<div>The exact morphology of interstellar ice is still an open question, and it is not yet established if they are porous or compact. The absence of -OH dangling bonds of water in observations does not conclusively prove that the ice is compact (Isokoski+2014). However, compact ice is more stable than porous one, and the latter tends to compact with time and temperature. Moreover, compaction is induced by energetic processes such as chemical reactions (Accolla+2011), UV radiation, or irradiation simulating cosmic rays (Mejía+2015). Finally, if water ice is formed on the grains, it is compact by nature (Accolla+2013). Therefore, experiments using compact ice films, even if they have not been used much in the past, are of paramount importance. <br></div><div><br></div><div>In this work, we present a series of targeted ice sublimation laboratory experiments designed to benchmark current astrochemical models. We conducted temperature-programmed desorption experiments with increasing levels of complexity of ice analogues of various chemical compositions and thicknesses (3-80 ML) using molecular beams in ultra-high vacuum conditions (10<sup>-10</sup> hPa) and low temperatures (10 K). We provide TPD curves of pure ices made of Ar, CO, CO<sub>2</sub>, NH<sub>3</sub>, CH<sub>3</sub>OH, H<sub>2</sub>O, and NH<sub>4</sub><sup>+</sup>HCOO<sup>-</sup>, their binary ice mixtures with compact amorphous H<sub>2</sub>O, ternary mixtures of H<sub>2</sub>O:CH<sub>3</sub>OH:CO, and water ice made in situ. <br></div><div><br></div><div>We found common trends in the desorption of molecules when their abundance is compared to water: compact amorphous water ices can trap up to 20% of high volatiles, around 5% of H-bonded molecules, and semi-volatile species are not trapped in the water ice matrix. <br></div><div><br></div><div>In summary, deposited or formed very compact, amorphous water ice of fewer than 100 layers cannot trap a large fraction of other gases, including CO or CO<sub>2</sub>. Astrochemical models should reproduce desorption kinetics and trapping fraction from our benchmark laboratory experiments to be reliable.</div>