February 1977

Document Type


Degree Name



Dept. of Applied Physics


Oregon Graduate Center


Determination of the sources of suspended particulate matter in the urban atmospheric environment is an important problem in both the study and control of this aspect of air pollution. Recent research has concentrated upon relating sources of airborne particulate matter to measurements of the chemical elements in the aerosol. Methods of source identification through elemental analysis can be divided into two classes: those methods that rely upon the observed mean elemental concentrations, and methods that utilize the intercorrelations of the observed elemental concentrations. The first class includes the mass balance techniques of source identification such as the enrichment factor method and the chemical element balance method. The second class incorporates the correlation techniques such as regression analysis, cluster analysis, and factor analysis. This dissertation presents a new method of aerosol source identification by elemental analysis that combines for the first time both mass balance and correlation techniques. This is accomplished by applying the mathematical formalism of factor analysis to the chemical element mass balance equations for the aerosol. It is shown that in the urban environment correlations of the chemical elements due to atmospheric dispersion rather than common source ancestry can be eliminated by dividing the elemental concentrations by the total aerosol mass. The correlation matrix of the elements normalized to total mass is subjected to the method of principal factor analysis. The chemical element balance equations are then put into a form that can be compared to the factor solution. The principal factor solution is rotated by two algorithms developed for this study so that the factors coincide as closely as possible with the initial sources assumed in the chemical element balance. A new source matrix of source elemental compositions is derived from the resulting factors. Finally, the fraction of the aerosol contributed by each source is calculated by reformulating the chemical element balance equations as a linear programming problem. The source contributions along with the source matrix constitute a model of the urban aerosol that accounts for the means, variances, and intercorrelations of the observed elements. This factor model possesses several important advantages over other techniques. The elemental composition of a source can be deduced from the knowledge of the amount of one element in the material emitted by the source. Also, the presence of an important unknown source can be inferred and its elemental composition estimated. All of this is accomplished by the mechanics of the model; there is very little reliance upon the intuition of the research work. In order to test the validity of a factor model derived in the manner above, fifteen aerosol samples were collected on .45 micron pore size cellulose acetate filters in the central business district of Portland, Oregon, during September, 1975. The concentrations of Al, Si, Cl, K, Ca, Ti, V, Cr, Mn, Fe, Ni, Cu, Zn, and Pb in the aerosol were determined by X-ray induced, X-ray fluorescence analysis of the samples. The intercorrelations of nine of the elements were analyzed by the principal factor method. Four factors were found to account for over 95% of the observed variance of the elements. These four principal factors were rotated so as to correspond to four sources of aerosol: street dust, metallurgical processes, plating processes, and a hypothetical zinc source whose presence was inferred from the model. Also included in the final model were the automotive and residual fuel oil sources. About 75% of the mass of the aerosol is accounted for by the factor model.





To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.