Thesis (Ph. D.)--University of Rochester. Institute of Optics, 2017
III-V semiconductors are increasingly used to produce high performance infrared photodetectors;
however a signi cant challenge inherent to working with these materials
is presented by unintended electrical conduction pathways that form along their surfaces.
Resulting leakage currents contribute to system noise and are ine ectively
mitigated by device cooling, and therefore limit ultimate performance. When the
mechanism of surface conduction is understood, the unipolar barrier device architecture
o ers a potential solution. III-V bulk unipolar barrier detectors that e ectively
suppress surface leakage have approached the performance of the best II-VI pn-based
structures.
This thesis begins with a review of empirically determined Schottky barrier heights
and uses this information to present a simple model of semiconductor surface conductivity.
The model is validated through measurements of degenerate n-type surface
conductivity on InAs pn junctions, and non-degenerate surface conductivity on
GaSb pn junctions. It is then extended, along with design principles inspired by the
InAs-based nBn detector, to create a flat-band pn-based unipolar barrier detector
possessing a conductive surface but free of detrimental surface leakage current.
Consideration is then given to the relative success of these and related bulk detectors in
suppressing surface leakage when compared to analogous superlattice-based designs,
and general limitations of unipolar barriers in suppressing surface leakage are proposed.
Finally, refinements to the molecular beam epitaxy crystal growth techniques
used to produce InAs-based unipolar barrier heterostructure devices are discussed.
Improvements leading to III-V device performance well within an order of magnitude
of the state-of-the-art are demonstrated.
