Fundamental Insights into the Electrochemical Performance of Chromium Doped LiCoPO4

High voltage cathode materials are promising technologies for increasing the energy density and thus increasing the range of electric vehicles. LiCoPO₄ is known as a possible high voltage material, which is electrochemically active at 4.8 V vs. Li/Li⁺ with a theoretical capacity of 167 mAh/g ^[1], corresponding to a theoretical energy density of ~800 Wh/kg. Unfortunately, this material suffers of bad rate capability because of its very low ionic and electronic conductivity.

To increase the ionic diffusion of the material, dopings of LiCoPO₄ with 3%_at Cr on the Li site and on the Co site are proposed. The syntheses are done by solid state method but different processes are used to dope on the Li site, nominally Li_0.97Cr_0.03CoPO₄ (Yang et al ^[2] synthesis method) or on the Co site, nominally LiCo_0.97Cr_0.03PO₄ (Wolfenstine et al ^[3]synthesis method). At the end of the synthesis, a ball milling step is added to reduce the particle size in order to improve the rate capability. Two different milling conditions are set: “soft” (400 rpm) and “hard” (1100 rpm). The first one seems to not damage the structure of the materials whereas the other one destroys the crystallinity and some impurities are identified by XRD. An additional annealing of the material ball milled under “hard” condition allows re-forming the phospho-olivine structure with good crystallinity. The synthetized materials are characterized by XRD, ICP and XAS (Cr K-edge and Co K- edge). Based on XAS analysis, no difference is noted in the Cr environment for the two syntheses and based on XRD, no side phase is detectable for all the samples. The Cr content of 3% is confirmed by ICP.

The discharge capacities of the various samples are presented Figure 1. It can be seen that for a sample ball milled at 400 rpm, the Cr doped sample (LiCo_0.97Cr_0.03PO₄; green bar in Fig. 1) is showing the best rate capability, whereas for a sample ball milled at 1100 rpm, it is the undoped LiCoPO₄ (black bar in Fig. 1). The higher capacity and rate capability of the latter compared to the undoped LiCoPO₄ ball milled at 400 rpm (red bar in Fig. 1) is due to its much smaller particle size and its higher specific surface area (»25 vs. »1 m²/g). Additionally, an unexpected huge difference is observed for the two chromium doped materials: LiCo_0.97Cr_0.03PO₄ (green bar in Fig. 1) has much higher capacity and rate capability than Li_0.97Cr_0.03CoPO₄ (orange bar in Fig. 1), even though XAS analysis suggests that the chromium dopant site is identical in both samples despite their different nominal dopant sites. Considering the recent work by Lee et al.⁴ on LiMO₂ active materials, an excess of lithium in the material LiCo_0.97Cr_0.03PO₄ compared to Li_0.97Cr_0.03CoPO₄could be responsible for the better capacities.

Acknowledgements: This work is financially supported by BMW.

References:

[1]   N. Bramnik, K. Bramnik, T. Buhrmester, C. Baehtz, H. Ehrenberg, H. Fuess, J. Solid State Electrochem(2004) 8, 558

[2]   I.C. Yang, H.H. Lim, S.B. Lee, K. Karthikeyan, V. Aravindan, K.S. Kang, W.S. Yoon, W.I. Cho, Y.S. Lee, Journal of Alloys and Compounds(2010) 497, 321-324.

[3]   J. Wolfenstine, U. Lee, B. Poese, J-L. Allen, J. Power Sources(2005) 144, 226

[4]   J. Lee, A. Urban, X. Li, D. Su, G. Hautier, G. Ceder, Science (2014) DOI:10.1126/science.1246432

Figure 1: First-cycle discharge capacities of doped and undoped lithium cobalt phosphate materials using ball milling (BM) under different conditions. Obtained from a rate capability test (3 cycles C/10, C/5, C/2, 1C, and 2C) in Swagelok cells in 1M LiPF₆ EC:DMC (active material loadings of ca. 4 mg/cm²).