Complete uniformity generates bland food; minute admixtures of spices provide tasty surprise and special quality. This statement holds even more strongly for semicounductors, our materials for modern electronics. The significance of defects in otherwise almost perfect solid materials is described by Hans J. Queisser (Max Planck Institute for Solid-State Research, Stuttgart, Germany) and Eugene E. Haller (University of California at Berkeley, USA) in a special issue of Science (14 August 1998, p. 945 - 950) on "Control and Use of Defects in Materials".
Semiconductors, such as germanium, silicon, or gallium arsenide were initially considered to be terribly irreproducible materials. Their physical properties, for example electrical conductivity, color, hardness, or magnetic parameters scattered widely and could not be definitely entered into handbook tables. Hence arose this contemptible name semiconductors as being somewhere dislodged between proper metals and electrical insulators. Around midcentury, with the birth of the first germanium transistors in 1947, it became indisputable that not the solid material itself controls all ist properties, its defects, however small in quantity, determine many of the crystal qualities. Tremendously useful variability is thus created.
The understanding of the defects within these materials was, and still is, absolutely essential for the development of modern semiconductor microelectronics. The idiosyncracy of germanium and silicon against foreign atoms was scientifically converted into the viable and indispensable method of doping. Conductivity can thus be controlled over huge ranges, an absolute prerequisite for electronics usage. The lifetime of an extra minority charge carrier (like an electron in an acceptor-doped crystal) became the vital figure of merit for crystal perfection. Long lifetimes of injected carriers enable transistors to function. Measurements of lifetime are easy, yet they provide an incorruptible index for crystal purity and crystallographic perfection.
Some defects are fatal; too much admixture renders the material useless. Many defects are vital, on the other hand. The controlled substitution of a host atom by a specific foreign atom is the key idea of semiconductor materials engineering. Profound knowledge of the physics and chemistry of the defects yields the basis for these applications. The article by Queisser and Haller cites many examples for this research and its technical consequences. A multitude of experimental techniques are today available for the characterization of defects. The detailed atomic structures are nowadays clearly resolved, using techniques such as electron microscopy or tunneling microscopes.
Defect dimension leads to a natural classification. Zero-dimensional defects are foreign atoms, vacant lattic sites, host atoms squeezed into interstitial lattice positions, or atoms on the wrong sublattice of a multi-atom crystal. Dislocations are one-dimensional line-defects. Areal defects arise from faults in the stacking order of lattice planes and from surfaces or interfaces. Precipitates, especially important for metallic impurities, are three-dimensional. All these configurations have their specific physical properties and critically affect device behavior. Analysis, suppression, and neutralization of deleterious defects are described in this review.
Queisser, queisser@quasix.mpi-stuttgart.mpg.de, and Haller, eehaller@Ibl.gov, have long been collaborators on these research topics, especially for germanium, silicon, and the opto-electronically important compound semiconductors. Powerful theory is available today: minimum energy configurations or defect wavefunctions are calculable; defect coupling to the host lattice and the diffusion through the host crystal are beginning to become quantitativeley accessible. An overview of these theoretical aspects - as well as enticing color illustrations are included, along with experimental and technological details, in this review paper in Science.
Published: 20-08-98
Contact: Hans-Joachim Queisser
Max Planck Institute for Solid-State Research,
Stuttgart, Germany
Phone: +49-711-689-1600
Fax: +49-711-689-1602
Journal
Science