AVS 52nd International Symposium
    Electronic Materials and Processing Wednesday Sessions
       Session EM-WeA

Paper EM-WeA7
Molecular Beam Epitaxial Growth of Sc@sub x@Er@sub 1-x@Sb on III-V Compound Semiconductors

Wednesday, November 2, 2005, 4:00 pm, Room 309

Session: Contacts to Semiconductors
Presenter: S.G. Choi, Univerrsity of Minnesota
Authors: S.G. Choi, Univerrsity of Minnesota
B.D. Schultz, University of Minnesota
C.J. Palmstrom, University of Minnesota
Correspondent: Click to Email
Epitaxial metallic or semimetallic layers in semiconductors have potential application in novel electronic devices. Rare-earth group-V (RE-V) compounds have received much attention since they are thermodynamically stable and epitaxial on III-V semiconductors. In particular, Sc@sub y@Er@sub 1-y@As alloys have been employed successfully in GaAs-based materials, however, application of Sc@sub y@Er@sub 1-y@As alloys to other III-V's has been less fruitful due to large lattice-mismatches. Therefore, RE-V alloys with lattice parameters close to InP or GaSb are of considerable interest and Sc@sub x@Er@sub 1-x@Sb is one of the promising candidates. The lattice parameter of Sc@sub x@Er@sub 1-x@Sb ranges from 5.85Å (ScSb) to 6.11Å (ErSb), and can therefore be lattice-matched to InP, InAs, and GaSb by controlling the Sc to Er ratio. In this work, Sc@sub x@Er@sub 1-x@Sb alloys have been grown on various III-V semiconductors by molecular beam epitaxy (MBE). The surface ordering was monitored in-situ by RHEED and LEED. ErSb grown on GaSb(100) exhibited a (1x1) surface ordering with high Sb incorporation and a mixed (1x4)/(4x1) with low Sb incorporation. In-situ XPS data showed no significant differences in Er and Sb coverage between the two surfaces, however, the amount of Ga riding on the two surfaces was different, which may cause the different surface ordering. One possible mechanism for Ga atoms to segregate on a 500Å-thick ErSb film can be explained in terms of an embedded growth model. Growth of Sc@sub x@Er@sub 1-x@Sb on InAs and InP is more challenging since intermixing of the different group-V elements at the interface would be expected to degrade the quality of interface. Different methods for minimizing the intermixing at the interface will be discussed. Supported by ONR, DARPA, and ARO.