Comparison of performance characteristics for different data-processing methods for kinetic determinations
Abstract
This thesis describes a comparative study of selected features using a selected group of data-processing approaches for kinetic-based determinations. The study focuses primarily on the relative abilities of selected data-processing approaches to compensate for random and systematic changes in experimental variables. The thesis is presented in five parts. In Chapters One and Two, the data-processing approaches are developed and described. In Chapter Three, the ability of many data-processing options to provide accurate and precise parameter estimates on simulated first-order data with added random noise is evaluated for fixed-time, rate, direct computational, and iterative, least-squares data-processing algorithms. An iterative, least-squares procedure and the method of successive integration provide estimates with similar accuracy and precision under demanding conditions. For a calibration study in the presence of random noise, the recommended options include two-point fixed-time, successive integration, partial sums, a method by Guggenheim, and iterative least-squares; those options not recommended include three-point fixed-time, four-point fixed-time, two-rate, a method by Kezdy and Swinbourne, and an iterative least-squares variable-order model. In Chapter Four, a similar set of data-processing algorithms is compared based on the ability to provide accurate estimates of concentration in the presence of systematic error in both the rate constant and reaction order. In the case of the rate constant, the multipoint algorithms which determine the rate constant from the transient response provide more accurate estimates of concentration than those algorithms which do not. In the case of reaction order, the iterative, least-squares, variable-order model consistently provides accurate estimates of concentration, whereas the other options provide estimates of concentration with lesser accuracy. In Chapter Five, many different models are applied to response data from the potentiometric ion-selective electrode for ammonia, a system for which the true kinetic form of the response is unknown. Several empirical models which use only transient data provide results similar to steady-state measurements and are found suitable for quantitation of ammonia over a concentration range covering three orders of magnitude.
Degree
Ph.D.
Advisors
Pardue, Purdue University.
Subject Area
Analytical chemistry|Computer science
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.