A test of Ca ii H & K photometry for isolating massive globular clusters below the metallicity floor

TL;DR

This study tackles the existence of massive globular clusters below the metallicity floor by testing integrated-light Ca II H & K CaHK photometry as a metallicity proxy to pre-select EXT8-like candidates. Using CFHT MegaCam imaging in CaHK, plus u, g, and i bands for EXT8 and a broad sample of M31 GCs across $-2.9 \leq [\mathrm{Fe/H}] \leq 0.4$, the authors calibrate CaHK colours against spectroscopic metallicities and evaluate HB morphology effects. They find CaHK-g exhibits a strong metallicity sensitivity with a near-linear, bilinear relation in the metal-poor regime, while CaHK-u is non-monotonic; both yield an ~0.3 dex metallicity precision, and CaHK-based selection reduces false positives by at least a factor of 2 relative to broadband colours. Horizontal branch morphology has limited impact on CaHK colours, supporting their use as a candidate-selection tool. Overall, CaHK photometry shows promise for efficiently identifying EXT8-like GCs in external galaxies, potentially informing GC formation scenarios and hierarchical assembly studies.

Abstract

The serendipitous discovery of the M31 globular cluster (GC) EXT8 has presented a significant challenge to current theories for GC formation. By finding other GCs similar to EXT8, it should become clear if and/or how EXT8 can fit into our current understanding of GC formation. We aim to test the potential of integrated-light narrow-band Ca II H & K photometry as a proxy for the metallicity of GCs to be able to provide effective candidate selection for massive GCs below the GC metallicity floor ([Fe/H] $\leq$ -2.5), such as EXT8. We investigate the behaviour of two colours involving the CaHK filter employed by the Pristine survey, CaHK-u and CaHK-g, as a function of metallicity through CFHT MegaCam imaging of EXT8 and a wide set of M31 GCs covering the metallicity range of -2.9 $\leq$ [Fe/H] $\leq$ +0.4. Additionally, we investigate if the CaHK colours are strongly influenced by horizontal branch morphology through available morphology measurements. In both of the CaHK colours, EXT8 and two other potential GCs below the metallicity floor can be selected from other metal-poor GCs ([Fe/H] $\leq$ -1.5) with CaHK-g showing the larger metallicity sensitivity. The RMS of linear fits to the metal-poor GCs show an uncertainty of 0.3 dex on metallicity estimations for both colours. Comparisons with u-g and g-z/F450W-F850L colours reinforce the notion that CaHK photometry can be used for effective candidate selection as they reduce false positive selection rates by at least a factor of 2. We find no strong influence of the horizontal branch morphology on the CaHK colours that would interfere with candidate selection, although the assessment is limited by quantity and quality of available data.

Figures (7)

• Figure 1: The observational footprint of this study. The GC sample has been marked with either orange diamonds or blue circle, signifying GCs with and without robust reddening estimates. EXT8 has been marked in red and the black circles are confirmed M31 GCs from the C&R16 catalogue which were not considered. The background colour is the reddening from the Schlegel1998 dust map with the correction of Schlafly2010, clearly highlighting M31 itself. The two rectangle with 'ears' are the observed fields that form the footprint for this study.
• Figure 2: Discretized calibration residuals before and after applying the FoV corrections determined by PhotCalib for the g and i filters. The first column shows the residuals for g (top) and i (bottom) before correction and the second column shows the residuals after the correction is applied.
• Figure 3: Differences in u-g (top) and g-i (bottom) between our photometry and the Peacock2010 photometry transformed to MegaCam colours. The text in the upper left of each plot is the average difference and scatter for the entire comparison sample.
• Figure 4: (CaHK-u)$_{o}$ (left) and (CaHK-g)$_{o}$ (right) plotted versus metallicity for the GC sample. The GCs with robust reddening estimates are marked with orange diamonds and the remaining are marked by blue circles. EXT8 can be seen separating itself from the other metal-poor GCs as the left most GC in each plot. Note the difference in vertical scale between the two panels (0.1 vs 0.2 mag major tick spacing).
• Figure 5: The broadband colours, $(u-g)_o$ (top) and $(g-z)_{\mathrm{SDSS},o}$ (bottom), plotted against metallicity for our GC sample. For $(g-z)_{\mathrm{SDSS},o}$, the spectroscopically calibrated CMRs from Peng2006 (dotted), Sinnott2010 (dashed) and Usher2012 (solid) are also plotted. For both colours, the GCs below the metallicity do not separate clearly from other GCs.