Nowadays, Lithium ion (Li-ion) batteries are present in many applications such as mobile phones, computers, power tools, hybrid electric (HEV) and electric (EV) vehicles. However abundance, cost and especially accessibility of lithium will be able to bring some issues for Li-ion development in the future [1]. This leads the scientist community to go back to sodium ion battery research, which decreased significantly after Li ion battery success in the 90’s. Sodium seems to be an interesting alternative to lithium because of its abundance and availability. However, in order to be competitive against other technologies, capacity, rate capability and cycle life should be improved [2]. Carbon for Na-ion battery applications has been extensively studied in the literature [3,4] whatever its form: hard carbon, carbon black, carbon nanospheres or nanotubes, carbon fibers and graphene. Especially for hard carbons the main problems reported in literature are a high first irreversible capacity and a poor cyclability. For example, Ding et al.[5] tested carbon made from peat (organic matter) pyrolysis and they obtained a reversible capacity of 255 mAh.g-1 after 200 cycles with a current density equal to 50 mA.g-1 and an initial coulombic efficiency reaching 60%. Recently Sun et al.[6] have synthesized hard carbon from shaddock peel pyrolysis. The pyrolysis temperature was varied from 800 to 1400°C and the best performances were obtained for the temperature 1200°C: the material exhibited a very high reversible capacity of around 400 mAh.g-1 with a good stability over 200 cycles and a first coulombic efficiency of 68%. Even though hard carbon structure and sodium ion insertion into these materials were studied, especially by Dahn’s group in 2000 [7,8,9], insertion mechanisms are still in debate in the literature. Dahn and coworkers developed a card-house model with two different mechanisms: in the galvanostatic curve, the sloppy part corresponds to sodium intercalation between graphene sheets and the low voltage plateau is related to the sodium insertion in micropores. Lately Bommier et al. [10] have proposed another point of view concerning this model based on ab initio calculations: the low voltage plateau is due to intercalation into sites around the defective carbon surface. In this presentation we proposed to develop two parts. The first part exhibits the dependence of electrochemical performances on hard carbon structure. All hard carbon samples are made from the same organic precursor, cellulose. The final pyrolysis temperature is varied from 700°C to 1600°C. The samples are fully characterized with both usual techniques such as X-ray diffraction, scanning electron microcopy and N2 adsorption and more advanced tools such as small-angle X-ray scattering (SAXS) and transmission electron microscopy. The different parameters obtained from all these characterization methods (d002, specific surface area, radius of gyration and crystallite size) can be linked to the electrochemical performances as indicated in the figure. Samples pyrolysed from 1300°C to 1600°C showed a reversible capacity equal to 300 mAh.g-1 at C/10 rate (where C=372 mA.g-1) with an excellent stability in cycling and a very good initial coulombic efficiency reaching 84%. The second part of the presentation aims to deeper understand sodium insertion mechanisms in these disordered structures. To distinguish the different mechanisms hard carbons are discharged at different potentials and analyzed with SAXS and impedance spectroscopy.