BUILDUP OF PHOSPHORUS in excessively fertilized soils is of environmental concern as P transfer from soils to surface and subsurface waters is increased (Sharpley et al., 1993; Pote et al., 1996; Heckrath et al., 1995; Eghball et al., 1996). This situation has led many countries to define critical threshold levels of soil test P to limit the potential negative effects of P on water quality (Sharpley et al., 1996; Centre de Référence en Agriculture et Agroalimentaire du Québec, 2003; Breeuwsma et al., 1995). At the same time, new soil tests that are designed to more directly predict the potential for P loss to waters have been proposed (e.g., ion exchange membranes, iron oxide–coated filter paper, easily desorbable P; Sims et al., 2000). Such invasive methods give a relative measure of the ease of removing phosphorus from soil solids. In contrast, deterministic approaches for predicting the potential loss of phosphorus from agricultural soils with different properties and fertilization histories could be based on knowledge of chemical species (forms) of soil phosphorus and a more fundamental chemical understanding of P release from these specific species.
The increase in soil P solubility that is often correlated to an increase in total soil P concentration may be explained in part by changes in solid-phase speciation or by the affinity of orthophosphate (PO4) for sorbing soil components. Adsorption appears to be the dominant retention mechanism that regulates dissolved phosphate at low concentrations, whereas phosphate mineral precipitation controls P solubility at high concentrations (Lindsay et al., 1989). Consequently, distinguishing between adsorbed phases and precipitates is critical to ascertain the long-term behavior of P in soils. Typically, adsorption isotherms from laboratory experiments are characterized by an L-curve (Sposito, 1984), which can be fitted with a Langmuir or Freundlich isotherm model. This type of isotherm predicts that as the PO4 loading rate approaches the maximum adsorption capacity of the soil, additional phosphate cannot be retained by the soil. Unlike surface-adsorbed chemical species, the solubility of a solid-phase precipitate is essentially independent of the amount of the solid phase present (Lindsay, 1979). Thus, precipitation of minerals such as Ca-, Al-, or Fe-phosphates at higher soil P concentrations may represent a sink for P that has a constant solubility under given chemical conditions. However, the type of phosphate mineral formed and soil conditions such as pH and presence of dissolved complexing species will determine the phosphate activity in solution (Lindsay, 1979).
In the past, thermodynamic models of mineral solubility predicted that dissolved PO4 would be controlled at equilibrium by Fe- and Al-phosphates in acidic soils and by Ca-phosphates in neutral and alkaline soils (Lindsay, 1979). However, kinetics of P transformations were not considered (Bohn et al., 1985), even though kinetic limitations often exert considerable influence on P speciation in natural environments. Also, adsorbed P phases were poorly understood and difficult to include in such models. For example, the concentrations of ammonium oxalate–extractable Al and Fe in soils have often been found to be the best variable to predict P sorption capacities of acidic (Laverdière and Karam, 1984; van der Zee and van Riemsdijk, 1986; Freese et al., 1992; Simard et al., 1994) and neutral to calcareous soils (Ryan et al., 1984; Tran and Giroux, 1987). This correlation suggests that oxide mineral surfaces are significant P-sorbing components at all pH levels. Likewise, results of energy-dispersive X-ray analyses of excessively fertilized soils showed that P-rich particles contained P predominantly associated with Al in amorphous solid phases, even for neutral to slightly alkaline soil samples (Pierzynski et al., 1990). Such observations illustrate the need for direct identification of soil P species, regardless of soil properties, when trying to understand and quantitatively model long-term changes in P solubility in P-enriched soils.
In the present study, X-ray absorption near-edge structure (XANES) spectroscopy was used in conjunction with sequential chemical fractionation to characterize the dominant solid-phase species of P in selected soils. Total-electron-yield XANES studies at the P K-edge of several commercial phosphate powders have shown that each compound had a unique spectrum that reflected the specific molecular environment of P (Franke and Hormes, 1995; Okude et al., 1999). Rose et al. (1997) determined the local structure of P during hydrolysis of FeCl3 in the presence of phosphate using P K-edge extended X-ray absorption fine structure (EXAFS) spectroscopy. For soils, XANES spectroscopy has been mainly applied to sulfur and metal speciation (Fendorf and Sparks, 1996), but Hesterberg et al. (1999) have shown the feasibility of using this approach for more direct identification of some soil P species. X-ray absorption near-edge structure spectroscopy has the main advantages of being element specific and nondestructive (no sample pretreatment required). It further provides information on the local molecular bonding environment of the element (Fendorf and Sparks, 1996). Unlike X-ray diffraction, poorly ordered mineral phases can also be characterized by XANES spectroscopy (Schulze and Bertsch, 1995). The objective of this study was to determine chemical speciation of P in long-term-fertilized, P-enriched soil samples using synchrotron XANES spectroscopy and sequential chemical fractionation. For this purpose, soil samples were selected to represent a range of properties such as pH, texture, organic matter content, and P source.