J. MATER. CHEM., 1991, 1(4), 595-596 Synthesis and Properties of a New Family of Phases, Li,XYO,: X=Al, Ga; Y =Si, Ge Chnoong Kheng Leet and Anthony R. West University of Aberdeen, Department of Chemistry, Meston Walk, Aberdeen AB9 2UE, UK A family of new phases, Li,XYO,, has been synthesised by solid-state reaction. They are all thermodynamically stable, melting incongruently in the range 1165-1250 "C. Electrically, they are insulators and show very low electronic conductivity at high temperatures. Unit cells are orthorhombic, typically 5.4 A x 15.8 A x 4.9 A, with space-group Pnam. The phase diagram of the binary join Li,GeO,-LiAIO, is given, showing the presence of incongruently melting Li,AIGeO, and a limited range of LiAIO, solid solutions. Keywords: Lithium aluminosilicate; Lithium aluminogermanate During a study of new lithium-ion conducting ceramics in the system Li4Si04-LiGaSi04,1*2 the presence of an additional phase of unknown structure was indicated from the X-ray powder diffraction data.A subsequent phase dia- gram study of the system Li20-Ga203-Si02 showed its formula to be approximately Li14Ga6Sis026, although it was not obtained as a single phase.3 Follow-up work on the Al- and Ge-containing analogous systems has shown that similar phases are also formed in these systems. We have now confirmed their stoichiometry to be Li3XYOs: X=Al, Ga; Y=Si, Ge. These results are presented here, together with crystallographic, phase diagram and electrical property data for the new phases. Experimental Starting materials were reagent-grade Li2C03, A1203, GeO,, and Ga203.All were used as received since tests showed that drying was not necessary. Reagents were weighed out, mixed using an agate mortar and pestle, and using acetone to form a paste, dried and heated in Au- or Pt-foil boats at 700 "C for 2 h to eliminate CO,. The temperature was then raised to 950-1 100 "C,depending on composition. Trial and error tests on a considerable number of compositions were required to find the optimum conditions (i.e. temperature and time of firing) for reaction to go to completion. At the outset, the major objective and difficulty, was to determine the composition(s) of the new phases. Initial experi- ments were carried out in the system Li,0-A1,03-Ge0,; various compositions in the region of 'Li14A16Ge5026' were prepared, reacted and analysed by X-ray powder diffraction.Gradually, it became clear that the true composition was displaced somewhat from the supposed composition and may be close to Li3A1Ge05. It was, however, very difficult to prepare a phase-pure sample for any composition, by using a single-step reaction of the three oxide/carbonate reagents. Since the composition Li3A1Ge05 may be regarded as an equimolar mixture of Li,Ge03 and LiAlO,, stock quantities of these two phases were prepared and reacted. Using this procedure, a phase-pure sample of the new phase Li,AlGeO,, was obtained. For all subsequent work, on this and the other systems, a similar two-step reaction procedure was used.Phase analysis was carried out by X-ray powder diffraction with a Stoe Stadi psd diffractometer, Cu-Kcc, radiation. For conductivity measurements, gold-paste electrodes were t Permanent address: Chemistry Department, Universiti Pertanian Malaysia, 43400 Serdang, Selangor, Malaysia attached to opposite pellet faces, dried, fired at 200-600 "C, and the assembly loaded into a conductivity jig; a.c. impedance measurements were made using two instruments: Solartron 1250/1286 for the frequency range 0.65 Hz-65 kHz; Hewlett- Packard 4192A RF bridge for the range 100 Hz-13 MHz. Melting points were determined approximately by placing pellets in a furnace at various preset temperatures and observing whether or not melting occurred. Results and Discussion Using the procedures described above, pure samples of the three new phases, Li3A1GeOs, Li3GaGe05 and Li3GaSiOS, were prepared. Reaction conditions, for the mixtures of Li2Y03 and LiXO,, were 2 days at 1050 "C, 2 days at 1050 "C and 2 days at 1100 "C (pelleted samples) for the above new phases, respectively.The X-ray powder diffraction data were indexed success- fully, by trial-and-error Visser methods. An abbreviated listing is given in Table 1; a more complete set is available from the authors and will be submitted to the JCPDS file. Unit-cell data are summarised in Table 2; systematic absences indicate the space group to be Pnarn. The fourth phase in this group, Li3A1Si05, has been synthe- sized, but never as a pure sample.This phase was detected Table 1 Indexed X-ray powder diffraction data Li,AlGeO , Li,GaGeO, Li,GaSiO, dobs/A I dobs/A I dobs/A I ~~ 020 7.9297 13.5 7.8305 58.3 110 5.1084 17.1 5.I728 9.2 011 4.6570 4.3 4.71 11 6.9 4.6282 43.6 120 4.4592 100.0 4.5031 100.0 4.4435 100.0 040 3.9583 13.1 3.96 11 4.5 130 3.7504 5.7 031 3.5796 85.5 3.6068 90.4 3.5506 76.4 111 3.5263 39.0 3.5666 26.7 3.51 10 4.9 121 3.2896 35.6 3.3236 36.5 3.2742 29.6 140 3.1905 1.5 3.1662 12.8 131 3.0106 8.6 2.9654 21.4 150 2.7302 49.0 2.7445 49.1 2.7038 90.7 200 2.6985 25.7 210 2.6608 2.9 2.6586 12.4 060 2.6377 5.0 2.5488 4.6 002 2.4362 43.4 2.4642 49.7 2.4227 52.5 230 2.4023 11.5 2.4277 15.0 2.3960 8.O 151 2.3819 26.1 2.3974 44.8 2.3620 43.7 201 2.3607 13.8 Table 2 Unit-cell data Li,AlGeO, 5.397(1) 15.825(1) 4.873(1) 416.17(6) Li,GaGeO, 5.467(1) 15.861(2) 4.929(1) 427.4q4) Li,GaSiO, 5.398(1) 15.633(4) 4.845(1) 408.8(2) after heating at 1050-1 100 "C for up to 4days.It is possible that longer heating times would yield a single-phase product but significant lithia loss by volatilisation may also become a problem, thereby making it difficult to obtain a single-phase sample. Further work is required, probably using alternative reaction pathways, in order to obtain single-phase samples of Li3A1Si05. In order to obtain more information on the thermal stability and melting behaviour of the new phases, a brief study of the phase diagram Li,GeO3-LiA1O2 was made; the results are shown in Fig.1. Melting information was obtained from the appearance of pelleted samples after they had been placed inside a furnace for a few minutes at various preset tempera- tures. From this, it was clear that Li3A1Ge05 melted incongru- ently to LiAlO, and liquid at 1165f10 "C and that a eutectic existed between Li3A1Ge05 and Li,Ge03 at 11 10 & 10 "C. To confirm the incongruent melting of Li,A1Ge05, a sample was heated briefly at 1180 "C and quenched. X-Ray diffraction of the product showed a rather poor-quality pattern with evidence of significant amounts of LiA102, together with smaller amounts of Li3A1Ge05 and Li2Ge03. This showed LiA10, to be present at the temperature of heating, together with liquid which subsequently crystallised to give a mixture of Li,Ge03 and Li3A1Ge05 on cooling.There was no evidence for any solid-solution formation with either Li3A1Ge05 or Li,Ge03, but LiAlO, does appear to form a limited range of solid solutions. The phase diagram, Fig. 1, may be regarded as a true binary join within the ternary system Li,0-A1203-Ge02. This is I I I1 I I I1 1 II Li,GeO, 20 40 60 80 Li A102 LiAIO, composition (mol%) Fig. 1 Phase diagram for the join Li,Ge03-LiA10,; compositions studied are marked 0 J. MATER. CHEM., 1991 VOL. 1 I I I I I I 0.8 1.0 1.2 1.4 1.6 1.8 103 KIT Fig.2 Arrhenius conductivity plot for Li,AlGeO, (a) and Li,Ga- GeO, (b) Activation energies are 1.02 and 1.08 eV, respectively, (a) 0,heating; @, cooling; (b) @, heating; 0,cooling. because all the phases that are present on the diagram have compositions which also lie on this join. The other two new phases, Li3GaGe05 and Li3GaSi05, also appear to melt incongruently, in similar fashion, at 1250&10 and 1210&10 "C, respectively. The phase diagrams of their corresponding Li,Y03-LiGa02 joins have not been determined, however, but almost certainly will be more com- plex owing to the appearance of phases such as Li5GaSi208.1*2 Conductivity data for two of the new phases are presented as Arrhenius plots in Fig. 2. The data are very similar, giving linear plots over the temperature range 300-900 "C with activation energies of 1.05 eV. The conductivity values are very low, 2x10-' Q-'cm-' at 300°C. The conducting species appear to be electronic since no evidence of electrode- polarisation effects was seen in the impedance data., This is therefore in complete contrast to the Al, Ga-doped Li4Si04/ Li,GeO, materials, which exhibit high lithium-ion conduc- tivity, lop6 0-' cm-' at 25 0C.1*596. References 1 P. Quintana, F. Velasco and A. R. West, Solid State lonics, 1988, 34, 149. 2 P. Quintana and A. R. West, J. Solid State Chem., 1989, 81, 257. 3 P. Quintana and A. R. West, Br. Ceram. Trans. J., 1989, 88, 17. 4 J. T. S. Irvine, D. C. Sinclair and A. R. West, Adv. Muter., 1990, 2, 132. 5 A. Garcia, G. Torres-Treviiio and A. R. West, Solid State lonics, 1990,40/41, 13. 6 C. K. Lee and A. R. West, in preparation. Paper 1/00564B; Received 6th February, 1991