25 Conducting Solids Covering Ionic and Electronic Conductors By P. R. SLATER School of Chemistry University of Birmingham Edgbaston Birmingham B 15 2TT UK 1 Introduction There have been a number of significant advances in 1994 particularly in the area of superconductivity. Research into intermetallic superconductors has been revived with the discovery of new quaternary boro-carbide and -nitride alloys showing Tc’sas high as 23 K (equalling the longstanding record for an intermetallic held by Nb,Ge). In the area of oxide superconductors several recent results have indicated the importance of flat CuO planes to obtain the maximum T with an apparent correlation between the distortion of these planes and the lowering of T,. A similar correlation involving Ni,B layers which appear to be a common feature in these phases also appears to exist in the intermetallic boro-carbides and -nitrides.The field of ionic conductivity continues to be dominated by research into new materials for solid oxide fuel cells and rechargeable Li battery applications with the latter area attracting the most interest as in previous years. A review on fullerene chemistry appears elsewhere in this volume. 2 High Temperature Oxide Superconductors Hg Cuprate Superconductors.-A considerable amount of effort has been expended in the area of Hg cuprate superconductors since they represent the phases with the highest superconducting transition temperatures T,. A problem with these materials has been the difficulty in synthesizing pure bulk samples and a considerable amount of work has been concerned with this dilemma with many subtle modifications of experimental conditions.Due to lack of space and the fact that even now the synthesis methods are far from perfect these studies are not included in this review. The problems associated with growing single crystals of HgBa,Ca,- 1C~n02n+ using a solid media high pressure technique have been overcome by using a gas-phase high pressure technique.’.2 The former method was limited to several hundred milligrams with no free space to grow single crystals. The use of an Ar gas atmosphere with a partial pressure of0 leaves free space for single crystal growth and allows large sample volumes (as large as several cm’). Using this method single crystals of Pb doped J.Karpinski,H. Schwer,I. Mangelschots K. Conder A. Morawski T. Lada and A. Paszewin Nature 1994 371,661. J. Karpinski H. Schwer I. Mangelschots K. Conder A. Morawski T. Lada and A. Paszewin Physica C 1994 234 10. 467 468 P. R. Slater HgBa,Ca,Cu,O,,+ (Hg-1234) of size up to 0.5 x 0.5 x 0.05mm3 with a high T onset of 129K have been prepared. Single crystals of the infinite layer compound CaCuO, without Sr doping have also been grown. Thin films of Hg,Ba,CalCU,06+ (Hg-1212) on a SrTiO substrate have been prepared by laser ablation starting with sintered pellets of target composition H~,B~,C~,CU,~,+,,~ or separate targets of HgO and Ba,Ca,Cu,05 The as-grown films were amorphous due to the necessity of using low substrate temperatures (< 250 "C) to incorporate Hg into the film.Annealing in a quartz tube with a Hg-1212 bulk sample was necessary in order to obtain a crystalline film. The resulting crystalline Hg-1212 films showed preferential orientation perpendicular to the (100) SrTiO substrate with superconducting transitions as high as 124 K. Gao et al. have re-examined the pressure dependence of T in the HgBa,Ca,-,Cu,02 + + system up to higher pressure^.^ They observed record Tc'sof 118K (n = l) 154 K (n = 2) and 164 K (n = 3). Moreover the partial substitution of Pb for Hg in HgBa,Ca,Cu,O,+ (Hg-1223) was found to suppress the pressure induced T enhancement suggesting that Hg plays an important role in these +, compounds. The effect ofpressure on the structure of HgBa,Ca,- 1Cun02n+2 (n = 1 2,3) has been examined by Hunter et al.using neutron diffraction. The compressibility along the c axis is nearly the same for all three compounds and is up to two times larger than that along the a axis. For n = 1 and 2 the Cu-0 apical bond distance showed the largest compressibility while for n = 3 it was the Hg-0 bond distance. A neutron diffraction study of the five Cu layer phase HgBa2Ca4Cu5OI2+ (Hg-1245) has been performed at room temperature and at 10K by Huang et aL7 As in other members ofthe homologous series all the CuO planes were found to be almost coplanar i.e. there is insignificant or no 'buckling' of the CuO planes. There was no evidence for the substitution of Hg by Cu or the presence of extra oxygen atoms associated with such a substitution.Hur et al. have reported enhancements in the critical current density and stability in air or 0 at high temperatures (> 400 "C) by partial substitution of Hg by TI in HgBa,Ca,- 1C~,02n+2+6 (n = 2,3) with similar T values being ~btained.~.~ The synthesis of phases containing double Hg layers has been demonstrated this year. The occurrence of double Hg layers had been identified during HREM studies of HgBa,Ca,-1C~n02n+ suggesting that it might be possible to synthesize bulk phases containing these double layers as observed in the related TI-containing system.lo Radaelli et al. have performed such a synthesis by a high temperature high oxygen pressure route (18 kbar 1000"C).ll As in the case of the single layer phase HgBa,YcU,O +, the prototype Ca-free double Hg layer material S.Miyashita H. Higuna and F. Uchikawa Jpn. J. Appl. Phys. Lett. 1994 33 L931. C. C. Tsuei A. Gupta G. Trafas and D. Mitzi Science 1994 263 1259. L. Gao Y. Y. Xue F. Chen Q. Xiong R. L. Meng D. Ramirez C. W. Chu J. H. Eggert and H. K. Mao Phys. Rev. B 1994 50 4260. ' B. A. Hunter J. D. Jorgensen J. L. Wagner P. G. Radaelli D. G. Hinks H. Shaked R. L. Hitterman and R.B. Von Dreele Physica C 1994 221 1. ' Q. Huang 0.Chmaissem J. J. Capponi C. Chaillout M. Marezo J. L. Tholence and A. Santoro Physica C 1994 227 1. N.H. Hur N. H. Kim K. W. Lee K. H. Yook Y. K. Park and J. C. Park Physica C 1994 234 19. N. H. Hur N. H. Kim K. W. Lee Y. K. Park and J. C. Park Physicu C 1994 231,4. lo G.Van Tendeloo C. Chaillout J. J. Capponi M. Marezio and E. V. Antipov Physica C 1994 223,219. P. G. Radaelli M. Marezio M. Perroux S. de Brion J. L. Tholence Q. Huang and A. Santoro Science 1994 265 380. Conducting Solids Covering Ionic and Electronic Conductors 469 Hg,Ba,YCu,O -6 is non superconducting but superconductivity can be induced by Ca doping for Y with a maximum T of -45 K. Rietveld refinement of powder neutron diffraction data confirmed the general features similar to Bi- and T1-2212 but revealed some differences. In particular the oxygen site in the Hg bilayer (O* Figure 1) is partially (20 to 25%) vacant. This means that some of the Hg atoms lose one of the neighbouring oxygen atoms along the z axis. Therefore they no longer have an apically compressed octahedral coordination but rather a pyramidal coordination with only one short Hg-0 bond along the z axis and three or four in-plane bonds.In addition the CuO planes are buckled (Cu-01-Cu 'buckling' angle is 168.8') making the Cu environment similar to YBa,Cu,O, unlike the single Hg layer compounds. It is interesting to note the low T and significant buckling of the CuO planes which is not observed for the higher T single Hg layer phase suggesting a possible correlation between these two features. Bryntse has reported a double layer phase containing a mixture of T1 and Hg in the double layer.' A phase of composition Hgo~,,7T1,,333Ba,CaCu208+, was prepared with a T of 100K which is 12K lower than for the corresponding TI-2212 phase.A three Cu layer phase of composition HgO,,T1,,,Ba,Ca,Cu,O,,~, with a T of 130K has also been reported.' The limits of Sr substitution for Ba have been examined in HgBa,CuO,+ by Subramanian and Whangbo.' Single phase HgBa -xSrxC~04+s has been achieved for 0 Ix I0.7 with T decreasing with increasing x such that superconductivity is lost when x > 0.5. These results together with conclusions from band structure calculations indicate that the hole source in this compound is the excess oxygen content (6) which is reduced by Sr substitution due to a reduction in the cell size. In order to stabilize the Ba-free compounds the partial replacement of Hg by other elements is required. The stabilization of HgSr,CuO,+ has been achieved by partial substitution of Mo for Hg with a maximum T of 78 K for 15% Mo substitution for Hg.I5 This represents the highest T for a Ba-free Hg-1201 phase and since high pressure oxygenation was found to reduce T,the samples were considered to be either optimally doped or overdoped.Cr substitution for Hg has also been achieved with a maximum T of 58 K for Hgo,,Cro~3Sr,Cu0,+6.'6 The attempted substitution ofCr on the Cu sites resulted in large impurities indicating that Cr prefers the Hg site. Mo substitution in Ba- free Hg-1212 has also been demonstrated with Tc'sin the region of loOK.' Re is also reported to stabilize Sr-based Hg compounds but it is tentatively assumed that it substitutes for the Cu site rather than the Hg site. Stabilization by Ce substitution i.e.Hg0,,Ce0~,Sr3~,M,Cu2~,0,~, (M = Ca Ce Nd) has also been demonstrated with a maximum T of 62 K for Hgo~,Ceo,,Sr,,,Cao~,Ndo~2Ceo,lCu2,1-0 As in the case of Pr substitution reported last year a superstructure resulting from cationic ordering between Ce and Hg (plus Cu) was observed. The Ca and Nd homogeneity range was very narrow and no Hg-1212 phase was observed when M = Ce. The stabilization of Ba-free Hg-1223 has however not yet been achieved. l2 1. Bryntse Physicu C 1994 226 184. l3 F. Goutenoire A. Maignan G.Van Tendeloo C. Martin C. Michel M. Hervieu and B. Raveau Solid State Cotnmun. 1994 90,47. l4 M.A. Subramanian and M.-H. Whangbo J. Solid State Chem. 1994 109 410. l5 K. K. Singh V. Kirtikar A. P. B. Sinha and D. E. Morris Physica C. 1994 231 9.J. Shimoyama S. Hahakura K. Kitazawa K. Yamafuji and K. Kishio Physica C 1994 224 1. l7 S. Hahakura J. Shimoyama 0.Shiino and K. Kishio Physica C 1994 233 1. A. Maignan M. Hervieu C. Martin C. Michel and B. Raveau Physica C 1994 232 15. 470 P. R. Slater Figure 1 The structure ofHg,Ba,YCu,O A 1 1 intergrowth of Hg-1201 and La,Cu04 has been synthesized by Huve et ~1.'~ The phase HgBa,La,Cu,O,,, is superconducting with a T of 53K (after Ar annealing) and represents the first member of the series [HgBa,CuO +,Irn b2CuO4In. Choy et al. have succeeded in intercalating mercury halides between the two BiO layers in Bi,Sr,CaCu,O which has been previously reported to intercalate halides.,' The intercalation compounds (HgX,),~,Bi,Sr2CaCu,0 (X = Br I) retain superconductivity although with a slightly reduced T (71 K for X = Br 68 K for X = I compared to 76K for the starting material) despite the large expansion along the c axis (7.2A).Superconductors Containing 0xyanions.-Cuprate systems have been shown in the past few years to exhibit a remarkable flexibility for the incorporation of C02 3 -groups l9 M. Huve C. Martin G. Van Tendeloo A. Maignan C. Michel M. Hervieu and B. Raveau Solid State Commun. 1994 90,37. 'O J.-H. Choy N.-G. Park S.-J. Hwang D.-H. Kim and N.H. Hur J. Am. Chem. Soc. 1994 116 11 564. Conducting Solids Covering Ionic and Electronic Conductors 471 as well as other oxyanions. A significant amount of further work has been performed in this area with the synthesis of several new intergrowth phases and other systems.Ono et al. have observed a new oxycarbonate phase Bi,(Bi,Sr,Ca),Ca(Cu,C),C,O, which contains a mixture of octahedral and square pyramidal Cu in some multiphase Bi-Sr-Ca-Cu-C-0 samples formed at high pressure (2 GPa).,’ Since they were unable to prepare the oxycarbonate as a major phase (and it was only detected in 3 out of 25 samples examined) the superconducting properties were not measured. The structure consists of alternating layers of half Bi,Sr,Cu,C,O,, CaCuO, and half Bi2Sr6Cu3C201 6. They also observed another oxycarbonate containing no Bi i.e. (Sr,Ca),,Cu +,C4 -x024 which consists of intergrowths of infinite layer SrCuO and Sr,CuO,CO type lamellae. Pelloquin et al. have prepared the third member (n = 3) of the oxycarbonate series (Bi -,P~,S~,CUO,)(S~,CUO~CO~)~ (0 5 x 2 0.25),, which had previously been seen as a defect in lower n members.This corresponds to an intergrowth of triple Sr,CuO,CO oxycarbonate layers and a single 2201 layer. The structure is modulated as with all bismuth cuprates and oxycarbonates. The T (x= 0)is 34 K compared with 30 K (n = 1) and 40 K (n = 2) and in contrast to the first two members T is not raised by Pb doping (x > 0). Thus T appears to go through a maximum as n increases although the metastable character of this phase due to the high stability of Sr,CuO,CO and the lower n members means that it is difficult to optimize the T,. Stage I iodine intercalation has been successfully achieved into the oxycarbonate B~,S~,CU,CO,O,.~~ The semiconducting underdoped host material became super- conducting with a T of 23 K after intercalation which is explained by hole doping from the intercalated iodine atoms into the CuO sheets.A new oxycarbonate Pbo~7Hg,,,Sr4Cu,C0,07 has been reported by Martin et The phase which consists of an intergrowth of the two non-superconducting compounds Sr,CuO,CO and Pb,~,Hg,,,Sr,CuO was superconducting after anneal- ing in a reducing atmosphere with a T of 70K. Uehara et al. have reported the oxycarbonate HgBa,Sr,O,+,Cu,(CO,) with a T of 66K.25 As in the case of the related T1-containing oxycarbonate T1Ba,Sr,Cu,07(C0,) reported last year the structure can be related to HgBa,CuO,+ with a periodic replacement of COi- for HgO along the (110) direction.The similar oxycarbonate Hg,~,Pbo~,Sr,~,Ba,Cu2CO,O,-,has also been prepared with a maximum T of 68 K for x = 2.26A modulation along the (110) direction is observed which is attributed to an ordering between (Pb,Hg)O and carbonate groups with the degree of ordering depending on among other factors the local CO content. The compound Bi,~,Hg,,,Sr,Cu,CO,O -,,consisting of an intergrowth of 1201 Bi,~,Hg,,,Sr,CuO,+ and Sr,CuO,CO, has been prepared with a T of 17K.27This phase exhibits large possibilities for non-stoichiometry such as partial substitution of Sr for (Bi Hg) and a shearing mechanism along c that leads to longitudinal ” A. Ono S. Horiuchi and M. Tsutsumi Physicu C 1994 226 360. ’‘ D. Pelloquin M. Hervieu A. Maignan C. Michel M.T. Caldes and B. Raveau Physicu C 1994,232,75. 23 Y. Muraoka H. Nameki M. Kikuchi S. Awaji N. Kobayashi and Y. Syono Physicu C 1994,233,209. 24 C. Martin M. Hervieu M. Huve C. Michel A. Maignan G. Van Tendeloo and B. Raveau Physicu C 1994 222 19. z5 M. Uehara S. Sahoda H. Nakata J. Akimitsu and Y. Matsui Physicu C 1994 222 27. 26 M. Huve G. Van Tendeloo M. Hervieu A. Maignan and B. Raveau Physicu C 1994 231 15. *’ D. Pelloquin M. Hervieu C. Michel A. Maignan and B. Raveau Physicu C 1994 227 215. 472 P. R. Slater intergrowths of 1201 and oxycarbonate which dramatically changes the local composition. Amamoto et a!. have prepared a new insulating oxyborate Nd,Sr,Cu,O,(BO,) consisting of alternate stacking of (Nd,Sr),CuO,BO and N~S~,CU,O,BO,.~~ The high room temperature resistance (> 2 MQ cm) is attributed to the Cu oxidation state of 2+ and the consequent absence of holes.Single crystals (2 x 2 x 2mm3) of the parent oxycarbonate (Ba -xSrx)2Cu1.102,2 +,(CO,),. have been prepared for the first time by means of the abnormal grain growth of ceramic samples.29 Although they show some voids and do not exhibit a perfect shielding effect the crystals are superconducting with an onset of T of 32K. Oxyanion substitutions in Y(Ba/Sr),Cu,O related systems have attracted further study. A structural study of the system (Cu,C)(Sr,Ca),(Y,Ca,Sr)Cu,O has been performed by Miyazaki et aL3’ As-prepared samples were semiconducting but superconductivity was induced by annealing at 1000°C in an oxygen pressure of 4 MPa.The structural study and wet chemical analysis suggested that hole doping is caused during high pressure oxygen annealing by the release of some Cog- groups and by the consequent introduction of oxygen atoms into the vacant apical oxygen site where the C0:- group has been lost. Ono and Horiuchi have synthesized superconducting samples of the Ba-free 1212 oxycarbonate phases (Cu,C)Sr,(Y,Ca)Cu,O for the first time by using high pressure ~ynthesis.~’ The Tc’sof the Ca-containing phases are -80 K regardless of considerable variations in Ca and C contents while the T is -45K for the Ca-free phase. The incorporation of the oxyanions SeOi-and Te0;-into YSr,Cu,O,- and (Y/Ce),Sr,Cu,O,- has been attempted., No evidence for the substitution of Te was found whereas Se substitution was successful (highest purity samples for 50% Se substitution of the ‘chain’ Cu sites) with superconductivity (T = 45 K) being observed in YSr,Cu,,,Se,.,O,- by Ca doping for Y.XPS measurements indicated that Se enters the structure as Se4+ as opposed to Se6+ and remains in the lower oxidation state even after annealing under high oxygen pressure. Superconductivity (T = 15-20K) has been observed in (Y/Ce),Sr,-,Ba,Cu,O,~ phases doped with BO;- SO:- PO:-on the ‘chain’ Cu sites after annealing under high oxygen pressure.33 Neutron diffraction studies showed significant (4-9%) vacancies in the (Y/Ce),O fluorite block which are filled after oxygen annealing. New phases La,~,Ce,~,Ba,Cu,BO,- and Nd,~,Ce,~,BaSrCu,BO with complete BO layers were also reported.Hiroi et a/.have observed a new metallic oxycarbonate Sr8Cu7COl in high pressure synthesis studies of the Sr-Cu-C-0 system.34 The structure consists of an oxygen deficient perovskite where 1/8 of the Cu are replaced by Cot- so as to form a 242a x 242a x c unit cell. The lack of superconductivity can be attributed to the fact that there are no intact CuO planes. ” Y. Amamoto H. Yamane T. Oku Y. Miyazaki and T. Hirai Physicu C 1994 227 245. 29 H. Shibata K. Kinoshita and T. Yamada Physica C 1994,232 181. 3” Y. Miyazaki H. Yamane T. Kajitani N. Kobayashi K. Hiraga Y. Morii S. Funahashi and T. Hirai Physica C 1994 230 89. 31 A. Ono and S. Horiuchi Jpn. J. Appl. Phys. 1994 33 L1149. 32 P. R. Slater C. Greaves M.Slaski E. Z. Kurmaev St. Uhlenbrock and M. Neumann Physica C 1994,231 109. 33 P. R. Slater C. Greaves and M. Slaski Physicu C 1994 235-240 741. 34 Z. Hiroi K. Yamaura M. Azuma and M. Takano Physica C 1994 235240 1005. Conducting Solids Covering Ionic and Electronic Conductors 473 High pressure synthesis techniques have resulted in a new series of oxycarbonate superconductors although the exact compositions of these phases remains somewhat controversial. In March 1994 Ihara et al. reported superconductivity with T > 117K in the system Ag -,Cu,Ba,Ca,- 1C~n02n+3-6.35 Their idea was to form similar phases to the Hg series with non-toxic Ag in place of Hg. The synthesis was performed at high temperatures and high pressures (5 GPa at 1100 "C for 1-3 hours) using a Ba2Can-1C~,0 precursor (n = 3 4 5) and Ag,O or Ago.XRD data showed the presence of major phases ascribed to Ag-1223 Ag-1234 and Ag-1245 with the T for Ag-1234 being 117 K and lower Tc'sbeing observed for the other two phases. EDX of platelike Ag-1234 crystals indicated that the content of Ag varied with the Cu content in a wide range of xfrom 0.25-1.0 demonstrating a mixing of Ag and Cu. These phases were observed to be more stable to atmospheric degradation than the corresponding Hg or T1 phases. Similar phases were subsequently reported by the same group but without the presence of Ag i.e. Cu -.Ba2Can- 1C~n02n+4-6.36 A superconducting sample containing mainly Cu-1234 phase gave a T of 116.4 K. A deficiency in Cu (x-0.4) was determined from an XRD analysis of a single crystal.(3-1223 and 1245 phases were also synthesized showing lower Tc's. Jin et al. also succeeded in synthesizing similar superconducting multiphasic samples in the Ba-Ca-Cu-0 sys- HREM confirmed structures similar to the type found in analogous single layer Hg or TI systems and showed the presence of intergrowths of other n members (up to n = 6),as are also observed for the related Hg and T1 system^.^' However some controversy has arisen over the exact nature of these superconducting phases. In particular as to whether the Ag-containing samples actually contain Ag in the superconducting phase. Wu et al. have examined the Ag-Ba-Ca-Cu-0 system and found no evidence by EDX analysis for Ag in the perovskite related phase.However samples which had Ag in the starting matrix were found to show a sharper superconducting transition and a larger Meissner signal suggesting that Ag20 could act as a reaction enhancer in the formation of the Cu-l2(n -1)n phases.39 Kawashima et al. also examined this system and found by chemical and structural analysis that no Ag was present but instead the phases contained carbon suggesting they were oxy~arbonates.~' The presence of CO; -was later confirmed by EELS investigations by Alario-Franco et aL4' Two members of the homologous series were (Cuo,5Co~5)Ba,Ca,~,Cu,0,,+3isolated with n = 3 (T = 67K) and n = 4 (T = 117 K). Electron microscopy indicated that Cu and C are in both cases located alternately along the a axis of the subcells resulting in superstructures with dimensions 2a x b x 2c.Most of the Ag was found to be present as Ag metal suggesting that Ag,O worked only to increase the oxygen pressure. The cell parameters were similar to those observed by Ihara et al. in their Ag-free samples suggesting that these samples are also 35 H. Thara K. Tokiwa H. Ozawa M. Hirabayashi H. Matuhata A. Negishi and Y.S. Song Jpn. J. Appl. Phys. 1994 33 L300. 36 H. Ihara K. Tokiwa H. Ozawa M. Hirabayashi A. Negishi H. Matuhata and Y. S. Song Jpn. J. Appl. Phys. 1994 33 L503. 37 C.-Q. Jin S. Adachi X.-J. Wu H. Yamauchi and S. Tanaka Physica C 1994 223 238. 38 X.-J. Wu S. Adachi C.-Q. Jin H. Yamauchi and S. Tanaka Physica C 1994 223 243. 39 X.-J. Wu C.-Q. Jin S. Adachi and H. Yamauchi Physica C 1994 224 175.40 T. Kawashima Y. Matsui and E. Takayama-Muromachi Physica C 1994 224 69. 41 M. A. Alario-Franco P. Bordet J.-J. Capponi C. Chaillout J. Chenavas T. Fournier M. Marezio B. Souletie A. Sulpice J.-L. Tholence C.Colliex R. Argoud J. L. Baldonedo M. F.Gorius and M. Perroux Physica C 1994 231 103. 474 P. R. Slater oxycarbonates. a conclusion supported by the observed Cu deficiency. An interesting point however is that the n = 4samples of Jin et al. and the Ag-containing samples of Ihara et al. have a larger c axis than those of Kawashima et al. (18.30A and 18.1 1 A compared to 17.93 A) suggesting that if COi- is also present for the former two cases then it is present in a lower concentration i.e. the phases are non-stoichiometric (Cu -xC,)Ba2Ca,- ,CU,O~,+~.The question as to why the Tc’sof all samples are so similar remains unanswered. Shimakawa et al. have determined the structure of the n = 4 phase by neutron diffra~tion,~ and observed a composition of (Cuo~,,Co~3,)Ba,Ca3Cu401 The structure has average tetragonal symmetry and is similar to the (Hg T1)-containing cuprate analogues. The structure has two types of inequivalent CuO planes square planar (inner) and square pyramidal (outer). The inner two CuO planes are less ‘buckled’ than the outer two. An unusual feature of these oxycarbonates is that the T,for the n = 4 phase is greater than the T,for the n = 3 phase. It is also unusual that such a high T is observed for such a distorted structure (the Cu-0 apical distance for the outer plane Cu atoms is different depending on whether Cu or C is present).One possible explanation proposed by the authors is that even though the presence of CO; -degrades superconductivity for the two outer CuO layers the two inner CuO layers support superconductivity with a high T,. Kawashima et al. have since discovered a new series of similar oxycarbonates containing two (Cu,C) layers instead of one the ideal composition being (Cuo~,Co~5)2Ba3Ca,-,Cu,0,,-5.43Then = 4(T -113K)andn = 5 (T,near 1lOK) phases were prepared in bulk and in addition an intermediate phase in which n = 4 and n = 5 slabs were stacked alternately was found from electron micro~copy.~~ As before the Cu and C were found to be located alternately along the a axis of the subcell resulting in a doubling of the cell along a.The n = 3 phase has also been prepared with a T of 91 K.,’ In this phase a slightly different ordering between Cu and C was observed by electron microscopy resulting in a superlattice 2a x b x 2c with a doubling along the c direction as well as the a direction. Similar oxycarbonate superconductors have been obtained in the Sr-based systems Ca-Sr-Cu-C-0 by Yamaura et A single phase sample (T -92K) has been obtained from a nominal composition (Cao~,Sro~7)3Cu2Co~7s0,, by heating in the presence of KClO oxidizer at 1200°C and 6GPa for 30 minutes. A supercell a x 4a x 2c most probably a result of the ordering of C vacancies (the sample is 25% C deficient) along the b axis was observed. No superconductivity was detected when the KClO oxidizer was not added.Equivalent systems with partial BO; -substitution for C0:-have also been reported by Uehara et al. i.e. Sr2(Ca,Sr),~,Cu,(C03),~,(B0,),0, (n = 1 x = 0.3 T = 50K; n = 2 x = 0.45 T = 105 K; n = 3 x = 0.6 T = 115K).47Samples without B0;- sometimes showed superconducting properties but with very low volume fraction ( < 1 YO),thus indicating that BOZ-was acting as a hole dopant. 42 Y. Shimakawa J.D. Jorgensen D.G. Hinks H. Shaked R. L. Hitterman F. Izumi T. Kawashima E. Takayama-Muromachi and T. Kamiyama Phys. Rev. B 1994 50 16008. 43 T. Kawashima Y. Matsui and E. Takayama-Muromachi Physica C 1994 227,95. 44 Y. Matsui T. Kawashima and E. Takayama-Muromachi Physica C 1994 235240 166. 45 T.Kawashima Y. Matsui and E. Takayama-Muromachi Physica C 1994 233 143. 46 K. Yamaura Z. Hiroi and M. Takano Physica C 1994,229 183. 47 M. Uehara M. Uoshima S. Ishiyama H. Nakata J. Akimitsu Y. Matsui T. Arima Y. Tokura and N. Mori Physica C 1994 229 310. Conducting Solids Covering Ionic and Electronic Conductors 475 The Main Superconducting Families and Related Phases.-Superconducting tetragonal Sr,CuO synthesized at high temperatures and high pressures has attracted further interest. The T of this phase has been raised from 70 K to 94 K by annealing in a N atm~sphere.~~ There was no sign of any weight loss during the annealing treatment and so the change in T is presumed to be due to a readjustment of the oxygen positions to achieve an optimum doping level.A study of the synthesis conditions of Sr,CuO + by Laffez et al. has shown the presence of two crystallographic domains (a -3.76 c -12.54A and a -3.79 c -12.46A) of the fundamental tetragonal structure.49 The first phase is formed if a small amount of oxidizer (KClO,) is added while the latter is formed for higher amounts of oxidizer. The Tc’sof 5 1K and 80 K have been assigned to these phases respectively. For the latter phase a new superlattice 5a/,/2 x 5a/,/2 x c was observed. Tetragonal Sr,CuO has also now been synthesized at low temperature (370 “C) in latm. 0 from the decomposition of the hydroxometallate precursor S~,CU(OH),.~~ Contrary to the phase prepared under high pressure this phase was not superconducting despite the fact that a similar supercell 4aJ2 x 4aJ2 x c was observed indicating that the material was structurally the same.Heating to 450 “C in 0 resulted in a conversion into the orthorhombic phase. Synthesis at moderately high pressure (50-3000 bar) also produced non-superconducting tetragonal Sr,CuO + with a similar 4J2a x 442a x c superstructure.’’ The structures of high pressure- moderate pressure- and ambient pressure-phases have been determined by neutron diffraction and shown to be the same with the latter having a slightly higher oxygen content.’ ’*’,Moreover in contrast to previous assumptions (see last years review) the structure does not appear to contain ‘intact’ 2D-CuO planes. The oxygen vacancies are located in these CuO planes and not in the Sr,O layers as previously thought and the exact origin of the superconductivity in these phases is not clear.One possible explanation is that oxygen clustering occurs thereby creating islands of superconduct- ing material with ‘intact’ CuO planes.’ Three very important results reported this year have been the observation of hole doped superconductivity in phases without apical oxygen atoms coordinated to the CuO planes. It had been previously thought that such apical oxygen atoms were necessary for hole doped superconductivity. Low temperature ( -200 “C) fluorination (by F gas) of Sr,CuO has been shown to insert fluorine atoms to give the oxyfluoride Sr,CuO,F with the La,Cu04 str~cture.’~ Superconductivity was observed after post synthesis reduction in N or H,/N to reduce the amount of interstitial fluorine (6) with a maximum T of 46K being recorded for 6 -0.3.The superconducting properties and Madelung energy calculations indicate that the fluorination results in O/F interchange to create electronically active CuO planes (Figure 2). Superconduc-tivity with a maximum T of 26K has been induced in Ca,CuO,Cl by partial substitution of Na for Ca under high oxygen pressure (900 “C,6 GPa 30 min~tes).’~ 48 P. D. Han L. Chang and D. A. Payne Physica C,1994 228 129. 49 P. Laffez X.J. Wu S. Adachi H. Yamauchi and N. Mori Physica C,1994 222 303. 50 J. F. Mitchell D. G. Hinks and J. L. Wagner Physica C,1994 227 279. 51 T. Ami M. K. Crawford,R. L. Harlow Z. G. Li T. Vogt Q.Zhu and D. E. Cox Physica C,1994,235-240 1003.52 Y. Shimakawa J.D. Jorgensen J. F. Mitchell B.A. Hunter H. Shaked D. G. Hinks R. L. Hitterman Z. Hiroi and M. Takano Physica C,1994 228 73. 53 M. Al-Mamouri C. Greaves P. P. Edwards and M. Slaski Nature 1994 369 382. 54 Z. Hiroi N. Kobayashi and M. Takano Nature 1994 371 139. 476 P.R. Slater 00 #F F* Sr cu Figure 2 The structure of(a) Sr,CuO and (b)Sr,CuO,F,, showing the idealized apical F(F)and interstitial F(F*) sites NaClO has been used both as a source of Na and to create an oxidizing atmosphere. The nature of the superconducting carriers has not yet been unequivocally assigned but the fact that a highly oxidative pressure was required for the appearance of superconductivity strongly suggests that the carriers are holes.No superconductivity was observed if KClO was used instead of NaClO, showing that Na is important. Hole doped superconductivity has been observed for the first time in a phase with the Nd,CuO structure. Tm,~,,Ca,~,,CuO, prepared at 1100"C under a pressure of 6GPa (using KClO to provide an oxidizing atmosphere) was shown to be superconducting with a T of -30K.55 Normally superconductivity can only be induced in Nd,CuO,-type phases by electron doping. This is related to the long Cu-0 bonds in the CuO planes which arises out of lattice mismatch between the CuO planes and the Nd,O block. The long Cu-0 bonds mean that it is difficult to introduce sufficient mobile holes in the CuO sheets and in fact favours the doping of electrons.In contrast the small size of Tm results in short Cu-0 bonds (1.916 A j such that hole doped superconductivity becomes possible. New methods are increasingly being required to prepare high temperature superconductors. This is particularly so for the phases containing Hg and T1 due to their high toxicity and volatility which causes safety problems under normal high temperature preparation conditions. An interesting paper by Chen et al. reports the preparation of multiphase T1-based cwprate superconductors by a hydrothermal method which avoid the problems of T1 ~olatility.'~ Mixtures of TlNO, BaO CuO 55 W. J. Zhu Y. S. Yao X. J. Zhou B. Yin C. Dong Y. Z. Huang and Z. X. Zhao Physica C 1994,230,385. 56 Q. W. Chen Y. T. Qian Z. Y. Chen K. B. Tang G. E. Zhou and Y.H. Zhang Physica C 1994,224,228. Conducting Solids Covering Ionic and Electronic Conductors 477 CaCO, and H,O (to convert TlNO into Tl,O,) were heated in a teflon vessel with H,O at 160 "C for 16 hours. The product showed the presence of small amounts of TI-2201 and 2212 phases as well as various impurities and exhibited superconductiv- ity with a T onset near 95 K. The amount of BaO and H,O was found to play a crucial role in the formation of superconducting phases. The partial substitution of fluorine for oxygen has been achieved in TlBa,CuO -6 using TlF as a source of fluorine.57 The parent phase is normally non-superconducting due to the high Cu oxidation state (close to Cu3+) and the effect of fluorination is to reduce the Cu oxidation state and consequently induce superconductivity.Single phase tetragonal T1Ba,CuO5-,-,F has been observed for 0.1 5 x I0.5 with supercon- ducting transitions as high as 75 K for x = 0.3. XPS measurements made to investigate the core-level electronic states for TlBa,Ca -xYxCu207-y (x = 0 0.1 0.8 1.O) have shown that the T1 core-level shifts towards the higher binding energy side implying that electrons donated by Y substitution do not only fill the Cu-0 hole states but are also partially transferred to the TI site which thus behaves as a charge reservoir., Sastry and West have developed a new synthetic route to phase pure Pb-free Bi 2233.59 The synthesis involves the reaction (over 15-30 days) between the precursors Bi,CuO and (Sr,Ca),CuO,. In this way Bi-2233 with variable compositions can be prepared from stoichiometric compositions with variable Ca :Sr ratio and Bi :(Ca + Sr) ratios.An optimum T of 107K was obtained which was a lot less dependent on composition and heat treatment than for Bi-2201 and Bi-2212 phases. This method can also be used to form Pb-doped samples by starting with a (Bi ,Pb),CuO precursor. Iodine intercalation into semiconducting Bi,Sr,Y,,,Ca,~,Cu,O + has been shown to induce superconductivity (T,onset close to 76 K) indicating a transfer of holes from iodine atoms located in the BiO layers to the CuO sheets6' The fact that Bi,Sr,Y,,,Ca,,,Cu,O + does not become superconducting indicates that the hole transfer on iodine intercalation is only small. Iodine and bromine have been intercalated into the Bi-2222 phase to give X,Bi2Sr2(Gd,~82Ce,~,8)2~U2~~o (X = Br 0 Iy I0.2; X = I 0 < y I 1).62 Non-superconducting samples became superconducting (-20 K) after light bromination although there was no evidence from XRD that Br was intercalated between the two BiO layers.Normally high pressure oxygen annealing is required to induce superconductivity and such supercon- ducting samples were found to intercalate less iodine and become poorer superconduc- tors suggesting the presence of excess oxygen atoms between the BiO layers in these materials. Superconductivity has also been induced in Bi-2222 phases by the partial substitution of Pb for Bi without the need for high oxygen pressure annealing although the T was found to be increased by such annealing.63 Unlike conventional BCS superconductors the cuprate superconductors exhibit only a small oxygen isotope effect.Of this small effect Zeck et al. have determined that 57 M.A. Subramanian Mat. Res. Bull. 1994 29 119. 58 T. Suzuki M. Nagoshi Y. Fukuda S. Nakajima M. Kikuchi Y. Syono and M. Tachiki Supercond. Sci. Technol. 1994 7,817. 59 P.V. P.S.S. Sastry and A.R. West Physica C 1994 232 63. 6o P. V. P. S. S. Sastry and A. R. West J. Muter. Chem. 1994 4 647. 61 M.A. Subramanian J. Solid State Chem. 1994 110 193. 62 Y. Koike T. Hisaki K. Sasaki A. Fujiwara T. Noji and Y. Saito Physica C 1994 224 31. 63 H. B. Liu X. D. Chen X. P. Jiang D. E. Morris and A. P.B. Sinha Physica C 1994 220 265. P. R. Slater 80% of the total (positive) isotope effect is associated with the CuO planes., This suggests that theories of electron pairing in high temperature superconductors have to consider a phononic contribution in which the planar tilting or buckling modes in the CuO layers play an essential role.The compound PrBa,Cu,O has been synthesized for the first time by annealing at or 957°C under 5atm. 02,65 990°C under 50atm Ar/20O/00,.~~ No sign of superconductivity was observed but the compound is metallic in contrast to PrBa,Cu,O, which is semiconducting. An anomalous behaviour (non-linear above 6GPa) of the 147cm-' Cu(2) Raman mode corresponding to the z-motion of the CuO planes has been observed in YBa,Cu,O under pressure., The observed behaviour is similar to the variation of T with increasing pressure.In contrast YBa,Cu,O shows no non-linear behaviour in either the Raman mode or T,. There thus appears to be a correlation between the pressure-induced increases in T and the change in the frequency of the Cu(2) mode which the authors attribute to changes in the electronic state of the CuO planes as a result of charge carrier redistribution between Cu and 0 in the planes. Nakamura et al. have prepared c axis orientated superlattices of (La,A),CaCu,O -x (A = Sr Ca)/(La,Sr),CuO,- by laser ablation on a SrTiO substrate., Underdoped semiconducting La ~6Sro~,CaCu,0 -and overdoped metallic La .,Sr,.,CuO - form a superconducting (T onset = 25 K) superlattice if they are alternately stacked with four half unit cells of the former and ten half unit cells of the latter.Superconductivity at 60 K with a low volume fraction ( < 1YO) has been induced in the double Cu layer Ruddlesden-Popper phase La .,Srl ,,Cu,O, by synthesis at ambient pressure in the presence of KC10 oxidizer.69 Previously superconductivity could only be induced in these double layer cuprates by heating under high oxygen pressure. Guloy et al. have succeeded in preparing the three Cu-layer Ruddlesden-Popper phase La,Ca,Cu,O under high pressure.,' Although the phase exhibited metallic conduc- tivity doping to induce superconductivity could not be achieved due to its thermal instability making it very difficult to optimize the hole concentration. The effect of hydrostatic pressure on the structure of orthorhombic La -,Sr,CuO has been examined by Takahashi et aL7' Increasing the pressure causes the tilt angle of the CuO octahedra to decrease leading to a transition to a tetragonal structure and a small increase in T,.In the tetragonal phase the T was found to be independent of pressure. T thus appears to vary inversely with the tilt angle and is maximum for the tetragonal structure i.e.for flat and square CuO planes. This feature of the highest T for flat CuO planes seems to be common to superconducting cuprates cf. the Hg cuprates which have flat CuO planes and the highest T,. The synthesis of intergrowth structures reported widely for oxycarbonates is now 64 D. Zech H. Keller K. Conder E. Kaldis E. Liarokapis N. Poulakis and K. A. Muller Nature 1994,371 681. 65 Y.Yarnada S. Horii N. Yarnada Z. Guo Y. Kodarna K. Kawarnoto U. Mizutani and 1. Hirabayashi Physica C 1994 231 131. 66 N. Seiji S. Adachi and H. Yarnauchi Physica C 1994 227 377. " M. Kakihana H. Arashi M. Yashima M. Yoshimura L. Borjesson and M. Ka11 Physica C 1994,230,199. 68 K. Nakarnura H. Noburnasa K. Shirnizu and T. Kawai Physica C 1994 221 387. 69 R. Mahesh R. Vijayaraghavan and C.N.R. Rao Mat. Res. Bull. 1994 29 303. 'O A.M. Guloy B.A. Scott and R. A. Figat J. Solid State Chem. 1994 113 54. " H. Takahashi H. Shaked B. A. Hunter P. G. Radaelli R. L. Hitterrnan D.G. Hinks and J. D. Jorgensen Phys. Rev. B 1994 50 3221. Conducting Solids Covering Ionic and Electronic Conductors 479 being extended to other systems. Barbey et al. have prepared Eu2~,Ba,~,Cu3Co,012 which corresponds to the first (n = rn = 1) member of the intergrowth series [LnBa(Cu,C0)~0,],,[LnBa~(Cu,Co)~0,]~.~~ HREM has shown several intergrowth defects corresponding to the local formation of other (n,rn)members suggesting that it should be possible to prepare other members of the series.Li Rukang has reported the synthesis of two phases in the homologous series of fluorite block-containing compounds (BaTi0,),(Gd,Ce)3Cu,07 with rn = 2 and rn = These compounds are built up by alternative stacking of multiple perovskite layers CuO,planes and double fluorite layers. Semiconducting behaviour was observed down to 15K. A new semiconducting layered cuprate Gd,CaBa,Cu,Ti,O,, has been synthesized by Fukuoka et al.74 It is proposed that this phase is related to the previously reported Gd,Ba,Cu,Ti,O, compound but contains an extra rock-salt (Gd,Ca)O sheet inserted between the two TiO octahedra.The lack of superconductivity is attributed to the fact that the Cu oxidation state is only 2.0+. Ag addition to Pb,~,Cu,,,Sr,(Ln -,Ca,)Cu,O (R = Y Gd Dy Ho Er Tm) has been shown to enhance T by up to 15K above those of the corresponding Ag-free samples with the compounds containing the larger rare earths (lower initial T,) showing the greatest T enhancement (highest T = 78K. for Ag-doped Pb,~,Cu,~,Sr,Y,~,Ca,,5Cu207).75 EDX studies have indicated that a small amount of Ag enters the structure and it has also been suggested that Ag facilitates the incorporation of oxygen in the grains thus promoting the formation of the fully oxidized phase.The versatility of high pressure synthesis techniques has been further demonstrated by the synthesis of a new series of high temperature superconductors MSr,Ca,-1C~,02n+3 (M = Al Ga) isomorphous to the corresponding M = Hg TI At ambient pressure only the n = 2 phase can be formed provided Y is in place of Ca which corresponds to (A1,Ga)-doped YSr,Cu30 In order to form the Y-free phase the high pressure synthesis is required., For the Al-containing phases Tc’s of 110K and 83K were obtained for n = 4 and 5 re~pectively.~~ Significant intergrowths of phases with different values of n were observed in the n = 5 sample (values of up to n = 8 were present). EPMA measurements suggest deviations from the ideal composition with partial substitution of Cu for A1 and Sr for Ca.For the Ga-containing phases samples with n = 3 and n = 4 have been prepared with Tc’sof 70 K and 107 K re~pectively.,~ Non-cuprate Systems.-The search for superconductivity in non-cuprate systems has been the subject of further intense research with a number of reports of new superconducting materials. Superconductivity albeit at very low temperature (0.93K) has been observed in single crystals of Sr,RuO (grown by a floating zone method) with the La,Cu04 structure.79 This is only the second layered perovskite without Cu 72 L. Barbey B. Domenges N. Nguyen and B. Raveau J. Solid State Chern. 1994 111 238. l3 R. Li J. Mater. Chern. 1994 4 773. 74 A. Fukuoka S. Adachi T.Sugano X.-J. Wu and H. Yamauchi Physica C 1994 231 372. 75 H. B. Liu D. E. Morris X. P. Jiang and A. P. B. Sinha Physica C 1994 222 119. 76 M. Isobe Y. Matsui and E. Takayama-Muromachi Physicu C 1994 222 310. 77 M. Isobe T. Kawashima K. Kosuda Y. Matsui and E. Takayama-Muromachi Physica C 1994,234,120. 7R E. Takayama-Muromachi and M. Isobe Jpn. J. Appl. Phys. 1994 33 L1399. 79 Y. Maeno H. Hashimoto K. Yoshida S. Nishizaki T. Fujita J. G. Bednorz and F. Lichtenberg Nature 1994,372 532. 480 P. R. Slater to exhibit superconductivity [the first being Ba,-,K,Pb -yBiy04-8 (T = 14 K x =0.1 y = 0.15) reported last yeargo]. Gasparov et al. have observed evidence for superconductivity above 20K in multiphase samples of Ba,NbO,- prepared from Nb and BaO rapidly heated to -500OC in air O,/Ar,H, or under vacuum.g' When samples in the composition range 0.2 <x < 1 were prepared in air two superconducting transitions 18.6K and 9.3 K were observed with the 9.3 K transition apparently due to metallic Nb.For x = 0.5 a broader signal with a higher T onset (-22 K) was observed. Nakamura observed broad superconducting signals in the defect perovskite-type reduced niobates A,Nb,+,O (A = Ca Sr) (9.2K for 0 Ix I1.0 and 8.3 K for 1.2-1.5 I x I 5.0) covering almost the whole valence state of Nb from Nbl+ to Nb5+ in the Ca system and a somewhat narrower range in the Sr system.g2 The highest shielding effect was observed for starting mixtures having initial mean Nb valence between 1+ and 2 +. The authors argue that although Nb metal exhibits a similar T, the superconducting transitions are not due to this.Manivannan et al. have reported possible high temperature superconductivity at 70 K in an oxygen-deficient cubic perovskite of nominal composition BaPbo~,5Tlo~25Cuo~,0, -y (y = 0.854.89).83The phase was prepared by reaction at high temperature (137CL1420K) followed by quenching then annealing in 0 at 107Ck1120 K. Superconductivity was then observed after annealing in N at 770 K. No evidence was found for any other phases such as Tl,Ba,CuO which could account for this superconductivity. Further research into La2-xNi0,+8 (0 Ix 5 0.15; 6 >0,6 <0)has shown that the resistivity change at temperatures >600K is not due to a semiconductor to metal transition as previously thought.g4 Careful measurement of the sample weight in this temperature range shows that the oxide does not maintain constant composition but undergoes a clear oxygen loss acceleration at -600 K.Consequently the hole concentration is reduced and the resistivity increased. Taking this into account the observed resistivity upturn is not due to a 'transition' to a metallic state but is instead due to a normal non-stoichiometric effect. The two Ni layer Ruddlesden-Popper phase La,Ni,O has been prepared as a single phase for the first time by Zhang et al. using a method involving the co-decomposition of organic precursor^.^ The fully oxygen- ated phase (6 =0) obtained by annealing under high oxygen pressure is metallic and the phase becomes semiconducting as oxygen is removed.Alkaline earth (Ba Sr Ca)-doped samples have also been prepared.g6 Significant oxygen deficiencies are observed in samples as-prepared but these can be filled by annealing under high oxygen pressure. The room temperature resistivity of the as-prepared substituted samples decreases relative to La,Ni,O,,, and a semiconductor to metal transition occurs for La -xM,Ni,O,- (M = Ba Sr Ca) at x =0.075. All the samples annealed under high oxygen pressure show metallic behaviour. Thus contrary to La -.SrXNiO4 where metallic behaviour is only achieved for approximately 50% alkaline earth So M. Licheron and F. Gervais J. Alloys Compd 1993 195 77. " V.A. Gasparov G. K. Strukova and S.S. Khassanov JETP Lett. 1994 60 440. 82 A.Nakamura Jpn. J. Appl. Phys. 1994 33 L583. 83 V. Manivannan J. Gopalakrishnan and C. N. R. Rao J. Solid State Chem. 1994 109 205. 84 J. M. Bassat P. Odier and J. P. Loup J. Solid State Chem. 1994 110 124. 85 Z. Zhang M. Greenblatt and J. B. Goodenough J. Solid State Chem. 1994 108 402. 86 Z. Zhang and M. Greenblatt J. Solid State Chem. 1994 111 141. Conducting Solids Covering Ionic and Electronic Conductors 48 1 substitution La -xMxNi,O becomes metallic with only very small amounts (-4%) of substitution. Reduction of the La,CuO type phase YSr,Ni,O under H has been shown to give a stoichiometric Ni’ oxide YSr,Ni,O, with a Sr,CuO type structure in which 1/3 of the bridging oxygen atoms are missing from chains of apex linked Ni-0 square planar Measurements of the conductivity of this phase have been hampered by the presence of Ni metal in these samples.3 Other Electronic Conductors and Superconductors Intermetallic Systems.-As reported last year a weak superconducting signal at 12K was observed in a study of the Y-Ni-B intermetallic system.,’ This report and subsequent discoveries by Cava et al. in particular of phases with higher Tc’s have aroused considerable interest in these intermetallic superconductor systems which are prepared by standard arc-melting techniques. Continuing their studies into the nature of the superconducting phase in the Y-Ni-B system Nagarajan et al. found that adding small quantities of C dramatically enhanced the superconducting fraction thus indicating that the superconducting phase was a quaternary Y-Ni-B-C system.” A phase of composition YNi,BC,.was nominally single phase with a T onset of 12.5K. Cava et al. reported superconductivity at 23 K in the Y-Pd-B-C system in January 1994 which equals the highest Tcfor an intermetallic although the materials were not single phase.’ The best superconducting transitions were observed for the composi- tions YPd,B,C,., and Yl~2sPd4~sB4,2s which had been cooled on a C block. Electron microscopy on YPd,B,C samples (x = 0 superconducting fraction = 0.2%; x = 0.35 superconducting fraction = 1@-20%) has been performed by Zandbergen et al. to try to identify the superconducting phase.92 For the x = 0.35 sample six different phases were observed with only two phases in the other sample.The authors proposed that the superconducting character was due to a phase of composition YPd,BC. However the exact composition of this superconductor is somewhat controversial. Another group has suggested that the phase has composition YPd,B,C,93 and later work on the Th-Pd-B-C system has suggested a different comp~sition.~~ Superconductivity was then reported in single phase LnNi,B,C (Ln = Y Tm Er Ho Lu) with Tc’sup to 16.6 K (Ln = Lu).~’ Superconductivity was observed not only for the non-magnetic rare earth elements Lu and Y but also for the magnetic rare earth elements Tm Er and Ho although for the latter the T was suppressed indicating that thefelectrons are pair breaking. For other larger rare earth elements no superconduc- ” M. James and J.P. Attfield Chem. Commun. 1994 1185. M. James and J. P. Attfield Physica C 1994 235-240 751. 89 C. Mazumdar R. Nagarajan C. Godart L. C. Gupta M. Latroche S. K. Dhar C. Levy-Clement B. D. Padalia and R. Vijaaraghavan Solid State Commun. 1993 87 413. 90 R. Nagarajan C. Mazumdar Z. Hossain S. K. Dhar K. V. Gopalakrishnan L. C. Gupta C. Godart B. D. Padalia and R. Vijayaraghavan Phys. Rev. Lett. 1994 72 214. 91 R. J. Cava H. Takagi B. Batlogg H. W. Zandbergen J. J. Krajewski W. F. Peck Jr. R. B. Van Dover R. J. Felder T. Siegrist K. Mizuhashi J. 0.Lee H. Eisaki S. A. Carter and S. Uchida Nature 1994,367 146. 92 H. W. Zandbergen W. G. Sloof R. J. Cava J. J. Krajewski and W. F. Peck Jr. Physica C 1994,226,365. 93 C. L. Jia Y.-H. Xu M. Beyss F. Peter and K.Urban Physica C 1994 229 325. 94 H. W. Zandbergen T. J. Gortenmulder J. L. Sarrac J. C. Harrison M. C. de Andrade J. Hermann S. H. Han Z. Fisk M. B. Maple and R. J. Cava Physica C 1994 232 328. 95 R. J. Cava H. Takagi H. W. Zandbergen J. J. Krajewski W. F. Peck Jr. T. Siegrist B. Batlogg R. B. Van Dover R. J. Felder K. Mizuhashi J.O. Lee H. Eisaki and S. Uchida Nature 1994 367 252. 482 P. R. Slater Figure 3 The structure ofLuNi,B,C tivity was observed above 4.2 K. The superconductivity in the Y-Ni-B-C system observed by Nagarajan et al." has been attributed to a phase of similar composition. The crystal structure of LuNi,B,C has been determined by a single crystal X-ray diffraction study and is found to be tetragonal and a variant on the layered ThCr,Si type structure (Figure 3) with LuC NaC1-type layers alternating with Ni,B layers.The Ni,B layers contain a square planar Ni array sandwiched between the B planes.96 The Ni-Ni in-plane distances of 2.45 A are slightly shorter than those found in metallic Ni (2.50 A) suggesting strong metal-metal bonding. Another closely related non-superconducting phase LuNiBC has also been observed which is derived from LuNi,B,C by adding another LuC layer. Powder neutron diffraction data on YNi,B,C showed similar results to the single crystal X-ray diffraction study of LuN~,B,C.'~ There was no sign of B t)Ni disorder and all sites were fully occupied except for possibly the C site. HREM studies on superconducting LuNi,B,C and the related compound LuNiBC has shown the occurrence of intergrowths of the two structures together with planar defects." Such defects are however rare and are only dominant in batches prepared under unfavourable conditions.These intermetallics can 96 T. Siegrist H. W. Zandbergen R. J. Cava J.J. Krajewski and W. F. Peck Jr. Nature 1994 367 254. 97 B. C. Chakoumakos and M. Paranthaman Physicu C 1994 227 143. 98 H. W Zandbergen R. J. Cava J. J. Krajewski and W. F. Peck Jr. Physicu C 1994 224 6. Conducting Solids Covering Ionic and Electronic Conductors be viewed as an homologous series (LuC),(Ni,B,) (m= 1 LuNi,B,C; m = 2 LuNiBC). In unsuccessful efforts to introduce an extra LuC layer a new phase of composition Lu,NiBC was ~bserved.~' This phase has a layered structure with a sequence (Ni,-B-LuC-Lu-C,-Lu-LuC-B),.Superconductivity at around 15-16 K has also been observed in multiphase samples containing the metastable compound ScNi,B,C which is isostructural to LUN~,B,C.'~~ No superconductivity or ScNi,B,C was observed after annealing at 1050"C indicating that the phase is entropy-stabilized during the arc-melting process. The variation of crystallographic cell size with Ln radius in LnNi,B,C is anomalous showing a decrease in c (distance perpendicular to the layers) with increasing Ln size."' The anomaly results from the rigidity of the B-C bond and the conservation of Ni-B bond length such that the B-Ni-B bond angle accommodates the change in LnC layer size. It has been suggested that this change in bond angle which greatly changes the electronic states at the Fermi level is the origin superconductivity in LnNi,B,C with small Ln.For the larger Ln such as La the bond angle is strained (102")and so the Ni,B layers are somewhat buckled. This apparent correlation of decreasing T with increasing Ni,B layer buckling may be compared to the buckling observed in the cuprate superconductors since for these phases the maximum T,seems to occur for the smallest buckling (i.e.flat CuO planes). The effect of this contraction along the c axis and buckling of the Ni,B layers is also observed in the variation of T with pressure. Thus despite the fact that Lu is smaller than Y and that LuNi,B,C has a higher T than YNi,B,C the application of pressure is found to suppress Tc.Io2 Arisawa et a!.have prepared YNi,B,C thin films by magnetron sputtering.lo3 The as-deposited films were completely amorphous although the films could be crystallized very uniformly after annealing at 1050 "Cin a vacuum. The annealed films were highly c axis orientated and showed a T onset of -15 K. Tunnelling measurements of the superconductor energy gap in YNi,B,C were shown to be in good agreement with the BCS prediction indicating that the superconductor mechanism appears to be BCS type.lo4 Band structure calculations for LuNi,B,C indicate a density of states peak near the top of the nearly filled Ni (3d) bands with only modest B and C orbital admixt~re.'~~.'~~ There is also an appreciable amount of Lu (5d)character present which may account for the depression of T for magnetic rare earth elements.Despite the layer structure the band structure exhibits 3D character in contrast to the 2D character of the cuprate superconductors. For the non-superconductor LuNiBC the same band is present but the electron count is such that the Fermi level no longer falls in that band. Thus it may be possible to induce superconductivity by a suitable adjustment of the electron count by chemical substitution. The results suggest that these intermetallics represent a family of conventional (rather than high T,) supercon-ductors. 99 H. W. Zandbergen R. J. Cava J. J. Krajewski and W. F. Peck Jr. J. Solid State Chem. 1994 110 196. loo H. C. Ku C. C. Lai Y.B. You J. H. Shieh and W. Y. Guan Phys. Rev. B 1994 50 351. Io1 T. Siegrist R.J. Cava J. J. Krajewski and W. F. Peck Jr. J. Alloys Compd 1994 216 135. C. Murayama N. Mori H. Takagi H. Eisaki K. Mizuhashi S. Uchida and R. J. Cava Physica C 1994 235240 2545. S. Arisawa T. Hatano K. Hirata T. Mochiku H. Kitaguchi H. Fujii H. Kumakura K. Kadowaki K. Nakamura and K. Togano Appl. Phys. Lett. 1994 65 1299. T. Ekino H. Fujii M. Kosugi Y. Zenitani and J. Akimitsu Physica C 1994 235240 2529. lo' L.F. Mattheiss Phys. Rev. B 1994 49 13279. lo6 W. E. Pickett and D. J. Singh Phys. Rev. Lett. 1994 72 3702. 484 P. R. Slater Superconductivity in isostructural compounds LnPt,B,C containing Pt instead of Ni have also been reported although the phases could not be prepared single phase.lo7 T values of 10 K were observed for Ln = La Y and 6 K for Ln = Pr reflecting the influence of magnetic pair breaking.These materials can apparently accommodate some non-stoichiometry which leads to slightly different Tc's.The density of states for LaPt,B,C is approximately 1/4 that observed for LnNi,B,C suggesting that the lower T is at least in part due to a lower density of states at the Fermi level.lo8 Partial substitution of Au for Pt yielded greatly improved phase purity with no significant change in TC.lo9Phases with Ln = Nd Ce could also be prepared by means of this substitution but both phases were non-superconducting above 1.8 K. The phases LaIr,B,C and LaRh,B,C have also been prepared.' lo Although isostructural to LnNi,B,C and despite the fact that the density of states at the Fermi level (determined from magnetic susceptibility measurements) are comparable to those of the supercon- ducting materials of the same structure type neither of these phases exhibited superconductivity above 1.4K.The range of these intermetallics has been extended from the lanthanide elements to the actinides. While no superconductivity was observed in U-containing compounds superconductivity has been observed in Th-based borocarbides.' l1 For ThM,B,C superconducting transitions of 6 K (M = Ni) 6.5 K (M = Pt) and 14.5 K (M = Pd) have been observed the compounds being isostructural to LnNi,B,C. For M = Ni and Pt essentially single phase samples were observed while for M = Pd the samples were multiphasic. Bulk superconductivity at 21.5 K has also been observed in an as-yet undetermined Th-Pd-B-C quaternary phase (with excess B and Pd e.g.ThPd,B,C) which represents the highest T for a Th-based intermetallic phase. The pressure dependence of the 21 K phase was found to be small and negative. HREM studies ofthe 14.5K superconductor ThPd,B,C showed no intergrowths or planar defects. The through-focus reconstructed image indicated that the C position is fully occupied whereas no definite conclusion could be drawn on the precise occupancy of the B site.'12 It has been suggested from electron microscopy studies of several samples in the Th-Pd-B-C system that the 21 K superconductor is due to a phase of composition ThPd0,,,B4, exhibiting a cubic 4.2 A cell with a short-range ordered superstr~cture.~~ HREM indicates that this phase adopts a modified CaB structure in which the cubic face 1/2,1/2,0 which is vacant in CaB is partially occupied by Pd with the removal of some B atoms.The superstructure indicates short range clustering of the Pd atoms. The exact B composition is not known. EPMA indicates little C present and unlike the other intermetallic phases C is not necessary for the structure although the question of whether it is necessary for superconductivity still remains open. A similar cubic phase YPd,.,B,,, is also observed in the Y-Pd-B-C system and so the authors have lo' R. J. Cava B. Batlogg T. Siegrist J. J. Krajewski W. F. Peck Jr. S. Carter R. J. Felder H. Takagi and R.B. Van Dover Phys. Rev. B,1994 49 12 384. lo8 D. J. Singh Phys. Rev. B 1994 50 6486.lo' R. J. Cava B. Batlogg J. J. Krajewski W. F. Peck Jr. T. Siegrist R. M. Fleming S. Carter H. Takagi R.J. Felder R. B. Van Dover and L. W. Rupp Jr. Physica C 1994 226 170. 'lo R. J. Cava T. Siegrist B. Batlogg H. Takagi H. Eisaki S. A. Carter J. J. Krajewski and W. F. Peck Jr. Phys. Rev. B 1994 50 12966. J. L. Sarrao M. C. de Andrade J. Hermann S.H. Han Z. Fisk M. B. Maple and R. J. Cava Physica C 1994 229 65. 112 H. W. Zandbergen E. J. van Zwet J. C. Sarrac M. C. de Andrade J. Hermann S. H. Han Z. Fisk M. B. Maple and R. J. Cava Physica C 1994 229 29. Conducting Solids Covering ionic and Electronic Conductors Figure 4 The structure of La,Ni,B,N suggested that the 23 K superconductor in this system may be due to this phase rather than to phases of composition YPd,BC or YPd,B,C as previously suggested.In late 1994 the observation of superconductivity in these intermetallic systems was extended to boronitrides with a T of 12-13K being reported for a phase of composition La,Ni,B,N,. l1 A related non-superconducting phase LaNiBN was also prepared. The structure of La,Ni,B,N (Figure 4) is related to that of LuNi,B,C but consists of three rock-salt type LaN layers alternating with tetrahedral Ni2B2 layers an arrangement considerably more two-dimensional than in the superconducting borocarbides. LaNiBN has a related structure with two LaN layers stacked with Ni2B layers and isostructural with LuNiBC.lI4 The phases can be considered as members of a homologous series (LaN),(Ni,B,) with n = 2 (LaNiBN) or 3 (La,Ni,B,N,).Planar defects indicative of the existence of phases with larger n are possible with n = 4,5 and 6 defects being observed for La,Ni,B,N,. The discovery of superconductivity in R. J. Cava H. W. Zandbergen B. Batlogg H. Eisaki H. Takagi J. J. Krajewski W.F. Peck Jr. E. M. Gyorgy and S. Uchida Nature 1994 372 245. H. W. Zandbergen J. Jansen R. J. Cava J. J. Krajewski and W. F. Peck Jr. Nature 1994 372 759. 486 P. R. Slater La,Ni,B,N, with similar Ni,B tetrahedral layers to superconducting LuNi,B,C establishes Ni,B layers as important building blocks for high temperature supercon- ductivity in these intermetallics in much the same way as CuO layers are important to the cuprate high temperature superconductors. Lee et al.have reported superconductivity in the first Ni-based ternary silicide LaNiSi with a T of 1.2&1.26K and a structure which appears to be of the LaPtSi type.'15 Subsequently superconductivity was also reported in Lu,Ni,Si with a slightly higher T (-2 K)." Interestingly it is known that even small amounts of Ni suppress T in Lu,Fe,Si but superconductivity is observed in the structurally different Lu,Ni,Si which suggests that the crystal structure plays an important role in sustaining superconductivity in this compound. Organic Superconductors.-Schlueter et al. have reported two superconducting organic charge-transfer salts derived from the electron donor molecule bis(ethy1enedithio)tetrathiafulvalene (ET) the organometallic anion [Cu(CF,),] -and the neutral solvent molecule 171,2-trichloroethane (TCE).' '7*1' Superconductiv-ity at 4 K was observed in plate-like crystals of composition K,-(ET),Cu(CF,),TCE.A second superconducting phase suggested to be K -(ET),Cu(CF,),(TCE), with needle-like crystals was subsequently found with a higher T of 9 K. The change in T is thought to be due to a change from the disordered crystallographic structure of the anion in the K phase to an ordered structure in the K phase. Other Systems.-A new low temperature synthesis route to the superconducting Chevrel phase PbMo,S has been reported by Rabiller et al.'" The route involves heating a mixture of Mo$ and PbS metal precursor in H at 440 "C. The H allows excess S to be removed as a H,/H,S mixture and simultaneously prevents oxidation.By this method the reaction time is only 3 days compared to 3 weeks for the reaction of Mo,S and Pb powder. High pressure synthesis techniques have been used to synthesize hydrides and deuterides of alloys of the Group 4 metals Hf Zr with the Group 8 metal Ru.' 2o As expected from theoretical studies superconductivity was observed in these phases HfRuH(D),,. (T = 4.5 K) ZrRuH(D),,,-,~ (T,= 3 K). Single phase samples of TiRu hydride could not be obtained. A new halide Hg,As,Br has been synthesized by Shevelkov et ~1.'~' Unlike the related semiconducting phase Cd,As,Br this phase is metallic. The metallic character is attributed to an overlapping of occupied and vacant bands. '15 W. H. Lee F. A. Yang C. R. Shih and H. D. Yang Phys.Rev. B 1994 50 6523. '16 C. Mazumdar K. Ghosh S. Ramakrishnan R. Nagarajan L. C. Gupta G. Chandra B. D. Padalia and R. Vijayaraghavan Phys. Rev. B 1994 50,13 879. ''' J. A. Schlueter U. Geiser J. M. Williams H. H. Wang W.-K. Kwok J. A. Fendrich K. D. Carlson C. A. Achenbach J. D. Dudek D. Naumann T. Roy J. E. Schirber and W. R. Bayless Chem. Commun. 1994 1599. '18 J.A. Schlueter K. D. Carlson J. M. Williams U. Geiser H. H. Wang U. Welp W.-K. Kwok J. A. Fendrich J. D. Dudek,C. A. Achenbach P. M. Keane A. S. Komosa D. Naumann T. Roy J. E. Schirber and W.R. Bayless Physica C 1994 230 378. 'I9 P. Rabiller M. Rabiller-Baudry S. Even-Boudjada L. Burel,R. Chevrel M. Sergent M. Decroux J. Cors and J.L. Maufras Mater. Res. Bull. 1994 29 567. lZo V. E. Antonov E.L. Bokhenkov A. I. Latynin V. I. Rashupkin B. Dorner M. Baier and F. E. Wagner J. Alloys Cmpd. 1994 209 291. 12' A.V. Shevelkov E.V. Dikarev and B.A. Popovkin J. Solid State Chem. 1994 113 116. Conducting Solids Covering Ionic and Electronic Conductors 4 Ionic Conductors and Intercalation Compounds Li Ion Conductors.-Considerable interest has continued in the search for improved Li intercalation electrodes for rechargeable battery applications both in the modification of existing materials and the synthesis of new intercalation hosts. Two useful reviews on Li batteries have been written by Brandt (Historical development of secondary Li batteries)' 22 and Goodenough (Design considerations). 123 Manganese- and vanadium-oxides continue to attract considerable attention as intercalation hosts for use as cathode materials.Hwang et al. have prepared thin films of LiMn,O deposited by rf magnetron sputtering. 24 The as-deposited films were amorphous but a rapid thermal annealing process (above 650°C) in 0 led to crystallization into LiMn,O spinel. The LiMn,O thin film cathodes showed improved characteristics compared to films previously prepared by other techniques demonstrating good intercalation kinetics and promising cycling behaviour. About 0.8Li could be intercalated reversibly at a nearly constant potential and the films maintained more than 98% of their original capacity after more than 100 cycles. Gummow et a/. have succeeded in improving the rechargeable capacity of 4V Li,Mn20 (0 < x 5 1)spinel cathodes by doping with mono- or multi-valent cations + (e.g.Li ,Mg2+ ,Zn2+ ) or alternatively with additional oxygen to marginally increase the average Mn ion oxidation state above 3.5 + .12' The improved rechargeability (stable rechargeable capacities in excess of 100mAhg-' can be achieved) which is gained at the expense of capacity is attributed primarily to the suppression of the Jahn-Teller effect on deep discharge of the doped spinel electrodes.Furthermore for the doped cathodes complete Li removal cannot be achieved and the authors believe that the presence of this residual Li is also significant in enhancing the stability to repeated insertion/extraction of Li. Huang and Bruce have reported a new 3 V Li-Mn-0 cathode LiMn,O,. (A defect spinel with a higher average Mn oxidation state than the stoichiometric LiMn,O prepared at high temperatures) for Li batteries prepared by a new patented solution-based route which can be carried out in air unlike previous sol-gel processes.'26 The addition of C to the solution during preparation was found to yield a material with a significantly enhanced cycling performance-an enhancement of the discharge capacity of 50% on the 300th cycle for the cathode when C was added.The same workers has also reported the preparation of a very promising 4V Li-Mn-0 cathode by the same solution-based technique but with the product fired at 600 "C to give a composition close to LiMn,O,.' 27 Unlike the previous case the addition of C to this material during the solution synthesis had no effect on the cycling performance in the 4V region.The LiMn,O,. cathode when cycled on the 4 V plateau under the same conditions showed much poorer characteristics. Therefore despite the fact that these materials both have similar spinel structures their performance as cathodes for rechargeable Li batteries is markedly different. A problem associated with the use of LiMn,O as a cathode for Li polymer 122 K. Brandt Solid State lonics 1994 69 173. 123 J. B. Goodenough Solid State lonics 1994 69 184. 124 K.-H. Hwang S.-H. Lee and S.-K. Joo J. Electrochem. Soc. 1994 141 3296. 125 R.J. Gummow A. de Kock and M. M. Thackeray Solid State Ionics 1994 69 59. 126 H. Huang and P.G. Bruce J. Electrochem. Soc. 1994 141 L76. 12' H. Huang and P.G.Bruce J. Electrochem. Soc. 1994 141 L106. 488 P. R. Slater electrolyte batteries concerns the upper voltage plateau being above the stability threshold of polymer electrolytes such as poly(ethy1ene oxide) (PEO) and although the lower voltage plateau may be cycled it is only at the expense of decreased energy density and increased capacity fade. Doeff et al. have therefore reported the use of orthorhombic Na,MnO as a cathode for alkali metal polymer electrolyte batteries.',' This compound has a tunnel structure and reversibly intercalates up to 0.6 alkali metal (Li+/Na+) ions per Mn at moderate current densities corresponding to capacities of 160 to 180mAhg-'. Li/PEO/Na,.,Li,MnO cells in particular showed excellent capacity retention over 90 cycles at 0.1 mAcmP2.Ferg et al. have evaluated the properties of the spinel-type anodes Li,Mn,O,, Li,Ti,O,, and Li,Mn,O in Li ion 'rocking-chair' cells with the stabilized spinels Lil~03Mnl,9704 Li ions and LiZno,o,,Mn,~,,04 and layered LiCoO as ~ath0des.I~~ were shown to shuttle between the two transition metal hosts at reasonably high voltage with an accompanying change in the oxidation state of the transition metal cations so that Li ions do not reduce to the metallic state at the anode during charge thus helping to eliminate the safety problems associated with cells containing metallic Li Li-alloy and Li-C anodes. The cells with Li4Ti,01 anodes exhibited the best characteristics and are possible alternatives to Ni/Cd cells delivering comparable energy densities at approximately twice the voltage.Electrochemical insertion of Li into SrVO has been demonstrated.' 30 After the insertion of 0.1 Li to give Li,.,SrVO metallic conductivity was maintained and good reversibility of Li insertion/extraction was obtained. The authors therefore suggest the possibility of using SrVO as a cathode without any conductive additives in Li rechargeable batteries. Garcia-Alvarado and Tarascon have investigated the electrochemical intercalation of Li in Ag,V4O1 and Ag,V4O1 -,,.131 These com- pounds intercalate 7 and 5.7 Li ions respectively through a multiphase intercalation process. Ag,V,O, has been previously reported and has been successfully used as a positive electrode in primary Li cell power sources for biomedical devices.The results of this study demonstrate the potential for it to be used as a positive electrode in rechargeable Li cells. The intercalation/deintercalation reaction showed a small degree ofirreversibility due to some Ag displacement by Li and demonstrated by the presence of free Ag in the XRD patterns. Li insertion into p-Cu,V,O has been investigated by Eguchi et The single phase region of the lithiated fi-phase oxide Li,~-Cu,V,O is metastable at 0 < x < 0.5 and is smaller than in the corresponding a-phase oxide. Up to x = 2 Cu" was reduced preferentially to Cu' and Cu' and half of the Vv were reduced to Cuo and V" respectively for x > 2. Fey et al. have shown that the Li atoms in the inverse spinel LiNiVO can be electrochemically removed and reinserted in non-aqueous elec- trochemical cells at -4.8 V Versus metallic Li representing the highest voltage known for a Li intercalation reaction and suggesting a possible use as a cathode in Li ion cells to provide very high voltages.' 33 Isostructural LiCoVO also shows reversible Li 12' M.M.Doeff M.Y. Peng Y. Ma and L.C. De Jonghe J. Electrochem. Soc. 1994 141 L145. 129 E. Ferg R. J. Gummow A. de Kock and M. M. Thackeray J. Electrochem. SOC. 1994 141 L147. Y. J. Shan L. Chen Y. Inaguma M. Shikano M. Ttoh and T. Nakamura Solid State lonics 1994,70/71 429. 131 F. Garcia-Alvarado and J. M. Tarascon Solid State lonics 1994 73 247. 13' M. Eguchi I. Furusawa T. Miura and T. Kishi Solid State lonics 1994 68 159. 133 G.T.-K. Fey W. Li and J.R.Dahn J.Electrochem. Soc. 1994 141 2279. Conducting Solids Covering Ionic and Electronic Conductors 489 intercalation but only near 4.2 V uersus Li. in situ X-Ray diffraction demonstrates that this is an intercalation reaction (cell parameters contract on Li removal returning to the original value when Li is reinserted). The reason for the high voltage of the Li/LiNiVO cell is not understood and it is also unclear how the Li diffuses in and out of this compound because the tunnels which are present in normal spinels are absent in this inverse spinel. Nishijima et al. have deintercalated Li,FeN electrochemically to form Li -,FeN (0 Ix I1.O) with a decrease in resistivity with increasing Li deintercalation (x).'34 The extraction of Li occurs smoothly up to x = 1 .O with evidence for decomposition for x > 1.0.The Li/Li,FeN cell showed good reversibility with a high current density (500pA~m-~). Li,FeN appears to have ideal properties not only as a cathode for Li secondary batteries but also as an anode if combined with a cathode having high electrode potential relative to Li metal. The same authors have also reported similar studies of the cubic compound Li7MnN,.'35 Li deintercalation can be readily achieved electrochemically to give Li,-,MnN (0 I x I1.25) exhibiting good reversibility under high current density (1200pAcm-2). A good cycling performance with low potential to Li metal (-1.2V) suggests that it is a good candidate as an anode for 'rocking-chair' type Li secondary batteries.LiFeO with a layered ordered rock-salt structure of the a-NaFeO type has been obtained from NaFeO via Li +Na exchange in a eutectic melt of LiCl and LiN0,.136 The crystal structure strongly correlates with the temperature of the melt. The use of the pure halides with higher melting points results in the formation of a-LiFeO with a disordered rock-salt structure. LiMnO (ordered rock-salt structure) has also been obtained from NaMnO in molten LiI as well as by Li intercalation via molten LiI into the spinel LiMn,O,.LiFeO with the layered structure is isotypic with LiCoO and LiNiO, which are much studied intercalation cathode materials. In view of their high electronic conductivity (including reports of superconductivity at 50 K) and stability in reducing atmospheres the possibility of using the magnesium titanate spinels Mg -yTil (y =0.33 0.57) as electrode materials has been examined.'37 For y =0.33 Li insertion proceeds largely via a two phase equilibrium reaction with a maximum Li content of 0.37.For y =0.57 a different behaviour is exhibited with a minimum in the unit cell parameters and a marked drop in cell potential being observed at a Li content of -0.2. The authors have attributed the insertion reaction as a competitive process between two different mechanisms. For Li contents below 0.18 Li is inserted into the octahedral sites with associated movement of Mg from the tetrahedral to octahedral sites (a gradual composition driven spinel to rock-salt type change). For Li contents above 0.18 Li is inserted into empty octahedral sites without major rearrangement of the spinel lattice.Morales et al. have prepared the monophasic layered structure type Li -,(Ni,Co -y)l +,02 materials (0 Iy I0.4 and 0.8 Iy I 0.9) by the reaction of (NiyCol-y)304 (0 I y I 0.4) and Ni,Co -,O (0.8 I y I0.9) with LiOH at low temperatures (LT) (450 0C).'38 As in the case of the high temperature (HT) synthesized 134 M. Nishijima Y. Takeda N. Irnanishi 0.Yamarnoto and M. Takano J. Solid State Chem. 1994,113,205. 135 M. Nishijima N. Tadorkoro Y. Takeda N. Irnanishi and 0.Yamamoto J. Electrochem. SOC.,1994,141 2966. 13' B. Fuchs and S. Kemrnler-Sack Solid State lonics 1994 68 279. 13' J. T. S.Irvine S. Thiemann,G.Mather A. A. Finch and H. Tukarnoto Solid State lonics 1994,70/71,445.138 J. Mordles R. Stoyanova J. L. Tirado and E. Zhecheva J. Solid State Chem. 1994 113 182. 490 P.R. Slater analogues Co substitution for Ni stabilizes the Ni3+ ions and improves the two-dimensionality of the crystal structure. Acid treatment of both LT and HT oxides leads to a partial dissolution of the solid and to removal of Li and 'impurity' Co and/or Ni from the LiO layers of the trigonal framework. For Co-rich samples the ion extraction is accompanied by partial Li +/H + exchange especially for the LT samples. The best reversibility in the 4-3.2 V range for Li intercalation/deintercalation was found for the HT Ni-rich samples. New data (EPR of low spin Ni3+ magnetic susceptibility thermal analysis) for acid delithiation of Li,Ni -xO have been reported by Stoyanova and Zhecheva.13' It is shown that short range and long range cation order governs the behaviour of the oxides towards acids.For 0.6 < x < 0.9 where partial cation order occurs acids remove Li and impurity Ni ions from the LiO layers but do not attack Ni2 + and Ni3 + ,which are segregated in neighbouring (1 11)cubic planes. For x > 0.9 where long range cation order occurs the extraction of Li and impurity ions proceeds simultaneously with partial exchange between Li from the LiO layers and protons from the acid solution. The different chemical compositions of the acid-treated oxides are clearly demonstrated by their thermochemical properties up to 200 "C. For the partially layered oxides 0.6 < x < 0.9 reduction of the highly oxidized Ni ions and cation redistribution occurs within the initial structural framework whereas for the layered oxides (x > 0.9) dehydration processes also take place.Croce et a!. have reported a promising model system of a novel type of solid state Li 'rocking-chair' battery employing TiS as the negative Li sink electrode LiCoO as the Li source positive electrode and a highly conductive gel-type polymer membrane as the ele~trolyte.'~' Normally TiS is used as a cathode in conventional Li batteries but in this case it was used as the anode to accept Li ions from LiCoO during the charge process and to release them during the discharge process. The cell showed excellent charge- discharge efficiency over a prolonged number of cycles and the fact that the Li intercalation in Li,TiS may proceed beyond the well known x = 1 step to a second step extending to x = 2 means that there is a buffer against incidental overcharge and short circuiting thus improving the battery safety and reliability.Improvements to the properties of polymer electrolytes are continuing to be sought with the search for new improved polymeric hosts and means of improving the properties of existing hosts. A new host polymer poly(tetrahydrofurfry1 acrylate) (PTHFA) containing organic solvent moieties as a pendant group has been reported by Takebe and Shir~ta.'~' The idea of the research is based on the fact that some organic solvents such as THF dissolve inorganic salts dissociating them into ions; systems having such solvent moieties as a pendant group would be expected to enhance conductivity.PTHFA was found to be completely amorphous and compatible with LiCF,SO,. The PTHFA-LiCF,SO hybrid system has been shown to exhibit ionic conductivities of ca. lop6Scm-l at 100"C and the temperature dependence of the ionic conductivity was found to be in accord with that of the dielectric relaxation (which is thought to represent the segmental motion of the polymer) of the host polymer. Although the ionic conductivity of PTHFA-LiCF,SO was lower than PEO-salt systems it should be possible to enhance the conductivity by the addition of appropriate plasticizers. 139 R. Stoyanova and E. Zhecheva J. Solid State Chem. 1994 108 211. 140 F. Croce S. Passarini and B. Scrosati J.Electrochem. Soc. 1994 141 1405. Y. Takebe and Y. Shirota Solid State Ionics 1994 68 1. Conducting Solids Covering Ionic and Electronic Conductors 49 1 Improvements in the mechanical consistency of polyvinylether (PVE) based electrolytic membranes was obtained by introducing silyl groups suitable for cross-linking during the membrane processing into the ~ide-chains.'~~ The PVE membranes containing triethoxysilyl groups in the side-chains were synthesized and cross-linked during membrane processing in the presence of LiClO,. The electrolytic membranes had excellent stability and showed promising ionic conductivity approach- ing the value of lO-'Scm- ' at 25 "C. The introduction of siloxane bonds in the network because of its flexibility contributed to a low glass transition temperature without depressing the conductivity with respect to analogous PVE unfunctionalized systems.Lee et al. have synthesized a new type of plasticizer for Li polymer electrolytes a modified carbonate made by attaching three ethylene oxide units to the 4-position of ethylene ~arb0nate.l~~ Adding 50% of this plasticizer by weight of PEO to the PEO-LiCF,SO system yielded an ionic conductivity of 5 x Scm-' at 25 "C which is two orders of magnitude higher than that found for a PEO-LiCF,SO electrolyte without plasticizer and one order of magnitude higher than that found when using propylene carbonate as a plasticizer. This new plasticizer exhibits two outstanding properties compared to most existing plasticizers; (i) rather than creating a high conductivity pathway through the plasticizer itself it appears to increase the ionic conductivity throughout the entire complex system; (ii) it enhances ion pair dissociation which in turn improves the ionic conductivity.The search for high temperature polymer electrolytes with good thermal and oxidative stability at temperatures >250 "Chas continued. Based on previous findings that the individual incorporation of 2,5-diaminobenzene sulfonic acid (DABSA) and polysiloxane (PSX) with polyimides improved the ionic conductivity Bradshaw et al. have investigated the incorporation of both DABSA and PSX with polyimide creating the quaternary PMDA-ODA-PSX-DABSA (4:3 :0.6 :0.4) copolymer system with Li dopant LiCF,SO (PMDA = 1,2,4,5-benzenetetracarboxylic dianhydride; ODA = 4-aminophenyl ether).'44 This resulted in high ionic conductivity Scm-'at 250 "C which is almost two orders of magnitude higher than that of LiCF,SO doped PMDA-ODA-PSX copolymer and four orders of magnitude higher than that of LiCF,SO doped PMDA-ODA-DABSA copolymer.The large improve- ments are achieved by the combined effect of both fixed anionic groups which increases mobile cation concentration and PSX which introduced chain flexibility. Angell et al. have reported a new type of electrolyte for use in Li batteries.'45 The electrolyte is essentially a low temperature molten salt reduced to a rubbery condition by addition of a small amount of high molecular weight polymer. High molecular weight PPO (MW lo6) or PEO (MW lo5) was dissolved in a solvent (MeCN or acetone) and the salt mixture (LiClO,/LiNO or LiClO,/LiClO,) dissolved in this solution.The solvent was then removed by careful heating. The resulting electrolyte exhibits the best characteristics of 'superionic glass' and 'salt in polymer' electrolytes in a single material i.e. cationic conduction and rubbery compliances respectively. The 14' M. Andrei L. Marchese A. Roggero and P. Prosperi Solid State lonics 1994 72 140. 143 H. S. Lee X. Q. Yang J. McBreen Z. S. Xu T. A. Skotheim and Y. Okamoto J. Electrochem. Soc. 1994 141 886. 144 J. Bradshaw S. B. Tian and G. Xu Solid State lonirs 1994 73 147. 145 C.A. Angell J. Fan C. Liu Q. Lu E. Sanchez and K. Xu Solid State lonics 1994 69 343. 492 P.R.Slater material not only showed conductivity comparable with the conductivity of the best non-aqueous Li salt electrolyte solutions but could be made comparable to that of the lead acid electrolyte by moderate increases in temperature (e,g to 100"C). High Li ion conductivities (0> 10-4Scm-1 at room temperature) have been reported in solid solutions of A site deficient perovskites La,., -xLi3,Ti03 (-0.15 > x > -0.04)by the substitution mechanism La = 3Li.'46 The dome shaped composition dependence of conductivity indicates that the conduction mechanism involves the movement of Li ions through the A site vacancies. The conductivities of the corresponding Nd perovskite with smaller lattice parameters were low (2 x Scm-') and the smaller rare earth elements Sm Gd Ho Yb did not form a single phase perovskite.Dissanayake et al. have reported high ionic conductivity in the 'system Li,- 2xSxTi -xO (0 < x < 0.2) with the Li,TiO structure; this is the first report of high Li+ ion conduction in this family.'47 Previous attempts to dope Li,XO (X = Ge Ti) always yielded so called y (or LISICON) phases rather than Li,XO solid solutions. The conductivity was shown to increase with increasing x up to a maximum (bulk) value of 8.3 x lop6Scm-l (x= 0.2) at 27 "C with E = 0.51 eV. The enhancement in conductivity is mainly attributable to a decrease in the activation energy for conduction rather than the introduction of Li vacancies. The structures of the Li-containing nasicon-type phases ~i ~11~111 ~ -x(S04)3 -y(SeO,)y (M" = Mg Ni Zn; M"' = Al Cr) have been determined bq neutron diffra~ti0n.l~' Surprisingly the Li ions were found to occupy the large (8-10 coordinate) sites.However in these sites they are shifted off-centre to give four short Li-0 bonds arranged approximately tetrahedrally. Na Ion Conductors.-Nasicon related phases continue to attract interest. Large single crystals of the nasicon phase NaZr,(PO,) have been synthesized by hydrothermal crystallization in a F-containing medium.'49 The presence of F-ions substantially lowers the temperature of hydrothermal crystallization (150 "C compared to 250 "C) and leads to larger crystals (ca.50 pm x 50 pm compared to ca. 0.5pm x 0.5 pm). The structures of the Na-containing nasicon-type phases Na,M~M~'~,(SO,) -y(SeO,) (M" = Mg; M"' = Fe In) have been determined by powder neutron diffraction.' 50 A splitting of the octahedral Na site into two sites is seen due to a reduction in symmetry (compared to the nasicon system Na,+,Zr,Si,P,-,O,,) from R3c to R3 and for x = 0.5 Na and vacancies order along the hexagonal c direction.Data collected at different temperatures showed an apparent change in symmetry from rhombohedra1 to monoclinic in the temperature range 100"C-200 "C which correlates with the observed change in activation energy for conduction in this range. Such a change from high to low symmetry with increasing temperature is odd and it is suggested that the true structure at room temperature is also monoclinic but that it is only at temperatures greater than 100 "C that significant changes from the pseudorhombohed- ral cell is observed.Mentre et al. have investigated the nasicon-type solid solution H. Kawai and J. Kuwano J. Electrochem. SOC. 1994 141 L78. 147 M. A. K. L. Dissanayake H. H. Sumathipala and A. R. West J. Muter. Chem. 1994 4 1075. 14' P. R. Slater and C. Greaves J. Mater. Chem. 1994 4 1463. 149 W. Heizeng and P. Wenqin J. Alloys Compd 1994 209 L9. 150 P.R. Slater and C. Greaves J. Mater. Chem. 1994 4 1469. 14' Conducting Solids Covering Ionic and Electronic Conductors 493 Ca,_,Na2,Ti,(P04),.'S1 The space group changes from R3 (x I0.5) to R3c (x > 0.5).For CaTi,(PO,), Ca and vacancies order along the hexagonal c direction as observed also for Na,,,M','~,M~~,(SO,) -,,(SeO,) (x = 0.5).The conductivity of CaTi,(PO,) is slightly better than that of Na,Ti,(PO,) despite the higher valence of Ca2+ compared to Na'. This is explained by the half occupancy of the cavity sites in CaTi,(PO,) so that the ionic pathway is not loaded with Ca2+ cations leading to an easier Ca2+ ion mobility. Introduction of the faster Na' ions into the framework improves the ionic conductivity up to a maximum at x = 0.5. Crystalline and glassy samples of the phase Na,NbP,O ,have been investigated by Wang et al. using ac complex impedance and FTIR.ls2The ionic conductivity of the glassy phase (cr673K = 3.1 x lo- Scm-') was at least two orders of magnitude higher than that of the crystalline phase due to large differences in the structure of the two modifications and the random distribution of Na+ ions over the equipotential sites in the glass network.In addition to the major powder X-ray diffraction lines of the nasicon-type phase many other intense lines were observed and the ionic conductivity of the crystalline phase was significantly lower than that of other nasicon analogs. The authors therefore suggested that this phase is not of the nasicon-type but is likely to be a new compound. The conductivity of the phase Na,Nb,P,O, has also been measured (06733 = 8.4 x lo-' Scm-') and is slightly higher than crystalline Na4NbP30,,. A new sodium titanogallate Na,Ti,~,Ga,+,O, (x -0.8) has been obtained from studies of the Na,O-Ti0,-Ga,O system.' s3 The structure determined by single crystal X-ray diffraction contains one-dimensional channels of octagonal cross-section extending along the b axis in which the Na ions are located.A similar phase containing channels of hexagonal cross-section (Na,~,Ti,,,,Ga,,,,O,) has been reported previ- ously. Fourier analysis revealed that the distribution of Na ions in the channels was broadened along the channel axis suggesting that these channels may be promising one-dimensional conduction pathways for Na. A range of new Na ion conductors Na,M,(X207) (M = Al Ga Cr Fe; X = P As) have been synthesized by Masquelier et ~1."~ These disphosphates and diarsenates exhibit high ionic conductivity (cT,~,~lop3to 10-1 Scm-') and good cation exchange properties especially with Ag+ ions. At temperatures varying between -20 "C and 240 "C (depending on the compound) they undergo a reversible phase transition a f-* p which is associated with a loss of long range ordering of the Na ions and an increase in conductivity.The crystal structures suggest a two-dimensional conduction pathway and this has been confirmed by conductivity measurements on single crystals. Balsys and Davis have determined the structure of the new layered alkali transition metal oxide Li,~,,Na,,,,CoO,~, by powder neutron diffra~ti0n.l~~ The structure consists of alternating Li and Na layers interleaved between octahedral COO sheets with Li in an octahedral environment and Na in a trigonal prismatic environment 0. Mentre F. Abraham B. Deffontaines and P. Vast Solid Stute lonics 1994 72 293.'" B. Wang M. Greenblatt and J. Yan Solid Stute lonics 1994 69 85. 153 Y. Michiue and M. Watanabe Solid Stute lonics 1994 70171 186. C. Masquelier F. d'Yvoire E. Bretey P. Berthet and C. Peytour-Chansac SolidState lonics 1994,67,183. R. J. Balsys and R. L. Davis Solid State Ionics 1994 69 69. 494 P. R. Slater (between the COO layers). The structural data suggests that there is a tunnel for Na ion migration through this material but not for Li and preliminary conductivity measurements suggest some ionic conductivity in this material although the relative amount of ionic to electronic conductivity is unknown. The structural features suggest that this phase may be a candidate for use as a battery cathode material like LiCoO and NaCoO,.H Ion Conductors.-Perovskite-type phases are still attracting interest as a result of + their good H+ conductivity. Thangadurai et al. have investigated the layered perovskite phases of the Ruddlesden-Popper type H,Ln,Ti,O, (Ln = La Nd Sm and Gd) and observed moderate conductivity (-lo-' Scm-' at 673 K) in a hydrogen atmosphere which is likely to be protonic in nature.'56 The conductivity is attributed to the anion deficient layered perovskite Ln,Ti,O,Vac formed in situ by dehydration which interacts with water vapour in a hydrogen atmosphere in a manner analogues to Y doped BaCeO, and SrCeO,. The phases HLa,Ti,NbOlo and HCa,Nb,O,, possessing a slightly different layered perovskite structure exhibit higher conductivity to lop3Scm-' at 673 K) under the same conditions.The Ca-containing phase shows the highest conductivity suggesting that the acidity of the protons plays an important role in determining the H conduction. + Good proton conduction in H20 atmospheres have been demonstrated in two different non-stoichiometric mixed perovskites Sr,(Sc +xNbl -x)O,- (x= 0.05 and 0.1) and Ba,(Ca,~,,Nb,,,2)0,-,.'57 Conductivity has been studied in the range 3W550K where the proton conductivity is frozen in and was found to be higher than Yb doped SrCeO for the former and Nd doped BaCeO for the latter which exhibited the highest conductivity. An advantage of the present systems over rare earth doped SrCeO and BaCeO in that they do not show electronic conduction after treatment in highly reducing atmospheres although there is evidence for a small degree of hole conduction after treatment in pure 0,.One of the problems associated with the use of polymer electrolytes in cells such as the direct methanol-air fuel cell is the need for a polymer which retains high ionic conductivity above the boiling point of water so as to combine the advantages of solid polymer electrolytes with the enhanced catalytic activity above 100"C. Another advantage of the higher temperature is that it should also help to depress the poisoning effects associated with strongly adsorbed intermediates and/or impurities in the fuel (e.g. CO). To achieve such a high temperature stable polymer Savinelli et al. have equilibrated standard perfluorosulfonic acid polymer electrolytes such as NafionTM with a high boiling point Brernsted base phosphoric acid.' 58 The Nafion/H,PO electrolyte showed reasonably high conductivity (>0.05 Scm-I) at elevated tempera- tures.Zima et al. have prepared the intercalation compound LixVOP0,.2H,0 (x= 0.4-1.0) by redox intercalation of VOP0,.2H20 with LiI in a~et0ne.l'~ The conductivity of the parent phase was relatively high (-4 x Scm-' at ambient temperature) and was attributed to proton conduction since the conductivity of ls6 V. Thangadurai A. K. Shukla and J. Gopalakrishnan Solid State lonics 1994 73 9. 15' K.C. Liang Y. Du and A. S. Nowick Solid State lonics 1994,69 117. 15' R. Savinelli E. Yeager D. Tryk U. Landau J. Wainwright D. Weng K. Lux M. Litt and C. Rogers J. Electrochem. Soc.1994 141 L46. V. Zima L. Benes R. Siskova P. Fatena and J. Votinsky Solid State lonics 1994 67,277. Conducting Solids Covering Ionic and Electronic Conductors 495 dehydrogenated VOPO was several orders of magnitude lower. The conductivity was reduced with increasing Li content which is associated with the blocking of the conduction pathways by Lif-V4+ pairs. A crystal structure determination of the hexagonal high temperature phase Cs,H,(Se0,),.H20 has been performed by Merinov et al.'" The phase exhibits a transition from orthorhombic to hexagonal symmetry at -345 K with a correspond- ing increase in conductivity which is due to the disordering of the hydrogen bond network. Other Cation Conductors.-Hong et al. have observed a modified Ca F-Al,O structure in polycrystalline Ca FlP-type Al,O electrolytes.' The non-stoichiomet- ric formula of this modified P" phase is (Mg2xAl '4-,,0,,)(Ca,0). The structure consists of alternating blocks of All 0, (four layer-oxygen spinel blocks) and AI,,02 (six layer-oxygen spinel blocks). The phase contains higher Ca levels than for the ordinary F phase (x = 0.99 compared to 0.953) and the Ca2 ions have a better + conducting environment. Thus the modified CaP"-Al,O has a high conductivity o(ionic) = 5 x lo- Scm-' at 897 K. Problems are encountered in the preparation of Cuf ion conducting glasses due to the tendency for Cu'to spontaneously disproportionate to Cuo and Cu2+ during melting in the process of glass preparation. Machida et al. have prevented this by increasing the acidity of the melt by employing the CUT-Cu,WO,-CuPO glass system to stabilize CU+.',~ ESCA spectra have indicated the presence of only Cu' and the glasses showed ionic conductivity in the range lop2 to 10-4Scm-1 at room temperature.02-Ion Conductors.-The search is continuing for solid oxide fuel cells (SOFCs) which operate at lower temperatures (below 1273K) to avoid the problems associated with fuel cell operation at high temperatures such as electrode sintering and mechanical stress which arise from different thermal expansions of the constituents. One problem with using a lower temperature is that the rate of chemical reaction at the electrodes decreases resulting in large electrode overpotentials which are the main cause of the voltage drops in the cell at lower temperatures.Ishihara et al. have examined the cathodic overpotentials of the systems Lno~,Sro~,MnO (Ln = La Pr Nd Sm Gd Yb and Y) in the relatively low temperature (-900-1200K) region.163 Normally cathodes with Ln = La are used in SOFCs but cathodes with Ln = Pr showed better characteristics maintaining low overpotential values despite the low temperature operation and exhibiting a thermal expansion coefficient closer in value to yttria-stabilized zirconia (YSZ).The same power density was obtained in a SOFC with a Pro~,Sro,,Mn03 cathode at 973K as was achieved with a Lao~,Sro~,MnO cathode at 1073 K. Therefore Pro~,Sro~,MnO appears to be a promising cathode for low temperature operation of SOFCs. Another problem associated with low tempera- ture SOFCs operation is the poorer conductivity of the electrolyte at the lower temperatures.Tshii and Tajima have demonstrated good performance at lower operating temperatures (1073-1 173 K) in a solid oxide fuel cell using cubic stabilized 160 B.V. Merinov A. I. Baranov L.A. Shuvalov J. Schneider and H. Schulz Solid State lonics,1994,69,153. 161 Y.R. Hong B.B. Tang X.F. Wu and L.S. Li Solid State lonics 1994 70/71 121. 162 N. Machida Y. Matsuda T. Shigematsu N. Nakanishi and T. Minami Solid State lonics 1994,73,63. T. Ishihara T. Kudo H. Matsuda and Y. Takita J. Am. Ceram. Soc. 1994 77 1682. 496 P. R. Slater zirconia in the Zr0,-Sc,O,-Al,O (0.89-0.10-0.01) system as an electr~lyte.'~~ The power density at 1153 K (1.O W/cm2) was approximately comparable to a fuel cell with YSZ.Fagg et al. have investigated the reduced magnesium titanates Mg -,,Til -,,04 (0 < y < 0.5) as a possible high temperature anode for fuel cell and electrochemical reactor applications and as an alternative to Ni/ZrO cermets which experience problems due to Ni sintering and thermal mismatch with the YSZ electr01yte.l~~ The reduced magnesium titanate spinels exhibited similar thermal expansion characteristi- cs to YSZ with the best match for higher Ti oxidation states (3.75-3.95+). They showed good stability in reducing atmospheres and no reaction with YSZ at 1000"C. At higher temperatures some reaction was found but the product was found to be electrically conductive indicating that problems at the electrode/electrolyte interface would not be encountered.However the conductivities at fuel cell operating temperatures were slightly lower than targeted values ( > 1 Scm-'at 1000"C compared to a targeted value of lOScm-' for fuel cell operation). In addition to use as cathode materials perovskite materials have continued to attract significant interest as electrolyte materials for SOFCs. The oxide ion conductivity of the perovskite NdAlO doped with various cations has been examined by Ishihara et uZ.'~~The partial substitution of Ca for Nd has been shown to increase the conductivity with a maximum value for Nd,.,Ca,. A10 which represents the maximum solubility limit. Substituting Ga for A1 also raised the conductivity with a maximum value of -4 x lop2Scm-' at 1223 K (slightly higher than Ca stabilized ZrO but lower than YSZ) for Nd,.,Ca,,,Al,,,Ga,.,O, which may be a result of the enlarged lattice.The same authors have also found improved conductivity for LaGaO doped by Sr (for La) and Mg (for Ga).'67 The oxide ion conductivity of La,.,Sr,.,Ga,.,Mg,.,O (-0.2 Scm-at 1223 K) was higher than that of Sc-doped ZrO but slightly lower than that of 25% Y,O,-doped Bi203. In addition p-type semiconductivity in high 0 partial pressures was found to be suppressed for both systems. Aurivillius phases (Bi202)(Am- 1) have attracted significant interest due 1Bm03m+ to the observation of high oxide ion conductivity in phases of this type in particular in metal doped Bi,VO,, (BIMEVOX).Joubert et al. have investigated the new BIMEVOX series Bi,V -xMxOl (M = Sb Nb) which involves the partial substitu- and flt)y were ,!3t)ation of isovalent Nb or Sb for V. Structural phase transitions found to occur as a function ofcomposition.'68 The highest conductivity (lop2Scm-') at low temperature (590 K) was observed for y-Bi,Vl,7Sb,,301 The performance of this BiSbVOX was similar and even slightly better than those reported for some other BIMEVOX phases when ME = Cu or Ti. The crystal structure of the low temperature (a)form of Bi,V,O, has also been studied by the same group and shown to be monoclinic and not orthorhombic as previously ~1aimed.I~~ The (V03,5Vac,.5)2 -layers are broken into rows made up of tetrahedra and oxygen deficient octahedra.Sharma et al. have investigated anion-deficient Aurivillius phases of the general 16' T. Ishii and Y. Tajima J. Electrochem. Soc. 1994 141 3450. *6s D. P. Fagg S. M. Fray and J.T.S. Irvine Solid State lonics 1994 72 235. 166 T. Ishihara H. Matsuda and Y. Takita 1.Electrochem. Soc. 1994 141 3444. lh7 T. Ishihara H. Matsuda and Y. Takita J. Am. Chem. Soc. 1994 116 3801. Ih8 0.Joubert A. Jouanneaux M. Ganne R. N. Vannier and G. Mairesse Solid State lonics 1994,73 309. 169 0.Joubert A. Jouanneaux and M. Ganne Muter. Res. Bull. 1994 29 175. Conducting Solids Covering Ionic and Electronic Conductors 497 formula Bi,W -xCux06 2x possessing orthorhombic or tetragonal Bi,WO,-like ~ structure^.'^^ The tetragonal phase is stabilized for 0.7 Ix I0.8 and exhibits good oxide ion conduction (an abrupt increase about one order of magnitude in conductivity occurs in moving across the orthorhombic to tetragonal transformation) in the temperature range 50&900 K (67003 -1 x 10-Scm-'for x = 0.7 higher than that of Cu/Ti substituted Bi,VO,,,).A new oxygen deficient Aurivillius phase (Bi,O,)(NaNb,O,,,) exhibiting oxide ion conductivity has been reported by Pham et al.' ' A reversible phase transition occurs at 860"C probably due to an ordering of oxygen vacancies in the low temperature form resulting in an increased conductivity in the high temperature form (611733 = lo- Scm-'). The conductivity of this two-layer Aurivillius phase is nearly 1.5 orders of magnitude lower than the 1 layer system Bi,VO,.,. The corresponding phase (Bi,O,)(CaNb,O,) with no oxygen vacancies has as expected a lower ionic conductivity.Similar order-disorder transitions (at 775-800 "C) have been observed in the related oxygen-deficient Aurivillius phases (Bi,O,)(Sr,Nb,MO,~,) (M = Ga Al) which consist of intergrowths of the Aurivillius phase (Bi,O,)(SrNb,O,) and brownmillerite-like SrMO,. layers (to introduce oxygen vacancies).' 72 The phases show quite good oxide ion conductivity 01173K -3.1 x lo- Scm-' for M = Ga and 2.3 x 10-2Scm-1 for M = Al. The same group have also synthesized other similar compounds (Bi,0,)(BaBi,Ti3M0 2,5) consisting of intergrowths between the Aur- ivillius phase Bi,Ti,O, and the brownmillerite-like BaMO,. (M = Sc In Ga) layers with similar oxide ion conductivities (6'173K -4.9 x lop2 Scm-' for M = In).'73 All three materials show oxygen vacancy order-disorder transitions.For M = Sc there is an observed hysteresis for this phase transition due presumably to a slow relaxation process. The phase Bi,BaO,. has been examined by powder neutron diffra~ti0n.l~~ Although the cationic positions and large anisotropic thermal parameters of the oxygen atoms resemble the anti a-AgI structure the anionic positions actually correspond to those observed in the perovskite structure. Therefore contrary to observations for the related phase Bio,79Cdo~2101,39 this phase cannot be considered as an anti a-AgI structure. It can be better described as the coexistence of the two configurations of the perovskite structure inside the crystals.High ionic conductivity ~~ CJ -~ ~ Scm-I Ea = 0.90eV was observed which was attributed to the oxygen vacancies present in the structure. F-Ion Conductors.-Ito et al. have examined the crystal structure of the tetragonal form of the high fluoride ion conductor PbSnF determined by single crystal XRD.' 75 An ordered arrangement of Pb and Sn was seen with disordering of some of the fluoride ions which accounts for the high ionic conductivity in PbSnF,. The same group has reported preliminary single crystal structural studies for the F ion conductor ~-Pbo,9Ko~,Fl~,.'76 The conductivity of this phase is lo3 times higher than P-PbF, V. Sharma A. K. Shukla and J. Gopalakrishnan J. Muter. Chem. 1994 4 703. A.-Q. Pham M. Puri J. F. DiCarlo and A.J. Jacobson Solid State lonics 1994 72 309. 17' K. R. Kendall J. K. Thomas and H.-C. zur Loye Solid State lonics 1994 70/71 221. 173 J. K. Thomas K. R. Kendall and H.-C. zur Loye Solid State lonics 1994 70/71 225. 174 C. Michel D. Pelloquin M. Hervieu B. Raveau F. Abbattista and M. Vallino J. Solid State Chem. 1994 109 122. 175 Y. Ito T. Mukoyama H. Funatomi S. Yoshikado and T. Tanaka Solid State lonics 1994 67 301. 176 Y. Ito T. Mukoyama F. Kanamaru and S. Yoshikado Solid State lonics 1994 73 283. 498 P. R. Slater due to the presence of vacancies in the tetrahedral fluorine sites. There is structural evidence that indicates that the anharmonic vibration of the fluoride ion does not play a significant role in connection to the ionic conductivity but rather the high ionic conductivity may be ascribed to normal vacancies with easy migration of fluoride ions between tetrahedra via the vacancies.