Abstract
Determination of the effective atomic number is extremely important in the density measurement of multicomponent objects. With regard to the densitometry value of a numeric effective atomic number (including fractional) must correspond to the approximate values of mass attenuation coefficient, for which the attenuation of radiation is the same as for multi-component object. It is shown that for continuous spectra of X-ray radiation concepts of mass attenuation coefficient of such objects and its effective atomic number can be entered only for infinitely thin samples. This coefficient determined by shares registered radiation of multicomponent object with finite thickness depends on the distribution of the average density and thickness of the object. However, this contradicts physical meaning of the mass attenuation coefficient, which may not depend on these characteristics. The same applies to the effective atomic number, defined by dual-energy method. It is shown that only an average mass attenuation coefficient for the spectrum is invariant to the density and thickness of multicomponent object. But this coefficient is not known a priori. It is shown that it can be evaluated by the mass attenuation coefficient for the average energy of the spectrum. Traditional two-energy method for estimating effective atomic number when using sources with a continuous spectrum is characterised by a significant systematic error. The study describes a method to eliminate this bias. The article presents the results of research for materials consist of light elements with atomic numbers: low Z = 6, intermediate Z = 10 and high Z = 20, 26.