Next: Experimental Procedure Up: Main Previous: Main
Recent work has shown that impedance spectroscopy is a useful tool for characterizing the electrical properties of composites [ 1−7]. This is especially true when certain conditions are met, i.e., the matrix is moderately conductive, the second phase particles or fibers are highly conductive, and a thin, but resistive interfacial "coating" surrounds the particles/fibers within the matrix phase. This layer can be a passive oxide film (e.g., steel particles/fibers in cement-based matrices), an electrochemical double layer/charge transfer resistance (e.g., carbon particles/fibers in cement-based matrices), or a Schottky barrier (e.g., SiC whiskers in Si3N4). The result is that the particles/fibers behave as insulating inclusions at dc and low ac frequencies. With increasing frequency, however, displacement currents through the coating short it out, and the particles/fibers now behave as conductive inclusions. A conceptual "frequency-switchable coating model" was developed to describe this behavior .
The characteristic feature in impedance spectra of composites with discontinuous conductive (but coated) particles/fibers is the subdivision of a single bulk arc without particles/fibers into two separate arcs with the particles/fibers present, as in Fig. 1. The Nyquist plots (negative imaginary vs. real impedance) shown are for plain ordinary Portland cement paste (OPC) and OPC with a volume fraction of 15% steel or glass ball bearings. Details of their preparation and measurement are given below. Frequency increases from right to left in each plot. To the right of 1640 for the OPC paste and ≈2000 for the two composites is the beginning of a large low frequency electrode arc (due to the embedded steel electrodes in each case). There is only a single bulk arc for the plain OPC (to the left of 1640 Ω) and for the glass ball/OPC composite, albeit shifted to higher resistance (to the left of 2000 Ω). For the steel ball/OPC composite, however, the bulk response (to the left of ≈2000 ) consists of two arcs, with an intersection at ≈1000 Ω. We will refer to the frequency and real resistance at this intersection as the "cusp frequency" and cusp resistance (Rcusp). The appearance of two arcs is as expected for a conducting sphere coated with an insulating layer  and is consistent with the frequency-switchable coating model. The present work describes experiments involving spherical inclusions, both insulating and conductive, in cement paste as a model system. Cement paste has a moderate ionic conductivity, is easy to process, and readily forms the requisite high resistivity passive oxide film on steel particles due to its intrinsically high pH pore solution . Furthermore, the oxide film should prevent the onset of percolation regardless of particle loading. Finally, the ability to switch the resistance of the coating on (at dc) or off (at the cusp frequency) allows us to investigate electrical mixing law behavior for both insulating and conductive particles. Experiments were also carried out with insulating glass spheres to compare with the dc result for the "coated" steel spheres. A single sphere size was employed in each case. Experiments with distributions of particle sizes will be reported separately.
Fig. 1. Typical Nyquist plots for 7 day 15% volume fraction steel and glass ball bearing/OPC composites and plain OPC. The numbers corresponding to darkened points indicate the logarithm (base 10) of the frequency (Hz) values. The minimum value of −Im(Z) for the steel, plain, and glass samples occurs at a frequency (Hz) value of approximately 104.
Next: Experimental Procedure Up:
Main Previous: Main