Figure 7 (ALL FIGURES ARE COURTESY OF THE AUTHORS.)

In the process of designing a new sensor that can detect a specific analyte, one method to choose which polymer substrate to use is to determine the chemical solubility of the analyte in the polymer (42). A wide variety of chemicals and solvents can be sensed once the solubility of each analyte is known. Dielectric sensors can be quickly designed by simply studying the solubility of each analyte of interest into each polymer matrix that is used. By using other tuning parameters of dielectric sensors (e.g., the the electric field excitation frequency) and the electrode geometry, dielectric sensors can be designed for both selectivity and sensitivity to many analytes of interest on a single substrate.

Real-time or not? Real-time measurements in a process may be required if the result of those measurements are used to control another step in that process in a time-sensitive manner. Depending on how quickly events occur within the process, measurements may need to be taken very quickly and coarsely or more slowly and rigorously. For example, if moisture must be known before entering a drying cycle, real-time operation implies that the measurement is performed before the material enters the drying cycle. Continuous measurements imply real-time measurements when each measurement must be performed at a sufficient rate so that information about the process is not lost. While dielectric relaxation processes can vary widely from many seconds to less than a nanosecond, faster relaxation processes are typically used to achieve real-time operation in a process.

Dielectric spectroscopy is a promising technology that has found uses in many fields and is now being realized as a useful tool for the pharmaceutical industry. This paper discusses a number of versatile adaptations made possible by advances in electrode materials, fabrication techniques, and measurement systems. Presently, dielectric spectroscopy offers process analysts a wide variety of options, much like an advanced tool, where parameters can be tweaked as desired to implement a "tailor-made" process control solution. For instance, the measurement frequency can be changed to optimize measurements for slow bulk process or for fast intermolecular interactions. Similarly, changing the structural properties of electrodes such as coating of sensor heads can result in improved sensitivity to a particular analyte, or, changing the physical spacing between electrodes can selectively measure specific regions in a sample. Dielectric measurements can be made in a noncontact, nondestructive mode; therefore, measurement instrumentation is relatively easy to integrate at different stages of the process. Measurement time issues not withstanding, a process engineer could chose to obtain data at an early stage of the process and implement a feed-forward control system with the same ease as more conventional feedback-based control. Further, for industrial dielectric measurements, the frequency range is generally below 100 MHz; therefore, instrumentation required for such measurements is relatively inexpensive and does not have signal integrity issues associated with high frequency instrumentation. Dielectric spectroscopy has been widely investigated for measurement of pharmaceutical properties; however, majority of the work has been focused on establishing the feasibility of this technology at a laboratory level. Further work, more inline with the PAT initiative, will focus on identifying specific process needs and customizing sensor designs and circuitry.