The influence of crystalline and poorly crystalline iron oxides on the surface area of Podzol and Luvisol B horizons

Iron oxides, oxyhydroxides, and hydroxides (hereafter oxides) have profound influence on soil physical and chemical properties, and greatly contribute to the specific surface area (SA) of the soils. Crystallinity and particle size determine the SA, and strongly depend on formation conditions. As in natural soil environment such conditions are extremely variable, the SA of Fe oxides covers wide ranges (ferrihydrite: 100-700 m2 g-1; goethite: 20-200 m2 g-1; hematite: 2-90 m2 g-1). However, the association of Fe oxides with other soil components may significantly affect their relative contribution to the bulk SA. To evaluate the influence of different Fe oxides on soil SA, we selected and characterized spodic and argic horizons from three Podzols (5 samples) and three Luvisols (7 samples). The samples were expected to differ in amounts and crystallinity of Fe oxides as result of differing pedogenic processes. The samples were treated with NaClO to remove organic matter, and with DCB-NaClO to dissolve both organics and pedogenic Fe oxides; the residue of the latter treatment represented the silicate phase. The SA was measured on untreated and treated samples, by applying the BET theory to N2 adsorption data. The difference between the SA of NaClO and DCB-NaClO treated samples gives an estimate of the Fe oxide SA. In Podzols, the mineral SA was affected by the oxalate-extractable Fe (FeO), representing poorly crystalline oxides (r2=0.65), in Luvisols Fe from crystalline oxides explained most of the variability (r2=0.99). This is in agreement with pedogenesis, resulting in a preferential formation of poorly crystalline oxides in Podzols (FeO/FeDCB=0.66±0.06) and crystalline oxides in Luvisols (FeO/FeDCB=0.11±0.03). Oxide phases contributed 80±17% to the mineral SA of Podzols, while for Luvisols the contribution of oxide and silicate fractions was comparable. Considering that the amount of FeDCB never exceeded 4% w/w, oxide phases had always larger SA than silicates. Unexpectedly, the SA of the oxide fraction did not significantly differ between soils with only slightly smaller values in Luvisols (158±20 m2 g-1 oxides) than in Podzols (182±34 m2 g-1 oxides). This indicates for Podzols a SA close to the lower limit of ferrihydrite, while for Luvisols suggests the occurrence of small microcrystalline goethite or/and hematite. Another explanation could be that the effects of both the larger SA of poorly crystalline oxides and the translocation of Fe and its recrystallization promote strong interactions of organic matter and oxides, e.g. due to co-precipitation in illuvial horizons, preventing complete removal of organics by NaOCl and causing underestimation of the SA. Our results suggest that in different natural environments, despite the Fe oxide types and contents, the SA of Fe oxide phases shows no major differences. That points at other effects than just crystallinity to influence the SA of Fe oxides, such as interactions during pedogenic processes.