NGC 5272

NGC 5272, also known as M3, is a globular cluster located in the constellation Canes Venatici.

NGC 5272
Observation data: 2000.0 epoch
Constellation	Canes Venatici
Right ascension	13h42.2m
Declination	+28o23'
Type	Globular cluster
Apparent dimensions	16.2
Apparent magnitude	6.4
Other designations	M 3

Scientific research

• We performed a detailed study of the pulsational and evolutionary characteristics of 133 RR Lyrae variables in M3, selected among those with the best quality light curves from the CC01 data set. The availability of additional data sets (Car98 and Kal98) at different epochs allowed us to study in good detail the characteristics of the Blazhko stars. Mean magnitudes and colors, along with periods, light-curve amplitudes, and rise times, have been used to discuss the pulsational properties of these stars. A critical discussion of the temperature determination process (i.e., temperature indicators and calibrations) has been presented, and the physical parameters and evolutionary characteristics of these stars have been estimated. The unusual richness of RR Lyrae stars in M3 and the excellent quality of the available data have allowed us to identify a good number of stars in a more evolved stage of evolution off the ZAHB and study their characteristics. Finally, we performed a Fourier analysis of the V light curves and estimated the pros and cons of this technique when applied to the study of RR Lyrae properties. Our main conclusions are the following:

1. The basic characteristics of the CMD already discussed by CC01 are here reconfirmed, namely,
   1. The blue and red edges of the instability strip are located at B - V = 0.18 and 0.42, respectively; the RRc and RRab stars overlap in color in the interval ∼0.24 to 0.30.
   2. The 〈V〉 distribution is bimodal, with a main peak around &angl0;V&angr0; = 15.64 and a secondary peak around &angl0;V&angr0; = 15.52. There is no significant evidence of four populations, as claimed by Jurcsik et al. (2003). The intrinsic magnitude thickness of the HB within the instability strip is ≤0.20 mag if we consider only the main (fainter) component, or ∼0.30 mag if we include also the brighter one.
2. At least one-third of the RR Lyrae stars in M3 are affected by Blazhko modulation; in the studied sample they all belong to the RRab group. More could be hidden in the sample that we have not taken into account in the present analysis because of a large scatter in the light curves. The presence of unidentified Blazhko stars causes a scatter in the relations among various observable parameters that may be large enough to hide the presence of subgroups with different characteristics. The properties of Blazhko stars at the Blazhko phase corresponding to the largest light curve amplitude are generally more similar to the characteristics of regular RRab stars than at smaller amplitude phases. The average 〈V〉 magnitude does not vary significantly with Blazhko phase. The 〈V〉 magnitude distribution of the Blazhko stars is the same as that of the regular RRab stars, including the bimodal shape. The average &angl0;B - V&angr0; color distribution is also similar to that of the RRab stars, but it is truncated at a bluer color; i.e., there are no Blazhko stars redder than &angl0;B - V&angr0; ∼ 0.39.
3. In the period-amplitude diagram both RRc and RRab stars are located on well-defined sequences that are more accurately represented by quadratic rather than linear relations (especially the sequence of the RRab stars), in agreement with theoretical models. There is clear evidence of nearly parallel sequences for both RRc and RRab stars, shifted toward longer periods and populated by systematically brighter stars than the respective main stellar groups. From our sample of 133 RR Lyrae stars we have identified 19 such objects (9 RRab, five Blazhko, and five RRc) that are all consistent with a more advanced stage of evolution off the ZAHB. Their distributions are similar to the mean distributions of OoII RRc and RRab variables. The dependence of the period-AV relation on Oosterhoff type and/or evolutionary status rather than metallicity supports the conclusion that the Oosterhoff dichotomy is due to evolution. The numbers of RR Lyrae stars we have found in M3 near the ZAHB and evolved off the ZAHB are consistent with evolutionary lifetimes according to well-established theoretical considerations. One of the three shortest-period and lowest-amplitude RRc stars is likely to be a second-overtone pulsator.
4. After a critical discussion of the most reliable mean color to be taken as an indicator of the equivalent static color for an RR Lyrae star, we decided to use the formulation (B - V)S = &angl0;B&angr0;int - &angl0;V&angr0;int plus amplitude-related corrections based on theoretical models (Bono et al. 1995), which are also quite consistent with empirical estimates (Sandage 1990). From these colors and using a few independent methods, we estimated a mean reddening of E(B - V) = 0.01 ± 0.01 for M3. A comparative evaluation of various temperature scales led us to identify two temperature scales that meet both theoretical (pulsational and evolutionary) requirements and observational evidence on mass and luminosity for the RR Lyrae stars using B - V colors (in absence of V - K colors). These scales are from M98 (theoretical calibration) and SF. They differ by ∼150 K (M98 being cooler) and yield on average pair values of mass(M⊙)/MV(mag) about 0.74/0.59 and 0.69/0.54, respectively. The temperature scale by CSJ is very similar to SF's and is independent of color, but it is defined only for RRab stars. Considering that the M98 calibrations based on the V - K colors are supposed to be more reliable and are both ∼100 K hotter than the corresponding calibrations based on B - V colors, we have adopted the hotter temperature scale by SF for our analysis (corresponding to a distance modulus of 15.07 for M3). However, our considerations would also hold with the cooler B - V-based M98 scale and a distance modulus of 15.02, within the errors. By using the SF temperature scale and the (B - V)S colors we have derived the stellar physical parameters (temperature, luminosity, mass, and gravity) for our stars and compared them with the most recent stellar evolution and pulsation models. The agreement is good, confirming that the adopted calibration is reliable and accurate, and it yields fully consistent results with the theoretical framework within the respective errors. The use of the CSJ temperature scale yields equally good or better (less-dispersed) results, but for the RRab variables only.
5. We applied the Fourier transform technique to our variables. The main aim was to exploit our excellent data set and investigate the reliability of this type of analysis. First, we derived the Dm parameter, defined by Jurcsik & Kovàcs (1996) as a quality indicator of the regularity of the light curve shape. Only for Dm < 3 the physical parameters derived from Fourier coefficients are considered "reliable," according to Kovàcs and collaborators' prescriptions. We have adopted Dm < 5 to increase the statistics with no significant loss of accuracy. We found that Dm is effectively unable to distinguish between Blazhko and non-Blazhko stars unless set to an unpractically low value (Dm ≤ 2). Even among Blazhko stars, one can find the recommended value Dm < 3 as frequently at small-amplitude as at large-amplitude Blazhko phases. About the Fourier analysis results (for stars with Dm < 5), we found the following:
   1. [Fe/H] estimates seem on average acceptable, but are ∼0.1 dex more metal-rich than the high-resolution spectroscopic abundances of red giant stars derived by KI. A recalibration of the [F/H]-period-&phis;31 relation using 287 RRab variables in 18 globular clusters performed by Sollima et al. (2005) using a nonparametric fitting method and KI metallicity scale yields similar average metallicity values to those derived from equation (10), within the errors. This indicates that M3 lies ∼0.1 dex off (on the metal-poor side) the mean relation defined by the calibrating globular clusters. However, the use of a different fitting method and/or metallicity scale, such as Sollima et al.'s, produces a different shape of the metallicity distribution, which might become relevant in the case of a composite population with a nonnegligible metallicity dispersion. Evolution off the ZAHB does not affect [Fe/H] determinations. The inclusion of Blazhko stars in a sample of regular stars does increase the scatter in the [Fe/H] determinations, as Blazhko stars at the low-amplitude phase appear as more metal-rich. If this effect is taken into account, there is no evidence of metallicity spread among the RR Lyrae stars in M3.
   2. Intrinsic colors and temperatures estimated from equations (11)–(15) show serious discrepancies with observed color distributions and theoretical (pulsational and evolutionary) requirements, and cannot be taken as reliable results.
   3. Absolute magnitudes are affected by a "compression" effect that reduces their scatter by a factor ∼2 compared with the observed 〈V〉 distribution. This makes them unreliable as accurate individual values, but they may provide useful averages for groups of stars, if applicable and after proper calibration. The rms errors, however, are significantly underestimated.
   4. The values of mass show a distribution with temperature that has the opposite trend with respect to the ZAHB. Consequently, the values of gravity are also affected by serious uncertainties. Neither estimate can be considered reliable.