The analysis of three-component mixtures requires the determination of two concentration variables, and two independent measurements are needed. Only two are needed because knowing two concentrations the third can be computed by subtracting from one, the sum of the fractional concentrations. Commonly one or more of the ingredients contributes differently to the conductivity than to the permittivity, making the permittivity and conductivity linearly independent and enabling the Proceptor to determine individual concentrations in three-component mixtures.

As an example of how the permittivity and conductivity can vary differently with frequency, the first plot shows the (real part of the) permittivity on the horizontal axis and the imaginary part of the permittivity, which is proportional to the conductivity, on the vertical axis for two different mixtures of nylon, methanol and water. These materials are intermediate products in the manufacture of flexible printing plates and were measured during routine production. (Data courtesy Toyobo Corp., Japan) The plot shows a large semicircular curve representing measurements with a mixture having relatively high solvent concentrations and a smaller semicircle for a similar mixture with lower solvent concentrations. The different points along the curves represent different measurement frequencies. The frequency is not explicitly shown on the graph but the curves begin on the right at low frequency and high permittivity. As the frequency increases, the real part of permittivity falls while the imaginary part increases to a maximum then decreases. Both the centers and the diameters of the circular arcs differ with concentration.

Permittivity and conductivity at a single frequency can be empirically correlated with individual concentrations in known three component mixtures to construct calibrations for subsequent in-line composition determinations. An example of the results is shown below for dispersions of polyurethane in mixed water / methanol solvents.

The first graph compares the water concentration values reported by the in-line dielectric analyzer with the water concentrations reported by the laboratory GC. The agreement is excellent. The second graph shows the in-line analyzer's simultaneous determinations of methanol concentrations for the same materials, again compared with the lab values. The methanol determinations have more scatter than the water determinations, but they are still good and the Proceptor provides useful real-time analyses of these three-component mixtures.