Raman and Fourier transform infrared measurements for LixM0.03Fe0.97PO4, M=Cr, Cu, Al, Ti were performed. The spectra for delithiated samples to a low content of lithium extraction are practically the same as those of LiFePO4. For a high content of lithium extraction, the spectra repeat that of FePO4. In the case of the Li0.11FePO4 oxide the spectra cannot be reproduced just by the superposition of the end member profiles. An additional broad band contribution is found in both Raman and infrared spectra probably due to a disordered structure present in the mixture. This suggests that the well accepted two-phase model for the delithiation process in LiFePO4 is incorrect. The model should be revised to include the new phase as detected here for a particular level of lithium extraction close to that of complete oxidation of the Fe ions to Fe (c) 2006 Elsevier B.V. All rights reserved.