Safe and sustainable management of nuclear waste poses major scientific challenges to make the environmental footprint of nuclear energy as small as possible for very long periods of time. As many other countries, France is considering the deep geological disposal (in the Callovo-Oxfordian (COx) argillite formations of the Paris basin) as a reliable way of storing high-level radioactive waste in order to provide adequate protection for humans and the environment. In addition to being proven geologically stable for million years, the natural and engineered clay barriers can benefit from many favorable properties, such as low permeability, high sorption capacity, etc. The mineralogical composition of the Callovo-Oxfordian argillite shows about 41% of clay minerals (23% of interstratified illite/smectite, 14% of illite-type minerals, 2% kaolinite and 2% chlorite) [1,2]. A non-negligible amount of organic matter is also present (~1%) [3], and it is known that the interaction of natural organic matter (NOM) with radionuclides and clays can affect the solubility and toxicity of trace elements in natural aqueous environments [4,5]. Reliable prediction of the behaviour of radionuclides and their transport and retention in clayey formations at nuclear waste repositories requires detailed molecular scale understanding of these complex multicomponent systems. Computational molecular modelling has already become an important tool in the study of thermodynamic, structural and transport properties of hydrated clays (e.g., [6-8]). As the first step in our study of the effects of organic molecules on the adsorption and transport of radionuclides in hydrated clay systems we have investigated the effects of the ordering in charge distributions on the swelling behavior of simulated clays. Montmorillonite was chosen as a model of smectite clay. Montmorillonite structure consists of aluminum-oxygen octahedral sheet sandwiched between two opposing silicon-oxygen tetrahedral sheets giving rise to a 2:1 clay mineral. Isomorphic substitutions in the tetrahedral and octahedral sheets are responsible of the negative layer charge of montmorillonite clay minerals having the chemical composition (Si8-xXx)(Al4-yYy)O20(OH)4 where X = Al, Y = Mg, Fe...[9]. The montmorillonite models for our study are based on a pyrophillite unit cell structure (5.16Å×8.966Å×9.347Å) obtained from the crystallographic data of Lee et al. [10]. The 4×4×2 simulation supercells were built and substitutions were made in the pyrophillite structure in order to approximate as close as possible the chemical composition of Wyoming montmorillonite M24(Si248Al8)(Al112Mg16)O640(OH)128, where M is either Cs+, Na+, or K+ [9]. We explored three different models of substitution distributions. In the first model, the substitutions were uniformly and orderly distributed within the tetrahedral and octahedral sheets. In the second model, the substituted positions were kept ordered in the octahedral sheets but made disordered in the tetrahedral one. In the third model, the substituted positions of the octahedral sites were additionally made disordered. In order to study the swelling behavior of these montmorillonites, molecular dynamics (MD) simulations were run in the NPzT statistical ensemble (T = 298 K, Pz = 1 bar) for each of the three different substitution models and with 22 different hydration states ranging from 0 to 700 mgwater/gclay (from 0 to 42 H2O molecules per one monovalent cation). All MD runs were performed for a total of 2 ns using the CLAYFF force field [11]. At the beginning of the simulations, the cations were placed at the midplane of the clay interlayer space and water molecules were added randomly. After the system reached equilibrium, the last 1ns of each MD trajectory was used to compute the clay basal spacing and the swelling thermodynamic properties: hydration energy, immersion energy, isosteric heat of adsorption. The MD simulation results indicate that in addition to the commonly observed 1-layer and 2-layer hydrates, stable hydration states corresponding to 3-layer and 4-layer hydrates can also be distinguished. The stable states corresponding to the minima of hydration energy were then selected to run further 500 ps NVT-ensemble MD simulations at the same temperature and with the volume fixed at the average value resulting from the corresponding previous NPzT simulation. The equilibrium parts of these NVT-simulated trajectories were then used to calculate the structural (radial distribution functions, atomic density profiles) and dynamical (diffusion coefficient) properties of the hydrated montmorillonite. References [1] ERM (1997) Echantillons d'argiles du forage EST104 : Etude minéralogique approfondie. Rapport ANDRA n° D.RP.0ERM.97.008 [2] ERM (1996b) Caractérisation d'échantillons d'argiles du forage EST103. Rapport ANDRA n° B.RP.0ERM.96.003 [3] ANDRA (2005) Dossier 2005 Argile, Référentiel du site de Meuse Haute Marne. C.R.P.ADS.04.0022 Andra : Paris [4] Buffle, J. (1988) Complexation Reactions in Aquatic Systems: An Analytical Approach; Ellis Horwood Ltd.:Chichester, p 692. [5] Tipping, E. (2002) Cation Binding by Humic Substances, Cambridge University Press: Cambridge, p 434. [6] Smith, D.E., Langmuir, 14, 5959-5967 (1998). [7] Rotenberg, B., Marry, V., Vuilleumier, R., Malikova, N., Simon, C., Turq, P., Geochim. Cosmochim. Acta, 71, 5089-5101 (2007). [8] Liu, X.D., Lu, X.C., Wang, R.C., Zhou, H.Q. Geochim. Cosmochim. Acta, 72, 1837-1847 (2008). [9] Tsipursky, S.I., Drits, V.A. Clay Minerals, 19, 177-193 (1984). [10] Lee, J.H. and Guggenheim, S. American Mineralogist, 66, 350-357 (1981). [11] Cygan, R.T., Liang, J.J., Kalinichev, A.G. Journal of Physical Chemistry B, 108, 1255-1266 (2004).