The purpose of this research was to discriminate soil fractions using mineralogical and elemental analyses and to show those fractions that present greater contribution to the soil mass attenuation coefficient (μ) as well as their partial cross-sections for photoelectric absorption (PA), coherent scattering (CS) and incoherent scattering (IS). Soil samples from different places of Brazil classified as Yellow Argisol, Yellow Latosol and Gray Argisol were submitted to elemental and mineralogical analyses through energy dispersive X-ray fluorescence (EDXRF) and Rietveld Method with X-ray diffraction data (RM-XRD). The mixture rule was utilized to calculate μ of each soil. The EDXRF analysis showed as predominant elements Si, Al, Fe and Ti oxides. The highest contents were Si (914.3 to 981.3 g kg -1 ) in the sand fractions, Al (507.9 to 543.7 g kg -1 ) and Fe (32.5 to 76.7 g kg -1 ) in the clay fractions, and Ti (18.0 to 59.0 g kg -1 ) in the silt fractions. The RM-XRD allowed identifying that the sand fractions are predominantly made of quartz (913.3 to 995.0 g kg -1 ), while the clay greatest portion is made of kaolinite (465.0 to 660.6 g kg -1 ) and halloysite (169.0 to 385.0 g kg -1 ). The main effect responsible for μ was IS (50 to 61.4%) followed by PA (28 to 40.1%) and CS (9.9 to 10.6%). By using the principal component analysis (PC-1: 57.5% and PC-2: 20.9%), the samples were differentiated through the discrimination between physical, chemical and mineralogical properties. The results obtained suggest that general information about the radiation interaction in soils can be obtained through the elemental and mineralogical analyses of their fractions.
HIGHLIGHTS Semi-quantitative elemental analysis of the soils was accomplished through EDXRF. The µ, PA, CS and IS of hardsetting soil fractions was calculated by using the XCOM computer code. The mineralogical composition was determined by the RM-XRD. With PCA was possible to discriminate soil fractions and to correlate µ, PA, CS and IS.