With the fall of the Berlin Wall, the enlargement of the European Union (EU) by countries from the former Central and Eastern Europe became possible. Science and education are among the best assets of the accession countries, but the unfettered free market represents an increasing threat to their intellectual capital. Science and education have long been instruments of international cooperation. Now they have a vital role to play in leading Europe towards a new age of enlightenment.

The novel approach to testing the validity of various models of the Schottky barriers is reviewed. It is based on expected different dependence of the barrier on the external perturbation (temperature or pressure shift) for the two competitive contemporary models of the barrier formation, i.e. The metal-induced-gap-states model and the model invoking the presence of real defect states at the interface. In the former the barrier shifts would reflect the individual shifts of the relevant band edges, while in the latter the influence of the external perturbation of the ionization energy of the defect that pins the Fermi level would be reflected on the barrier shifts. The available data on Si-silicide barriers and MBE grown Al/AlGaAs junctions are in a direct conflict with the predictions of the MIGS theory, while at the same time can be positively explained within a frame of the defect model.

A standard textbook description of the crystalline solid is a perfect periodic arrangement of atoms. In real life how­ ever, all solids are imperfect, and it is the defects that make them useful. It is well known that their presence in insulators at even trace concentrations influences dramatically all their properties — from hardness to conductivity. The latter is just what made semiconductors (bad in­ sulators) the backbone of modern elec­ tronics. On the other hand, radiationinduced defects create headaches for nuclear and space engineers. It is almost impossible even to men­ tion all aspects of defect physics and technology in such a short article, and I shall therefore concentrate on a small, but quite important class of defects in solids : deep level point imperfections in semiconductors. It is one of the bran­ ches of semiconductor research which is developing rapidly and is still very far away from saturation. Defects, perturbing the ideal periodi­ city of the crystal, must introduce some extra states in the host crystal energy spectrum. Suppose at the beginning that the defect is just a foreign atom (do­ pant) replacing one of the host atoms and differing from it by a unit valence charge (like arsenic or gallium in silicon). This extra (or deficient) charge is likely to be weakly bound to the impurity atom owing to the host response, and the defect will resemble a hydrogen atom, whose energy levels are scaled down by the host dielectric constant ε, and an effective change m*/m of the electron (or hole) mass. For example, in GaAs — considered to be the silicon of the next century — the hydrogen-like donor bin­ ding energy{ED = 13.6 (m*/me)*ε-2 eV} is only about 6 meV (a 2000-fold reduc­ tion) and the effective Bohr radius is as large as about 100 A. Such a large spread of the electron wavefunction about the donor makes it fairly insen­ sitive to the chemical nature of the dopant. This is especially evident for the excited p-like states, having a node on the impurity site. Because of that, the Zeeman splitting of these states is often used to obtain a highly accurate deter­ mination of the effective mass of the host semiconductor. When the dielectric screening is not so strong, and especially when the ef­ fective mass is not so small, the impurity Bohr radius shrinks and the differences in energy levels introduced by the diffe­ rent isovalent impurities become larger. It may happen, however, that a strong binding impurity (or better to say defect) potential supplemented by the electronphonon interaction will lead to a collapse of the impurity electron wavefunction on to the defect, making it strongly loca­ lized with a mean electron radius of the order of the nearest neighbour distance in the host. An almost immediate conse­ quence of this is a sinking of the defect levels (although not always) towards the centre of the semiconductor gap. At this stage a fairly simple hydrogen effectivemass model description tends to fail and a more complex description taking into account not only the detailed defect chemistry, but especially the arrange­ ment of all neighbour atoms around the defect becomes necessary. This beco­ mes a formidable task and is solved satisfactorily in only very rare cases. Un­ fortunately, most defects fall into this category, and this is why for many years deep level defect research could hardly be considered to be an exact science. A big change started in the mid-seventies when major progress was made in semiconductor technology, computa­ tional techniques for the microscopic theory of point-defects in semiconduc­ tors and, especially, in the development of highly complex characterization tech­ niques. They allow us now to trace even single-defect recombination processes as well as single-defect imaging. Some five years ago, Paul Dean re­ viewed 1) the basic properties of "deeplevels" in semiconductors. I shall con­ centrate, therefore on those aspects which have undergone the most rapid development in recent years, pointing out also how "deep" we are and what the problems are that still await a satis­ factory explanation.

The talk gives a brief introduction to the large lattice relaxation (LLR) phenomena in solids with main emphasis given to the deep levels in semiconductors, and effects caused by a charge exchange between defects and bands (ionization and capture). Several experimental examples serving as the evidence for LLR are presented and discussed.