Characterization of AlGaAs/GaAs Heterojunction Bipolar Transistors Using Photoreflectance and Spectral Ellipsometry
- PDF / 1,156,837 Bytes
- 6 Pages / 414.72 x 648 pts Page_size
- 24 Downloads / 212 Views
ABSTRACT Photoreflectance spectroscopy (PR) and spectral ellipsometry (SE) have been used to characterize the doping and structure of heterojunction bipolar transistors (HBT). This information provides a more complete description of the epitaxial HBT structure than is possible by relying solely on electrical characterization of specially processed test structures. Additional benefit is derived from the nondestructive nature of both SE and PR. The measurements are fast enough to be implemented on all production-bound HBT material. We describe our recent results comparing capacitance-voltage measurements with PRderived doping levels in the emitter layer of the HBT. We also describe some work comparing SE fit results with Auger electron spectroscopy depth profiles for InGaAs contact layer composition and thickness. INTRODUCTION Characterization of AlGaAs/GaAs heterojunction bipolar transistors (HBT) is necessary in order to screen the material prior to device processing. Identification of material grown outside of specifications of layer thickness and alloy composition, and removal of this material from the processing stream, can lead to significant yield enhancement and reduced product cost. Although several nondestructive characterization methods are available, many of the epitaxial growth chambers currently in use cannot accommodate in situ probes. In addition, ex situ, nondestructive analysis of fully grown device structures has not enjoyed widespread use in production facilities due to a lack of straightforward methods for treating the experimental data. The combined use of photoreflectance (PR) and spectral ellipsometry (SE) nondestructive techniques with the C-V destructive tests provides a useful combination for material screening. We have found good agreement of the photoreflectance (PR) determined dopant level in the n-GaAs collector layer with C-V derived values.' The dopant values are not absolute concentrations. This was illustrated several years ago by J. L. Shay. 2 More recently, F. H. Pollak et al. have described experiments which show the magnitude of the photovoltaic effect in HBTs with similar structures to those described in our work. Pollak's work clearly shows that the magnitude of the photovoltaic effect depends upon the intensity of the probe and pump beams. 253 Mat. Res. Soc. Symp. Proc. Vol. 406 0 1996 Materials Research Society
Sample-to-sample comparisons of calculated dopant concentrations should be valid, so long as the light intensities are kept constant.3 C-V emitter layer doping concentration analysis accuracy is limited by the AIGaAs layer thickness, which is of the same order of magnitude as the width of the depletion region. In order to compare the C-V and PR emitter dopant concentration values, we grew HBT structures with 200 nm thick emitter layers. With the thicker AIGaAs layer, a more reliable comparison between the two techniques is possible. The results of this comparison will be described. We have previously described some of our work using SE to determine the structure of
Data Loading...
