Ariel is an ESA space telescope planned for launch in 2029. It is aimed at observing ∼1,000 exoplanets using transit spectroscopy at infrared wavelengths between 0.5-7.8μm, to characterise the chemical composition and thermal structure of their atmospheres. To detect minute flux variations and confidently extract the exoplanet signal, Ariel requires good instrumental flux stability. Moreover, the instrumental response must be monitored through time, and instrumental effects and stellar noise must be estimated and minimised to get as close as possible to the photon noise limit. To do this, Ariel must be calibrated in-flight by observing a set of stellar calibrators, stars whose spectral energy distribution is assumed to be well-studied and constant. This project is aimed at building the stellar calibrator catalogue for Ariel and looking for possible correlations between stellar properties and flux stability. This can also serve as a base study for future space missions other than Ariel. A starting candidate sample mainly including G dwarfs that were observed by TESS is filtered with a three-step procedure using the Lomb-Scargle periodogram and the reduced chi-squared statistic. Assuming an Ariel 3σ flux stability requirement of 100ppm over 6 hours, the defined Ariel Catalogue comprises 581 stars. These are fairly homogeneously distributed over the sky, without periodic variability, and without a large flux excess dispersion. It is found that when selecting the correct effective temperature ranges (5, 000 < Teff < 6, 300K), the probability of finding stable stars is almost constant at 30%. Only the dimmer stars seem to have a higher probability of passing the selection, which is possibly due to the higher noise in the data which makes variability more difficult to identify. Note, however, that the catalogue defined in this study still comprises calibrator candidates. The noise in TESS light curves is probably too high to effectively analyse the stability within the current requirement. So, follow-up observations of these stars with a lower noise are recommended, for instance with PLATO, to greenlight them as final Ariel calibrators.