Rachel S. Goldman

Professor

rsgold@umich.edu

2094 H.H. Dow Building

T: (734) 647-6821

Bio

Projects

Publications

Research Facilities

Group


Fundamental Phenomena in Semiconductors: Alloy Formation

Sponsor: National Renewable Energy Laboratory ACQ-1-31619, Air Force of Scientific Research F49620-00-0328, Office of Naval Research N000-021-0899
We are interested in the mechanisms of solute atom incorporation and alloy decomposition in dilute and concentrated semiconductor alloys, respectively. For example, we are exploring the effects of surface reconstruction during thin film growth on the incorporation of solute atoms and other point defects in semiconductor alloys with dilute concentrations of unusual impurities. In dilute GaAsN alloys, conflicting results have been reported regarding the mechanism of N incorporation. Our studies reveal a surface reconstruction-dependent incorporation of N, where substitutional N incorporation is maximized for those reconstructions with a high number of group V sites per unit area, presumably due to the increased availability of group V sites for N-As surface exchange. Indeed, preliminary in-situ STM studies suggest that N atoms tend to occupy interstitial sites at the earliest stages of growth. Interestingly, our measurements of stress evolution in GaAsN alloys indicate significant bowing of the elastic properties of GaAsN, presumably due to the small N atomic size. In dilute GaMnAs alloys, we have quantified the Mn-composition dependence of the concentrations and distributions of point defects in alloys grown at low temperatures. Our cross-sectional STM studies reveal anti-clustering of nearest M-M pairs, suggesting that the attractive interaction expected from the neighboring ionized impurities may be overcome by a magnetic interaction. Our future plans include examining the effects of various point defect concentrations (controlled via growth conditions) on the distributions of Mn in GaMnAs and N in GaAsN, for example.In concentrated (non-dilute) semiconductor alloys, including films AB and superlattices A/B/A/B, spontaneous lateral phase separation often leads to the formation of lateral superlattices consisting of alternating A- and B-rich layers. To date, the relative roles of morphological undulations and random compositional non-uniformities in the initiation of alloy phase separation are the subject of continued debate. In heteroepitaxial semiconductor alloy films, we suggested that phase separation is a misfit-driven kinetic process, initiated by random compositional variations that later develop into coupled compositional variations and morphological undulations. In the case of short-period superlattices, we reported the first direct observation that phase separation is initiated in the first atomic layer in contact with a buffer. These mechanisms are likely to be applicable to a wide range of lattice-mismatched thin film systems. Our future plans include examining the detailed mechanisms of alloy decomposition in other materials systems in order to develop predictive models of alloy phase separation.
Highlights (Click an image for more information)
  • P Vacancy Assisted In-Ga Interdiffusion

    Recently, low-dimensional semiconductor structures have been achieved by alloy decomposition of heteroepitaxial films.  For example, in an alloy film AB or a superlattice A/B/A/B, spontaneous lateral phase separation often leads to the formation of lateral superlattices consisting of alternating A-rich and B-rich layers. The relative roles of morphological undulations and random compositional non-uniformities in the initiation of alloy phase separation are the subject of continued debate. In heteroepitaxial semiconductor alloy films, we suggest that phase separation is a misfit-driven kinetic process, initiated by random compositional variations which later develop into coupled compositional variations and morphological undulations. In short-period superlattice (SPS) structures, we report the first direct observation that phase separation is initiated in the first atomic layer in contact with a buffer.  We also report the first direct evidence for anti-site vacancy diffusion, originally predicted by Van Vechten nearly 30 years ago. These mechanisms are likely to be applicable to a wide range of lattice-mismatched thin film systems. The figure shows XSTM image of 1.7 ML InP/GaP vertical SPS, with lateral contrast modulations, presumably due to alloy phase separation.