Dissertation submitted in partial compliance with the requirements of the Masters Degree in Technology: Chemistry, in the Faculty of Applied Sciences at the Durban University of Technology, 2008. / In this work, a holistic approach to validate analytical methods was assessed by virtue of Monte Carlo simulations. This approach involves a statement of the methodsâ s scope (i.e. analytes, matrices and concentration levels) and requisites (internal or external); selection of the methodâ s (fit-for-purpose) features; pre-validation and validation of the intermediate accuracy and its assessment by means of Monte Carlo simulations. Validation of the other methodâ s features and a validity statement in terms of a â fit-for-purposeâ decision making, harmonized validation-control-uncertainty statistics and short-term routine work with the aim of proposing virtually â ready-to-useâ methods. The protocol could be transferred to other methods. The main aim is to harmonize the work to be done by research teams and routine laboratories assuming that different aims, strategies and practical viewpoints exist. As a result, the recommended protocol should be seen as a starting point. It is necessary to propose definitive (harmonized) protocols that must be established by international normalisation/accreditation entities. The Quality Assurance (Method verification and Internal Quality Control, IQC) limits, as well as sample uncertainty were estimated consistently with the validated accuracy statistics i.e. E U (E) and RSDi + U (RSDi). Two case studies were used to assess Monte Carlo simulation as a tool for method validation in analytical laboratories, the first involves an indirect polarographic method for determining nitrate in waste water and the second involves a direct determination of heavy metals in sea water by differential pulse anodic stripping voltammetry, as an example of the application of the protocol. In this sense the uncertainty obtained could be used for decision making purposes as it is very tempting to use uncertainty as a commercial argument and in this work it has been shown that the smaller the uncertainty, the better the measurement of the instrument or the laboratoryâ s reputation.
| Identifer | oai:union.ndltd.org:netd.ac.za/oai:union.ndltd.org:dut/oai:localhost:10321/431 |
| Date | January 2008 |
| Creators | Gumede, Njabulo Joyfull |
| Contributors | Bisetty, Krishna, Redhi, Gyanasivan Govindsamy |
| Source Sets | South African National ETD Portal |
| Language | English |
| Detected Language | English |
| Type | Thesis |
| Format | 260 p |
Page generated in 0.0022 seconds