Earth atmospheric O<sub>2</sub> is mainly produced by biosphere photosynthesis, and biosphere respiration is also one of the main consumers of this gas. The evolution of the elemental and isotopic composition of atmospheric O<sub>2</sub> is thus linked to global biosphere productivity. Quantitative interpretation of the isotopic composition of O<sub>2 </sub>in the past as archived in polar ice cores relies on robust estimates of oxygen fractionation factors associated with the relevant biological processes: photosynthesis and respiration. In the past decades, some determinations of these biological fractionation factors were performed in uncontrolled large-scale environments or at the scale of the micro-organisms in conditions very different from the natural environment. This leads to uncertainties in the applicability of these determinations for the interpretation of isotopic composition of atmospheric O<sub>2</sub>. In order to come up with coherent estimates of oxygen biological fractionation factors applicable to the scale of plants or ecosystems, we developed a set-up of 4 closed biological chambers as a scaled-down replica of aquatic biosphere, with controlled environment parameters, and measured the evolution of O<sub>2</sub> concentration and of its isotopic composition. We present here 3 measurement series using this set-up run with the freshwater species <em>Chlorella vulgaris</em> and lasting between 2 and 9 months. These measurement series enabled us to validate and optimise our newly developed system for aquatic closed biological chambers. We also determined the isotopic discrimination associated with <sup>18</sup>O/<sup>16</sup>O of O<sub>2</sub> during respiration as -30 ‰ which is in the upper part of the distribution of the previously published values.