Aims. We aim to analyze the chemical evolution of the Small Magellanic Cloud, adding 12 additional clusters to our existing sample, based on accurate and homogeneously derived metallicities. We are particularly interested in seeing if there is any correlation between age and metallicity for the different structural components to which the clusters belong, taking into account their positions relative to the different tidal structures present in the galaxy. Methods. The spectroscopic metallicities of red giant stars were derived from the measurement of the equivalent width of the near-IR calcium triplet lines. Our cluster membership analysis was carried out using criteria that include radial velocities, metallicities, proper motions, and distances from the cluster center. Results. The mean cluster radial velocity and metallicity were determined with a typical error of 2.1 km s−1 and 0.03 dex, respectively. We added this information to that available in the literature for other clusters studied with the same method, compiling a final sample of 48 clusters with metallicities that were homogeneously determined. The clusters of the final sample are distributed across an area of ∼70 deg2 and cover an age range from 0.4 Gyr to 10.5 Gyr. This is the largest sample of spectroscopically analyzed SMC clusters available to date. Conclusions. We confirm the large cluster metallicity dispersion (∼0.6 dex) at any given age in the inner region of the SMC. The metallicity distribution of our new cluster sample shows a lower probability of being bimodal than suggested in previous studies. The separate chemical analysis of clusters in the six components (Main Body, Counter-Bridge, West Halo, Wing/Bridge, Northern Bridge, and Southern Bridge) shows that only clusters belonging to the Northern Bridge appear to trace a V-Shape, showing a clear inversion of the metallicity gradient in the outer regions. There is a suggestion of a metallicity gradient in the West Halo, similar to that previously found for field stars. It presents, however, a very broad uncertainty. Also, clusters belonging to the West Halo, Wing/Bridge, and Southern Bridge exhibit a well-defined age-metallicity relation with relatively little scatter in terms of abundance at a fixed age compared to other regions.

Context. The line strength of the near-infrared Ca II triplet (CaT) lines are a proxy for measuring metallicity from integrated and individual stellar spectra of bright red giant stars. In the latter case it is a mandatory step to remove the magnitude (proxy for gravity, temperature, and luminosity) dependence from the equivalent width (EW) of the lines before converting them into metallicities. For decades the working empirical procedure has been to use the relative magnitude with respect to the horizontal branch level or red clump, with the advantage that it is independent from distance and extinction. Aims. The V filter is broadly adopted as the reference magnitude, although a few works have used different filters (I and Ks, for example). In this work we investigate the dependence of the CaT calibration using the griz filters from the Dark Energy Camera (DECam) and the Gemini Multi-Object Spectrograph (GMOS), the G filter from Gaia, the BVI filters from the Magellanic Clouds photometric survey (MCPS), and the YJKs filters from the Visible and Infrared Survey Telescope for Astronomy (VISTA) InfraRed CAMera (VIRCAM). We use as a reference the FOcal Reducer and low dispersion Spectrograph 2 (FORS2) V filter used in the original analysis of the sample. Methods. Red giant stars from clusters with known metallicity and available CaT EWs were used as reference. Public photometric catalogues were taken from the Survey of the MAgellanic Stellar History (SMASH) second data release, VISTA survey of the Magellanic Clouds system (VMC), Gaia, MCPS surveys, plus VIsible Soar photometry of star Clusters in tApi’i and Coxi HuguA (VISCACHA) GMOS data, for a selection of Small Magellanic Cloud clusters. The slopes were fitted using two and three lines to be applicable to most of the metallicity scales. Results. The magnitude dependence of the CaT EWs is well described by a linear relation using any filter analysed in this work. The slope increases with wavelength of the filters. The zero point (i.e. reduced EW), which is the metallicity indicator, remains the same. Conclusions. If the same line profile function is used with the same bandpasses and continuum regions, and the total EW comes from the same number of lines (2 or 3), then the reduced EW is the same regardless the filter used. Therefore, any filter can be used to convert the CaT equivalent widths into metallicity for a given CaT calibration.