The Optical Gravitational Lensing Experiment. The OGLE-III Catalog of Variable Stars. I. Classical Cepheids in the Large Magellanic Cloud
I. Soszynski, R. Poleski, A. Udalski, M.K. Szymanski, M. Kubiak, G. Pietrzynski, L. Wyrzykowski, O. Szewczyk, K. Ulaczyk
aa r X i v : . [ a s t r o - ph ] O c t ACTA ASTRONOMICA
Vol. (2008) pp. 163–185 The Optical Gravitational Lensing Experiment.The OGLE-III Catalog of Variable Stars.I. Classical Cepheids in the Large Magellanic Cloud ∗ I. S o s z y ´n s k i , R. P o l e s k i , A. U d a l s k i , M. K. S z y m a ´n s k i ,M. K u b i a k , G. P i e t r z y ´n s k i , , Ł. W y r z y k o w s k i , ,O. S z e w c z y k , and K. U l a c z y k Warsaw University Observatory, Al. Ujazdowskie 4, 00-478 Warszawa, Polande-mail: (soszynsk,rpoleski,udalski,msz,mk,pietrzyn,wyrzykow,szewczyk,kulaczyk)@astrouw.edu.pl Universidad de Concepción, Departamento de Fisica, Casilla 160–C, Concepción, Chile Institute of Astronomy, University of Cambridge, Madingley Road, Cambridge CB30HA, UK
Received August 15, 2008
ABSTRACTWe present the first part of a new catalog of variable stars (OIII-CVS) compiled from the datacollected in the course of the third phase of the Optical Gravitational Lensing Experiment (OGLE-III). In this paper we describe the catalog of 3361 classical Cepheids detected in the ≈
40 squaredegrees area in the Large Magellanic Cloud. The sample consists of 1848 fundamental-mode (F),1228 first-overtone (1O), 14 second-overtone (2O), 61 double-mode F/1O, 203 double-mode 1O/2O,2 double-mode 1O/3O, and 5 triple-mode classical Cepheids. This sample is supplemented by the listof 23 ultra-low amplitude variable stars which may be Cepheids entering or exiting instability strip.The catalog data include VI high-quality photometry collected since 2001, and for some starssupplemented by the OGLE-II photometry obtained between 1997 and 2000. We provide basicparameters of the stars: coordinates, periods, mean magnitudes, amplitudes and parameters of theFourier light curve decompositions. Our sample of Cepheids is cross-identified with previously pub-lished catalogs of these variables in the LMC. Individual objects of particular interest are discussed,including single-mode second-overtone Cepheids, multiperiodic pulsators with unusual period ratiosor Cepheids in eclipsing binary systems.We discuss the variations of the Fourier coefficients with periods and point out on the sharpfeature for periods around 0.35 days of first-overtone Cepheids, which can be explained by the occur-rence of 2:1 resonance between the first and fifth overtones. Similar behavior at P ≈ P ≈
10 days for F Cepheids are also interpreted as an effect of resonances between tworadial modes. We fit the period–luminosity relations to our sample of Cepheids and compare thesefunctions with previous determinations.
Key words:
Cepheids – Stars: oscillations – Magellanic Clouds ∗ Based on observations obtained with the 1.3-m Warsaw telescope at the Las Campanas Observa-tory of the Carnegie Institution of Washington. A. A.1. Introduction
The Optical Gravitational Lensing Experiment (OGLE) is a wide-field sky sur-vey originally motivated by search for microlensing events (Paczy´nski 1986). Theobserving strategy of the project is to regularly monitor brightness of about 200 mil-lion stars in the Magellanic Clouds and Galactic bulge in the time-scales of years.A by-product of these observations is an enormous database of photometric mea-surements, which can be used for selecting long lists of newly discovered variablestars.The OGLE project yielded a wealth of information about variable stars. Thesecond phase of the survey, conducted between 1997 and 2000, resulted in catalogsof thousands Cepheids, RR Lyr stars, eclipsing binaries and long period variablesin the Magellanic Clouds. Moreover, the huge catalogs of variable sources foundin the OGLE-II fields in the Magellanic Clouds ( ˙Zebru´n et al. et al. d Cep stars, type I Cepheids), as the primary distance in-dicator, are among the most important variable stars. The Large Magellanic Cloud(LMC) is one of the most fundamental extragalactic targets of modern astrophysics,because it is our nearest non-dwarf neighbor galaxy. For this reason we begin theOIII-CVS with the catalog of classical Cepheids in the LMC.Large number of variable stars in the LMC, including classical Cepheids, werediscovered by Leavitt (1908). However, the first period derivation and the plot ofthe period–luminosity (PL) diagram for 40 LMC Cepheids was made by Shapley(1931). In 1955 the periods of 550 classical Cepheids in the LMC were published(see Shapley and McKibben Nail 1955 for the bibliography). Then, a consider-able survey for LMC Cepheids was done by Woolley et al. (1962). The catalogprepared by Payne-Gaposchkin (1971) on the basis of Harvard photographic platescontained about 1100 Cepheids in the LMC. After hiatus, a number of Cepheids inthe LMC were also discovered by Hodge and Lee (1984), Kurochkin et al. (1989),van Genderen and Hadiyanto Nitihardjo (1989) and Mateo et al. (1990). In the late1990’s very large catalogs of Cepheids were published as a by-product of gravi-tational microlensing surveys: EROS (Beaulieu et al. et al. et al. et al. et al. et al. ol. 58
2. Observations and Data Reduction
The photometric data were obtained with the 1.3-m Warsaw telescope located atLas Campanas Observatory in Chile. The observatory is operated by the CarnegieInstitution of Washington. The telescope is equipped with the “second generation”camera consisting of eight SITe 2048 × µ m pixelswhat corresponds to 0.26 arcsec/pixel scale. The gain of the chips is adjusted to beabout 1.3 e − / ADU with the readout noise from 6 to 9 e − depending on the chip.For details of the instrumental setup we refer to Udalski (2003).Observations of 116 OGLE-III fields covering 39.7 square degrees of the LMCstarted in July 2001. The data presented in this paper were collected up to March2008. In the future, photometry provided with the catalog will be supplemented byobservations obtained after this date, up to the end of the third phase of the OGLEsurvey.The photometry was obtained using Difference Image Analysis (DIA) tech-nique (Alard and Lupton 1998, Alard 2000, Wo´zniak 2000), which is able to per-form in dense stellar fields considerably better photometry than the traditional PSF-fitting programs. We emphasize that even though there are small gaps between thechips of the CCD mosaic, our final DIA photometry pipeline (Udalski et al. I photometric band closely re-sembling standard filter. The remaining frames were observed in the V bandpass.The data were calibrated to the standard system using hundreds of thousands starsobserved during the OGLE-II project (1997–2000) and common with the OGLE-IIIobservations. The procedure of calibration of the data and astrometric transforma-tions were described by Udalski et al. (2008a). The accuracy of the photometriccalibrations is better than 0.02 mag, while equatorial coordinates in the cataloghave uncertainties of about 0. ′′ A. A. estimated. Here we corrected the error bars using technique derived for OGLE-IILMC microlensing events search and described in detail in Wyrzykowski et al. (in preparation). In brief, the method compares observed photometric scatter ofconstant stars in a given field with their mean error bars and fits the coefficientsof relation between original and corrected error bar: s corr = (( gs ) + e ) / . ForOGLE-III data g and e were derived independently for each field and CCD chip.On the average g = .
204 and e = . I band. In V filter: g = .
996 and e = . I -band data in the OGLE-III database,usually due to exceeding the CCD saturation limit (which is about I =
13 mag).In a few cases our photometric techniques failed in the centers of the clusters. For16 of these objects the OGLE-II I -band photometry is available, where the level ofsaturation was slightly higher ( I = . V fil-ter, because our V -band data for classical Cepheids saturate for periods longer than ≈
50 days. Only for the five longest-period Cepheids, neither I nor V -band pho-tometry are available. However, even for these stars we could measure periods andshapes of their light curves, because in the DIA technique every bright variable starproduces a number of artificial objects in the closest neighborhood that mimic thevariability of this bright star. This is because the DIA method does not subtract theprofiles of the neighboring objects while doing photometry. We used these artificiallight curves for measuring periods of the long-period Cepheids. The calibrated VI photometry for brightest stars in the LMC will be published soon, when the shal-low survey conducted on the Warsaw telescope will be finished. At that momentwe will be able to supplement our catalog with the data for the brightest Cepheids.
3. Selection of Cepheids
The search for classical Cepheids in the LMC was preceded by a massive pe-riod search performed using supercomputers at the Interdisciplinary Centre forMathematical and Computational Modeling of Warsaw University (ICM UW). Wesearched for periodicity all 32 million stars in the LMC using program F
NPEAKS by Z. Kołaczkowski. We tested the range of frequencies from 0.0 to 24.0 cycles ol. 58 I , V , Wesenheit index, andnear-infrared K band from the 2MASS project (Cutri et al. ( V − I ) colors.We removed from the list objects bluer than ( V − I ) = . ( V − I ) = . d Sct stars (HADS). The long period classical Cepheids were relatively easy todistinguish, due to their narrow PL relations and characteristic shapes of the lightcurves. However, in the short-period domain (for P < P < i.e. , the ratio of periods vs. logarithm of thelonger period plot, Petersen 1973) the stars pulsating simultaneously in the fun-damental mode and first overtone (F/1O) are naturally separated into two groups,with the gap between (fundamental mode) periods in the range 0.4–1 days. Amongpulsators with the first two overtones excited (1O/2O) we detected only one groupwith the shortest first overtone period of about 0.24 days. Thus, P = .
24 days wastaken for the first-overtone pulsators as a boundary between HADS and classicalCepheids. For the fundamental-mode Cepheids we cut our sample of d Cep starsat P = .
995 days, i.e. , the shortest F period of double-mode F/1O Cepheids (withexception of a peculiar object OGLE-LMC-CEP-0083). We notice, that only a few68
A. A. pulsating-like variables with shorter periods were found on the continuation of thefundamental-mode Cepheids PL relation, but the shapes of their light curves wereconsiderably different than for F Cepheids. The preliminary distinction betweenfundamental (F) and first overtone (1O) classical Cepheids was done using theirpositions in the W I –log P diagram, but the final classification utilized the Fourierparameters R and f (see Section 6). The exemplary light curves of single-modeCepheids from the whole range of periods and luminosities covered by OGLE arepresented in Fig. 1. Fig. 1. Illustrative light curves of fundamental-mode ( left panel ), first-overtone ( middle panel ) andsecond-overtone ( right panel ) Cepheids. Small numbers at the right side of each panel show therounded periods in days of the light curves presented in panels. ol. 58
Classical Cepheids pulsating solely in the second overtone are very rare butastrophysically interesting objects, because they can be used as an independent testof pulsational and evolutionary models (Antonello and Kanbur 1997, Bono et al. et al. et al. (1999a) undertook a searchfor single-mode 2O Cepheids in the LMC using double-mode Cepheids pulsatingin the first and the second overtones as templates. They separated both modesin beat Cepheids and studied the 2O variations. Singly-periodic second overtoneCepheids should have nearly sinusoidal light curves, small amplitudes and meanluminosities slightly higher than 1O Cepheids of the same periods. In the color–magnitude diagram these stars should occupy the blue edge of the instability strip.Alcock et al. (1999a) proposed one candidate for 2O Cepheid in the LMC.Udalski et al. (1999b) followed the same strategy as MACHO group for Cephe-ids in the Small Magellanic Cloud (SMC). Theoretical investigations suggest thatmetal-poor environments favor Cepheids pulsating purely in the second overtonemode. As a result, Udalski et al. (1999b) found 13 firm candidates for 2O pulsators– the largest such sample detected to date.We started a search for singly-periodic second overtone Cepheids using allLMC stars in our database. We selected stars with S/N of periods larger than 9and located just above the period – I -band magnitude relation for 1O Cepheids(spreading from the upper edge of the PL sequence to 0.75 mag above this line).Then, the light curves were visually inspected and obvious eclipsing binaries wererejected. At this stage we noticed a distinct group of stars located at the blue edge ofthe instability strip (0 . < ( V − I ) < . V and I bands, what allowed us to remove a couple of ellipsoidal variables fromour list. Ellipsoidal modulation is mainly a geometrical effect, so the amplitudesin two filters are very similar, while for Cepheids, the V -band amplitudes are largerthan for I band by a factor of about 1.7.In total 14 objects passed our selection criteria, all of them in the period range0.58–1.2 days and in the low-amplitude domain. It is a very homogeneous groupwhich delineates additional PL sequence located above the relation of the first-overtone pulsators. The shapes of the light curves (see Fig. 1 for examples) are ofthe same type as in the 2O Cepheids found by Udalski et al. (1999b), with Fourierparameter R smaller than 0.1. The MACHO candidate for 2O Cepheid in theLMC (Alcock et al. Double-mode or beat Cepheids pulsate simultaneously in two radial modes –usually fundamental mode and first overtone (F/1O) or first and second overtones70
A. A. (1O/2O). Very few such objects had been known before the large microlensingsurveys era. The situation changed when the MACHO project announced the dis-covery of 45 beat Cepheids in the LMC (Alcock et al. et al. et al. et al. d Sct andRR Lyr stars, which will be presented in the next parts of the OIII-CVS.Second, additional search for double-mode variables was performed for all pre-viously selected Cepheids. Each light curve was fitted by a Fourier series with anumber of harmonics minimizing the c per degree of freedom and the functionwas subtracted from the observational data. The residuals were searched for otherperiodic signals and, if detected, such a candidate was marked for visual inspec-tion. We discovered several new double-mode Cepheids and a number of othermulti-periodic variables, such as Cepheids with periodicities very close to the dom-inant periods, Cepheids in eclipsing binary systems, Cepheids blended with othervariable stars.Fig. 2 shows the Petersen diagram for multi-mode Cepheids in our list. Blackpoints represent the Cepheids pulsating in two radial modes. It is worth emphasiz-ing that our sample of double-mode Cepheids covers much larger range of periodsthan presented to date. We especially point out for one long-period F/1O CepheidOGLE-LMC-CEP-1082 with fundamental-mode period equal to 7.86 days, how-ever connected with exceptionally low amplitude of pulsations. In the list of F/1Odouble-mode Cepheid we included also one double-periodic pulsator, OGLE-LMC-CEP-0083, with exceptionally short periods (0.581 d and 0.440 d) and the periodratio placing this object in the Petersen diagram somewhat above the line connect-ing F/1O double-mode Cepheids and HADS. On the other hand, we found a con-siderable number of short-period 1O/2O Cepheids, with the first-overtone periodseven as short as 0.24 days. It is remarkable that in the range of 1O periods 0.5–0.75 days double-mode pulsators are significantly more common than single-modeCepheids.We also performed a search for triple mode Cepheids. Three new stars of thattype were discovered in addition to two already known triple-mode Cepheids in theLMC (Moskalik et al. et al. (2008). In the same work we also announced the discovery oftwo double-mode Cepheids pulsating simultaneously in the first and third overtone ol. 58 Fig. 2. Petersen diagram for multiperiodic Cepheids in the LMC. Solid dots represent double-mode(F/1O, 1O/2O and 1O/3O) Cepheids, grey triangles show triple-mode Cepheids (three points per star)and empty circles show selected other stars with significant secondary periods. modes. This is a new class of double-mode Cepheids.During the search for multiperiodicity we detected a significant number of Ce-pheids with the secondary periods very close to the primary ones. Such ratios ofperiods close to 1 are usually connected with a long-term amplitude and/or phasemodulation, and, by analogy to RR Lyr stars, these objects are called Blazhko Ce-pheids. Using our simple analysis we detected such behavior in about 4% of FCepheids, 28% of 1O and in the same fraction of 2O single-mode Cepheids, 18%of F/1O beat Cepheids, and 36% of 1O/2O double-mode Cepheids. For single-mode pulsators this is somewhat larger fraction than detected by Moskalik andKołaczkowski (2008b) in the OGLE-II data, what can be explained by the longertime span of our new photometry. Among double-mode Cepheids, the fraction ofnon-radial pulsators seem to be comparable with values determined by Moskalikand Kołaczkowski (2008a,b).We also noticed in the Petersen diagram a number of objects with period ratiosthat do not match any known phenomena. Especially numerous group was foundfor period ratios of 0.60–0.63 days and (longer) periods in the range 1.7–2.6 days.It is interesting that the longer periods are always related to the first overtone mode,while the shorter periods are connected to low-amplitude variations. A few such72
A. A.
Cepheids were recently listed by Moskalik et al. (2008b) who suggested that sucha behavior is connected with a kind of non-radial oscillations. In our Petersendiagram these stars seem to follow two sequences, with period ratios in the ranges0.60–0.61 and 0.62–0.63.
Fig. 3. Light curves of Cepheids with additional eclipsing variability.
Left panels show the originalphotometric data folded with the Cepheid periods.
Right panels show the eclipsing light curves aftersubtracting the Cepheid component. In the case of OGLE-LMC-CEP-1718 two Cepheid light curvesare presented.
As a by-product of searching for double-mode pulsators we detected three Ce-pheids with eclipsing modulation imposed on the pulsational light curves: OGLE-LMC-CEP-0227, OGLE-LMC-CEP-1812 and OGLE-LMC-CEP-2532. Only thelast one was already reported by Udalski et al. (1999d – LMC_SC16 119952). ol. 58 et al. (1995– MACHO*05:21:54.8-69:21:50) and noticed by Udalski et al. (1999d), but theeclipses were overlooked. Fig. 3 shows the light curves of this object. Besides, wealso detected two already known double Cepheids (Alcock et al. e.g. , cataclysmic-like chan-ges, semiregular variations, long secondary periods, etc. The majority of thesecases can be explained as an effect of physical or optical binarity with other vari-able star.Since the time base of the OGLE-III observations is 2400 days, and for themerged OGLE-II and OGLE-III data is longer than 4000 days, our data offer apossibility of studying various long-term effects in the Cepheids – period changes,amplitude and phase modulation, or mean magnitude changes. The full study ofthe period changing Cepheids will be published in the forthcoming paper (Poleski2008, in preparation).
4. Catalog of Classical Cepheids in the LMC
In total 3361 classical Cepheids were found in the LMC OGLE-III fields. Thesample consists of 1848 fundamental-mode, 1228 first-overtone, 14 second-over-tone, 61 double mode F/1O, 203 double mode 1O/2O, 2 double-mode 1O/3O,2 triple-mode F/1O/2O and 3 triple-mode 1O/2O/3O classical Cepheids.In addition we prepared a list of 23 low amplitude variables which can be re-lated to classical Cepheids, because they lie in the vicinity of Cepheids in the color–magnitude and PL diagrams. Such low-amplitude Cepheids are expected by theevolutionary models as pulsators entering or exiting the instability strip. A . A . T a b l e 1
Exemplary part of the ident.dat file
Cepheid ID OGLE-III ID Mode RA DEC OGLE-II ID MACHO ID ASAS ID GCVS ID OtherField No [J2000.0] [J2000.0] LMC... designation...OGLE-LMC-CEP-0931 LMC118.5 23110 F 05:05:57.39 − − − − − − − − − − − − − − − − − − − − − − − − − ol. 58 et al. (2005). We stress that our sample of low amplitude variables notnecessarily consist of Cepheids only. Such light curves can be a product of ellip-soidal or rotational modulation, or blending with other type of variable star.The OIII-CVS is available in the electronic version only from the OGLE Inter-net archive: http://ogle.astrouw.edu.pl/ftp://ftp.astrouw.edu.pl/ogle/ogle3/OIII-CVS/lmc/cep/ The catalog is accessible through a user-friendly WWW interface or via
FTP site.In the FTP the full list of our sample of classical Cepheids is given in the file ident.dat . Part of this file is shown in Table 1. All objects are arranged accordingto their right ascension. The Cepheids have designations of the form OGLE-LMC-CEP-NNNN, where NNNN is a four digit consecutive number. In the followingcolumns of Table 1 we provide: Cepheid ID, OGLE-III field and the database num-ber of star (consistent with the LMC photometric maps of Udalski et al.
T a b l e 2
First 20 lines of the cepF.dat file
Cepheid ID h I i h V i P s P T max A ( I ) R f R f [mag] [mag] [days] [days] (HJD-2450000) [mag]OGLE-LMC-CEP-0002 15.672 16.412 3.1181195 0.0000161 2171.23921 0.257 0.296 4.705 0.101 2.962OGLE-LMC-CEP-0005 14.661 15.413 5.6120581 0.0000135 2171.78078 0.521 0.431 4.971 0.167 3.391OGLE-LMC-CEP-0012 15.469 16.067 2.6601882 0.0000022 2162.43751 0.688 0.522 4.402 0.314 2.611OGLE-LMC-CEP-0016 13.707 14.787 10.5064564 0.0015553 2157.57180 0.115 0.028 6.015 0.106 0.923OGLE-LMC-CEP-0017 15.345 15.992 3.6772562 0.0000299 2169.48050 0.611 0.494 4.594 0.287 2.983OGLE-LMC-CEP-0018 15.222 16.051 4.0478526 0.0000275 2165.39166 0.265 0.284 4.567 0.086 2.998OGLE-LMC-CEP-0021 14.722 15.491 5.4579746 0.0000259 2165.34485 0.431 0.416 4.969 0.145 3.555OGLE-LMC-CEP-0023 16.325 17.044 1.7018254 0.0000019 2164.75709 0.403 0.441 4.352 0.261 2.493OGLE-LMC-CEP-0025 15.343 16.157 3.7334998 0.0000099 2165.69444 0.360 0.409 4.694 0.173 3.131OGLE-LMC-CEP-0026 15.466 16.081 2.5706764 0.0000033 2164.25802 0.622 0.442 4.275 0.228 2.263OGLE-LMC-CEP-0027 15.039 15.641 3.5229468 0.0000052 2165.27241 0.647 0.449 4.377 0.254 2.509OGLE-LMC-CEP-0028 16.620 17.276 1.2629545 0.0000007 2171.08146 0.513 0.475 4.174 0.280 2.120OGLE-LMC-CEP-0033 14.400 15.233 7.1807757 0.0000217 2160.83200 0.488 0.339 5.377 0.213 3.723OGLE-LMC-CEP-0034 13.737 14.630 11.2546276 0.0000595 2180.63850 0.478 0.170 4.758 0.107 5.295OGLE-LMC-CEP-0035 14.375 15.157 6.9436898 0.0000448 2165.30738 0.393 0.359 5.365 0.138 3.768OGLE-LMC-CEP-0037 15.525 16.308 3.0669047 0.0000082 2165.55271 0.271 0.345 4.576 0.136 2.958OGLE-LMC-CEP-0039 15.415 16.069 3.1477368 0.0000291 2164.95088 0.338 0.411 4.568 0.180 2.783OGLE-LMC-CEP-0040 14.662 15.408 5.1651454 0.0000080 2182.37433 0.574 0.436 4.633 0.204 3.009OGLE-LMC-CEP-0041 15.522 16.270 2.9106624 0.0000235 2165.81479 0.181 0.245 4.451 0.043 2.401OGLE-LMC-CEP-0042 15.673 16.367 2.5770292 0.0000034 2183.19453 0.599 0.485 4.377 0.272 2.568 A. A.
Files cepF.dat , cep1O.dat , cep2O.dat , cepF1O.dat , cep1O2O.dat , cep1O3O.dat , cepF1O2O.dat , and cep1O2O3O.dat list basic parameters of the sin-gle-, double- and triple-mode Cepheids with the appropriate modes excited. Forsingle-mode objects the consecutive columns contain: intensity mean magnitudesin the I and V bands, periods in days and their uncertainties, moments of the zerophase corresponding to maximum light, amplitudes in the I -band, and Fourier pa-rameters R , f , R , f derived for the I -band light curves. For double-modeand triple-mode pulsators the format of tables is longer including additional peri-odicities. First rows of the file cepF.dat are shown in Table 2.The file remarks.txt contains additional information on some Cepheids. We pro-vide here remarks about uncertain classification, interesting features as additionalvariability, variable mean magnitudes or amplitude modulation, and informationabout differences compared to other catalogs (for example different periods). Di-rectory phot/ contains I - and V -band OGLE photometry of the stars in our cata-log. If available, OGLE-II data are merged with the OGLE-III data. Finally, thedirectory fcharts/ contains finding charts of all objects. These are the 60 ′′ × ′′ segments of the I -band DIA reference images, oriented with N up, and E to the left. Fig. 4. OGLE-III fields in the LMC. Dots indicate positions of classical Cepheids from the OIII-CVScatalog. The background image of the LMC is originated from the ASAS wide field sky survey. ol. 58 i.e. , a bump moving toward earlier phases with increasing periods.
5. Cross-Identification with Previous Catalogs
The catalog includes matches of our objects with previously published list ofclassical Cepheids in the LMC. We queried the following catalogs: OGLE-II cata-logs of Cepheids in the LMC (Udalski et al. et al. † , the extragalactic part of the General Catalogueof Variable Stars (Artyukhina et al. ‡ (Pojma´nski 2002). In our sample of Cepheids 2367 stars were already published inany of these catalogs, 994 objects are new detections.With the aim of testing the completeness of our catalog we carefully checkedall objects not present in our list, but potentially covered by the OGLE-III fields.Compared to the OGLE-II catalog of Cepheids in the LMC published by Udalski et al. (1999d – single-mode Cepheids) and Soszy´nski et al. (2000 – double-modeCepheids) we missed 8 classical Cepheids (not counting 16 objects saturated in theOGLE-III data). Five of these Cepheids are members of LMC clusters and severecrowding affected the OGLE-III photometry. The three remaining objects wereclose to the edges of the fields and were affected by a small number of observations.We supplemented our catalog with these missing objects.The MACHO project released the list of about 1800 Cepheids in the LMC.1721 of them could potentially be found in the OGLE-III fields. We did not findcounterparts for 15 variables. Lack of five of these stars can be explained by asmall number of points at the field edges or problems with the photometry in denseregions of the sky. The other objects were classified by us as different type ofvariable stars, usually eclipsing binaries.The General Catalogue of Variable Stars contains 873 stars classified as DCEPor DCEPS and potentially present in our fields in the LMC. We performed an ex-tensive searching for the counterparts of these stars in our sample. In a numberof cases the cross-identification between our sample and the GCVS was uncertain,because of a large discrepancy (sometimes larger than 2 ′ ) in coordinates of the ob-jects. Sometimes, the stars were positionally coincident, but the periods disagreed.For many of these objects we noticed that the period provided in the GCVS was analias of the true period. One missing object, LMC V0477 (DV53), we classified aseclipsing or ellipsoidal binary. Another star, LMC V0857 (HV 2284), was recog- † ‡ A. A. nized as Galactic RR Lyr star. LMC V0566 (HV 12509) is likely a type II Cepheidin the eclipsing binary system.We found no counterparts for nine stars classified as classical Cepheids in theGCVS, namely LMC V0224 (HV 12496), V0452 (DV44), V1620 (HV 13018),V2175 (HV 5767), V2329, V2368 (HV 13032), V3972, V4366, and V4441 (HV12909). Since periods and coordinates provided by the GCVS are sometimes er-roneous, it is possible that these objects are present in our sample, but each caseshould be directly checked using finding charts from the literature. Such investiga-tion should also reveal possible Cepheids which stopped pulsating.
6. Basic Parameters
The periods of variable stars and uncertainties of periods provided in the cata-log were calculated with the T
ATRY program using multiharmonic periodogram ofSchwarzenberg-Czerny (1996). To determine mean luminosities, amplitudes andFourier parameters, each of the light curve was fitted by a Fourier series of the or-der depending on the shape and scatter of the light curve. The number of harmonics(maximum 12) was adjusted to minimize the value of c per degree of freedom. Inthis procedure we used the program J-23 written by T. Mizerski.The VI intensity mean magnitudes were derived by integrating the light curvesconverted to intensity units and transforming the result back to the magnitude scale.The amplitudes provided with the catalog are the differences between the maximumand minimum values of the function fitted to the light curves. Fourier parameters– amplitude ratios R k = A k / A and phase differences f k = f k − k f (Simon andLee 1981) – were derived using the same method of Fourier decomposition. Forlight curves with insignificant higher harmonics the amplitude ratios are equal tozero, while the appropriate phase differences are not defined.Fig. 5 shows Fourier parameters R , f , R , and f of the I -band lightcurves plotted against log P . For clarity fundamental-mode and overtone Cephe-ids are presented in separate panels. Fourier coefficients are widely used tool forquantitative description of the structure of Cepheid light curves. Complex patternvisible in the diagrams reflects the Hertzsprung progression. The minimum of R at P ≈
10 days for fundamental-mode Cepheids is interpreted as a signature of2:1 resonance between the fundamental and the second overtone mode of pulsation(Andreasen and Petersen 1987). In the vicinity of 10 days periods the f parame-ter rises sharply to 2 p and appears again in the lower part of the diagram, what iscaused by rotation of f modulo 2 p . Note that such sharp feature in f is notobserved for some of the Cepheids around this period, but all of them cross the line f = P ≈
10 days.Similar behavior seem to appear twice for the first-overtone pulsators – at pe-riods ≈ .
35 days and ≈ ol. 58 Fig. 5. Fourier parameters of the fundamental-mode ( left panels ) and overtone ( right panels ) Cephe-ids. A. A. ⊙ (W. Dziembowski, private communication).
7. Period-Luminosity Relation
Period–luminosity relation of classical Cepheids plays a crucial role as an in-dicator of the cosmic distance scale. The LMC is of special interest in this field,because the extragalactic distances are calibrated with the distance to this galaxy.Classical Cepheids found during the second phase of the OGLE survey in the LMCwere widely used in various programs aiming at distance determinations, e.g. , HSTKey Project (Freedman et al. et al. V and I magnitudes and in the reddening-free Wesen-heit index W I = I − . ( V − I ) (Madore 1982) are shown in Fig. 6. It is strikingthat substantial fraction of points in the first two plots are located considerably be-low the mean PL relations, but these stars generally follow the narrow sequencesin the period – Wesenheit index plane. We interpret such behavior as an effect ofinterstellar extinction – very variable from star to star. Considerable reddening forsome Cepheids is visible on the color–magnitude diagram plotted in Fig. 7.We also carefully checked the stars that do not obey the PL relations in thelog P – W I plane. In the majority of these cases this discrepancy can be explainedby blending with other star unresolved in our data. Most of the Cepheids with su-perimposed additional type of variations (including Cepheids with eclipsing mod-ulations) do not match the period- W I laws. Notice that the PL sequence for thelongest-period fundamental-mode Cepheids seem to be more scattered than for theshorter-period variables. It may be due to saturation effects, but the definitive an-swer will be given after publication of the OGLE shallow survey to the LMC.The PL relations provided below are not compensated for interstellar extinction.We leave it to the reader to de-redden magnitudes using either average extinctioncorrection for the whole LMC or individual values for each object. We also take noaccount for possible break of the linearity suggested for the PL relation of funda-mental mode Cepheids (see Ngeow et al. s clipping yields the following PL relationsfor single-mode fundamental-mode classical Cepheids: V = − . ( ± . ) log P + . ( ± . ) s = .
22 mag I = − . ( ± . ) log P + . ( ± . ) s = .
15 mag W I = − . ( ± . ) log P + . ( ± . ) s = .
08 mag ol. 58
Fig. 6. Period–luminosity diagrams for classical Cepheids in the LMC. Blue and cyan points showfundamental-mode pulsators, red and magenta – first-overtone, green – second overtone. Solid dotsare single-mode Cepheids, while empty circles represent double-mode pulsators (two points per star). A. A.
Fig. 7. Color–magnitude diagram for classical Cepheids in the LMC. In the background stars fromthe subfield LMC100.1 are shown. The significant number of very red Cepheids are clearly locatedto the red side of the respective instability strips for the various pulsation modes indicating that largereddening is not unusual in the LMC. and for the first-overtone mode: V = − . ( ± . ) log P + . ( ± . ) s = .
23 mag I = − . ( ± . ) log P + . ( ± . ) s = .
16 mag W I = − . ( ± . ) log P + . ( ± . ) s = .
07 mag . The slopes of the PL relations for fundamental-mode Cepheids agree within 1 s with previous determinations by Udalski et al. (1999c) § , and Fouqué et al. (2007).Of course, in the V and I domains we can only compare the slopes, because wedid not applied the reddening correction, but in the W I extinction-free index wecan consider both, the slope and the zero point, of the PL relation. There is largerdiscrepancy ( > s ) in the zero point of the log P – W I relation. We also obtained § updated coefficients are available from: ftp://ftp.astrouw.edu.pl/ogle/ogle2/var_stars/lmc/cep/catalog/README.PL ol. 58 et al. (1999c) and Fouqué et al. (2007) used the same definition of the Wesenheit index as we did, but this definitiondepends on the adopted reddening law. Using different coefficient in the Wesenheitindex results in different slope and zero point of the fitted PL relation.The agreement in slopes is much worse for first-overtone Cepheids, althoughstill within 3 s . We suspect that a considerable number of very short-period first-overtone pulsators, not present in the OGLE-II sample of Cepheids, which changedthe fitted function could be an origin of this discrepancy. The non-linearity ofthe PL relation may take place for the first-overtone Cepheids, with the break at P ≈ .
8. Summary
In this paper we present the largest catalog of classical Cepheids in the LMCand probably the largest sample of such stars identified to date in any environment.Our list of Cepheids is supplemented by the high quality, long-term standard pho-tometry enabling precise analysis of these stars. These data are ideal for studyingmany fundamental problems, such as interpretation of the pulsational and evolu-tionary models of Cepheids, non-radial oscillations in the pulsating stars, possiblenon-linearity of the PL relation, structure and history of the LMC.The catalog contains very rare objects, such as Cepheids with three radialmodes excited, 1O/3O double-mode Cepheids, single-mode second-overtone pul-sators, Blazhko Cepheids, eclipsing binary systems containing Cepheids includingsystem of two Cepheids eclipsing each other. Our data show that first-overtoneclassical Cepheids and high amplitude d Sct stars follow continuous PL relation.Distribution of the Fourier parameters suggests that the internal resonance betweenradial modes may occur twice for the first-overtone pulsators: for periods of about0.35 days and 3 days. The PL relation for first-overtone Cepheids is possibly non-linear, with a discontinuity in the slope around P = . Acknowledgements.
The authors wish to thank Prof. W.A. Dziembowski,Prof. M. Feast, and Prof. W. Gieren for many helpful suggestions which improvedthe paper. We thank Drs. Z. Kołaczkowski, T. Mizerski, G. Pojma´nski, A. Schwar-zenberg-Czerny and J. Skowron for providing the software and data which enabledus to prepare this study.84
A. A.
This work has been supported by the Foundation for Polish Science throughthe Homing (Powroty) Program and by MNiSW grants: NN203293533 to IS andN20303032/4275 to AU.The massive period searching was performed at the Interdisciplinary Centre forMathematical and Computational Modeling of Warsaw University (ICM UW). Weare grateful to Dr. M. Cytowski for helping us in this analysis.REFERENCES
Afonso, C., et al. (EROS) 1999, astro-ph/9907355.Alard, C., and Lupton, R.H. 1998,
ApJ , , 325.Alard, C. 2000, A&AS , , 363.Alcock, C., et al. (MACHO) 1995, AJ , , 1653.Alcock, C., et al. (MACHO) 1999a, ApJ , , 185.Alcock, C., et al. (MACHO) 1999b, AJ , , 920.Andreasen, G.K., and Petersen, J.O. 1987, A&A , , 129.Antonello, E., and Poretti, E. 1986, A&A , , 149.Antonello, E., and Kanbur, S.M. 1997, MNRAS , , L33.Artyukhina N.M. et al. et al. A&A , , 137.Bono, G., Caputo, F., and Marconi, M. 2001, MNRAS , , 1353.Buchler, J.R., Wood, P.R., Keller, S., and Soszy´nski, I. 2005, ApJ , , L151.Burki, G., Schmidt, E.G., Arellano Ferro, A., Fernie, J.D., Sasselov, D., Simon, N.R., Percy, J.R., andSzabados, L. 1986, A&A , , 139.Cutri, R.M., et al. A&A , , 73.Freedman, W.L., et al. ApJ , , 47.Gieren, W., Pietrzy´nski, G., Walker, A., Bresolin, F., Minniti, D., Kudritzki, R.-P., Udalski, A.,Soszy´nski, I., Fouqué, P., Storm, J., and Bono, G. 2004, AJ , , 1167.Hertzsprung, E. 1926, Bull. Astr. Inst. Netherlands , , 115.Hodge, P.W., and Lee, S.-O. 1984, ApJ , , 509.Kurochkin, N.E., Tokovinin, A.A., and Loggins, A. 1989, IBVS , 3365.Leavitt, H.S 1908,
Harvard Obs. Ann. , , 87.Madore, B.F. 1982, ApJ , , 575.Mateo, M., Olszewski, E.W., and Madore, B.F. 1990, ApJ , , L11.Moskalik, P., Kołaczkowski, Z., and Mizerski, T. 2004, in: “Variable Stars in the Local Group”, Ed.D.W. Kurtz, and K. Pollard, ASP Conf. Ser., , p. 498.Moskalik, P., and Kołaczkowski, Z. 2008a, arXiv:0807.0615.Moskalik, P., and Kołaczkowski, Z. 2008b, arXiv:0807.0623.Ngeow, C., Kanbur, S.M., and Nanthakumar, A. 2008,
A&A , , 621.Paczy´nski, B. 1986, ApJ , , 1.Payne-Gaposchkin, C.H. 1971, Smithsonian Contrib. Astrophys. , .Petersen, J.O. 1973, A&A , , 89.Pojma´nski, G. 1997, Acta Astron. , , 467.Pojma´nski, G. 2002, Acta Astron. , , 397.Schwarzenberg-Czerny, A. 1996, ApJ , , L107.Shapley, H. 1931, Harvard College Observatory Bulletin , , 16.Shapley, H., and McKibben Nail, V. 1955, Proceedings of the National Academy of Science , , 829. ol. 58 Simon, N.R., and Lee, A.S. 1981,
ApJ , , 291.Soszy´nski, I., Udalski, A., Szyma´nski, M., Kubiak, M., Pietrzy´nski, G., Wo´zniak, P., and ˙Zebru´n, K.2000, Acta Astron. , , 451.Soszy´nski, I., Poleski, R., Udalski, A., Kubiak, M., Szyma´nski, M.K., Pietrzy´nski, G.,Wyrzykowski, Ł., Szewczyk, O., and Ulaczyk, K. 2008, Acta Astron. , , 153.Szyma´nski, M.K. 2005, Acta Astron. , , 43.Udalski, A. 2003, Acta Astron. , , 291.Udalski, A., Soszy´nski, I., Szyma´nski, M., Kubiak, M., Pietrzy´nski, G., Wo´zniak, P., and ˙Zebru´n, K.1999a, Acta Astron. , , 1.Udalski, A., Soszy´nski, I., Szyma´nski, M., Kubiak, M., Pietrzy´nski, G., Wo´zniak, P., and ˙Zebru´n, K.1999b, Acta Astron. , , 45.Udalski, A., Szyma´nski, M., Kubiak, M., Pietrzy´nski, G., Soszy´nski, I., Wo´zniak, P., and ˙Zebru´n, K.1999c, Acta Astron. , , 201.Udalski, A., Soszy´nski, I., Szyma´nski, M., Kubiak, M., Pietrzy´nski, G., Wo´zniak, P., and ˙Zebru´n, K.1999d, Acta Astron. , , 223.Udalski, A., Szyma´nski, M.K., Soszy´nski, I., and Poleski. R. 2008a, Acta Astron. , , 69.Udalski, A., Soszy´nski, I., Szyma´nski, M.K., Kubiak, M., Pietrzy´nski, G., Wyrzykowski, Ł.,Szewczyk, O., Ulaczyk, K., and Poleski. R. 2008b, Acta Astron. , , 89.van Genderen A.M., and Hadiyanto Nitihardjo, G. 1989, A&A , , 230.Welch, D.L., et al. (MACHO) 1997, in: “Variable Stars and the Astrophysical Returns ofMicrolensing Surveys”, Ed. R. Ferlet, J.P. Maillard, and B. Raban (Éditions Frontiéres),p. 205.Woolley, R.vdR., Sandage, A.R., Eggen, O.J., Alexander, J.B., Mather, L., Epps, E., and Jones, S.1962, R. Obs. Bull. , .Wo´zniak, P.R. 2000, Acta Astron. , , 421.Wo´zniak, P.R., Udalski, A., Szyma´nski, M., Kubiak, M., Pietrzy´nski, G., Soszy´nski, I., and ˙Zebru´n, K.2002, Acta Astron. , , 129.˙Zebru´n, K., Soszy´nski, I., Wo´zniak, P.R., Udalski, A., Kubiak, M., Szyma´nski, M., Pietrzy´nski, G.,Szewczyk, O., and Wyrzykowski, Ł. 2001, Acta Astron. ,51