4.1. Soil sample conditioning
The standard protocol in soil science, with sieving sift and air drying, cannot be used. Instead, soil samples must be sampled in a large volume with the surrounding soil solution and maintained in anoxic conditions in an airtight box. The permanence of the original blue or green blue colors during all experiments can be used as a preservation criterion. Analyses of concretions or spots of Fe concentrations, physically separated by microsampling under a binocular in an airtight box, completed the bulk analysis of the samples (Trolardet al., 1993).
4.2. Characterization of iron in the solid fraction
Many physical and chemical techniques, such as XRD, electron micros- copy, Mo¨ssbauer (Murad, 1988; Murad and Cashion, 2004, and references therein), Raman, or extended X-ray absorption fine structure (EXAFS)
Geochemistry of Green Rusts and Fougerite 239
spectroscopies, thermal analysis, and selective extractions ( Mehra and Jackson, 1960; Borggaard, 1988; Trolard et al., 1995, and references therein), have been used to study iron minerals in soils and sediments.
4.2.1. Physical techniques
In many soils where total Fe content is small, 5%, XRD is not sensitive enough to characterize iron fraction, unless special data treatment is carried (see below). It has been shown that several iron phases, especially GRs and ferrous clay minerals, are very reactive because iron can be oxidized or reduced inside their structures. Their lability makes the study difficult, and special care must be taken, both in the field and in the laboratory, for the conservation of these minerals ( Badaut et al., 1985, Trolard et al., 1996).
Mo¨ ssbauer spectroscopy has emerged as a key tool to study precipita- tions, transformations, substitutions, reactivities, and so on of iron solid or colloidal fractions (Schwertmann and Cornell, 2003), with a detection limit of 1% (absolute; Murad, 1988), and to define the conditions of syntheses of GRs (Refait and Ge´ nin, 1993; Drissi et al., 1994; Hansen and Koch, 1995; Refait et al., 1998 a,b, 2000; Simon et al., 2003). Well-crystallized, Al-free goethite and hematite show sextets, but soil goethite and hematite are frequently small sized and Al substituted, and show only one doublet at room or field temperature ( Cornell and Schwertmann, 2003); lepidocrocite is paramagnetic at room or field temperature and its spectrum shows a doublet. Fe(III) oxides, thus, show only ferric doublets, each doublet being characterized by two hyperfine parameters.
As synthetic GRs, fougerite contains both Fe(II) and Fe(III), and Mo¨ ssbauer spectra show two or three doublets ( Trolard et al., 1997), one ferric and one or two ferrous doublets. The presence of a ferrous doublet makes, thus, clear the distinction between fougerite and all other Fe(III) oxides, lepidocrocite, paramagnetic goethite, or hematite, in soils. The distinction between fougerite and Fe(II)–Fe(III) clay minerals is more difficult and needs to take into account the values of both hyperfine para- meters at field temperature and spectra obtained at low temperature in the laboratory (Federet al., 2005).
4.2.2. Selective chemical extractions
The presence of fougerite can be confirmed by selective chemical extrac- tions. Chemical extractions have been commonly used in soil science to extract different fractions and then to compare soils from different origins.
Soil Survey Staff (1999) and IUSS Working Group WRB (2006) us ed several chemical reagents to quantify organic or mineral compounds and then to define diagnostic horizons and properties. However, these extractions are not specific for one mineral with a given crystallinity and particle size (Borggaard, 1988) but have been used to operationally define fractions of solids. For example, Fe extracted by Mehra and Jackson’s method using the
240 Fabienne Trolard and Guilhem Bourrie´
dithionite-citrate-bicarbonate (DCB) reagent has been ascribed to all oxides (s.l.), excluding silicated Fe, while Fe extracted by an acid ammonium oxalate solution has been ascribed to poorly ordered oxides (s.l.) (T amm, 19 22;
Schwertmann, 1979). Nguyen Kha and Duchaufour (1969) tried to extract Fe(II) fraction in hydromorphic soils by using oxalate reagent, but this reagent does not dissolve fougerite and GRs because oxalate can be intercalated as the in te rl ay er a nio n t o f or m a st ab le ox al at e- GR ( Re fai t et al., 1998b).
Trolard et al., (1996) showed that citrate-bicarbonate (CB) reagent, that is, DCB without the reductive effect of dithionite, dissolves in a few hours synthetic GRs and fougerite. However, CB-extracted Fe does not consist solely of fougerite, but also includes organo-Fe complexes, for example, as coatings on clay minerals, which can be dissolved by the oxalate reagent. Nanoparticles of Fe(III) oxides can be partly dissolved by oxalate but not by CB. CB extraction is, thus, a better estimator of fougerite fraction and organo-Fe complexes than is oxalate extraction. To separate fougerite from organo-Fe complexes, extractions can be processed kineti- cally and/or sequentially with mineralogical controls made step by step (Trolard et al., 1996, 2007; Feder et al., 2005): fougerite is more labile than organo-Fe complexes, so that kinetics of CB extraction shows two successive steps. Fougerite is quasi-entirely dissolved after a few hours, while organo-Fe complexes are more resistant.
This method can be illustrated by results obtained in gleysol developed on granite in Brittany (France) (Photo 1 ), where seven microsampled soil features containing iron were distinguished from surface toward the depth (Table 2andFig. 5) (Trolard, 2006).
In summary, three of them (LV, GRG, and GRB) were associated with coatings located around fine roots and on surface of aggregates, two (BTO, SO) were ochre spots more or less diffuse in a clear gray or white matrix, one (PO) consisted of large rusty spots inside the zone of fluctuation of the water table, and the last one (AR) was the blue green matrix, which turns into ochre when in contact with air.
CB selective extraction showed the occurrence of a large Fe labile fraction available without reduction; it is larger in reduced than in oxidized
Table 2 Selective extraction of Fe by citrate-bicarbonate (CB) in the different features distinguished in the soil profile in Quintin
Sample LV BTO PO GRG GRB
AR
(N2) FEM FeCBt ẳ1h
FeCB to tal
=% 21.8 10.5 11.5 18.0 19.7 39.5 3.6
Geochemistry of Green Rusts and Fougerite 241
milieu (Fig. 6); 10–40% of this fraction is extracted after only 1 h. This fraction is labile, as even a short contact (a few hours) with air resulted in the immobilization of a large Fe fraction, for example, up to 40% for AR sample (Fig. 6), with respect to the amount extractible by CB under N2atmosphere for samples showing the characteristic blue green color before CB extraction.
This behavior is typical of labile compounds such as GRs and fougerite.
Classically, the difference DCB minus oxalate has been used to quantify Fe(III) oxides and explain association, or substitution, of elements (Al and transition metals) in Fe oxides ( Cornell et al., 1976; Borggaard, 1988; Singh et al., 1992). H owever, t he oxalate reagent appreciably dissolves gibbsite due to the low pHð’3ị, while CB does not dissolve gibbsite due to the neutral pH.
We can, thus, conclude that DCB minus CB extraction is a better estimator of Fe(III) oxides (Trolardet al., 1995), including goethite, lepido- crocite, hematite, and magnetite, but excluding silicated Fe (not dissolved by DCB) (Mitchellet al., 1971), fougerite, and organo-Fe complexes (dissolved by CB). Fougerite can be separated from organo-Fe complexes by using CB extraction kinetically.