Electrochimica Acta 48 (2003) 4205–4211 Long-term cyclability of nanostructured LiFePO 4 Pier Paolo Prosini a,b,∗ , Maria Carewska a , Silvera Scaccia a , Pawel Wisniewski a , Mauro Pasquali b a ENEA, IDROCOMB, C.R. Casaccia, Via Anguillarese 301, Rome 00060, Italy b Dipartimento ICMMPM, Facoltà di Ingegneria, Università di Roma “La Sapienza”, Rome, Italy Received 4 February 2003; received in revised form 6 May 2003 Abstract Amorphous LiFePO 4 was obtained by lithiation of FePO 4 synthesized by spontaneous precipitation from equimolar aqueous solutions of Fe(NH 4 ) 2 (SO 4 ) 2 ·6H 2 O and NH 4 H 2 PO 4 , using hydrogen peroxide as oxidizing agent. Nano-crystalline LiFePO 4 was obtained by heating amorphous nano-sized LiFePO 4 for different periods of time. The materials were characterized by TG, DTA, X-ray powder diffraction, scanning electron microscopy (SEM) and BET. All materials showed very good electrochemical performance in terms of energy and power density. Upon cycling, a capacity fading affected the materials, thus reducing the electrochemical performance. Nevertheless, the fading decreased upon cycling and after the 200th cycle the cell was able to cycle for more than 500 cycles without further fading. © 2003 Elsevier Ltd. All rights reserved. Keywords: Lithium iron phosphate; Cyclability; Lithium battery 1. Introduction Lithium iron phosphate (LiFePO 4 ), with a phospho-olivine structure is emerging as a promising cathode for lithium-ion batteries. Since the first report on the electrochemical prop- erties of this material, due to Padhi et al. [1], a large number of research groups are devoted to improving its electrochem- ical properties. Padhi showed that lithium can be chemi- cally extracted from the structure thus leaving a new phase, FePO 4 , isostructural with heterosite, Fe 0.65 Mn 0.35 PO 4 . The electrochemical lithium extraction proceeds via a two-phase process and the FePO 4 framework of the ordered olivine LiFePO 4 is retained with minor adjustments. Nevertheless, the electrochemical insertion/extraction of lithium con- ducted at a specific current as low as 2.1 mA g -1 (C/81 rate), was limited to about 0.6 Li per formula unit. A “ra- dial model” for the lithium motion was proposed to explain the poor electrochemical performance of the material. This was associated with a diffusion-limited transfer of lithium across the two-phase interface. Ravet et al. [2] renewed interest in LiFePO 4 , reporting a new synthetic route leading to electronically conductive par- ∗ Corresponding author. Tel.: +39-06-3048-6768; fax: +39-06-3048-6357. E-mail address: prosini@casaccia.enea.it (P.P. Prosini). ticles and outstanding electrochemical performance. They added sucrose during the synthesis of the material as a source of carbon, obtaining carbon-coated particles directly. In electrochemical cycling tests made at 80 ◦ C using a poly- mer electrolyte, almost the full theoretical capacity was re- ported for a cell discharged at a specific current as high as 170 mA g -1 . Andersson et al. [3] followed the electrochemical delithi- ation and subsequent relithiation of LiFePO 4 by in situ X ray diffraction and Mössbauer spectroscopy. They found “... that about 20–25% of LiFePO 4 remains unconverted and ... that this figure can be reduced by appropriate manipu- lation of particle size and particle-surface morphology”. In further work [4], they showed that the capacity during the first lithium extraction is higher than the capacity recovered during the following discharge cycle. Furthermore, the ca- pacity increases with temperature, supporting the notion that the diffusion of lithium within each particle is the limiting step. In addition to the “radial model”, they proposed a “mo- saic model” to explain both the source of the first-cycle ca- pacity loss as well as the poor electrochemical performance of the material [5]. Similar results were obtained by our group [6]. We synthesized LiFePO 4 in the presence of high surface area carbon-black, added to the precursors before the formation of the crystalline phase. SEM micrographs confirmed that 0013-4686/$ – see front matter © 2003 Elsevier Ltd. All rights reserved. doi:10.1016/S0013-4686(03)00606-6