Comparison of three sequential extraction procedures used to study trace metal distribution in an acidic sandy soil

On an acid sandy soil contaminated with trace metals (Fe, Mn, Cu, Pb and Zn), three sequential extraction procedures were compared to determine the efficiency of the reagents used and the effects of the step order on the fractionation of metal species. In all cases, a magnesium nitrate solution (MgNIT) was previously used to extract exchangeable forms. In the first procedure (I), the next extraction step was performed with sodium acetate buffer (NaOAc), as used on calcareous soils, to dissolve active calcium carbonate. Then trace metals bound to different forms of oxi-hydroxides (NH(2)OH, TAMOx and TAMAs fractions) were extracted before organic matter/sulfide oxidation with hydrogen peroxide at pH 2.0 in nitric acid medium (OMHyd). Finally, residual bound metals (RESID) in each procedure were extracted with a nitric-hydrofluoric-perchloric acid mixture. The second procedure (II) was the same as I, but without the NaOAc step, because of the absence of carbonate in the study soil. In procedure III, the NaOAc step was omitted and the oxidizable organic/sulfide fraction was extracted with sodium hypochlorite at pH 8.5 (OMOCl) before the reducible fractions. This study first showed that NaOAc may remove considerable amounts of metals (especially Mn and Zn) in other forms than exchangeable ones. Procedures II and III give similar results for Fe, Mn and Zn forms, which were mainly found in fractions of inorganic soil components, but not for Cu and Pb. Copper distribution was affected by the position of the oxidation step in the sequence. In procedure II, where the oxidation step (OMHyd) ended the sequence, Cu was mainly recovered in the TAMOx fraction. However, in procedure III, where the oxidation step (OMOCl) preceded the NH(2)OH, TAMOx and TAMAs steps, Cu was found in both OMOCl and TAMOx fractions. Lead distribution varied with oxidation reagent: it was partly removed in the OMHyd fraction of procedures I and II, and to a much lower extent in the OMOCl fraction of procedure III, probably due to the alkaline pH of the reagent in the latter case.