Publication Date


Document Type


First Advisor

Ballantine, David Stephen

Degree Name

Ph.D. (Doctor of Philosophy)

Legacy Department

Department of Chemistry and Biochemistry


Flavor; Food--Sensory evaluation; Neural networks (Computer science)--Data processing


Arrays of polymer-coated surface acoustic wave microsensors (SAW) are used in conjunction with a variety of signal-processing algorithms known as artificial neural networks (ANN). This format of data analysis has the capacity to characterize complex mixtures of volatiles and semivolatile organic compounds found in common flavoring agents and natural products. This complementary data analysis technique requires the use of a variety of primary analysis techniques currently found within the flavor and aroma industries (i.e., GC-MS, GCO, sensory panels). The results from these primary analysis techniques can be used to successfully train the ANN where it is subsequently used as a secondary, more robust/flexible, quantitative or qualitative technique. The initial study, which minimizes the number of training sets while retaining the robustness of an ANN, utilizes a 2-D bitmap matrix. The matrix is obtained by converting the time domain kinetics of sensor response into a bitmap. The high data throughput of this approach enables the ANN to reach considerably lower detection limits while retaining a relatively small predictive error for contaminated base flavors. The detection limits for this technique are 150 ppm for base flavor adulterants with an overall predictive error of 0.34% ± 0.01%. Industrial processing such as heating, fermenting, and blending can have a dramatic impact on aroma and flavor control. Real-time analysis of processing methods is typically limited to single-parameter measurements (i.e., colorimetry, pH). Sensory panels or various chromatographic techniques, while limited to postreaction characterization, are unable to participate in real-time reaction monitoring because of the time restraints of the panel evaluation and chromatographic separation process. This creates the demand for new dynamic forms of real-time monitoring techniques. The SAW-ANN approach is used in this second investigation to monitor the formation and degradation of key impact components within a Maillard reaction. The obtained predictive results had an average relative standard deviation of 2.37%. This error fits well within the range of acceptable flavor analysis errors (<5%) for the examined processing technique.


Includes bibliographical references ([201]-204).


xiii, 204 pages (some color pages)




Northern Illinois University

Rights Statement

In Copyright

Rights Statement 2

NIU theses are protected by copyright. They may be viewed from Huskie Commons for any purpose, but reproduction or distribution in any format is prohibited without the written permission of the authors.

Media Type