Available online at www.sciencedirect.com Journal of Power Sources 179 (2008) 754–762 High-throughput studies of Li 1-x Mg x/2 FePO 4 and LiFe 1-y Mg y PO 4 and the effect of carbon coating Matthew R. Roberts a , Girts Vitins b , John R. Owen a,∗ a School of Chemistry, University of Southampton, Southampton, SO17 1BJ, UK b Now at QinetiQ, Haslar Road, Gosport, PO12 2AG, UK Received 25 September 2007; received in revised form 3 December 2007; accepted 10 January 2008 Available online 30 January 2008 Abstract A two-dimensional sample array synthesis has been used to screen carbon-coated Li (1-x) Mg x/2 FePO 4 and LiFe (1-y) Mg y PO 4 powders as potential positive electrode materials in lithium ion batteries with respect to x, y and carbon content. The synthesis route, using sucrose as a carbon source as well as a viscosity-enhancing additive, allowed introduction of the Mg dopant from solution into the sol–gel pyrolysis precursor. High-throughput XRD and cyclic voltammetry confirmed the formation of the olivine phase and percolation of the electronic conduction path at sucrose to phosphate ratios between 0.15 and 0.20. Measurements of the charge passed per discharge cycle showed that the capacity deteriorated on increasing magnesium in Li (1-x) Mg x/2 FePO 4 , but improved with increasing magnesium in LiFe (1-y) Mg y PO 4 , especially at high scan rates. Rietveld-refined XRD results on samples of LiFe (1-y) Mg y PO 4 prepared by a solid-state route showed a single phase up to y = 0.1 according to progressive increases in unit cell volume with increases in y. Carbon-free samples of the same materials showed conductivity increases from 10 -10 to 10 -8 S cm -1 and a decrease of activation energy from 0.62 to 0.51 eV. Galvanostatic cycling showed near theoretical capacity for y = 0.1 compared with only 80% capacity for undoped material under the same conditions. © 2008 Elsevier B.V. All rights reserved. Keywords: LiFePO 4 ; Lithium battery; Mg doping; High-throughput; Conductivity; Carbon coating 1. Introduction LiFePO4 is of great interest as a safe, environmentally accept- able positive electrode for lithium ion batteries [1,2], but suffers from a low intrinsic electronic conductivity. Several authors have successfully introduced a surface coating of pyrolytic carbon from sucrose [3] and other precursors [4–6] after syn- thesis as a means of enhancing the surface conductivity, and therefore the rate performance. Recent work in our labora- tory focused on a one-step synthesis method which includes sucrose in a pyrolytic sol–gel synthesis of LiFePO 4 from a mixed salt solution precursor [7]. In this work sucrose also acted as a viscosity-enhancing additive to suppress individ- ual crystal growth of precursor components during the initial drying process. The resulting uniformity of the elemental dis- tribution is an advantage in the synthesis of doped materials ∗ Corresponding author. Tel.: +44 2380 592184; fax: +44 2380 593781. E-mail address: jro@soton.ac.uk (J.R. Owen). by introducing a soluble form of the dopant into the precur- sor. Several papers have suggested metal ion doping as a method for improving performance [8–10]. Substituting Mg and other species on the Li site was reported by Chung et al. [8] to give a ∼10 8 fold increase in conductivity of LiFePO 4 and a greatly improved electrode performance. However, the interpretation of these results was questioned by Ravet et al. [11] who suggested that the improved conductivity may be a result of carbon residues rather than the effect of the dopant ion. To the best of our knowl- edge further investigation into the substitution of Li by Mg has been limited to just two further publications [12,13], which also report improved conductivity and performance. The first reported work on LiFe 1-y Mg y PO 4 materials was presented by Barker et al. [14] who successfully synthesized LiFe 0.9 Mg 0.1 PO 4 via a carbothermal reaction and reported “outstanding ionic reversibility”. Later several publications con- firmed the formation of this material under different synthesis conditions and reported improved capacity, conductivity and rate capability [15–19]. The explanation for these improvements var- 0378-7753/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.jpowsour.2008.01.034