CSIRO PUBLISHING www.publish.csiro.au/journals/ajsr Australian Journal of Soil Research, 2007, 45, 224–232 Rietveld-based mineralogical quantitation of deferrified oxisol clays M. E. Alves A,E , Y. P. Mascarenhas B , D. H. French C , and C. P. M. Vaz D A Departamento de Ciˆ encias Exatas, Escola Superior de Agricultura ‘Luiz de Queiroz’ – ESALQ/USP, Caixa Postal 09, 13418-900 Piracicaba (SP), Brasil. B Departamento de F´ ısica e Inform´ atica, Instituto de F´ ısica de S˜ ao Carlos – IFSC/USP, Caixa Postal 369, 13560-970 S˜ ao Carlos (SP), Brasil. C CSIRO Energy Technology, Lucas Heights Science and Technology Centre, PMB 7, Bangor NSW 2234, Australia. D Embrapa Instrumentac ¸˜ ao Agropecu´ aria, Caixa Postal 741, 13560-970 S˜ ao Carlos (SP), Brasil. E Corresponding author. Email: mealves@esalq.usp.br Abstract. Although the mineralogical quantitative analysis of the soil clay fraction can provide useful information for the improvement of soil management practices, the quantitation of all clay components normally requires a combination of different analytical techniques, which makes this determination expensive and time-consuming. One alternative for more expeditious mineralogical quantitations consists of using the Rietveld method for the treatment of X-ray diffraction (XRD) data. In this study we evaluate the accuracy of the mineralogical quantitative analyses of oxisol deferrified clays carried out with the application of the Rietveld method to XRD data obtained for both non-spray- and spray-dried samples. Linear regression analyses were carried out for comparing the XRD-Rietveld results with those calculated from X-ray fluorescence spectroscopy (XRF) data. Correspondence was observed between the XRD-Rietveld and XRF-derived data, confirming the potential utility of the Rietveld method for soil clay mineralogical quantitative analysis. Although sample preparation by using the spray drying procedure tended to improve XRD mineralogical quantitation, accurate results can be also achieved when this procedure is not available in the XRD laboratory. Additional keywords: disordered clay minerals, kaolinite, weathered soils, X-ray diffraction. Introduction Oxisols cover almost 60% of the Brazilian land and comprise 40% of the mapping units of the soil map of the S˜ ao Paulo State (de Oliveira 1999). The clay mineralogy of these soils contains the usual components, namely kaolinite, gibbsite, haematite, goethite, and titanium oxides (Schwertmann and Herbillon 1992). Besides these minerals, it is not unusual to find maghemite as part of the clay fraction of the basalt-derived red oxisols (Fontes and Weed 1991). The less weathered oxisols can also present small amounts of hydroxyinterlayered vermiculite and/or illite. On the other hand, weathering promotes the destabilisation of kaolinite and, by consequence, the accumulation of oxides such as gibbsite, haematite, and goethite, which are responsible for net positive surface charge usually found in the subsurface layers of the deeply weathered oxisols. Although organic matter plays an important role in the physical-chemical behaviour of weathered soils, the soil inorganic fraction presenting particles ≤ 2 μm, i.e. the clay fraction, also exerts influence on soil properties such as water retention and resistance to root growth, and on the magnitudes and mechanisms of the sorption processes that take place on the soil-solution interface. Therefore, clay mineral characterisation can provide important data for the improvement of the soil management practices related not only to farming but also to environmental conservation. In general, combinations of 2 or more methods such as selective dissolution, chemical, thermal, and spectroscopic analyses, and X-ray diffraction (XRD) must be used for quantitative soil mineralogical analyses. For this reason, the conventional approaches used for quantitation are often expensive and time-consuming. One alternative to expedite soil quantitative mineralogical analysis is the application of the Rietveld method in the treatment of XRD data. The Rietveld method was proposed in 1969 (Rietveld 1969) for the refinement of crystal structures from neutron diffraction data. Later, the method was extended to XRD (Young et al. 1977) and to quantitative phase analyses (Hill and Howard 1987; Taylor and Matulis 1991). In the Rietveld method, a calculated diffraction profile of a mineral mixture, which corresponds to the sum of theoretical patterns calculated for each phase considering its refined unit cell features, geometric factors dependent on the Bragg angle, and thermal effects, is fitted to the measured pattern by means of a non-linear least-square procedure. For this fitting, unit cell dimensions, diffraction peaks shapes and widths, instrumental effects, and effects related to preferred orientation of diffracting planes, absorption, and extinction of X-rays are refined or corrected in order to minimise the differences between the calculated and measured XRD patterns. © CSIRO 2007 10.1071/SR06123 0004-9573/07/030224