The hot interstellar medium towards 4U 1820-30: a Bayesian analysis
High ionisation lines in the soft X-ray band are generally associated to either interstellar hot gas along the line of sight or to photoionised gas intrinsic to the source. In the low-mass X-ray binary 4U 1820-30, the nature of these lines is not well understood. We characterised the ionised gas present along the line of sight towards the source producing the X-ray absorption lines of Mg XI, Ne IX, Fe XVII, O VII, and O VIII. We analysed all the observations available for this source in the XMM-Newton and Chandra archives, taken with the RGS, HETG, and LETG spectrometers. The high-resolution grating spectra have been accurately examined through a standard X-ray analysis based on the $C$-statistic and through the Bayesian parameter inference. We tested two physical models which describe a plasma in either collisional ionisation or photoionisation equilibrium. We adopted the Bayesian model comparison to statistically compare the different combinations of the models used for the analysis.} We found that the lines are consistent with hot gas in the interstellar medium rather than being associated with the intrinsic gas of the X-ray binary. Our best-fit model reveals the presence of a collisionally ionised plasma with a temperature of $T=(1.98\pm0.05)\times10^6 $ K. The photoionisation model fails to fit the Fe XVII line (which is detected with a significance of $6.5σ$) due to the low column density predicted by the model. Moreover, the low inclination of the binary system is likely the reason for the non-detection of ionised gas intrinsic to the source.}