Nowadays, subsurface waste disposal is based on the concept of compacted clay barriers, characterized by a low permeability and high cation retention. The aim of this experimental study is not only to reproduce at the laboratory scale the in-situ compaction of saturated clays, but also to focus on the evolution of the clay properties. Therefore, percolation experiments have been performed using an oedometer cell equipped with an injection system under constant pressure (Jullien et al., 2002). The chosen clay is the hectorite from Hector (California, USA), with the structural formula (Na0.25Ca0.04)(Mg2.74Li0.31)Si3.97O10(OH,F)2.nH2O. This hectorite is associated with minor calcite. For all experiments, the hectorite sample was compacted at 0.5 MPa, and then soaked at 0.3 MPa with deionised water or Ni(NO3)2 0.01 M solution. Moreover, Ni(NO3)2 solutions were injected in clays for two different initial water saturation degrees: Sr=0.3 (dry clay) and Sr=1 (water-saturated). Some mechanical unloadings were also performed along the experiments, in order to test the swelling behaviour of the clay at different exchange rates. Axial strain variations were monitored during testing. Moreover, output solutions from the oedometer, i.e. leachates, have been regularly analysed for Ni2+, Na+, K+, Ca2+, Mg2+ and Li+. These analyses allowed to calculate the exchange ratio, KNa+, defined as the fraction of Na+ released to the solution, compared to the bulk Na+ content in the starting clay, in atomic proportion. Two samples, run in the oedometer with Ni(NO3)2 solution, were also obtained, with an exchange ratio KNa+ of 0.19 for Sr=0.3 and 0.34 for Sr=1. Whatever the initial saturation degree, neither Ni2+ nor Li+ have been detected in the leachates, and the Na+ content in the solutions is around 100 times higher than Ca2+, Mg2+ and K+: Ni2+ has been totally fixed by hectorite and exchanged to Na+. Hydraulic conductivity, k, was calculated from the Darcy law, using the expression elaborated for a saturated medium: k = Q / i S, with Q the measured volume flow (m3.s-1), i the hydraulic gradient and S the section area of the oedometer (m2). The obtained hydraulic conductivity k of the compacted hectorite drops in the same range, 4.2 10-13 m.s-1 to 4.9 10-13 m.s-1, when soaked with water or Ni2+ solution, respectively. Swelling strain values determined at the permanent regim are similar: 28% with water and 23% with Ni(NO3)2 solution. A mechanical unloading of experiment with pure water shows that the swelling strain incrreases up to 44%. Similarly, the Ni2+ exchanged hectorite with KNa+ = 0.19 exhibits a swelling strain of 38% after unloading. By contrast, the swelling strain of the same hectorite with KNa+ = 0.34 stands to the same value, 24%. These results highlight that the swelling ability of hectorite has been significantly reduced by the Na+-Ni2+ exchange. The microstructural evolution of hectorite has also been investigated, in order to better undertand the influence of nickel on macroscopic properties. For this purpose, a diffractometer equipped with a controlled atmospher cell was used. The three hectorite samples (soaked with water, soaked with Ni2+ solution with KNa+ = 0.19 or KNa+ = 0.34, respectively) have undergone the same treatments: on the one hand, a thermal cycle (50°C, 100°C, 150°C), and on the other hand, an hydration cycle (relative humidity at 50%, 70%, 80%, 90%, with a constant temperature of 30°C). When increasing temperature, deshydration of interlayer space is delayed when Na+ is exhanged by Ni2+. Indeed, d(001) reaches 10Ǻ since 100°C for hectorite soaked with water, whereas the same value is obtained at 150°C for hectorite with KNa+ = 0.34. Besides, when increasing relative humidity, hectorite with KNa+ = 0.34 swells less than both other samples. When increasing the interaction with nickel solution, the macroscopic behaviour of hectorite, a smectite known for its swelling properties, is getting close to that of a non-expansive clay. Moreover, nickel has significantly changed hydration properties of hectorite layers. So, evolution of swelling properties at a macroscopic scale can be linked to evolution of hydration properties at a nanoscopic scale.