Thèse de doctorat en Chimie - Physique
Sous la direction de Daniel Lemordant et de Bernard Carre.
Soutenue en 2007
à Tours .
This work aims at improving the security linked to the use of Li-batteries electrolytes. First, we tested the use of redox shuttles as an electrochemical system of internal protection against overcharge. We studied the possibility of employing ionic liquids as electrolytes for Li-ion accumulator. Finally, we tested two fluoro-alkylcarbonates as flame-retardant additives. Three classes of molecules were studied as redox shuttles : metallic complexes, aromatic and radical systems. Only a few molecules of the radical system allow an efficient protection against overcharge of the cell Li4Ti5O12 / (EC/DEC) + LiPF6 / LiFePO4. These molecules are the TEMPO (tetramethylpiperidinyl-1-oxy) and the Cyano-TEMPO, they limit the recharge voltage to 3. 5-3. 7 V vs. Li+/Li. We chose two ionic liquids whose anion is the bis(trifluoromethanesulfonyl)imide (TFSI-) ion and the cation N-trimethyl-N-hexylammonium (HTMA+) or N-trimethyl-N-propylammonium (PTMA+). Two electrolytes have been optimized regarding their transport properties and their thermal stability : HTMA+TFSI-/(EC:DEC) [ = 0. 3] + LiTFSI 1 mol. L-1 and PTMA+TFSI-/(EC:DEC) [ = 0. 25] + LiTFSI 1 mol. L-1 ( with the weight fraction of the ionic liquid in the mix). The electrochemical study shows that the electrolyte cannot be cycled on the graphite electrode. Cyclic voltammetry performed on the negative electrode demonstrates the absence of formation of the SEI and reveals the intercalation of ammonium cations into the graphene layers. The analysis of the electrode, after cycling, by X-Ray Diffraction (XRD) and NMR spectroscopy, confirms the insertion of the ammonium cations. To improve the cycling results, we replaced the EC by the Cl-EC (chloro-ethylene carbonate). Cl-EC allows the formation of the passivation layer that protects graphite against exfoliation, but does not prevent the intercalation of the ammonium cations. The cycling tests on the negative electrode Li4Ti5O12 and the positive LixCoO2 were conducted successfully. In order to improve the ininflamability of the electrolyte (EC/DMC/DEC + LiPF6 1 mol. L-1), we tested the use of two fluoro-alkylcarbonates as flame-retardant additives : the 2,2,3,3,4,4,4-hepatafluorobutyl methyl carbonate (FC2) and the 2,2,3,3,4,4,5,5,6,6,7,7,8,8,8-pentadecafluorooctyl methyl carbonate (FC1). These two additives were synthetized at the laboratory. The electrochemical study demonstrates that the FC1 compound does not modify the cyclability of the positive electrode LixCoO2. Nevertheless, it generates an important loss of capacity of the graphite electrode. The cycling tests conducted on the LixCoO2 cathode and the graphite anode show the good performances of the FC2 additive.
Additives for safe Li-batteries electrolyte : redox shuttle, flame-retardant and wetting agents
Le travail mené vise à améliorer la sécurité liée à l’utilisation des électrolytes à base d’alkylcarbonates. Nous avons tout d’abord testé l’usage des navettes rédox comme système électrochimique de protection interne contre la surcharge. Nous avons étudiée la possibilité d’utiliser les liquides ioniques comme électrolyte pour batterie Li-ion. Enfin, nous avons testé deux fluoro-alkyl-carbonates comme additifs retardateur de flamme. Trois classes de molécules ont été étudiées comme navettes rédox : des complexes métalliques, des systèmes aromatiques et des systèmes radicalaires. Uniquement quelques molécules du système radicalaire permettent une protection efficace contre la surcharge d’un accumulateur de type Li4Ti5O12 / (EC/DEC) + LiPF6 / LiFePO4. Ces molécules sont la TEMPO (tétraméthylpipéridinyl-1-oxy) et la Cyano-TEMPO, elles limitent la tension de recharge à 3,5-3,7 V vs. Li+/Li. Nous avons sélectionnées deux liquides ioniques dont l'anion est l'ion bis(trifluorométhylsulfonyl)amidure (TFSI-) et le cation N-triméthyl-N-hexylammonium (HTMA+) ou N-triméthyl-N-propylammonium (PTMA+). Nous avons tout d’abord montré que les mélanges (liquide ioniques + sel de lithium) ne sont pas assez conducteurs à température ambiante pour pouvoir être utilisé comme électrolyte dans les accumulateurs Li-ion. Pour augmenter la conductivité, nous avons ajouté au mélange un co-solvant de faible viscosité, un mélange binaire de carbonates EC/DEC, le carbonate d’éthylène (EC) et le carbonate de diéthyle (DEC) dans une proportion 2/3 en masse. Deux électrolytes ont été optimisés aux vues de leurs propriétés de transport et de leurs stabilités thermiques : HTMA+TFSI-/(EC:DEC) [ = 0,3] + LiTFSI 1 mol. L-1 et PTMA+TFSI-/(EC:DEC) [ = 0,25] + LiTFSI 1 mol. L-1 ( avec la fraction massique du liquide ionique dans le mélange). L’étude électrochimique montre qu’aucun électrolyte ne permet d’effectuer des cycles de charge et de décharge sur électrode négative en graphite. La voltammétrie cyclique réalisée sur l’électrode de graphite montre l’absence de formation de la couche de passivation et révèle l’intercalation des cations ammonium entre les couches de graphènes. L’analyse de l’électrode, après cyclage, par diffraction des rayons-X (XRD) et résonance magnétique nucléaire (RMN) confirme l’insertion des cations ammonium. A fin d’amélioré les résultats, nous avons remplacée l’EC par la Cl-EC (carbonate de chloro-éthylène). Cette dernière permet la formation d’une couche de passivation qui protège le graphite contre l’exfoliation, mais n’empêche pas l’intercalation des cations ammonium. Les tests de cyclage sur électrode négative Li4Ti5O12 et sur électrode positive LixCoO2 sont très satisfaisants. Afin d’améliorer l’ininflammabilité de électrolyte de référence (EC/DMC/DEC + LiPF6 1 mol. L-1), nous avons testé l’usage de deux composés fluoré comme additif retardateur de flamme : le carbonate de 2,2,3,3,4,4,4-hépatafluorobutyle méthyl (FC2) et le carbonate de 2,2,3,3,4,4,5,5,6,6,7,7,8,8,8-péntadécafluorooctyl méthyl (FC1). Ces deux additifs ont été synthétisés au laboratoire. La caractérisation physico-chimique des mélanges FC1/(EC/DMC/DEC + LiPF6 1 mol. L-1) et FC2/(EC/DMC/DEC + LiPF6 1 mol. L-1) montre que le FC1 (contrairement au FC2) devient rapidement non miscible avec l’électrolyte de référence. L’ajout de ces deux additifs, à faible concentration, à l’électrolyte de référence permet d’améliorer l’ininflammabilité de ce dernier. Les mesures des tensions superficielles et des angles de contact confirment que l’ajout de FC1 et de FC2 permet d’améliorer la mouillabilité du séparateur en Celgard®. L’étude électrochimique montre que l’additif FC1 ne modifie pas la cyclabilité de l’électrode positive LixCoO2. En Revanche, il provoque des pertes de capacités importantes de l’électrode négative en graphite. Les tests de cyclages réalisé sur l’électrode positive LixCoO2 et sur l’électrode négative en graphite montre les bonnes performances de l’additif FC2.