Exoplanets in ancient stellar populations: occurrence constraints and hot-Jupiter candidates in the Galactic halo
The Galactic halo preserves a record of the Milky Way's earliest assembly and contains both in-situ stars and stars accreted from dwarf galaxies. Possible planets around these stars, therefore, probe formation in ancient, metal-poor environments, including systems of extragalactic origin. We present a search for short-period transiting planets around kinematically selected halo dwarfs using Gaia DR3 and TESS, focusing on planets with periods of $1 < P < 10$ days. We identify two hot-Jupiter (HJ) candidates, one in the in-situ and one in the accreted halo, although the latter is highly grazing and excluded from the occurrence analysis. The accreted candidate, if confirmed, would orbit the most metal-poor HJ host known ([Fe/H] $\approx -1$). Using injection--recovery tests and automated vetting, we constrain occurrence in the full halo, in-situ, and accreted samples. In the HJ regime ($8\,R_\oplus < R_{\rm p} < 22\,R_\oplus$, $1\,\text{day}\ < P < 10$ days), the non-grazing candidate implies an overall halo occurrence rate of $0.13^{+0.12}_{-0.07}\%$ if planetary, while the absence of confirmed detections gives a corresponding $1\sigma$ upper limit of $<0.14\%$. For the in-situ halo, we infer $0.17^{+0.17}_{-0.10}\%$ (or $<0.19\%$ assuming no detections), while for the accreted halo we derive an upper limit of $<0.56\%$. These rates lie well below the corresponding short-period giant-planet occurrence measured in the Galactic disc. A forward model assuming Kepler-like occurrence also predicts $10 \pm 3$ detections compared with at most one observed. We find no significant occurrence difference between the in-situ and accreted halo populations, strengthening the evidence that close-in giant planets are rare across the old, metal-poor halo.