APOGEE's target selection algorithm includes many red-clump (RC) stars in the APOGEE sample. A new method for identifying red-clump stars from spectrophotometric data and its application to the APOGEE data was presented in Bovy et al. (2014). This technique creates a very pure sample of red-clump stars with minimal contamination from red-giant-branch, secondary-red-clump, and asymptotic-giant-branch stars. The narrowness of the RC locus in color-metallicity-luminosity space allows distances to the stars to be assigned with an accuracy of 5%-10%. The purity is estimated to be about 95%. The DR17 catalog is the latest version of the APOGEE RC catalog, and contains 50,837 stars.
The DR17 catalog is created in the same manner as the DR16 and DR14 catalogs. In APOGEE DR17, log g is calibrated for both the red giant branch (RGB) stars and the RC, so the DR17 RC catalog just makes use of the catalog (calibrated) log g to select RC stars. However, as in DR16 and DR14, we have made the upper log g cut 0.1 dex more stringent (equation 2 in the paper above) to account for some changes in how log g was calibrated. We also use the additional, conservative log g cut that was used in both DR11 and DR12.
Section 4 of Bovy et al. (2014) described how to correct for observational biases due to the APOGEE target selection (for the whole (J-Ks)0 > 0.5 sample, not just the red-clump sample). Specifically, these are the biases due to the incomplete sampling of the available targets with (J-Ks)0 > 0.5 as a function of position on the sky and magnitude. Tools to correct for these biases are included in the apogee python package and extensive documentation on how to use this package is given on the package webpage. This selection-function code has been updated for DR13 (from the DR11 version described in the paper). DR13, like DR11 and DR12, contains only stars observed in APOGEE-1. DR14 is the first release containing stars observed in APOGEE-2, which has different targeting criteria than APOGEE-1.
Section 5 of Bovy et al. (2014) discussed astrophysical biases due to the red-clump selection. Specifically, they investigated the biases in recovering the underlying age and metallicity distribution by only selecting red-clump stars. The RC selection preferentially select stars with ages between 1 and 4 Gyr and with metallicities > -0.9. Please consult section 5 of the paper above for full details. The apogee python package also has tools for working with the astrophysical biases.
We recommend that you also use the latest updated catalog (DR17), but previous versions of the APOGEE RC catalog are still available. The DR16 catalog contains 39,675 stars and can be found here, with its content described in the data model available here. Similarly, the DR14 catalog contains 29,502 stars and can be found here, with its data model available here. The DR13 catalog was selected based on the more accurate set of spectroscopic parameters from the overall DR13 APOGEE release and therefore suffered from less contamination than previously released catalogs, leading to a slight decrease in the number of RC stars in the DR13 catalog compared to the DR12 catalog. The DR13 catalog contains 19,291 stars and is available on the SAS; its contents are described in the datamodel.
The older DR11 and DR12 catalog versions are also still available. For DR11 (Niveder et al. 2014), containing 10,341 objects, the catalog is available from the SAS here, with its datamodel available here. The DR12 catalogue (Bovy et al. 2014), containing 19,937 objects, is available from the SAS here, with its datamodel available here.