Impedance response and modeling of composites containing aligned semiconductor whiskers: Effects of dc-bias partitioning and percolated-cluster length, topology, and filler interfaces

Abstract

Impedance spectroscopy and modeling were used to investigate the partitioning of 0-40 V dc bias in composites containing alumina and different volume fractions of silicon carbide whiskers (SiCw), which formed low-connectivity percolated clusters. Differences in response between long (∼25 cm) composite rods and thin (∼1.7 mm) slices thereof were interpreted in terms of the relative contributions to the impedance from the electrodes and SiCw-percolated clusters of the composite samples. Bias had minimal effect on the impedance of rods, because its distribution across the long percolated clusters within translated to low electric fields at the SiCw-SiCw interfaces. The impedance of thin slices was more sensitive to bias and was mainly due to such interfaces. The associated dc resistance and effective capacitance decreased significantly with increasing dc bias. A model for symmetrical Schottky energy barriers at interfaces fit the capacitance trend and outputted a parameter Φi/κ¯i. Different models for the non-linear current-voltage behavior were related to each other and indicated weak varistor-like behavior, i.e., 1.15 ≤ αV ≤ 2.57. With increasing SiCw content and composite dc conductivity, Φi/κ¯i increased and the varistor non-linearity strength αV decreased. Also, the exponent t describing conductivity divergence at percolation was reduced at large dc bias. A new model of percolated clusters was proposed and correctly predicted the qualitative character and some quantitative aspects of these experimental results. The model is based on tendencies of the current distribution, which are expected from the topological structure and contrast between interface/whisker electrical behavior. Accordingly, it outputs the voltage-distribution tendencies.