Masters Theses

Date of Award

8-2007

Degree Type

Thesis

Degree Name

Master of Science

Major

Chemistry

Major Professor

Kelsey D. Cook

Committee Members

Frank Vogt, Robert Compton

Abstract

Process mass spectrometry was used for the simultaneous quantitation of hydrocarbon mixtures with either six or four components. The four component hydrocarbon mixture contains isomers to increase the complexity of the mixture spectrum. The differences in relative intensities of pure component electron ionization mass spectra provided a basis for quantitation. Quantitation accuracy and precision were found to decrease as the spectral similarities among the components increased.

Selection of which ions to monitor (parameterization) was critical to optimize the analysis accuracy, precision, and speed for all mixtures tested. An empirical parameterization algorithm based on comparison of pure reference spectra of mixture components was developed for ion selection. A parameterization which monitors all the masses was used as a basis for comparison. The empirical algorithm parameterization gave analysis accuracy and precision statistically equal to the all-mass parameterization. Empirical algorithm square matrices (each component is assigned a single ion) were compared to the square matrices that were determined by the software provided with the mass spectrometer.

The six component hydrocarbon mixture which contained methane, ethane, propane, propene, isobutane, and isobutene had an r2 value of 0.99920 ± 0.00007, an average correlation coefficient value () of 0.43% ± 0.051, and a root mean squared error (ERMS) of 0.43 when analyzing all masses in the spectra. Using the empirical algorithm and selecting 35% of the masses in the spectra, the accuracy and precision are statistically similar with r2 = 0.99913 ± 0.00001, = 0.49% ± 0.59, and ERMS = 0.47. A more complex mixture consisting of the four butene isomers also gave similar results. When analyzing all of the masses in the butene isomers spectra the figures of merit were r2 = 0.986 ± 0.007, = 5.54% ± 6.66, and ERMS = 1.86. When using the empirical algorithm and selecting 26% of the masses in the spectra, that accuracy and precision were statistically similar with r2 = 0.983 ± 0.008, = 6.20% ± 7.38, and ERMS = 2.03. As shown, the empirical algorithm successfully chooses a portion of the mass spectra with excellent precision and accuracy while decreasing analysis time.

Files over 3MB may be slow to open. For best results, right-click and select "save as..."

Included in

Chemistry Commons

Share

COinS