2012, Astrophysics Source Code Library, ascl:1208.004Szabados, L., Kiss, L. L., Derekas, A., 2007, A&A , , 613Szab´o, R., et al., 2011, MNRAS , , 2709Szab´o, R., et al., 2014, A&A , to be submittedWo´zniak, P. R., et al., 2004, AJ ,127