No abstract

No abstract

X-RAY FLUORESCENCE IdETHODS have been widely used in the analysis of multicomponent mixtures. It is unfortunate, however, that accurate analyses are often not attained in practice because of sample matrix effects and particle size effects. Mitchell (I) describes the problem in elaborate detail. In a recent paper, Alley and Myers (2) discuss ways of using inverse estimation in linear regression to account for these effects. Also, Campbell and Brown (3) have reviewed mathematical and empirical methods.The consideration of particle size of the components is extremely important in x-ray fluorescence analysis for the case of granular matwials. Claisse and Sampson (4)) and Bernstein (5) discuss the particle size-intensity relationship.This paper describes a procedure which makes use of certain statistical techniques to approximate the relationship that exists between the x-ray intensity of each component and the concentration of all of' the components in the mixture. The method described permits the estimation of component concentrations even though measurement errors and particle size effects exist.Special attention is given to "Tichloral" igniter mixtures. These mixtures are 'comprised of potassium perchlorate, titanium and aluminum powders, and a small percentage of a binder such as polyisotlutylene. THEORETICAL JJEVELOPMENT OF PROCEDURELucas-Tooth and E'yne (6) developed a theoretical concentration-intensity model accounting for matrix effects. It is this model that serves as the basis for our development. This model can be expressed as:where V , is the percentage of component n in the mixture; I, is the x-ray intensity for component j ; the dn)'s are constant parameters related to mass absorption coefficients (7). adn)

No abstract