Title

Laser beam induced current studies of Hg1-xCdxTe photodiodes

Date of Award

1-1-2006

Degree Type

Thesis

Degree Name

Ph.D (Engineering)

School

School of Engineering and Mathematics

Faculty

Faculty of Computing, Health and Science

Abstract

Laser-beam-induced-current (LBIC) is being investigated as a combined electro-optical (EO) alternative to individual electrical and optical based measurements of p-n junction photodiodes. The technique is non-destructive and employs a focused low-power laser beam that is scanned across the semiconductor surface, between two shorted ohmic contacts located at remote positions on either side of the scanned area. The LBIC image is then a two-dimensional map of the steady-state current-flow as a function of laser position. However, despite its simplicity, its application has thus far been limited to either examining photodiodes in a qualitative manner, or as mathematical analogy. This problem is largely due to the complexity of the physical process, and the difficulty in isolating the multitude of semiconductor material and device parameters that influence the LBIC profile. This work significantly develops the physical theory, and quantitative application of LBIC for the characterization of photodiodes.

Numerical modeling of the LBIC problem using the full drift-diffusion and φ,ψ treatment of an n-on-p photodiode together with supporting experimental measurements have revealed a wealth of hidden physical mechanisms, quantifiable relationships, and conditions with which to characterize the p-n junction. A further revision to Nui's theoretical model for the lateral photovoltaic effect (LPE) has also been developed such that bulk recombination, surface recombination and junction leakage are taken into consideration.

A study of the forward problem has revealed that the LBIC of a diode is the product of two simultaneous physical current-flow mechanisms that are dependent upon the generation profile interaction with the p-n junction. Under ideal conditions, the net current signal is solely due to localized changes in the junction dipole. However, under non-ideal conditions, the presence of surface recombination, localized defects and/or junction leakage can either disrupt this ideal junction potential, or introduce additional parasitic current-flows. This work has shown that the LBIC of a diode can be directly related to properties such as junction geometry, uniformity, minority- and majority carrier diffusion length, and the junction dynamic resistance. With the aid of developed theory, the diode zero-bias resistance area product, R0, and other related quantities can be obtained by analysing the majority photocarrier spreading length, Ls, from the LBIC profile. However, this correlation is limited to higher temperatures and large diodes sizes, which avoid photocarrier spreading saturation of the LBIC profile. Temperature dependent studies of LBIC were carried out, which reflect the changing behaviour of junction resistance, R0. At low-temperatures, below some threshold temperature, Tth, the peak-to-peak LBIC signal was found to reach a saturation state that is effectively independent of bulk material parameters. This low-temperature current saturation is particularly useful as it allows for the examination of junction properties in greater detail.

The application of LBIC for the quantitative extraction of p-n junction material and device parameters in the inverse problem has also been demonstrated. Reconstruction of the semiconductor junction doping profile, N, from temperature dependent measurements of the peak-to-peak LBIC is performed when a sufficient number of material, junction and measurement parameters are known or can be determined. Measurement of the minority-carrier diffusion length, Lp, from part of the LBIC profile has revealed a complex relationship between surface recombination velocity, junction geometry, and the generation profile. A variation in device geometry has shown that the fitted decay length, despite the influence of surface recombination, closely resembles the bulk diffusion length. Alternatively, use of a variable depth generation profile has shown that depending on its relationship with junction depth, the extracted length can be either an effective diffusion length, and sensitive to a range of surface recombination velocities, or, representative of the bulk parameter. By applying the full drift-diffusion model, it has been determined that the peak-to-peak LBIC dependence with junction depth is largely determined by two factors. First, the absence of current-constriction on the illuminated vertical junction, the extent of lateral current spreading over the device region, and lastly, the contact-to-contact series resistance. Diode geometries containing shallow junction depths, and small device lengths compared to minority-carrier diffusion and photocarrier spreading lengths respectively, encourage a strong current coupling, and hence sensitivity to junction depth. However, for a fixed device geometry containing an unknown p-n junction depth, the application of multiple wavelength LBIC can be implemented to adjust the generation profile to maximize the carrier collection probability over the p-n junction depth.

http://library.ecu.edu.au/record=b1604982

Share

 
COinS