This issue of the MRS Bulletin, on the crystal engineering of high Tc and related oxide films, essentially responds to four questions. The first three are

∎ Why has a high Tc superconducting tunnel junction not yet been fabricated?

∎ Is it possible to fabricate a high Tc tunnel junction and if so, how?

∎ How can thin-film technology contribute to the progress of high Tc superconductor research beyond the fabrication of electronic devices?

Experimental efforts to answer these questions and solve associated problems must focus on atomic-scale control and characterization of surfaces and interfaces, which we designate here as crystal engineering of high Tc and related oxide films. Meanwhile, we have started to realize that the crystal engineering technology of oxides need not be limited to high Tc super-conductor research; it can be extended to electronic materials research, where it may have a somewhat more profound significance. In this light, I would like to pose the fourth question, one that may be especially relevant to this particular topic in the near future:

∎ Will we find or develop an electronic material superior to silicon?

Ever since the first success in superconducting La-Sr-Cu-O film deposition in early 1987, extensive studies have been conducted on the fabrication of high Tc oxide films and devices. Success has been achieved to some extent in relatively simple devices using plain superconducting films or grain boundary and step edge junctions. Nobody, however, has succeeded in fabricating an SIS (superconductor-insulator-superconductor) runnel junction that shows a clear hysteresis in I-V characteristics at a temperature above 77 K. Such a junction is presumably the key to the wide electronic device application of high Tc superconductors, for example, for switching devices expected to work much faster than conventional semiconductor transistors.

The discovery of superconductivity in yttrium barium cuprate above the boiling point of liquid nitrogen by Wu et al. initiated a race in the search for higher critical temperatures (Tc) which was singular in the history of science. Since the ceramic superconducting pellets could be prepared nearly overnight and superconductivity easily demonstrated with liquid nitrogen and a levitating magnet, a great deal of effort went into garnering funds for a “new technological era.” High Tc superconductivity (HTSC) was expected to have applications in many areas of our life, be it for energy transport and high-power generators, transformers, current limiters, or for ultrafast computers and communication technology, or in medicine for NMR tomography and for SQUIDs in magnetic encephalography. Large-scale uses were also foreseen, such as levitation for trains, magnetohydrodynamic propulsion in ships, and applications for fusion reactors. Thousands of physicists and engineers searched for new materials with high Tc, and explored thin-film preparation by physical vapor deposition. In the early phase only a few scientists recognized that materials engineering problems would greatly hinder the development of applications of the cuprate superconductors. Most early efforts on HTSC went toward physical investigations of samples (crystals, layers, sheets, wires, etc.) which due to their complexity could not be reproducibly prepared, and which were not adequately characterized (see the section on Characterization). The Bi-, Tl-, and Hg-containing cuprates with Tc up to 135 K, for example, were found empirically, not by physical understanding of the HTSC phenomenon. Applications of HTSC have been limited to very specific areas, but a concerted effort should reveal opportunities for the materials research and superconductivity communities. Significant applications could be developed

which in the long run would also benefit fundamental physical investigations of the HTSC phenomenon.

Epitaxial thin-film growth of high-critical-temperature (Tc) superconductors has been intensively studied not only because it is one of the key technologies for electronic application but also because it provides suitable specimens for elucidating the superconducting mechanism. For simply making thick (>100 nm) epitaxial films, various deposition techniques such as sputtering, pulsed laser deposition (PLD), evaporation, including molecular beam epitaxy (MBE), and chemical vapor deposition (CVD) have been verified as applicable. For instance, high-quality YBa2Cu3O7–δ (YBCO) films, in terms of superconducting properties (Tc and critical current Jc), can be made by adjusting the cationic composition and choosing the right deposition conditions, e.g., oxygen pressure and temperature close to the decomposition line in the phase diagram. The knowledge and techniques accumulated in the high Tc field have been successfully transferred for the film growth of such oxides as dielectric, ferroelectric, magnetic, and optically functional materials. Pulsed laser deposition, especially, is now widely used for those materials and was addressed in a previous issue of the MRS Bulletin. However, as the demand for film quality increases, allowing films to be used in complex heterostructures like Josephson tunnel junctions and in well-designed physics studies, the meaning of the term “highquality film” has been changing.

Since the discovery of superconductivity in layered cuprates there has been a strong desire to grow thin films of these materials in a so-called layer-by-layer manner, where each layer represents one atomic monolayer. There has been worldwide interest in this problem and numerous groups have invested time and equipment in an attempt, first, to grow the known materials using a sequential deposition technique so that the atomic monolayers are stacked artificially on top of one another in a sequence defined by the structure of the given unit cell, and second, to search for new stacking sequences of such atomic monolayers, thus making metastable compounds.

Understanding the growth of thin films using any deposition technique requires information about two fundamental processes, the nucleation of a crystal and its growth. Both are strongly affected by the three thermodynamic parameters, temperature, pressure, and chemical potential (composition), at the growth front. The third parameter, the surface composition from which the desired compound can nucleate and propagate, probably provides the greatest degree of freedom in the growth process. The compositional phase diagrams of most cuprates show that several phases can coexist with the superconducting phase at a certain temperature and pressure. This is also the main reason why a widely changing surface composition, such as that present during layer-by-layer or block-by-block deposition, still produces high-quality thin films.

Oxides exhibit most of the interesting phenomena known to occur in solid-state systems. As a class of materials they may be richer in phenomenology than any other comparable class. Oxides can be insulators, semiconductors, or metals. The copperoxide-based compounds we have studied are superconductors with the highest critical temperatures. In some oxides, electrons manifest simple single-particle transport properties, with a high mobility; in others, they show strongly correlated behavior resulting in a Mott-Hubbard transition, localization, and charge- or spin-density waves. In some oxides, electron-phonon coupling leads to polaronic transport. Others show collective states such as magnetism; in some there are large local magnetic moments that can couple to form ferromagnetic or antiferromagnetic phases that exist up to high temperatures. Yet others have large nonlinear dielectric and optical properties. In fact, it would seem there is very little that some such oxide couldn't do for or to the experimenter.

Interfaces between metal oxides have not been studied as extensively as metal or semiconductor interfaces. Even in magnetic metallic systems, new phenomena are still being discovered, such as giant magnetoresistance, which has been attributed to an interface phenomenon. As general interest in metal oxides increases, researchers are studying a large variety of heterostructures consisting of superconducting oxides, conducting oxides, ferroelectric oxides, magnetic oxides, and optical oxides. As the complexity and level of integration increase, scientists need a better understanding of the interfaces in metal oxide systems. For example, interface issues in grain boundaries in high Tc oxides and in Josephson junctions of the superconductor/normal material/superconductor (SNS) type have been identified as important technological barriers. In heterostructures based on ferroelectric materials, fatigue problems are believed to be associated with the interface between the conducting electrodes and the ferroelectric metal oxides, and with grain boundaries in the ferroelectric materials.

In this article we will focus on the interface issues related to YBCO superconductors. We will describe interface phenomena observed in several systems, such as YBCO/metal contacts, YBCO/YBCO grain boundaries, and YBCO/epitaxial metal oxides. From their seemingly universal behavior, we will try to identify the origin of the interface phenomena and will describe some recent efforts to control such phenomena.

The Center for Electron Microscopy of Materials Science (EMAT) at the University of Antwerp, Belgium, started in 1967 under the initiative of S. Amelinckx. While most materials science laboratories choose a particular research theme (e.g., superconducting materials) and study it through several available techniques, the EMAT group at Antwerp has taken a different approach. From the beginning, EMAT has concentrated on a single technique: transmission electron microscopy. Since the laboratory, as well as the university, is very small, we felt it would be better to be excellent in a single technique than to have limited proficiency in several techniques. Until recently, EMAT has been able to maintain this strategy, although the group has grown considerably; at present it consists of four academic staff members, four postdocs, and 12 PhD students, supported by a technical and administrative staff of eight people.

Instrumentation at EMAT

The range of electron microscopes and related instruments in the EMAT laboratory is quite extensive and includes several instruments dedicated to specific techniques. The instruments include:

∎ A 1,250 kV high-voltage electron microscope, used mainly for studying processed semiconductor devices and for in situ radiation experiments.

∎ A 400 kV and a 200 kV high-resolution electron microscope with resolving powers of 1.7 Å and 2.5 Å, respectively. Both instruments are equipped with a TV imaging system and a video camera.

∎ A 200 kV and a 100 kV analytical electron microscope with large tilting possibilities, cooling and heating facilities, and capability for energy dispersive x-ray analysis for measuring the local composition.