Randall M.
Feenstra did undergraduate work at the University of British Columbia, and received a
Ph.D. degree from California Institute of Technology in 1982. He worked at the IBM T.J.
Watson Research Center in New York for 13 years applying the technique of scanning
tunneling microscopy to the study of semiconductor surfaces and heterostructures. He
joined the Carnegie Mellon faculty in 1995.Advances in semiconductor physics in the past several decades have transformed daily life. Silicon, the basic material from which integrated circuits are fabricated, dominates the semiconductor market in terms of volume and revenue. Other semiconducting materials, such as gallium arsenide, fill important niche applications in high-speed electronics and optical devices.
What will be the next hot material? Materials such as zinc selenide and gallium nitride have large band gaps, which should prove useful in electronics that function at high temperatures (for example sensors in automobile engines). Large band gap materials also may find application as light emitting diodes and lasers in the blue to ultraviolet spectral range. The recent announcement of blue lasers (see http://nsr.mij.mrs.org/news/flash.html and http://www.sony.co.jp/CorporateCruise/News/96D-014E.html on the World Wide Web) demonstrate that the future is bright for this field of research.
Most optical semiconductor devices are heterostructures, containing thin layers of differing materials grown one on top of the other. The different band gaps of the respective materials localize the electron or hole carriers in specific layers. Adding another dimension to such band gap engineering is strain between the layers. Although strain allows the tailoring of otherwise inaccessible band alignments, it creates some undesirable side-effects such as rough layer growth and dislocation formation. These often lead to reduced operating efficiency of devices and may produce catastrophic device failure. In addition there are a number of physically interesting consequences of strain on the geometric structure of surfaces, such as novel surface reconstructions (different arrangements of atoms on a surface compared to the bulk) and unique morphology of steps and islands on the surface. These effects are especially common in alloys containing three or four different chemical constituents.
My research studies the geometric and electronic
structure of semiconductor heterostructures on an atomic scale. I use a scanning tunneling
microscope (STM), an instrument that allows one to image the atoms arranged on a surface
and also to probe the spectroscopy of electronic states at specific spatial locations.
With its unprecedented capability for imaging a surface on a nanometer (10-9 m) scale, the
STM has revolutionized the field of surface physics. Whereas conventional surface probes
such as low-energy electron diffraction reveal mainly the periodic aspects of surface
structure, the STM provides an detailed view of non-periodic features such as defects,
dislocations, and interfaces between different structures or materials.
In the STM, a sharp metal probe tip is brought within about 1.0 nm of the surface of a sample to be studied. A voltage is applied between tip and sample, and a current flows through the vacuum barrier by means of quantum-mechanical tunneling. The magnitude of the current depends exponentially on the separation between the tip and sample, so that by raster scanning the probe tip over the surface a map of the surface morphology can be constructed. Since the current flows through only the outermost few atoms of the probe tip, atomic resolution can be obtained in the images, as illustrated in the accompanying figure. Varying the voltage between tip and sample allows spectroscopic measurements of the density of electronic levels lying a few volts on either side of the Fermi level.
When I moved to Carnegie Mellon from IBM research labs, IBM donated much of the apparatus for performing our STM studies. With this equipment, my research group can continue our research program using cross-sectional STM to study semiconductor heterostructures. In this work, the heterostructure is cleaved apart to expose a cross-sectional face, thereby permitting observation of the interfaces between layers and the unique properties of the thin layers themselves. At IBM we demonstrated the capability to atomically resolve interface structure and to spectrally resolve both two-dimensional subbands of electron confined in thin quantum wells and states introduced into the semiconductor band gap by atomic impurities. At Carnegie Mellon, together with a graduate student and postdoctoral researcher, we continue this type of study on layers of gallium nitride buried between adjacent layers of gallium arsenide in an effort to determine the composition and homogeneity of the material. Future work is planned on zinc selenide films, where the use of gallium arsenide substrates creates dislocations and eventual device failure. A detailed understanding of the interface structure may lead to improved device reliability. Both of these projects are conducted with external collaborators (at Oklahoma State University, and at Philips Laboratories) who grow the layers of semiconductor crystal.
In a major new initiative, we are constructing a molecular beam epitaxy (MBE) system to grow semiconductors for in situ characterization by STM and other techniques. This work is done in collaboration with Prof. David Greve of the Electrical and Computer Engineering Department, and is funded by the Office of Naval Research. The system will be used for studies of gallium nitride and aluminum nitride growth on various substrates. A significant problem affecting the growth of this material is lack of a suitable substrate with a similar size lattice constant, so a very high dislocation density occurs in the grown films. Remarkably, devices still function with reasonable efficiency and lifetime. We hope to understand film growth and strain relaxation mechanisms to explain why the gallium nitride devices are so much more resistant to degradation by dislocations compared to those made from zinc selenide. The MBE system being constructed will be suitable for growth of a wide range of other materials as well. We intend in the future to probe general mechanisms of epitaxial growth in strained semiconductor alloys.
