Alternative measurement and data-processing options for transient-based analytical methods with emphasis on membrane-based amperometric sensors and overlapped chromatographic peaks
This thesis is presented in two parts. The first part describes data-processing methods intended to improve the ruggedness of results obtained using membrane-based amperometric sensors. Using an oxygen sensor as a model system, the new method is evaluated for both batch and flow systems. In the new approach, a pulsed-voltage source is used to permit measurement of currents corresponding to near-equilibrium conditions between solutions inside and outside the membrane that isolates the working electrode from the sample solution. In the batch mode, ruggedness of the new method to changes in selected variables (membrane thickness, stirring rate and temperature) is better by factors of 40- to 100-fold than for the conventional steady-state option. In the flow mode, the pulsed amperometric option is approximately 5-fold less dependent on membrane thickness and flow rate than the steady-state option.^ The second part describes an alternative approach to the deconvolution of overlapped chromatographic peaks. The approach studied here involved the use of curve-fitting methods to resolve unidirectional data for peak area vs. time. An empirical mathematical model which takes the form of consecutive sigmoid shapes was developed and tested on overlapped chromatographic data for polynuclear aromatic hydrocarbons (PAH). Preliminary results demonstrated the feasibility of using peak areas rather than peak data to resolve components that are not fully resolved in chromatographic process. For large degrees of overlap, results for two- and three-component mixtures are much better than results obtained using a perpendicular-drop method and are at least as good as results obtained by applying an Exponentially Modified Gaussian (EMG) model to the peak data from which the area data were obtained. ^
Major Professor: Harry L. Pardue, Purdue University.
Off-Campus Purdue Users:
To access this dissertation, please log in to our