ABSTRACT
We have conducted a ground-based BVI synoptic survey of the Local Group galaxy M 33 which covers most of its disk and spans a period of 7 years. The survey targets luminous, long-period variables such as Cepheids and Miras and combines images from the DIRECT project and follow-up observations at the WIYN 3.5 m telescope. This paper, the first in a series, presents the discovery and characterization of 564 Cepheid variables, which represent a factor of two increase over previous samples with calibrated point-spread function (PSF) photometry. We also describe the details of the observations and analysis of the survey data, including the use of archival Hubble Space Telescope images to characterize biases in our ground-based PSF photometry.
Export citation and abstract BibTeX RIS
1. INTRODUCTION
The study of Cepheid variables in nearby galaxies has been a very active field for nearly 100 years (Leavitt & Pickering 1912), due to the crucial role they played in establishing the extragalactic nature of "spiral nebulae" (Hubble 1925) and later in the determination of the Hubble constant, H0. Over the past two decades, observations of Cepheids with the Hubble Space Telescope (HST) have resulted in a significant improvement on the accuracy and precision in the determination of H0. The first generation of Hubble projects aimed at measuring H0 had uncertainties of ∼10% (Freedman et al. 2001; Sandage et al. 2006), while more recent work has reduced that to 5% (Riess et al. 2009). The accuracy and precision in the measurement of the Hubble constant via Cepheid variables and secondary distance indicators has a direct impact on the constraints in the equation of state of dark energy (Komatsu et al. 2011). Thus, improving our understanding of the properties of Cepheid variables remains an important and active area of observational and theoretical research.
M 33 has been the target of numerous surveys for Cepheids and other variable stars, starting with Hubble (1926), due to its relative proximity and its moderate inclination angle. The DIRECT survey (Stanek et al. 1998) carried out the first large-area CCD-based synoptic survey of M 31 and M 33 with the aim of discovering detached eclipsing binaries and Cepheid variables. Results for the central region of M 33 were presented by Macri et al. (2001b) who published calibrated point-spread function (PSF) photometry for 251 Cepheids. Follow-up observations by Mochejska et al. (2001a, 2001b), analyzed via difference-imaging photometry, yielded additional short-period Cepheids. More recently, Hartman et al. (2006) carried out a survey of the entire disk of M 33 and identified ∼2500 Cepheid candidates via difference imaging, although the photometric calibration of this sample is still pending.
Among the many eclipsing binaries discovered by DIRECT in M 33 there are at least two detached systems well suited for distance determination. Bonanos et al. (2006) analyzed one of these systems and derived a distance of 964 ± 54 kpc, equivalent to a distance modulus μ0 = 24.92 ± 0.12 mag. However, their results are somewhat at odds (2.7σ discrepancy) with previous Cepheid-based distance moduli for M33: 24.65 ± 0.12 mag (Macri 2001), 24.62 ± 0.08 mag (Freedman et al. 2001), and 24.53 ± 0.11 mag (Scowcroft et al. 2009). The analysis of additional detached eclipsing binary systems in M33 will hopefully resolve this discrepancy and further reduce the uncertainty in its distance. M33 would then become an important contributor to the "first rung" of the Extragalactic Distance Scale, provided its Cepheid sample is properly calibrated. This was one of the main motivations for our survey. An additional motivation for studying the Cepheids in M 33 was the steep abundance gradient measured across its disk by Zaritsky et al. (1994) and Magrini et al. (2007), which would enable a robust determination of the "metallicity dependence" of the Cepheid distance scale. However, Bresolin et al. (2010) and Bresolin (2011) recently presented evidence for a shallow abundance gradient in the disk of M 33, with a mean value very similar to the Large Magellanic Cloud (LMC).
The relative proximity of M 33 minimizes the impact of unresolved blends and crowding in PSF photometry, which can lead to an overestimation of Cepheid magnitudes and biases in their measured colors. Both effects must be properly taken into account in any study of Cepheid variables, since they can contribute to an underestimation of the host galaxy distance. Extensive work must be performed to properly quantify these effects. In the case of nearby (D ≲ 2 Mpc) galaxies, archival HST images can greatly help with this task, as shown by Mochejska et al. (2000, 2001c) and Bresolin et al. (2005).
This paper presents the discovery and characterization of a sample of 564 Cepheids in M 33 resulting from a ground-based BVI synoptic survey which covers most of the disk of M 33 and spans a period of 7 years. We describe our survey in Section 2 and discuss the details of the photometry and Cepheid light curve fitting in Section 3. Color–magnitude diagrams (CMDs) and Cepheid period–luminosity (PL) relations are presented in Section 4. We address the issue of stellar crowding and our method for characterizing this effect using archival HST images in Section 5. We present our catalog in Section 6 before concluding.
2. OBSERVATIONS OF M 33
This survey combines the data obtained by the DIRECT project with follow-up observations obtained at the Wisconsin–Indiana–Yale–NOAO (WIYN) Observatory. The observations span a minimum of 7 years (up to 10 years for a few fields). In this section, we describe the two data sets in detail.
2.1. Images from the DIRECT Project
As part of an effort to improve the absolute calibration of the Extragalactic Distance Scale, the DIRECT project2 obtained BVI observations of M 33 over ∼200 nights between 1996 September and 1999 November. Most of the survey was carried out at the F. L. Whipple Observatory (FLWO) 1.2 m telescope on Mount Hopkins, Arizona, with additional images obtained at the Michigan–Dartmouth–MIT 1.3 m telescope on Kitt Peak, Arizona. The project observed 11 different fields, each 11
5 × 115 in size, which covered most of the M 33 disk (see Figure 1). Most of the DIRECT observations were obtained in the V band, which provided sensitivity to both early-type detached eclipsing binaries and Cepheid variables. The typical seeing at the FLWO telescope was ∼1
5, and the three different cameras used by the project had pixel scales of ∼032 pixel−1. Exposure times were 1200, 900, and 600 s per image in the B, V, and I bands, respectively. A more detailed description of the DIRECT observations of M 33 can be found in Macri et al. (2001a, 2001b). The latter publication presented the discovery of 251 Cepheids and additional variables in the three central DIRECT fields within M 33.
2.2. Images from WIYN
While the images obtained by the DIRECT project provided an excellent synoptic data set for the discovery of Cepheids and eclipsing binaries, they were less than ideal for a proper characterization of the photometric properties of these variables. The DIRECT project focused on delivering the best possible phase coverage for the variables, and therefore spent very little telescope time observing standard star fields. This made it difficult to obtain a very accurate photometric calibration. Additionally, the relatively poor seeing at the FLWO 1.2 m made the DIRECT light curves more susceptible to blending and crowding biases. To remedy these issues and further extend the temporal baseline offered by the DIRECT project, we collected additional images at the WIYN 3.5 m telescope between 2002 August and 2006 December. We used the Mini-Mosaic imager, which consists of two CCDs of 4096 × 2048 pixel2 each, separated by a small gap. The native pixel scale of the camera is 014 pixel−1, leading to a field of view of 96 × 96. Eighteen separate pointings were required to cover the same area as the DIRECT project. Each pointing was offset in declination by about half the field of view, relative to the preceding pointing. This observing strategy led to 29 different fields of 96 × 48 each (see Figure 1), with slight overlaps in both directions. The images were binned by a factor of two to reduce the readout time, leading to an effective sampling of 028 pixel−1. Exposure times were 90 s in B, 60 or 90 s in V, and 30 or 60 s in I, to match the depth of the DIRECT images. A detailed observation log for the WIYN data is presented in Table 1. The 29 WIYN fields are identified from 0 to 9 and a to s, as labeled in Figure 1.
Figure 1. Digitized Sky Survey image of M 33 showing the location of the DIRECT fields (red) and our follow-up WIYN observations (blue). The image is about 40' × 50' in size. The WIYN fields are also labeled in blue or white. North is up and east is to the left.
Download figure:
Standard image High-resolution imageTable 1. WIYN Observation Log
| Date | MJD | Field | Band | Exp. Time | Seeing |
|---|---|---|---|---|---|
| (days) | (sec) | (arcsec) | |||
| 2002 Aug 12 | 2499.40241 | c | I | 30 | 1.57 |
| 2002 Aug 12 | 2499.40241 | d | I | 30 | 1.55 |
| 2002 Aug 12 | 2499.40552 | d | V | 60 | 2.12 |
| 2002 Aug 12 | 2499.40552 | c | V | 60 | 1.85 |
| 2002 Aug 12 | 2499.40749 | d | B | 90 | 1.65 |
| 2002 Aug 12 | 2499.40749 | c | B | 90 | 2.01 |
| 2002 Aug 12 | 2499.41333 | c | I | 30 | 0.69 |
| 2002 Aug 12 | 2499.41333 | b | I | 30 | 0.60 |
| 2002 Aug 12 | 2499.41571 | b | V | 60 | 0.71 |
| 2002 Aug 12 | 2499.41571 | c | V | 60 | 0.83 |
Only a portion of this table is shown here to demonstrate its form and content. A machine-readable version of the full table is available.
Download table as: DataTypeset image
3. DATA REDUCTION AND ANALYSIS
All images were subject to overscan correction, bias subtraction, and flat fielding using the MSCRED package in IRAF.3 Given the high density of stars in our images, particularly in the central part of M 33, we performed PSF photometry using the DAOPHOT, ALLSTAR, and ALLFRAME (Stetson 1987, 1994) programs. Due to the large number of images in our data set, we used a series of Tcl/Tk-based scripts originally developed by the DIRECT project to run these programs in batch mode. A detailed description of the individual steps that were performed is presented below.
3.1. PSF Photometry
We performed PSF photometry on every individual image as follows. We used the DAOPHOT task FIND to identify point-like sources 5σ above background and obtained magnitudes through a 7 pixel (∼21) radius aperture using PHOT. We used PICK to obtain a list of 200 bright, isolated stars that might be suitable to determine the PSF. Due to the high stellar density of our fields, the list of stars returned by this task contained a large fraction of blended or crowded objects. Given the large number of images we had to process, it was not practical to remove these unsuitable stars by hand as it is often done, nor to rely on an iterative application of the DAOPHOT PSF task. Instead, we developed an IDL-based routine that fit a two-dimensional Gaussian profile to every star in the initial list of PSF stars. We rejected objects whose full width at half-maximum (FWHM) values in either the x- or y-axis were more than 1σ from the median, as well as any objects whose ellipticity was larger than 3σ from the median. This removed elongated sources that could be two or more stars as well as background sources such as galaxies. We also calculated the residuals from the Gaussian fit and eliminated sources with residuals that were 3σ or greater than the median value. This removed sources located in very crowded regions. We kept a minimum of 25 stars in order to create a reliable PSF model, increasing the initial sample returned by PICK if necessary. Once we obtained the desired sample of bright and isolated stars, we ran the PSF task iteratively with the ALLSTAR program as prescribed in the DAOPHOT manual. Typical PSF FWHM values were 07 and 15 for the WIYN and DIRECT images, respectively. Once all model PSFs had been determined, we performed PSF photometry on every star in every frame using ALLSTAR.
We created master images for each of the 29 fields in each of the three bands, derived frame-to-frame transformations, and carried out the final PSF photometry as follows. We selected 4–7 images from the WIYN observations, with the smallest FWHM and low sky values, to make the master image. We used the DAOMATCH and DAOMASTER programs to derive the positional transformations between these images and combined them using the MONTAGE program. We performed PSF photometry as described above to obtain a master star list, which we matched to the star list of every frame using DAOMATCH and DAOMASTER to derive frame-to-frame transformations. Next, we ran ALLFRAME using the master star list and the derived transformations to obtain consistent PSF photometry for all stars in all images of a given field and filter. We will refer to these measurements as "WIYN+DIRECT photometry." We performed artificial star tests on the master frames and derived a 90% completeness limit at typical magnitudes of 23.3, 23.1, and 21.5 mag in the B, V, and I bands, respectively. These completeness limits change slightly from field to field depending on the stellar density.
After we carried out the previous analysis, we found that the DIRECT images have significant distortion in their corners, which in a few cases was beyond the ability of DAOMASTER and ALLFRAME to compensate for. This seriously degraded the quality of light curves in small regions within our survey area. Hence, we created a second set of master images based only on DIRECT data and performed a separate round of ALLFRAME photometry on DIRECT images only. We will refer to these measurements as "DIRECT-only photometry." This allowed us to obtain good light curves at the corners of the DIRECT fields, at the price of brighter limiting magnitudes and higher risk of stellar blends. In order to mitigate these effects, we performed a third round of ALLFRAME photometry on the WIYN images only (hereafter the "WIYN-only photometry"), based on the original WIYN-based master frames described above. We matched the stars between the DIRECT-only and WIYN-only sets and used the light curves from the latter set in our final analysis. This will be discussed in more detail in Section 3.2. Note however that, as shown in Figure 1, some DIRECT fields extend to areas not covered by the WIYN fields. Consequently, the DIRECT-only photometry gave us an additional ∼170 arcmin2 area for the Cepheid search, which can be visualized in Figure 1.
3.2. Variable Search, Period Determination, and Light Curve Fitting
We identified variable stars and extracted their light curves using the TRIAL program (Stetson 1996). This program derived final frame-to-frame photometric offsets using a list of bright, isolated, and constant-magnitude stars, which we selected based on the ALLFRAME output, and calculated a variability index J for each star. We extracted light curves for all sources displaying J>0.75 and searched for periodicity using an IDL implementation of the CLEAN algorithm4 (Högbom 1974; Roberts et al. 1987). We selected up to eight trial periods that were at least 10σ above the noise level of the periodogram. We simultaneously fit the phased BVI light curves using the Cepheid templates of Yoachim et al. (2009) and selected the trial period that returned the lowest χ2 value for further inspection. The B-band templates were kindly provided to us by P. Yoachim (2009, private communication), since they were not part of their original work. The fit returned mean magnitudes and light curve amplitudes for each band. We did not constrain the light curve amplitudes to obey the "canonical" 1.4:1:0.7 BVI ratios, since we planned to use the observed amplitude ratios to identify possible blends as described below. The V and I light curves are usually better fitted by the templates than the B-band data because the latter contain a significantly lower number of epochs. This leads to slightly larger uncertainties in the mean magnitude and amplitude values in B.
We gave preference to the "WIYN+DIRECT" photometry for the variability and period search, except for the regions that were affected by the optical distortion described in Section 3.1 where we used the "DIRECT-only" light curves. There are 153 out of 667 stars for which the variability and period search was based on "DIRECT-only" images. The final magnitudes and pulsation amplitudes we report are based on light curve template fits to the "WIYN-only" photometry, to minimize the effects of blending and crowding. Figure 2 shows a comparison of the light curves for one Cepheid derived from the "WIYN+DIRECT" and "WIYN-only" observations. The former tends to have flatter minima and slightly reduced amplitudes, which are signs of blending in the DIRECT data.
Figure 2. Example of BVI light curves for one of the Cepheids in our sample (J013302.0+303634.3, P = 30.501 days) illustrating the different sampling provided by the WIYN+DIRECT and WIYN-only observations (black and red symbols, respectively). The black and red lines indicate the best-fit template light curves for the combined and WIYN-only photometry, respectively. Photometric uncertainties are only plotted for the WIYN observations to ensure clarity.
Download figure:
Standard image High-resolution imageIn a few cases where the WIYN coverage was too sparse or non-existent, we were forced to use "DIRECT-only" light curves; these exceptions are properly identified in our final catalogs. Thirty-six variables have no WIYN image counterpart, hence have DIRECT-only light curves.
3.3. Astrometric and Photometric Calibration
We based our astrometric and photometric calibrations on the stellar catalog of M 33 published by Massey et al. (2006). We carried out the astrometric calibration using the WCSTools package.5 We used the initial WCS provided by the telescope control system to identify bright stars in common between the master frame of each field and filter and the reference catalog. We constructed a paired list of master frame (x, y) positions and Massey et al. celestial coordinates and used the imwcs task to determine the astrometric solution. We used eight parameters for the astrometric solutions, with the exception of a few outer fields where the lower stellar density did not yield a sufficient number of bright stars in common. In those cases, we carried out a six parameter fit. The typical astrometric residuals were less than 02.
Once we had established a common WCS between our catalog and Massey et al., we expanded our search to select common stars that could be used to derive the photometric calibration of our instrumental magnitudes. We derived photometric zero points and color terms based on stars detected in all three bands in both catalogs, which were brighter than 20.3 mag and displayed no significant variability (J < 0.75). In the case of a few outer fields with lower stellar density, we extended the range to 21 mag to increase the sample size. We carried out the photometric calibration of each of our 29 fields separately, and then compared the derived magnitudes for stars in common between adjacent fields to ensure their consistency. Figure 3 shows a comparison of our final calibrated magnitudes relative to the input Massey et al. catalog for one of the 29 fields located at an intermediate galactocentric radius. Note that this comparison includes stars that are significantly fainter than those used to derive the actual transformations, but we maintained the requirements of three-band photometry, variability index J < 0.75, and photometric uncertainties below 0.05 mag. The median differences are −0.015 ± 0.043, −0.016 ± 0.033, and −0.010 ± 0.039 mag in the B, V, and I bands, respectively. The median differences are slightly larger for more heavily crowded fields and even smaller in the outermost regions of the disk.
Figure 3. Comparison of photometry between our final calibrated catalog and the reference catalog of Massey et al. (2006) for one of the 29 fields in our survey, located at an intermediate galactocentric radius. The top three panels show the magnitude differences in B, V, and I as a function of the reference catalog magnitudes in the respective band. The bottom two panels present the difference in V and I as a function of B − V or V − I color of the reference catalog, respectively. In all panels, the comparison was restricted to stars with three-band photometry, a low variability index (J < 0.75), and photometric errors below 0.05 mag.
Download figure:
Standard image High-resolution image4. COLOR–MAGNITUDE DIAGRAMS AND CEPHEID PERIOD–LUMINOSITY RELATIONS
Figure 4 shows CMDs of ∼793,000 stars detected in the WIYN-only master frames. They are very similar to those presented by Massey et al. (2006), though our photometry is slightly deeper. We can clearly see the main sequence of massive stars along B − V = 0 and V − I = 0 mag, a sequence of foreground Milky Way dwarfs to the red of the M 33 main sequence, and the red giant and asymptotic giant branches of M 33. The Cepheid variables in our "main sample" (see below) are plotted using red symbols and delineate the instability strip.
Figure 4. Color–magnitude diagrams of all stars detected by our survey. Small black points: all stars detected in the WIYN master images. Filled red symbols: Cepheid variables. Green contours show the stellar density in the saturated regions of the diagrams. The median photometric uncertainties in magnitude and color for each bin are indicated by blue error bars on the right of each diagram.
Download figure:
Standard image High-resolution imageFigure 5 shows PL relations for 705 variables that were adequately fit by a Cepheid template light curve (Section 3.2), regardless of their magnitude. The top three panels depict the BVI relations, uncorrected for extinction, while the bottom panel shows the dust-free or "Wesenheit" relation for those objects which had V and I photometry. The Wesenheit magnitude was calculated assuming a standard Cardelli et al. (1989) extinction law (see also Schlegel et al. 1998) as WVI = (2.375 × I) − (1.375 × V).
Figure 5. PL relations for M 33 Cepheids. The B-band, V-band, and I-band PL relations (top three panels) are uncorrected for extinction. The bottom panel is the Wesenheit (i.e., dust-free) PL relation, and based on the extinction law of Cardelli et al. (1989). All PL relations have been corrected for crowding (see Section 5). Full circles: main Cepheid sample. Open circles: abnormally bright variables, likely to be blended Cepheids. Open triangles: abnormally faint variables, likely to be blended Cepheids or population II variables. Crosses: variables with very sparse or extremely noisy light curves. Solid lines: best fit to the main Cepheid sample with periods above 8 days, keeping the slopes fixed to the values derived from the OGLE-II LMC sample (Udalski et al. 1999). Dashed lines: 2σ deviation from the fit.
Download figure:
Standard image High-resolution imageThe relations presented in Figure 5 contain obvious outliers. We visually examined each individual light curve and we inspected the corresponding master image at the location of every variable. We rejected 38 objects which exhibited poorly sampled and/or extremely noisy light curves, which raised doubts about the reliability of the period determination or the nature of the stellar variability. These rejected objects are plotted using "x" symbols. We identified additional outliers in the Wesenheit PL relation, using as fiducial the corresponding OGLE-II LMC PL relation (Udalski et al. 1999). Forty-six objects lying 2.5σ or more below the mean are plotted using open triangles; they are likely to be blended Cepheids (at shorter periods) or population II Cepheids (at longer periods). Fifty-seven variables lying 3σ or more above the mean are plotted using open circles; they are likely to be blended Cepheids. We rejected outliers at different significance above and below the mean to preserve overtone Cepheids in the main sample, which consists of 564 variables plotted using filled symbols.
The top three panels of Figure 5 also show fits to the cleaned BVI main sample, keeping the slopes fixed to the values derived using the OGLE-II LMC Cepheid sample (Udalski et al. 1999). To prevent contamination from overtones, we restricted the fit to Cepheids with P>8 days. The results from the fits are given in Table 2, which are still preliminary values. We defer the determination of a new Cepheid distance to M 33 to a forthcoming publication (A. Pellerin & L. M. Macri 2011, in preparation). That publication will combine our ground-based measurements and HST photometry (J. Chavez et al. 2011, in preparation). Our preliminary distance moduli presented in Table 2 have only been corrected for crowding effects (see Section 5) and are consistent with previous results (Macri 2001; Freedman et al. 2001; Scowcroft et al. 2009).
Table 2. Preliminary PL-Relation Fitting Parameters
| Band | Distance Modulusa | Distance Modulusb | Slopesc |
|---|---|---|---|
| (All Cepheids) | (Cleaned Sample) | ||
| B | 25.39 ± 0.03 | 25.35 ± 0.02 | −2.439 |
| V | 25.32 ± 0.03 | 25.26 ± 0.02 | −2.779 |
| I | 25.05 ± 0.03 | 25.00 ± 0.03 | −2.979 |
| WVI | 24.77 ± 0.04 | 24.76 ± 0.02 | −3.309 |
Notes. The fits were performed for Cepheids with periods longer than 8 days assuming an LMC distance modulus of 18.5 mag. aDistance moduli, in magnitudes, include all Cepheids from Tables 3 to 5. bDistance moduli, in magnitudes, include only the cleaned Cepheid sample as reported in Table 3. cThe slope values are those derived for the LMC by OGLE-II (Udalski et al. 1999).
Download table as: ASCIITypeset image
Table 3 lists the coordinates, magnitudes, pulsation amplitudes, and other information for the main sample. Tables 4–6 contain the different samples of rejected outliers.
Table 3. M 33 Cepheids—Main Sample (Abridged)
| Designation [M33SSS] | R.A. | Decl. | Period | Va | I | B | σV | σI | σB | Amp V | Amp I | Amp B | CV | CI | CB | Data Set |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| J013425.57+310029.2 | 23.60654 | 31.00812 | 2.052 | 22.202 | 21.590 | 22.613 | 0.002 | 0.006 | 0.011 | 0.085 | 0.054 | 0.119 | 0.183 | 0.158 | 0.088 | WD |
| J013319.01+302512.3 | 23.32920 | 30.42008 | 2.096 | 21.963 | 21.484 | 22.438 | 0.007 | 0.037 | 0.033 | 0.103 | 0.100 | 0.183 | 0.173 | 0.154 | 0.080 | WO |
| J013438.46+310155.9 | 23.66026 | 31.03220 | 2.319 | 21.835 | 21.370 | 99.999 | 0.001 | 0.004 | 9.999 | 0.151 | 0.147 | 0.121 | 0.167 | 0.149 | 9.999 | WD |
| J013256.63+301849.8 | 23.23594 | 30.31382 | 2.370 | 22.024 | 21.405 | 22.340 | 0.013 | 0.043 | 0.033 | 0.496 | 0.313 | 0.718 | 0.176 | 0.150 | 0.077 | WO |
| J013258.16+302505.2 | 23.24235 | 30.41810 | 2.387 | 21.830 | 21.191 | 22.230 | 0.014 | 0.030 | 0.043 | 0.186 | 0.148 | 0.238 | 0.167 | 0.142 | 0.072 | WO |
| J013243.21+302653.2 | 23.18006 | 30.44811 | 2.483 | 22.274 | 21.799 | 22.684 | 0.017 | 0.125 | 0.053 | 0.389 | 0.252 | 0.533 | 0.187 | 0.167 | 0.091 | WO |
| J013349.33+310219.8 | 23.45556 | 31.03884 | 2.548 | 21.826 | 21.441 | 22.214 | 0.001 | 0.007 | 0.012 | 0.086 | 0.054 | 0.122 | 0.167 | 0.152 | 0.072 | DO |
| J013330.16+303417.7 | 23.37568 | 30.57158 | 2.560 | 21.662 | 21.200 | 21.974 | 0.009 | 0.040 | 0.026 | 0.152 | 0.169 | 0.185 | 0.160 | 0.142 | 0.063 | WO |
| J013454.32+305929.4 | 23.72635 | 30.99150 | 2.640 | 22.448 | 21.880 | 22.966 | 0.025 | 0.023 | 0.100 | 0.439 | 0.311 | 0.684 | 0.195 | 0.170 | 0.103 | WO |
| J013324.20+302248.9 | 23.35084 | 30.38026 | 2.686 | 21.779 | 21.013 | 22.375 | 0.008 | 0.023 | 0.028 | 0.175 | 0.081 | 0.220 | 0.165 | 0.135 | 0.078 | WO |
| J013332.36+302819.8 | 23.38483 | 30.47216 | 2.689 | 22.004 | 21.152 | 22.472 | 0.006 | 0.044 | 0.042 | 0.333 | 0.081 | 0.339 | 0.175 | 0.140 | 0.082 | WO |
| J013447.66+305805.7 | 23.69860 | 30.96824 | 2.748 | 21.778 | 21.127 | 22.123 | 0.006 | 0.026 | 0.016 | 0.245 | 0.190 | 0.346 | 0.165 | 0.139 | 0.068 | WO |
| J013221.78+302059.5 | 23.09077 | 30.34985 | 2.765 | 22.610 | 21.733 | 23.139 | 0.029 | 0.035 | 0.096 | 0.404 | 0.262 | 0.641 | 0.202 | 0.164 | 0.111 | WO |
| J013421.36+305806.2 | 23.58901 | 30.96839 | 2.769 | 22.377 | 21.785 | 22.860 | 0.013 | 0.055 | 0.044 | 0.450 | 0.322 | 0.633 | 0.191 | 0.166 | 0.098 | WO |
| J013335.18+305425.0 | 23.39660 | 30.90694 | 2.773 | 21.750 | 21.188 | 22.181 | 0.016 | 0.042 | 0.064 | 0.115 | 0.076 | 0.161 | 0.164 | 0.142 | 0.070 | WO |
Notes. aThe mean magnitudes are in the Vega system and corrected for crowding, with crowding values reported in the three columns CV, CI, and CB.
Only a portion of this table is shown here to demonstrate its form and content. A machine-readable version of the full table is available.
Download table as: DataTypeset image
Table 4. Abnormally Faint/Likely Type II Cepheids (Abridged)
| Designation [M33SSS] | R.A. | Decl. | Period | Va | I | B | σV | σI | σB | Amp V | Amp I | Amp B | CV | CI | CB | Data Set |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| J013353.74+305848.6 | 23.47390 | 30.98017 | 2.472 | 21.607 | 21.794 | 99.999 | 0.021 | 0.040 | 9.999 | 0.490 | 0.302 | 9.999 | 0.158 | 0.167 | 9.999 | WO |
| J013229.99+303141.9 | 23.12494 | 30.52830 | 2.477 | 22.454 | 99.999 | 22.939 | 0.008 | 9.999 | 0.036 | 0.441 | 9.999 | 0.531 | 0.195 | 9.999 | 0.102 | WO |
| J013328.84+301942.6 | 23.37015 | 30.32849 | 2.508 | 22.740 | 99.999 | 23.104 | 0.031 | 9.999 | 0.102 | 0.500 | 9.999 | 0.720 | 0.208 | 9.999 | 0.109 | WO |
| J013232.55+302534.5 | 23.13562 | 30.42626 | 2.649 | 22.646 | 99.999 | 23.238 | 0.020 | 9.999 | 0.088 | 0.510 | 9.999 | 0.729 | 0.204 | 9.999 | 0.116 | WO |
| J013244.23+303936.4 | 23.18431 | 30.66011 | 3.179 | 22.134 | 99.999 | 22.856 | 0.007 | 9.999 | 0.018 | 0.461 | 9.999 | 0.688 | 0.180 | 9.999 | 0.098 | WO |
| J013417.35+304505.4 | 23.57228 | 30.75151 | 3.655 | 22.254 | 21.956 | 22.619 | 0.001 | 0.011 | 0.008 | 0.414 | 0.055 | 0.625 | 0.186 | 0.174 | 0.088 | WD |
| J013308.61+302503.5 | 23.28586 | 30.41764 | 4.085 | 22.309 | 21.693 | 23.234 | 0.011 | 0.042 | 0.026 | 0.402 | 0.271 | 0.573 | 0.188 | 0.162 | 0.115 | WO |
| J013452.65+305836.7 | 23.71936 | 30.97685 | 4.290 | 22.103 | 21.664 | 22.738 | 0.018 | 0.108 | 0.074 | 0.514 | 0.323 | 0.760 | 0.179 | 0.161 | 0.093 | WO |
| J013353.38+304743.1 | 23.47240 | 30.79531 | 4.312 | 21.325 | 99.999 | 21.702 | 0.003 | 9.999 | 0.007 | 0.239 | 9.999 | 0.329 | 0.146 | 9.999 | 0.053 | WO |
| J013326.22+303219.1 | 23.35926 | 30.53865 | 4.391 | 21.714 | 99.999 | 22.393 | 0.000 | 9.999 | 0.004 | 0.223 | 9.999 | 0.283 | 0.162 | 9.999 | 0.079 | WD |
| J013330.48+304417.6 | 23.37698 | 30.73823 | 4.474 | 22.043 | 21.491 | 22.713 | 0.019 | 0.072 | 0.051 | 0.368 | 0.247 | 0.495 | 0.176 | 0.154 | 0.092 | WO |
| J013355.92+304720.2 | 23.48300 | 30.78893 | 4.541 | 22.146 | 21.554 | 23.014 | 0.030 | 0.041 | 0.111 | 0.497 | 0.316 | 0.720 | 0.181 | 0.156 | 0.105 | WO |
| J013320.35+304539.3 | 23.33478 | 30.76092 | 4.550 | 22.440 | 21.878 | 23.064 | 0.021 | 0.060 | 0.080 | 0.311 | 0.172 | 0.223 | 0.194 | 0.170 | 0.107 | WO |
| J013415.34+305523.5 | 23.56392 | 30.92320 | 4.617 | 22.132 | 21.735 | 22.936 | 0.029 | 0.071 | 0.096 | 0.432 | 0.270 | 0.690 | 0.180 | 0.164 | 0.102 | WO |
| J013301.76+302607.4 | 23.25732 | 30.43540 | 5.317 | 21.725 | 21.169 | 22.189 | 0.025 | 0.039 | 0.089 | 0.319 | 0.198 | 0.483 | 0.163 | 0.141 | 0.071 | WD |
Notes. aThe mean magnitudes are in the Vega system and corrected for crowding, with crowding values reported in the three columns CV, CI, and CB.
Only a portion of this table is shown here to demonstrate its form and content. A machine-readable version of the full table is available.
Download table as: DataTypeset image
5. CROWDING CORRECTIONS
Despite the fact that our WIYN images provide a very good spatial resolution of ∼3.25 pc (075 FWHM at a distance of 890 kpc), we cannot neglect the effects of crowding and blending in our photometry. Crowding can be defined as the partial overlap of two or more stellar profiles which can still be spatially resolved into individual PSFs. Blending is the more severe case where the stars are separated by less than 0.375 FWHM (following the criterion used in ALLSTAR) and cannot be spatially resolved, resulting in their detection as a single PSF. The crowding level is high in M 33, especially in the central region and in the young stellar population regions where Cepheids are often found. Furthermore, it is impossible to eliminate random projection effects within the galactic disk or binary systems. For these reasons, we need to quantitatively characterize the effects of both blending and crowding, ideally in a separate way. A detailed study of these effects in M 31 and M 33 was carried out by Mochejska et al. (2000, 2001c) and we followed their approach as discussed below.
In this publication, we address crowding corrections, which can be derived via artificial star simulations and applied to our entire Cepheid sample. Corrections for blends cannot be applied statistically, since they require individual imaging of every variable at higher angular resolution. Fortunately, the disk of M 33 has been extensively covered by the HST, which has a diffraction-limited spatial resolution ∼10× better than our WIYN images. A companion paper (J. Chavez et al. 2011, in preparation) presents PSF photometry of all available HST images that overlap with our Cepheid sample. In the absence of higher-resolution images, blends can sometimes be identified via abnormal light curve amplitude ratios or colors. We will explore the effectiveness of these cuts in a forthcoming publication (A. Pellerin & L. M. Macri 2011, in preparation).
We derived the crowding corrections using images of M 33 obtained with the Advanced Camera for Surveys (ACS) on board HST. We used the Multimission Archive at STScI6 (MAST) to select observations obtained in the F475W, F606W, and F814W filters (equivalent to B, V, and I, respectively) which had significant overlap with our fields. The raw images were pipeline-processed by MAST. We further processed the reduced images using the PyRAF/STSDAS7 MultiDrizzle task (Fruchter et al. 2009) to correct for optical distortion, remove cosmic rays, and combine multiple images of the same field. We carried out PSF photometry using DAOPHOT and ALLSTAR as previously described in Section 3.1. We verified that the photometry was consistent with the work of Williams et al. (2009). More details on the HST photometry will be presented by J. Chavez et al. (2011, in preparation).
Since we wanted to characterize the effect of crowding in our WIYN frames in a manner that was decoupled from the effects of blending, we started by identifying stars in the HST photometry which had no detectable neighbors within a radius of 038. This is equivalent to the half-width at half-maximum of the PSF in the WIYN master frames. Next, we simulated a WIYN master frame using the DAOPHOT task ADDSTAR, placing all stars detected in the HST photometry at suitably rescaled CCD coordinates and with the proper instrumental magnitudes, using the ground-based PSF derived in 3.1. Lastly, we added a sky background with the same mean and standard deviation as the actual WIYN master frame we were simulating. We visually inspected the simulated images to ensure they were comparable to the actual master frames. Given the smaller field of view of ACS relative to WIYN, there were relatively few isolated stars at the brightest magnitude levels. In order to obtain better statistics, we randomly removed faint isolated stars in the HST photometry and replaced them with bright ones. We generated several simulated images with these artificial bright stars to allow statistically meaningful results without increasing the crowding level of the images.
Once the simulated images were generated, we carried out PSF photometry in the same manner as we had done on the actual master images. We then compared the recovered and input magnitudes of the unblended stars identified in the first step. The results of this exercise are presented in Figure 6, separately for each band. We found that the crowding bias increases for fainter stars, reaching ∼0.2 mag at the faint end. We fit a second-order polynomial to the results and corrected our photometry accordingly. We conducted these simulations and measurements using all the available ACS fields which overlapped with our Cepheids, to probe regions of different surface brightness, ranging from 20 to 23 mag arcsec−2. We found no statistically significant trend as a function of surface brightness.
Figure 6. Crowding bias as a function of magnitude, derived using stars known to be free of blends through higher-resolution HST imaging. The filled symbols indicate the median value of each magnitude bin, while the triangular symbols indicate the 25% and 75% quartile ranges. The solid lines indicate the second-order polynomial fit that was used to correct all our photometry for crowding.
Download figure:
Standard image High-resolution image6. CATALOG OF M 33 CEPHEIDS
Tables 3–5 present the coordinates and photometric properties of Cepheid variables discovered in this survey, separated into three categories: main sample, abnormally faint/likely population II variables, and abnormally bright/likely blended Cepheids. Only a few lines of each catalog are presented in the printed version; the complete catalogs are available in the electronic edition of the journal and in the project Web site8. Column 1 contains the designation of each variable following the IAU standard; we have selected the acronym "M33SSS" for our project. Celestial coordinates (J2000.0 R.A. and decl.) are provided in decimal degrees in Columns 2 and 3. The period of variability (in days) is reported in Column 4. Given the exceptionally long time coverage of our survey, we were able to derive periods with very high precision. The mean magnitudes in the V, I, and B bands already corrected for crowding bias are listed in Columns 5, 6, and 7, respectively. The 1σ uncertainties on the mean magnitudes are provided in Columns 8, 9, and 10 for the V, I, and B bands, respectively. The best-fit light curve amplitude in V, I, and B are given in Columns 11, 12, and 13, respectively. The crowding corrections already applied to the mean magnitudes are reported in Columns 14, 15, and 16 for V, I and B, respectively. The last column indicates the data set on which the photometric values are based, i.e., "WIYN+DIRECT" (WD), "WIYN-only" (WO), or "DIRECT-only" (DO) photometry, as discussed in Section 3.1. Since the variability search was conducted in the V band (Section 3.2), where the largest number of images was obtained, all variables in this catalog have V-band properties but some of the fainter variables may lack B or I data.
Table 5. Abnormally Bright/Likely Blended Cepheids (Abridged)
| Designation [M33SSS] | R.A. | Decl. | Period | Va | I | B | σV | σI | σB | Amp V | Amp I | Amp B | CV | CI | CB | Data Set |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| J013325.08+305807.6 | 23.35448 | 30.96879 | 2.398 | 21.795 | 21.040 | 22.205 | 0.001 | 0.005 | 0.009 | 0.158 | 0.063 | 0.252 | 0.166 | 0.136 | 0.071 | DO |
| J013319.28+304831.5 | 23.33033 | 30.80876 | 2.485 | 21.563 | 20.710 | 22.025 | 0.007 | 0.025 | 0.031 | 0.085 | 0.055 | 0.127 | 0.156 | 0.123 | 0.065 | WO |
| J013444.34+305217.2 | 23.68477 | 30.87145 | 2.508 | 21.239 | 20.501 | 21.953 | 0.044 | 0.025 | 0.042 | 0.475 | 0.417 | 0.315 | 0.143 | 0.116 | 0.062 | WO |
| J013354.21+305859.4 | 23.47588 | 30.98318 | 2.548 | 21.687 | 20.936 | 22.230 | 0.013 | 0.027 | 0.052 | 0.129 | 0.089 | 0.148 | 0.161 | 0.132 | 0.072 | WO |
| J013446.14+304928.2 | 23.69226 | 30.82451 | 2.749 | 22.034 | 21.089 | 22.511 | 0.026 | 0.032 | 0.075 | 0.343 | 0.216 | 0.493 | 0.176 | 0.138 | 0.083 | WD |
| J013440.10+305905.7 | 23.66709 | 30.98492 | 3.149 | 22.216 | 21.006 | 22.929 | 0.009 | 0.022 | 0.015 | 0.462 | 0.179 | 0.899 | 0.184 | 0.135 | 0.101 | WO |
| J013243.67+302610.2 | 23.18197 | 30.43618 | 3.225 | 22.365 | 21.160 | 23.137 | 0.001 | 0.024 | 0.009 | 0.372 | 0.177 | 0.476 | 0.191 | 0.141 | 0.111 | WD |
| J013310.12+303355.3 | 23.29215 | 30.56535 | 4.027 | 21.361 | 20.317 | 22.164 | 0.018 | 0.021 | 0.084 | 0.343 | 0.205 | 0.673 | 0.148 | 0.109 | 0.070 | WO |
| J013451.98+304318.3 | 23.71660 | 30.72175 | 4.144 | 21.502 | 20.475 | 22.245 | 0.010 | 0.019 | 0.029 | 0.104 | 0.119 | 0.193 | 0.153 | 0.115 | 0.073 | WO |
| J013339.32+304725.2 | 23.41382 | 30.79033 | 4.228 | 22.090 | 20.709 | 22.789 | 0.021 | 0.024 | 0.079 | 0.346 | 0.217 | 0.517 | 0.178 | 0.123 | 0.095 | WO |
| J013351.94+304658.8 | 23.46640 | 30.78299 | 4.384 | 21.393 | 19.922 | 22.075 | 0.008 | 0.012 | 0.018 | 0.487 | 0.322 | 0.738 | 0.149 | 0.096 | 0.066 | WO |
| J013306.29+303050.3 | 23.27619 | 30.51396 | 4.689 | 21.428 | 20.444 | 21.957 | 0.000 | 0.002 | 0.005 | 0.131 | 0.053 | 0.202 | 0.150 | 0.114 | 0.062 | DO |
| J013406.94+304258.0 | 23.52893 | 30.71611 | 4.711 | 21.496 | 19.662 | 22.166 | 0.009 | 0.012 | 0.031 | 0.271 | 0.136 | 0.296 | 0.153 | 0.088 | 0.070 | WO |
| J013359.39+304213.8 | 23.49744 | 30.70384 | 4.782 | 21.521 | 20.277 | 22.469 | 0.005 | 0.009 | 0.037 | 0.496 | 0.303 | 0.749 | 0.154 | 0.108 | 0.082 | WO |
| J013400.30+304808.0 | 23.50127 | 30.80223 | 5.101 | 21.592 | 20.472 | 22.371 | 0.003 | 0.006 | 0.018 | 0.429 | 0.268 | 0.644 | 0.157 | 0.115 | 0.078 | WO |
Notes. aThe mean magnitudes are in the Vega system and corrected for crowding, with crowding values reported in the three columns CV, CI, and CB.
Only a portion of this table is shown here to demonstrate its form and content. A machine-readable version of the full table is available.
Download table as: DataTypeset image
Table 6. Rejected Variables
| Designation [M33SSS] | R.A. | Decl. | Perioda | Va,b | Ia,b | Ba | σV | σI | σB | Amp V | Amp I | Amp B | CV | CI | CB | Data Set |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| J013308.82+304608.7 | 23.28673 | 30.76907 | 2.310 | 20.964 | 19.390 | 22.448 | 0.008 | 0.006 | 0.016 | 0.086 | 0.067 | 0.344 | 0.132 | 0.079 | 0.081 | WD |
| J013300.57+304255.9 | 23.25239 | 30.71554 | 2.563 | 21.466 | 20.394 | 99.999 | 0.060 | 0.025 | 9.999 | 0.210 | 0.079 | 9.999 | 0.152 | 0.112 | 9.999 | WO |
| J013426.55+304334.6 | 23.61061 | 30.72628 | 3.105 | 19.946 | 20.250 | 19.729 | 0.007 | 0.016 | 0.059 | 0.118 | 0.076 | 0.167 | 0.096 | 0.107 | 0.005 | WO |
| J013412.00+304020.0 | 23.55002 | 30.67222 | 3.509 | 21.411 | 20.080 | 21.641 | 0.012 | 0.019 | 0.033 | 0.148 | 0.067 | 0.238 | 0.150 | 0.101 | 0.051 | WO |
| J013413.38+303618.0 | 23.55575 | 30.60501 | 3.738 | 21.369 | 19.998 | 22.108 | 0.015 | 0.010 | 0.043 | 0.131 | 0.082 | 0.239 | 0.148 | 0.098 | 0.068 | WO |
| J013351.26+304504.5 | 23.46360 | 30.75126 | 3.824 | 21.326 | 20.250 | 21.920 | 0.007 | 0.007 | 0.023 | 0.144 | 0.126 | 0.221 | 0.146 | 0.107 | 0.061 | WO |
| J013354.37+310101.3 | 23.47655 | 31.01702 | 3.918 | 20.648 | 20.805 | 99.999 | 0.011 | 0.022 | 9.999 | 0.099 | 0.067 | 9.999 | 0.120 | 0.127 | 9.999 | WO |
| J013251.03+303535.1 | 23.21263 | 30.59309 | 4.173 | 19.413 | 19.418 | 19.232 | 0.007 | 0.020 | 0.020 | 0.226 | 0.211 | 0.207 | 0.079 | 0.080 | 0.000 | WO |
| J013401.43+304631.8 | 23.50596 | 30.77551 | 4.537 | 20.122 | 19.484 | 19.980 | 0.009 | 0.009 | 0.008 | 0.111 | 0.071 | 0.162 | 0.102 | 0.082 | 0.009 | WO |
| J013336.91+303020.1 | 23.40378 | 30.50558 | 4.564 | 19.348 | 19.177 | 19.250 | 0.005 | 0.006 | 0.013 | 0.121 | 0.058 | 0.110 | 0.077 | 0.073 | 0.000 | WO |
| J013223.81+303023.9 | 23.09922 | 30.50664 | 4.880 | 23.389 | 21.845 | 24.615 | 0.034 | 0.106 | 0.226 | 0.481 | 0.289 | 0.733 | 0.240 | 0.169 | 0.189 | WO |
| J013303.85+302120.3 | 23.26603 | 30.35565 | 8.050 | 20.531 | 20.359 | 99.999 | 0.001 | 0.003 | 9.999 | 0.147 | 0.164 | 9.999 | 0.116 | 0.111 | 9.999 | WD |
| J013228.14+303414.0 | 23.11724 | 30.57056 | 8.183 | 20.880 | 21.577 | 99.999 | 0.018 | 0.040 | 9.999 | 0.288 | 0.179 | 9.999 | 0.129 | 0.157 | 9.999 | WD |
| J013327.04+304209.7 | 23.36266 | 30.70270 | 8.217 | 21.199 | 99.999 | 23.032 | 0.001 | 9.999 | 0.012 | 0.211 | 9.999 | 0.111 | 0.141 | 9.999 | 0.106 | WD |
| J013319.91+302018.6 | 23.33297 | 30.33851 | 11.325 | 21.516 | 20.801 | 21.885 | 0.008 | 0.018 | 0.031 | 0.202 | 0.074 | 0.284 | 0.154 | 0.127 | 0.060 | WO |
| J013427.14+303452.0 | 23.61309 | 30.58112 | 12.080 | 22.051 | 20.864 | 99.999 | 0.035 | 0.064 | 9.999 | 0.573 | 0.185 | 9.999 | 0.177 | 0.129 | 9.999 | WO |
| J013425.66+303538.5 | 23.60692 | 30.59404 | 12.168 | 21.412 | 20.520 | 99.999 | 0.061 | 0.037 | 9.999 | 0.402 | 0.219 | 9.999 | 0.150 | 0.116 | 9.999 | WO |
| J013424.59+303448.1 | 23.60246 | 30.58002 | 12.369 | 22.990 | 99.999 | 24.017 | 0.015 | 9.999 | 0.009 | 0.261 | 9.999 | 0.531 | 0.220 | 9.999 | 0.155 | WD |
| J013257.31+304317.6 | 23.23878 | 30.72156 | 12.701 | 23.472 | 99.999 | 24.304 | 0.106 | 9.999 | 0.134 | 0.637 | 9.999 | 0.574 | 0.245 | 9.999 | 0.171 | WO |
| J013218.18+304039.1 | 23.07577 | 30.67754 | 15.420 | 22.595 | 21.675 | 23.052 | 0.002 | 0.008 | 0.011 | 0.434 | 0.195 | 0.671 | 0.201 | 0.162 | 0.107 | DO |
| J013433.08+303524.2 | 23.63785 | 30.59006 | 16.413 | 22.793 | 20.656 | 24.424 | 0.026 | 0.016 | 0.045 | 0.651 | 0.564 | 0.665 | 0.211 | 0.121 | 0.178 | WD |
| J013319.46+304546.7 | 23.33110 | 30.76296 | 20.810 | 22.629 | 20.512 | 24.343 | 0.035 | 0.022 | 0.155 | 0.532 | 0.307 | 0.811 | 0.203 | 0.116 | 0.173 | WO |
| J013420.25+305049.7 | 23.58438 | 30.84715 | 21.240 | 21.210 | 19.517 | 22.300 | 0.003 | 0.007 | 0.020 | 0.416 | 0.403 | 0.696 | 0.142 | 0.083 | 0.075 | WD |
| J013306.75+304938.4 | 23.27814 | 30.82732 | 21.886 | 22.146 | 21.381 | 22.733 | 0.188 | 0.109 | 0.283 | 0.723 | 0.244 | 0.816 | 0.181 | 0.149 | 0.093 | WO |
| J013420.18+305132.1 | 23.58409 | 30.85891 | 30.044 | 22.231 | 20.134 | 99.999 | 0.031 | 0.019 | 9.999 | 0.839 | 0.406 | 9.999 | 0.185 | 0.103 | 9.999 | WO |
| J013454.42+305122.9 | 23.72677 | 30.85635 | 34.925 | 21.494 | 19.066 | 99.999 | 0.016 | 0.010 | 9.999 | 0.444 | 0.295 | 9.999 | 0.153 | 0.070 | 9.999 | WO |
| J013454.89+305421.5 | 23.72869 | 30.90598 | 35.400 | 22.011 | 20.421 | 99.999 | 0.004 | 0.003 | 9.999 | 0.557 | 0.369 | 0.545 | 0.175 | 0.113 | 9.999 | WD |
| J013251.87+305003.1 | 23.21611 | 30.83418 | 35.666 | 22.306 | 21.449 | 22.891 | 0.212 | 0.083 | 0.264 | 0.605 | 0.323 | 0.755 | 0.188 | 0.152 | 0.100 | WO |
| J013321.54+304445.6 | 23.33974 | 30.74599 | 38.500 | 22.081 | 20.172 | 23.287 | 0.035 | 0.011 | 0.078 | 0.689 | 0.495 | 0.317 | 0.178 | 0.104 | 0.118 | WO |
| J013435.99+304656.4 | 23.64997 | 30.78234 | 39.351 | 21.683 | 20.673 | 99.999 | 0.044 | 0.010 | 9.999 | 0.462 | 0.328 | 9.999 | 0.161 | 0.122 | 9.999 | WO |
| J013308.88+302402.4 | 23.28699 | 30.40068 | 49.303 | 20.844 | 20.282 | 21.290 | 0.012 | 0.022 | 0.045 | 0.453 | 0.408 | 0.591 | 0.127 | 0.108 | 0.040 | WO |
| J013255.21+304802.6 | 23.23005 | 30.80072 | 50.915 | 21.445 | 20.477 | 99.999 | 0.079 | 0.052 | 9.999 | 0.522 | 0.395 | 9.999 | 0.151 | 0.115 | 9.999 | WO |
| J013251.24+304939.8 | 23.21350 | 30.82772 | 56.467 | 21.272 | 20.124 | 99.999 | 0.020 | 0.011 | 9.999 | 0.417 | 0.232 | 9.999 | 0.144 | 0.103 | 9.999 | WD |
| J013322.15+302114.2 | 23.34230 | 30.35395 | 87.691 | 20.816 | 19.951 | 21.415 | 0.000 | 0.002 | 0.004 | 0.229 | 0.163 | 0.388 | 0.126 | 0.097 | 0.044 | DO |
| J013412.45+303350.2 | 23.55186 | 30.56394 | 113.000 | 20.608 | 20.353 | 99.999 | 0.032 | 0.024 | 9.999 | 0.474 | 0.310 | 9.999 | 0.119 | 0.111 | 9.999 | WO |
Notes. aPeriods and magnitude values are subject to changes, since their light curves have not been well defined. bAll magnitudes are in the Vega system and corrected for crowding, with crowding values reported in the three columns CV, CI, and CB.
Download table as: ASCIITypeset image
Examples of Cepheid light curves covering the entire range of periods are plotted in Figure 7, along with their best-fit templates. All light curve data are available to the community at the Web site previously mentioned, or by contacting the authors.
Figure 7. Examples of light curves of M 33 Cepheids. The three B-band (blue), V-band (black), and I-band (red) light curves are shown. The points represent the photometric observations, while the lines are the best-fit templates from Yoachim et al. (2009). The periods, in days, are indicated within each panel. The Cepheids shown here are J013314.77+303907.2, J013305.58+303825.2, J013324.63+303517.8, J013252.21+303715.1, J013254.34+303805.4, and J013253.81+303525.3, in order of increasing period.
Download figure:
Standard image High-resolution image
We compared our Cepheid catalog with the variable-star catalog of Hartman et al. (2006). Five hundred forty-four of our variables were located within 5'' of a variable star in that catalog, while 123 had no entries. We also compared the periods found for Cepheids in common between both catalogs, using periods kindly provided by J. Marquette et al. (2010, private communication). The vast majority of the periods were in excellent agreement, with a mean difference of 0.05%.
7. CONCLUSION
We have presented the details and first results of a long-term synoptic BVI survey of the stellar content of M 33. We have carried out time-series PSF photometry of ∼793, 000 stars with V < 25 mag and have identified 667 variables whose light curves are consistent with Cepheid pulsation. After rejecting abnormally faint and bright objects, we obtain a main sample of 564 Cepheid variables, which represents an increase of a factor of two relative to the DIRECT project sample of Macri et al. (2001b). We present catalogs which include the periods, mean magnitudes, and light curve amplitudes for each object. Light curve data will be made available online.
We carried out detailed simulations based on archival HST images to quantify biases in our photometry due to crowding effects. A companion paper (J. Chavez et al. 2011, in preparation) will quantify the level of blending for a sub-sample of our Cepheids, using HST data. A detailed analysis of the effects of blending and crowding on the derivation of a Cepheid distance to M 33 will be presented by A. Pellerin & L. M. Macri (2011, in preparation).
L.M.M. acknowledges initial support for this project by NASA through Hubble Fellowship grant HST-HF-01153 from the Space Telescope Science Institute and by the National Science Foundation through a Goldberg Fellowship from the National Optical Astronomy Observatory. We acknowledge financial support from a Texas A&M University faculty startup fund. We thank P. Yoachim for generating B-band Cepheid light curve templates. This research has made use of NASA's Astrophysics Data System.
Footnotes
- 2
- 3
IRAF is distributed by the National Optical Astronomy Observatory, which is operated by the Association of Universities for Research in Astronomy (AURA) under cooperative agreement with the National Science Foundation.
- 4
- 5
- 6
- 7
PyRAF and STSDAS are products of the Space Telescope Science Institute, which is operated by AURA for NASA.
- 8