The recent revolution in high temperature superconducting materials has generated a wave of intense excitement and activity that has swept through the scientific community, attracting the attention of the news media and general public as well. The reason for this is twofold: the unexpected occurrence of superconductivity at such high temperatures is of immense scientific interest, and the new high temperature oxide superconductors may have important technological applications.

Based on a large amount of experimental information and (presumed!) theoretical understanding, the prevailing view prior to 1986, when high temperature superconductivity in oxides was discovered, was that the maximum value of the superconducting transition temperature Tc of any material would not increase much above ˜23 K, the high Tc record held since 1973 by the A15 compound Nb3Ge. In fact, between 1911 (the year H. Kammerlingh Onnes discovered superconductivity) and 1986, Tc only increased at an average rate of ˜0.25 K per year. However, within the last two years the maximum Tc value of the new copper oxide super-conductors has risen at an average rate of ˜50 K per year to its present value of ˜125 K! Thus, superconductivity near or above room temperature no longer seems out of the question, as it did a few short years ago! Moreover, the oxides were generally regarded as the least likely candidates for high Tc superconductivity due to their low concentrations of charge carriers. An understanding of the origin and nature of high Tc superconductivity in the new oxide compounds constitutes one of the most important and challenging scientific problems that has emerged in recent years.

Our understanding of the superconducting mechanism in the newly discovered oxide superconductors is closely connected with a detailed knowledge of the structural properties displayed by these materials. Moreover, it is important to explore possible connections between structural and superconducting properties and to establish general structural principles which can guide our search for new compounds.

We present here a survey of the structures of the recently discovered oxide superconductors. The common feature of these structures (with the exception of cubic Ba1-x Kx BiO3-y) is the presence of one or more planes of copper atoms with four strongly bonded oxygen atom neighbors in a square planar arrangement at a distance of approximately 1.90 Å (“CuO2 planes”). In some structures these planes occur in groups, with the individual planes inside a group being separated by one Y or Ca layer and the groups being intercalated by a variable number of LaO, BaO, CuO, TIO and/or BiO layers. (The intercalated CuO layer is crystallographically distinct from the “CuO2 plane”). A natural model for superconductivity which arises from these structures is that the conductivity occurs predominantly in the Cu02 planes, while the other (intercalated) layers provide in some fashion carriers and/or the coupling mechanism necessary for the superconductivity.

La2CuO4, crystallizes in an orthorhombic distortion of the tetragonal K2NiF4 structure, a so-called layered perovskite. The important question here concerns the stoichiometry of La2CuO4 prepared by various techniques. Regarding the La and Cu sites, recent neutron diffraction studies of powders have shown that the La:Cu ratio can vary from 2:1 by ≳1% in various preparations., Single crystals have been grown variously from CuO, PbO and Li2O-B2O3 fluxes. The last two fluxes incorporate some Pb and Li, respectively, into the crystals, but the cation ratio is found to be 2:1 within the accuracy of x-ray refinement.

It is also known that the oxygen stoichiometry corresponds to La2CuO4 within ≃ 1%. Extensive and precise magnetic susceptibility (χ) and other studies of powder samples showed, however, that the physical properties are extremely sensitive to < 1% changes in the oxygen content. A clear indication from electric and magnetic measurements is that the oxygen stoichiometry is always that, or richer than that, corresponding to a Mott insulator with an exactly half-filled band. Consistent with this, the stoichiometry is more properly written La2-z CuO4-y with small, positive y and z.

Several groups found that bulk superconductivity above 30 K can be produced in La2CuO4 by annealing powders or single crystals at high oxygen pressure. The weight gain in this process corresponds to producing La2CuO4.13, and these authors believe the excess oxygens go in as O-2, based on iodometric titration measurements giving the amount of [Cu-O]⊥ in the sample. Jorgensen et al. found that there is a phase separation below 320 K of La2CuO4 and the O2-loaded phase in oxygen-rich samples; the latter structure is also orthorhombic, but the position of the excess oxygens could not be determined unambiguously.

Twelve compounds with the formula RBa2Cu3O7-δ where δ≍0.1 and R=Y or a lanthanide element except Ce, Pm, or Tb, crystallize in the same layered orthorhombic perovskite-like structure (see the article by I.K. Schuller and J.D. Jorgensen in this issue). All such compounds are superconducting with superconducting transition temperatures Tc≍92-94 K, except for R=Pr. The compound PrBa2Cu3O7-δ appears to be a special case, since it does not exhibit metallic behavior and is not superconducting. Historically, the prototype YBa2Cu3O7-δ compound is singularly important since it was the first superconducting material with a Tc greater than 77 K, the boiling point of liquid nitrogen. In the meantime, yet higher Tc's have been obtained in YBa2Cu3O7-δ at very high pressure (Tc˜107 K at 150 kbar)5 and in new layered compounds in the Bi-Sr-Ca-Cu-O (maximum Tc≍110 K)6 and Tl-Ba-Ca-Cu-O (maximum Tc≍125 K)7 systems (see the article by A.W. Sleight, M.A. Subramanian, and C.C. Torardi in this issue).

The origin and nature of the high Tc superconductivity of the RBa2Cu3O7-δ compounds, and the other layered copper oxide compounds as well, are presently not understood and constitute a formidable challenge to experimentalists and theorists alike. One of the most intriguing possibilities is that a magnetic mechanism, rather than the electronphonon interaction, is responsible for the formation of the superconducting electron pairs in the high Tc copper oxides. The primary evidence for a magnetic pairing mechanism is the proximity of antiferromagnetism and superconductivity as the concentration of holes in the conducting CuO2 planes is varied,8 as discussed below for the RBa2Cu3O7-δ compounds.

The first report of a superconductor in this family came from the laboratory of B. Raveau. The system was Bi/Sr/Cu/O. The Tc's were in the range 10-20 K, and this system therefore did not attract much attention when there was intense effort on 90 K superconductors. Neither the structure nor the composition of the superconductor in the Bi/Sr/Cu/O system was known at the time of the first publication, a common feature with all discoveries of new high Tc superconductors in the past couple of years. We now know that the first discovered member of this new family of superconductors can be represented as Bi2Sr2CuO6. However, the composition and structure of this phase are still not precisely established.

Subsequent to the French work, the Bi/Sr/Ca/Cu/O system was studied in Germany, and higher Tc's were obtained. Although patents were filed; this work was not publicly announced. In January 1988, Maeda announced high Tc in the Bi/Sr/Ca/Cu/O system and Hermann announced high Tc in the Tl/Ba/Ca/Cu/O system.

It is now known that there exists a very large family of compounds of the type (AO)mM2Can-1CunO2n+2. The A cation can be Tl, Pb, Bi or mixtures of these elements. The value of m may be one or two but is only two when A is Bi. The M cation is Ba or Sr, and substitution of Ca by Sr is frequently observed. The number of consecutively stacked CuO2 layers is indicated by n. The highest temperature at which zero resistivity has been obtained in this family is 122 K, and this has been obtained for several distinctly different compositions (Table I).

Before the pioneering work of Bednorz and Müller in finding superconductivity near 30 K in lanthanum-barium-copper oxide,. oxide superconductors were well known, but perhaps not fully appreciated. Most anomalous among those superconductors was the perovskite structure material BaPb0.75Bi0.25O3, with a superconducting transition temperature (Tc) of 12 K2 and a surprisingly low density of states at the Fermi level. The increases in Tc for copper-oxide-based materials continue to generate worldwide excitement, but from both a chemical and theoretical point of view, high Tc superconductivity observed in a noncopper containing material is also of considerable interest.

Recently we found that the simple cubic perovskite compound Ba0.6K0.4BiO3 displays a superconducting transition temperature near 30K—a Tc considerably higher than that of conventional superconductors and surpassed only by copper containing compounds. This material is in stark contrast to the now well-known copper oxides for two reasons: (1) superconductivity occurs within the framework of a three dimensionally connected bismuth-oxygen array (and not a 2-d array as in the Cu-O based compounds) and: (2) there are no magnetic fluctuations present in the chemical system, either in the superconductor itself or in the nonsuperconducting end member compound, eliminating the possibility that the high Tc might be caused by magnetic interactions. The parent compound BaBiO3 is, however, of considerable interest due to the presence of a structurally frozen charge disproportionation of the bismuth atoms, considered by many to be the electronic equivalent of the antiferromagnetism observed in the nonsuperconducting cuprate host compounds.

The ideal undistorted perovskite ABO3 structure consists of a regular array of equally dimensioned BO6 octahedra sharing all corner oxygens with neighboring equivalent octahedra, with 180° B-O-B angles.

Oxide compounds have been extensively studied through the years because they exhibit a broad spectrum of electrical, magnetic, and optical properties providing both scientific and technological interest. Most oxides are insulators, but a few of them (e.g., LiTi2O4 or BaPb1−x BixO3 show metallic conductivity and even superconductivity at low temperatures. The discovery of superconductivity at 35 K by Bednorz and Müller in the cuprate La-Ba-Cu-O system prompted the search for other high Tc compounds among this oxide family. Superconductivity above liquid nitrogen was then rapidly achieved with the Y-Ba-Cu-O system (Tc=90 K) and subsequently, with the Bi-Sr-Ca-Bu-O and Tl-Ba-Ca-Cu-O systems, Tc was raised to 110 K and then 125 K.

A common feature of these new high Tc cuprates is that they belong to the large family of materials, termed perovskites, which have been studied over the years because of their ability to absorb or lose oxygen reversibly (i.e., for their nonstoichiometry in oxygen). It had been previously established in the field of superconductivity that Tc is extremely sensitive to compositional stoichiometry.

While high Tc superconductivity was first discovered in bulk material, it was apparent that thin films of these materials, particularly the compound YBa2Cu3O7-δ, would be of great interest to both science and technology. In this sense the development of these materials parallels a similar history in the low Tc materials. Initially, most of the low Tc materials of interest were single element metals such as Nb, Pb and Al in bulk form. Later work, mostly in magnets, led to the development of compounds or alloys of such metals as Nb-Sn, Nb-Ti, and many others. However, many physical and technological investigations required thin films with thicknesses in the range of 0.1-10μm. Microwave, infrared, and critical current studies are examples of some of the scientific uses of thin films. A few examples of the applications would include josephson junction-based digital computer circuits, SQUID (Superconducting Quantum Interference Devices), transmission lines, and interconnects and rf mixers. These studies are also of great interest in the high Tc materials. It is readily apparent that scientific and technological developments in superconductivity are closely interwoven.

The high level of interest in thin films can be appreciated by observing that it was barely a few months after the announcement of superconductivity above 77 K that the first films of these complex, multi-element materials, superconducting at about 86 K were announced. These early efforts at thin film fabrication were generally accomplished using multi-element deposition techniques but subsequent development has seen many varieties of film fabrication techniques used quite successfully to fabricate high-quality films.

High Tc oxide superconductors can be prepared by a variety of novel metallurgical or ceramic processing techniques. For example, forming metallic precursors of YBa2Cu3 or REBa2Cu3, where RE is a rare earth element, followed by oxidation has yielded the corresponding ceramic oxide in pellet, wire, and thin film forms. Ceramic fabrication techniques (which often involve incorporating fine particles of the ceramic in a suitable vehicle) such as doctor-blade tape casting and screen printing have been utilized to synthesize both tapes and thin films. Due to intense worldwide effort to develop these ceramics into useable forms, many unique fabrication techniques have been employed, such as the melt-textured growth method of Jin et al. and dynamic compaction. In addition, it is likely that numerous laboratories, particularly industrial, are developing proprietary methods to fabricate wire and other forms but have not published these results due to the intensely competitive nature of this area. In this article, we review some of the methods reported so far for preparing high Tc ceramic superconductors, and we discuss one example of a proprietary method being developed at one of the author's (LDW) laboratories.

Tapes of YBa2cu3O7-x fabricated using the doctor-blade tape casting technique were first displayed by AT&T Bell Laboratories scientists at the 1987 March meeting of the American Physical Society. This method involves mixing ceramic powder, a binder to provide green (unfired ceramic) strength, a plasticizer to provide flexibility, a dispersant to prevent particle agglomeration and a solvent. After mixing all these components together, the solution is poured into a “doctor blade” device which allows a controlled thickness of material to be deposited on a plastic film carrier that passes underneath the doctor blade.

There is a widespread feeling that the discovery of high temperature superconductors will force us to change our way of thinking about superconductivity in solids. It has steadily emerged that the simple free-electron picture is inadequate, that a new mechanism of superconductivity is most likely at work, and that modifications (if not outright revisions of the many-body theory) are needed. The situation is not entirely without precedent: much the same could have been said of superfluidity in liquid He. That, however, did not cause such a stir: it had been anticipated some 12 years before its eventual discovery, and the transition temperature of liquid He is so low that experiments have been confined to the few laboratories around the world with a milli-kelvin capability. Finally, the normal state of He was already wellunderstood, so that theorists were poised and ready to tackle the problems posed by the superfluid.

Contrast the oxides. Even the rather extensive earlier studies of Ba1-x PbxBiO3 and other superconducting oxides did not prepare us for the advances of the past two years. Laboratories all over the world have been able to prepare and study the new superconductors rather easily, although well-characterized samples and incisive experiments have not been so easy to come by. The flood of new information poses a particular challenge for condensed matter theory — to distil the essence of these complicated multicomponent materials and to explain how the genie of high temperature superconductivity has escaped after so many years. Despite the existing understanding of the properties of oxides, much work remains to be done before we have a good grip on the many-body theory of these strongly correlated systems.