Benchmark stars are crucial as validating standards for current as well as future large stellar surveys of the Milky Way. However, the number of suitable metal-poor benchmarks is currently limited. We aim to construct a new set of metal-poor benchmarks, based on reliable interferometric effective temperature ($T_text{eff}$) determinations and a homogeneous analysis with a desired precision of $1%$ in $T_text{eff}$. We observed ten late-type metal-poor dwarf and giants: HD2665, HD6755, HD6833, HD103095, HD122563, HD127243, HD140283, HD175305, HD221170, and HD224930. Only three of the ten stars (HD103095, HD122563, and HD140283) have previously been used as benchmarks. For the observations, we used the high angular resolution optical interferometric instrument PAVO at the CHARA array. We modelled angular diameters using 3D limb darkening models and determined $T_text{eff}$ directly from the Stefan-Boltzmann relation, with an iterative procedure to interpolate over tables of bolometric corrections. Surface gravities ($log(g)$) were estimated from comparisons to Dartmouth stellar evolution model tracks. We collected spectroscopic observations from the ELODIE and FIES spectrographs and estimated metallicities ($mathrm{[Fe/H]}$) from a 1D non-LTE abundance analysis of unblended lines of neutral and singly ionized iron. We inferred $T_text{eff}$ to better than $1%$ for five of the stars (HD103095, HD122563, HD127243, HD140283, and HD224930). The $T_text{eff}$ of the other five stars are reliable to between $2-3%$; the higher uncertainty on the $T_text{eff}$ for those stars is mainly due to their having a larger uncertainty in the bolometric fluxes. We also determined $log(g)$ and $mathrm{[Fe/H]}$ with median uncertainties of $0.03,mathrm{dex}$ and $0.09,mathrm{dex}$, respectively. These ten stars can, therefore, be adopted as a new, reliable set of metal-poor benchmarks.
Download