Does the linear conversion between calcium infrared triplet and metallicity of simple stellar populations hold in the whole range of metallicity?

The calcium infrared triplet (CaT) is one of the prominent absorption features in the infrared wavelength regime. Recently, this absorption feature has been getting attention in the prediction of metallicity of stellar populations because of its strong sensitivity to the calcium abundance and metallicity of a star. However, we find that measuring metallicity directly from CaT is very dangerous because the formation mechanism of CaT is very inefficient in the cool stars which are abundant in metal-rich populations. This characteristics of CaT make the CaT-metallicity relation to converge around ~ $8{\AA}$ in the metal-rich regime. Our results suggest that, because of the converging CaT-metallicity relation in the metal-rich regime, the metallicity of simple stellar populations greater than [Fe/H]~-0.5 is unreliable when the linear conversion between CaT and metallicity is applied to derive metallicity. Based on these results, we suggest that CaT is not a good metallicity indicator for the metal-rich stellar populations.