Exploring the Li–Ga room temperature phase diagram and the electrochemical performances of the Li x Ga y alloys vs. Li J. Saint * , M. Morcrette, D. Larcher, J.M. Tarascon Universite ´ de Picardie Jules Verne, Laboratoire de Re ´activite ´ et Chimie des Solides, UMR-6007, 33 rue Saint Leu, 80039, Amiens, France Received 10 February 2004; accepted 18 May 2004 Abstract Li 2 Ga 7 , LiGa, and Li 2 Ga alloys were synthesized by ball-milling, from Li powders and Ga ingots, and characterized for their structural and electrochemical performances. Special attention was devoted to the Li-driven 3D(LiGa) ! 2D(Li 2 Ga) transformation that occurs at 0.02 V vs. Li + /Li on discharge. By in situ X-ray measurements we demonstrated that such a transition, accompanied by a 60% volume change, is initially nicely reversible, but rapidly becomes detrimental to the cell capacity retention upon cycling. Bypassing the (LiGa) ! (Li 2 Ga) transition was shown to considerably increase our cell performances, with capacity retention curves levelling at 300 mA h/g for at least 20 cycles. D 2004 Published by Elsevier B.V. Keywords: Li – Ga; Electrochemical performance; Li x Ga y alloys 1. Introduction Most of today’s Li-ion cells utilize electrochemical couples consisting of layered LiCoO 2 as the positive electrode and negative electrodes of carbonaceous prod- ucts, typically graphite. Within such cells, upon charge, the incoming Li + ions from the positive electrode are interca- lated into the carbonaceous materials up to compositions of Li x C 6 (where x can reach values as high as 2) at voltage near zero vs. Li + /Li. Despite several phase transitions, lithium intercalation into graphite occurs without causing any drastic change to its basic structure so that the cycle life and capacity retention of LiCoO 2 /C cells are excellent. However, ongoing research efforts are placed on searching for carbon alternatives in the hope of finding materials with both larger gravimetric and/or volumetric capacities and slightly more positive intercalation voltages as com- pared to Li + /Li, so as to minimize safety problems (e.g., any risks of high surface area Li plating when Li-ion cells are fast recharged). Through the years, considerable research has been un- dertaken to develop a viable lithium alloy or lithium- intermetallic compound as negative electrode, and some binary lithium-metals systems have been extensively inves- tigated. Among them, Li x Al [1,2], Li x Si and Li x Sn appeared, at first, to be promising materials since they have high a theoretical capacity of 1000, 4200 and 1000 mA h/g, respectively. Despite these large and attractive capacities, they were rapidly abandoned due to their poor capacity retention associated with their large volume expansion/ contraction upon cycling and strains due to phase transitions leading to loss of electrical contact. Different strategies have been proposed to overcome problems linked to the large volume changes that occur during cycling. One of these approaches consists in using a composite microstructure in which active alloy particles are finely dispersed in a solid matrix, the latter acting as a matrix that ‘‘buffers’’ the expansion of the active phase [3]. In a more recent approach, intermetallic electrodes that operate by a lithium insertion/metal extrusion such as InSb [4,5], Cu 6 Sn 5 [6,7] and Cu 2 Sb [8] have been investigated as possible negative materials. These materials present strong structural relationships with their lithiated products, and the volume change during cycling is less than that observed for pure metals/semi-metals when cycled in limited voltage windows. Results obtained by these various approaches 0167-2738/$ - see front matter D 2004 Published by Elsevier B.V. doi:10.1016/j.ssi.2004.05.021 * Corresponding author. Tel.: +33-3-22-82-75-84; fax: +33-3-22-82- 75-90. E-mail address: juliette.saint@u-picardie.fr (J. Saint). www.elsevier.com/locate/ssi Solid State Ionics 176 (2005) 189 – 197