Structural change of the LiMn,O, spinel structure induced by extraction of lithium Kiyoshi Kanamura,* Hidetoshi Naito, Takeshi Yao and Zen-ichiro Takehara Division of Energy and Hydrocarbon Chemistry, Graduate School of Engineering, Kyoto University, Yoshida-honmachi, Sakyo-ku, Kyoto 606-01, Japan A structural change of the Li,Mn,O, spinel induced by extraction of lithium was investigated using the Rietveld refinement method for its X-ray diffraction pattern change. Two cubic phases of the spinel Li,Mn,04 were observed in the range 0.5 >x >0.13 and their lattice parameters were found to decrease with decreasing x. If both phases were thermodynamically stable, the lattice parameters should not change during the extraction of lithium. Therefore, our X-ray diffraction (XRD) patterns suggest the destruction of the rigid Li,Mn,O, spinel structure which provides a high reversibility for the insertion and extraction of lithium.The possible mechanisms corresponding to this structural degradation are the compositional change of Mn or 0 atoms in Li,Mn,O, through the extraction of lithium. On the other hand, the separation of Mn2+ ions from Li,Mn,04 (0.5>x >0.13) was detected by electron paramagnetic resonance (EPR). From these results, it is concluded that Li,Mn,O, decomposes through the extraction of lithium to form Mn2+ compounds as a separate phase. Various kinds of transition-metal oxides have been investigated Japan; Mo-Ka radiation) equipped with a curved-crystal as materials for lithium insertion and extraction processes, which are very important electrochemical reactions for active materials in rechargeable lithium batteries;lP8 for example, LiCoO, has been utilized in a practical rechargeable lithium battery.'-* Recently, alternative cathodes, in particular the LiMn,O, spinel, have also been investigated as good candi- dates for electrochemical lithium insertion materials. Since the electrochemical characteristics of LiMn,O, depend on its crystal nature, size and shape, various kinds of preparation methods have been developed to improve the characteristics of the LiMn,O, spinel.'^^.^-^^ Fundamental studies on this structure have been performed using X-ray diffraction (XRD) and electrochemical methods to examine the electrochemical characteristics associated with its structural changes.1*2,9*10.18v21 Several studies have proposed that the structural changes take place when lithium ions are inserted into or extracted from LiMn204.These changes are: the phase transition of Li,Mn,O, from a cubic phase to a tetragonal phase in the region from x=2 to x =1, and the homogeneous phase reaction of the Li,Mn,O, spinel with a continuous lattice parameter change in the region from x= 1 to In the region x<O.5, the presence of two cubic phases has been proposed, but the detailed structural change has not yet been reported. In this study, the Rietveld method was used to refine the structural changes of Li,Mn,O, in the region from LiMn,O, to Lio.13Mn,0,. Experiment a1 Several different preparation methods of LiMn,O, have been rep~rted;'.~.~-~'Li,C03 and MnO, were the most commonly used starting materials.Recently, Momchilov et ul. suggested that LiMn,O, prepared from LiNO, and chemical manganese dioxide (CMD) has a large specific surface area and a high electrical conductivity.21 This preparation method was adopted in this study in order to realize a more uniform electrochemical reaction and to obtain equilibrium states of Li,Mn,O,. A mixture of LiNO, and CMD in a mole ratio of 1:2 was heated at 450°C for 36 h and then at 750°C for 72 h in air, according to ref. 21. The crystal structure of the product was determined by XRD using Mo-Ka (or Cu-Ka) radiation, to confirm the formation of a well characterized LiMn,O,.For the Rietveld analysis, a powder XRD pattern was collected using a Model RAD-B powder X-ray diffractometer (Rigaku Co., Tokyo, graphite monochromator at the High Intensity X-ray Laboratory, Kyoto University. The step-scanning technique with steps of 28=0.01" and a stepping time of 15 s was used over the range 5"<28<60". The Rietveld calculation was performed on the vector processor (Cray Y-MP2E/264) at the Institute for Chemical Research, Kyoto University, by using the 'Rievec' computer program for Rietveld refinement.,,-,' The atomic ratio of Li :Mn was determined by atomic adsorp- tion spectroscopy (AAS) to be 1:2. Electrochemical extraction of lithium from LiMn204 was performed using a pellet-type electrode to prepare several Li,Mn,O, samples (0.13 <x ~0.5).The cathode pellet was prepared by pressing a mixture (100 mg) of the prepared LiMn204, and acetylene black in a mass ratio of 70: 30 at a pressure of 5 x lo5 Pa for 15 min. The diameter of the cathode pellet was 1.0 cm. Propylene carbonate contain- ing 1.0 mol dm-3 LiClO, was used as the electrolyte. Lithium metal was used for the reference and counter electrodes. After the electrochemical extraction, the pellet electrode was washed with pure propylene carbonate and then with tetrahydrofuran. The prepared sample was dried in vacuum for 1 h. The current used for the electrochemical extraction of lithium was 0.03 mA, which was low enough to attain an equilibrium state within the sample.After the electrochemical extraction process, the samples were kept in electrolytes for several days until the electrode potential became constant. The crystal structures of the Li,Mn,O, samples were also analysed with the Rietveld method to determine the lattice parameters. Electron paramagnetic resonance (EPR) was also used to analyse the reaction products of the extraction of lithium from LiMn,O,; the EPR signal was obtained at room temperature using JES-RE equipment (JEOL). An X-band microwave was used, with an output power of 1mW. The amplitude of the magnetic field modulation was 0.5 mT and its frequency was 100 Hz. The response time was 0.3 s. Results and Discussion Fig. 1 shows the XRD patterns of Li,Mn20, (x= 1.0, 0.5, 0.3 and 0.13) in the 28 region from 43" to 47" (Cu-Ka). At x= 1.0, 0.5 and 0.13, only a single peak was observed at 28=43.87, 44.34 and 44.93", respectively.At x=0.3, two peaks were observed, at 20 =44.55 and 44.82". When a two-phase reaction takes place during the extraction of lithium, the peak positions should not change with changing x value in Li,Mn,O,. J. Muter. Chem., 1996, 6(1), 33-36 0.3 Phase I + Phase I1A I 'i 0.25 x=0.5 v) Y5 0.2 0 0 $I 0.15 \ .-a v)c a, 0.1 c. c.-0.05 0 40 42 44 46 48 50 281degrees (Cu-Ka) Fig. 1 XRD patterns of Li,Mn,O, (x= 1.0, 0.5, 0.3 and 0.13) prepared by the electrochemical extraction of lithium from LiMn,O, using the LiMn,O, disc cathode with an apparent electrode area of 0.785 cm2 under galvanostatic conditions at 0.03 mA in propylene carbonate containing 1.0 mol dm-3 LiClO,.LiMn,O, was prepared by heating a mixture of LiN03 and CMD. XRD patterns were obtained with an X-ray diffractometer using Cu-Ka radiation. However, both peak positions in the XRD pattern at x=O.3 were different from those of Li0.,Mn,O4 and Li0.,,Mn,O4, which are two different single phases of Li,Mn,O,. These XRD patterns suggest that the electrochemical reaction of LiMn,O, is not explained by the extraction of lithium via a two-phase reaction. More detailed analyses on the XRD patterns were performed using the Rietveld method to clarify the structural change undergone by Li,Mn204 (0.5 >x > 0.13). The lattice parameters, oxygen positional parameters, frac- tions of two cubic phases, Rwp, R,, RF and R, values for the Rietveld analysis on the XRD patterns of Li,Mn,O, are summarized in Table 1.Fig. 2 shows the observed and simu- lated XRD patterns of Li0.,Mn204, which consist of two different phases. The observed pattern agreed well with the simulated pattern obtained from the Rietveld method; such good agreement between the observed and simulated patterns was also obtained for the other x values in Li,Mn20,. The LiMn,O, prepared in this study possessed a cubic structure with Fd3m symmetry, and with the lattice parameter calculated as 0.824219 nm. The electrochemical extraction of lithium from this LiMn,O, was performed under galvanostatic conditions at 0.03mA.The crystal structure of Li,Mn20, in the region from x= 1.0 to x=O.5 was also determined to be cubic. The lattice parameter decreased from 0.824219 nm to 0.815041 nm during the extraction of lithium from the LiMn,O,. This structural change indicates that the electrochemical extraction 5 10 20 30 40 50 2t)(Mo-Ka)/degrees Fig.2 XRD pattern for Li,.,Mn,O, prepared under the same conditions as those in Fig. 1; the solid curve is the simulated pattern and the broken curve is the observed one. The XRD pattern was obtained with an X-ray diffractometer using Mo-Ka radiation. of lithium from LiMn,O, in the region from x= 1.0 to x=O.5 proceeds via a homogeneous phase reaction with a decrease in the host matrix size. This result was in good agreement with those reported previously,2,18-20 but we have now provided more precise lattice parameter changes of Li,Mn,O, in the region from x=1.0 to x=O.5.Fig. 3 shows the electrode potential change during the electrochemical extraction of lithium from LiMn,O, in the region from x = 1 to x =0.1 at 0.03 mA in propylene carbonate containing 1.0 mol dmP3 LiClO,. When homogeneous lithium extraction takes place, the free energy of Li,Mn,O, changes with changing x value, which causes the electrode potential change. This is typical behaviour for a solid solution. The electrode potential change in the region from x = 1 to x =0.5 was in good agreement with the reaction scheme for the homogeneous extraction of lithium from LiMn,O,.Similar results have been reported elsewhere.2,18-20 When the electro- chemical extraction of lithium was performed in the region of 0.5 >x, two cubic phases were observed in the XRD pattern. The lattice parameter of one cubic phase (phase I) was slightly larger than that of the other phase (phase 11). The relative amounts of the two phases changed according to the amount of lithium extracted from LiO.,Mn2O4, as shown in Table 1. At x =0.13, phase I almost disappeared and phase I1 only was observed, as shown in Fig. 1, suggesting that the extraction of lithium from Li,.,Mn,O, proceeds via a two-phase reaction in Table 1 Lattice parameters, oxygen positional parameters and R-factors of the Li,Mn,O, spinel (space group Fd3m) obtained from the Rietveld refinement method for the X-ray diffraction patterns lattice parameters (phase I) (phase 11) oxygen positional (phase I) parameters (phase 11) RWP RP RF (phase I) (phase 11) RB (phase I) (phase 11)4, (phase 1)(phase 11) fraction of phase I fraction of phase I1 x= 1.0 x=0.5 x =0.4 x =0.3 x=0.15 0.8242 19 ( 8) 0.38701 (6) 0.0895 0.815041( 1) 0.3871 1( 13) 0.0775 0.8 12510( 2) 0.810254( 3) 0.38730( 24) 0.38750( 15) 0.0563 0.8 10779( 4) 0.807956( 2) 0.38750( 35) 0.38765( 50) 0.0548 0.804834( 9) 0.38770(7) 0.0781 0.1181 0.0875 0.0738 0.0726 0.0936 0.0393 0.0309 0.0273 0.0256 0.0305 0.0341 0.0319 0.0372 0.0392 0.0341 0.03 19 0.0310 0.0199 0.0249 0.17( 1) 0.56(2) 0.42 ( 1) 0.30(2) 0.697 0.52(8) 0.14( 8) 0.412 0.57( 1) 0.303 0.588 Rwp, R-weighted pattern; R,, R-pattern; R,, R-structure factor; RB, R-Bragg factor; Bsm, temperature factor.34 J. Muter. Chem., 1996, 6(l), 33-36 4.50 4.25 h ?-2 -! Lu 4.00 3.750 10 0.20 0.40 0.60 0.80 1.00 x in Li,Mn204 Fig. 3 Electrode potential change of the LiMn20, disc cathode with an apparent electrode area of 0.785 cm2 during the charge process at 0.03 mA in propylene carbonate containing 1.0 mol dm-3 LiClO, in the region of 1.O >x >0.1 which the phases are Li,.,Mn,O, and Li,~,,Mn,O,. If the electrode reaction proceeds uiu a two-phase reaction without any lattice parameter changes, the electrode potential should be independent of the concentration of lithium in Li,Mn,O,. The flat potential change (no free-energy change) has been reported in several previous paper^.'.'^^*^^ This is typical thermodynamic behaviour for a two-phase reaction.Our results also show a relatively flat potential change, which differs from that in the region of 1>x >0.5. However, from our XRD pattern at x=O.3 in Li,Mn,O,, it can be seen that both peaks corresponding to the two cubic phases move with changing x in Li,Mn,O,. This result is inconsistent with the flat potential curve in this region. Note that a flat potential change (no free-energy change) is not necessary for a two-phase reaction because, when some unexpected physical prop- erty changes occurs during the extraction of lithium from Li,.,Mn,O,, the electrode potential change can be compen- sated by the unexpected physical property change. In the case of LiMn204, the flat potential change might be obscured by some physical property change which is related to our XRD pattern change.Since the Rietveld refinement method enables one to deter- mine the precise lattice parameter, the structural change of Li,Mn,O, in the region of OS>x was analysed using this method. Two phases appeared in the region from x=O.5 to x=O.13 with cubic structures, and their space groups were assigned to Fd3rn, as shown in Table 1. The lattice parameters of phase I at x=0.4 and 0.3 were 0.812510 and 0.810779 nm, respectively; those of phase I1 at x=0.3 and 0.15 were 0.807956 and 0.804834nm7 respectively.From these results, it is clear that the lattice parameter decreases during the electrochemical extraction of lithium, indicating that the host matrix of Li,Mn,O, is influenced by the extraction of lithium. The decrease in the lattice parameters of the two cubic phases is probably explained by a chemical composition change in the host matrix of Li,Mn,O,. If both cubic phases have different chemical compositions from Li,,,,Mn,O, and Lio.,Mn,O,, the structural change and electrode potential change in the region from x=OS to x=0.13 can be understood. Such chemical composition changes must result in the formation of a third acetylene black/ "VJ \ I I I I 3 178 278 378 478 magnetic field / mT Fig. 4 EPR signals for (a) Lio.3Mn20, and (b) LiMn20, phase, corresponding to the degradation of the host matrix.No third phase could be detected by XRD in this study. The third phase may have an amorphous nature, or it may dissolve in an electrolyte solution. If this is correct, the third phase prevents the reversible insertion and extraction of lithium. Since the lattice parameter change is very small, it can be expected that the amount of the third phase formed during the first extraction of lithium is very small. Fig. 4 shows the EPR signals for (a) Li,,,Mn,O, and (b) LiMn,04 at room temperature. A broad peak was observed and was assigned to the electron spins of Mn3+ or Mn4+ ions in LiMn,O,. Simultaneously, six sharp signals were observed. The same signal pattern were observed for an Mn2+/Mg0 marker sample, indicating that Mn2+ ions are formed in Lio.,Mn,O,.This result is a direct indication that the reaction of Li,Mn,O, in the region x=O.5-0.13 cannot be explained by a two-phase reaction. The EPR signal corresponding to Mn2+ ions shows that Mn2+ ions are present in a phase separated from the two cubic systems. Conclusion The Rietveld refinement method for the X-ray diffraction patterns proved the existence of a very fine structural change of LiMn,O,. We can conclude that an irreversible structural change takes place during the extraction process in the region of x <0.5. This result was also confirmed by EPR spectrometry. This study was supported by a Grant-Aid for Scientific Research from the Society for Promotion of Space Science and a Grand-in-Aid for Scientific Research (C) from the Ministry of Education Science and Culture of Japan (no.06650745). Computation time was provided by the Supercomputer Laboratory, Institute for Chemical Research, Kyoto University. References J. M. Tarascon and D. Guyomard, Electrochim. Acta, 1993, 38, 1221. T. Ohzuku, M. Kitagawa and T. Hirai, J. Electrochem. Soc., 1990, 137,769. T. Ohzuku, A. Ueda and M. Nagayama, J. Electrochem. Soc., 1993, 140,1862. J. R. Dahn, U. von Sacken, M. W. Juzkow and H. Al-Janaby, J. Electrochem. Soc., 1991, 138,2207. J. N. Reimers and J. R. Dahn, J. Electrochem. Soc., 1992,139,2091. E. Plichta, S. Slane, M. Uchiyama and M. Salomon, J. Electrochem.Soc., 1989, 136, 1865. M. Antaya, J. R. Dahn, J. S. Preston, E. Rossen and J. N. Reimers, J. Electrochem. SOC., 1993, 140, 575. T. Ohzuku and A. Ueda, J. Electrochem. Soc., 1994,141,2972. J. Muter. Chern., 1996, 6(l),33-36 9 J. M. Tarascon, E. Wang and F. K. Shokoohi, J. Electrochem. SOC., 20 M. M. Thackeray, W. I. F. David, P. G. Bruce and J. B. 10 1991,138,2859. D. Guyomard and J. M. Tarascon, J. Electrochem. SOC., 1992, 21 Goodenough, Muter. Res. Bull., 1983,18,461. A. Momchilov, V. Manev and A. Nassalevska, J. Power Sources, 139, 937. 1993,41,305. 11 J. M. Tarascon and D. Guyomard, J. Electrochem. SOC., 1991, 138,2864. 22 23 H. M. Rietveld, J. Appl. Crystallogr., 1969,2, 65. Y. Oka, T. Yao and N. Yamamoto, J. Solid State Chem., 1990, 12 J. C. Hunter, J. Solid State Chem., 1981,39, 142. 86, 1 16. 13 W. J. Macklin, R. J. Neat and R. J. Powell, J. Power Sources, 1991, 24 Y. Oka, T. Yao and N. Yamarnoto, J. Solid State Chem., 1990, 34,39. 89, 372. 14 M. H. ROSSOUW, A. de Kock, L. A. de Picciotto, M. M. Thackeray, W. I. F. David and R. M. Ibberson, Muter. Res. Bull., 1990,25, 173. 25 26 T. Yao, Y. Oka and N. Yamamoto, Mater. Res. Bull., 1992,27,669. Y. Oka, T. Yao, N. Yamamoto, Y. Ueda and A. Hayashi, J. Solid 15 L. Li and G. Pistoia, Solid State Ionics, 1991,47,231. State Chem., 1993, 105,271. 16 L. Li and G. Pistoia, Solid State Ionics, 1991,47,241. 27 T. Yao, Y. Oka and N. Yamamoto, J. Mater. Chem., 1992,32,331. 17 18 G. Pistoia, G. Wang and C. Wang, Solid State Ionics, 1992,58,285. A. Mosbah, A. Verbaere and M. Tournoux, Muter. Res. Bull., 1983, 18, 1375. Paper 5/04666A; Received 17th July, 1995 19 M. M. Thackeray, P. J. Johnson, L. A. de Picciotto, P. G. Bruce and J. B. Goodenough, Muter. Res. Bull., 1984,19, 179. 36 J. Mater. Chem., 1996, 6(l), 33-36