Strain Relaxation Mechanisms in Lattice Mismatched Epitaxy
- PDF / 820,621 Bytes
- 13 Pages / 420.48 x 639 pts Page_size
- 65 Downloads / 211 Views
STRAIN RELAXATION MECHANISMS IN LATTICE MISMATCHED EPITAXY R. HULL, J.C. BEAN, J.M. BONAR and L. PETICOLAS, AT&T Bell Laboratories, 600 Mountain Avenue, Murray Hill, NJ 07974 ABSTRACT The relaxation of strained epitaxial layers by the introduction of misfit dislocations is reviewed. Current theoretical and experimental understanding of the nucleation, propagation and interaction of misfit dislocations are summarized. The ramifications for applicability of strained layer epitaxy to practical device structures are discussed. INTRODUCTION Lattice-mismatched materials combinations are receiving increasing interest for practical device applications. The advantages of lattice-mismatched heteroepitaxy for enhanced electronic or optical performance are: (i) The opportunity to grow high quality epilayers of materials that have desirable electronic/optical properties, but that are difficult, expensive or impossible to grow as single crystal boules for high quality substrate preparation. Example of such materials might be InAsSb alloys (for long wavelength optical applications) and II-VI compounds grown onto Ill-V (primarily GaAs substrates). (ii) The ability to vary band gap (with concomitant ability to fabricate quantum well devices or to suppress minority carrier recombination effects) in closely isostructural materials. Examples here include the GeSi/Si and InGaAs/GaAs systems (note that AlGaAs is widely used for heterojunction engineering on GaAs substrates as AlAs is closely lattice-matched to GaAs. The high affinity of Al for oxygen and the inherent high carrier mobility of InAs, however, are spurring interest in the InGaAs/GGaAs system). (iii) The potential to integrate different materials technologies onto the most attractive substrate materials from the commercial processing viewpoint: the classic example here is the attempts to grow GaAs epilayers onto Si substrates. For successful commercial application of any of the above technologies it is necessary to be able to predict and understand the limits of defect-free strained layer growth, or to be able to control and limit misfit dislocations to tolerable densities. These requirements extend not only to growth of the heterostructures, but also to subsequent processing (annealing, implantation etc.). STRAINED LAYER EPITAXY If an epitaxial layer is grown upon a substrate.with a different lattice parameter, it may be possible for the epilaiyer to adopt the substrate lattice parameter in the interfacial plane by distortion of the epilayer inter-atomic bonds. Elasticity theory then predicts that the bond length components normal to the interface will be distorted in the opposite sense to the interfacial bond components, causing a tetragonal distortion of the unit cell. This configuration contains extremely high elastic energy densities due to the individual bond distortions, particularly in covalently bonded materials where bond rotations are significant. Strained layer growth to finite epilayer thicknesses is possible only because the misfit dislocations that are necessa
Data Loading...
