It is well recognized that soils and sediments contain not only aero-anaerobic (facultative anaerobic) and anaerobic well identified bacterial strains, but also large ill-defined bacterial communities able to reduce ferric to ferrous iron. Such bacteria can dissolve and weather efficiently (at least for some populations and communities) ferric (oxyhydr)oxides, e.g., ferrihydrite, goethite, hematite, lepidocrocite, by ferrireduction [1,2,5–7,9,16,25,28,32–34].

Although the molecular mechanisms of the bacterial iron reduction remain not completely understood, two bacterial groups should be differentiated on the basis of their metabolism: the non-fermentative (or iron respiring) bacteria and the fermentative iron reducing bacteria. In the first group, facultative anaerobic bacteria such as Shewanella sp. or strict anaerobic bacteria such as Geobacter sp. use ferric iron as terminal electron acceptor for their anaerobic respiration as indicated in simplified reactions (2) and (3):

These IRBs use mainly organic compounds as carbon and energy sources (reaction (2)), but they can also use inorganic compounds, e.g., hydrogen, as indicated in reaction (3). For fermentative bacteria, iron reduction may accompany fermentation of facultative or strict anaerobic bacteria such as Bacillus sp. or Clostridium sp., respectively, and in this case, ferric iron serves as supplementary terminal acceptor as shown in simplified reactions (4) and (5).

Works performed using Most Probable Number techniques (MPN) to evaluate the pools of IRB involved in the dissolution of ferric (oxyhydr)oxides have shown that large communities are present in different types of soils. Depending on the horizon depth, the site (rhizospheric or non-rhizospheric soil), the soil type, the nature and amount of available energetic and carbon sources, communities ranging from 10² to 10⁸ IRB per gram of dry soil have been evaluated in upper soil horizons [14,32] (Table 1). Fermentative bacteria occur to be more abundant than non-fermentative.

Several authors have reported that IRB can use different well-defined organic compounds (glucose, acetate, methanoate...) as carbon and energy sources [9,17,22]. However, it was only recently that a significant correlation between the bacterial degradation (oxidation and mineralization) of soil organic matter and the bacterial iron reduction and dissolution of ferric (oxyhydr)oxides has been determined (Fig. 2) [32]. Different stages appeared in the correlation indicating that in the three sampled soil horizons, the limiting factor was the ferric iron availability. However, the discrepancy that is often observed between the amount of mineralized organic matter (O.M.) and the extent of iron reduction and dissolution underlines the fundamental role played by the quality and biodegradability of O.M. during the process [32].

The addition of plant origin material, e.g., cellulose, promoted IRB activity [28]. These IRB reduced soluble and solid forms of ferric iron (e.g., ferric citrate, ferric (oxyhydr)oxides, phyllosilicates) [6,25,28,35].

Different physical, chemical and biological parameters control these dissolution and weathering processes. Poor crystalline features and small particle size (large specific surface area) promote reduction and dissolution [20,29]. Metal substitution in the mineral structure has effects that depend on the nature and substitution rate of the metal: aluminium substitution decreases the goethite dissolution kinetics but manganese or cobalt has no effect on [6].

The presence of ligands (complexing agents) of ferric iron [9], as well as that of an electron shuttle such as anthroquinone-2,6-disulfonate (AQDS) [9,25] promote iron reduction. Quinone bearing compounds like humic substances seem to be able to increase electron transfer [18]. The contact between bacteria and oxyhydroxides [21,30] promotes bacterial reduction and dissolution. Interactions at the interface (including contact) seem to be of major importance that has to be more clearly defined.

Bacterial reduction of solid ferric species seems also able to take place not only in waterlogged but also in better drained soils when microbial activities consume oxygen during organic matter biodegradation, thus promoting the development of anoxic conditions and anaerobic microsites [33]. Acidification of soil by fermentative bacteria can also increase Fe(III) solubility and stabilise Fe²⁺ in aqueous solution [2,5].

Bacterial dissolution of Fe and Mn oxides promotes solubilisation of substituted metals (Co, Ni, Cr...), as observed during waterlogging of a Ferralsol of New Caledonia [28]. Moreover, such bacterial weathering modifies the metal distribution within the geochemical soil compartments. Metal content (Fe, Mn, Co, Ni and Cr) of most labile water soluble and exchangeable compartments increased. It has also been observed that metal concentrations of amorphous and poorly crystallised Fe oxides increased significantly [28]. Table 2 summarizes these results after weathering of oxides by aero-anaerobic and anaerobic bacteria that have mineralised 1.4% of soil organic carbon and solubilised 0.19% Fe and 17% Mn.

When weathering was more intense after bacterial solubilisation of 0.77% Fe and 32% Mn, a more pronounced amorphisation of mineral constituents was noted: transmission electron microscope observations and microanalysis then revealed the formation of amorphous FeSi gel containing also Mn, Ni, and Cr [28].

The mixed Fe(II)–Fe(III) compound green rust was primarily found to be formed during corrosion processes of iron-based materials [10,23,31]. Recently its natural occurrence was demonstrated in hydromorphic soils [36,37] and in particular in a Reductisol of the forest of Fougères (Brittany, France), where IRB activity can occur significantly. Works have been done to screen and evaluate the presence of IRBs that are able to reduce and dissolve ferric iron from oxihydroxides in different horizons. Simultaneously, laboratory experiments were performed to determine the soil IRB activity using soil organic matter (SOM) as carbon and energy sources and soil ferric species as ferric substrate. The presence of relatively large communities of IRBs (fermentative and non-fermentative) has been observed in upper soil horizons (from 0- to 80-cm depth) using the Most Probable Number technique in different liquid media [32] in microplates in anaerobic conditions (Table 3). After incubation of suspension of the soil horizon samples (soil/water ratio: 1/10) in anaerobic conditions, the number of IRBs increased strongly, particularly for fermentative bacteria (Table 3). So IRBs are able to grow in Fougères Reductisol, using SOM as carbon and energy sources.

Simultaneously to IRB counting before and after soil sample incubations, measurements have concerned the biodegradation of SOM evaluated by organic carbon mineralization, i.e. CO₂ evolution and iron dissolution as Fe²⁺ ions in the same incubation devices. Neither a significant dissolution of iron nor a production of CO₂ was observed in abiotic controls. The mineralisation of SOM by anaerobic bacteria occurs in all soil horizons from 0- to 100-cm depth (Fig. 3). This organic matter biodegradation, expressed by carbon mineralisation (C–CO₂ evolved) per thousand of total soil organic carbon, appears to be more important in upper horizons where organic matter biodegradability is relatively higher than in deeper horizons.

Fig. 4 shows the dissolution of iron as Fe²⁺, occurring simultaneously to SOM biodegradation. This indicates relatively high IRB activity in upper horizons and particularly in 20–30-cm depth horizon where fougerite was observed [36,37]. This reductic horizon, where fougerite formed, presents large IRB communities able to grow significantly using soil organic matter as source of carbon and energy and having a very efficient iron-reducing activity. Such IRB activity can be involved in the formation or in the control of the chemical and physicochemical parameters controlling fougerite formation.

Experiments were performed in laboratory-controlled conditions in batch devices as described by Ona-Nguema et al. to determine the possible role of IRBs in green rust (GR) formation [25]. The non-fermentative bacterium Shewanella putrefaciens was cultivated in anaerobic conditions (N₂ atmosphere) in 100-ml flasks receiving a basic liquid nutrient medium (pH 7.5). Each flask received 80 mM of Fe(III), the source of ferric iron and electron acceptor as oxyhydroxyde (lepidocrocite γ-FeOOH). Electron donor was provided by 23–75 mM of methanoate (HCO2−) (sodium formate) in each flask and finally an electron shuttle for increasing electron transfer was added using 100 μM of anthraquinone-2,6-disulfonate (AQDS). The inoculum of S. putrefaciens was prepared by culture in trypcase soy broth in the presence of oxygen and brought in experimental flasks to have a concentration of 2×109 cells ml⁻¹. Abiotic controls were also performed. Measurements were accomplished during experiments with the soluble phase on Fe(II), methanoate, pH, carbonate, sulphate... and with the solid phase as well for mineral characterisation (Transmission Mössbauer Spectroscopy, TMS; X-ray Diffraction, XRD; Diffuse Reflectance Infrared Fourier Transform Spectroscopy, DRIFTS; Scanning Electron Microscopy, SEM; Transmission Electron Microscopy, TEM). All experimental and analytical methods were presented by Ona-Nguema et al. [25,26]. Results presented display the correlative production of Fe(II) and consumption of methanoate during the bacterial reduction of lepidocrocite (Fig. 5). Fe(II) production and methanoate consumption are obviously linked and the beginning of stationary state of Fe(II) production corresponds to the total consumption of methanoate. An increase of pH (not shown) from 7.2 to 9.0 can be related to iron reduction (proton consumption). Redox equation (6) between HCO2− and γ-FeOOH indicates that one mole of methanoate is used to reduce 2 mol of Fe(III):

After 15 days of incubation, 62 mM of Fe(III) are reduced in Fe(II) and almost all methanoate was used (Fig. 5). The balance indicates that 12 mM of methanoate have been used or removed from the solution in excess of the 31 mM necessary to reduce 61 mM of Fe(III). So, part of methanoate can be absorbed on green rust and/or on siderite that have been formed but another part is likely incorporated into the bacterial biomass.

In abiotic control iron reduction was not observed (Fig. 5). Bacterial reduction and GR formation occurred efficiently without AQDS in the medium [25].

The colour of the solid phase changed during the bacterial reduction of lepidocrocite. From orange (lepidocrocite colour), the medium became orange-brown then dark green testifying the formation of green rust. Green rust was clearly characterised using different methods indicated here. In these experimental conditions, hydroxycarbonate green rust of type 1, was formed. XRD analysis performed after 15 days of incubation indicated a mixture of a GR1 and of few siderite (Fig. 6). The observed dhkl interplanar distances were those of GR1(CO32−). The parameters a (0.317 nm) and c (2.272 nm) of the hexagonal structure are also comparable to those of GR1(CO32−) obtained by chemical synthesis [8]. Transmission electron microscopic observation of the dark green solid phase showed hexagonal GR crystals (Fig. 7a) and the electron diffraction provided the same values of parameter as those indicated by XRD. Optical microscopy (Fig. 7b) displayed hexagonal crystals of green rust of approximately 13-μm diameter. Formation of hydroxysulphate green rust 2 in Shewanella putrefaciens cultures have been also observed by Ona-Nguema et al. [26] only for very low values of the {[HCO3−]/[SO4−]} ratio (i.e. <0.17). The results of this laboratory experiment, comparatively to other works using Shewanella sp. and ferrihydrite as ferric iron [11,27,38] showed the major role of IRB in the formation of well defined green rust due to their involvement in the reduction processes of ferric (oxyhydr)oxide.

Soils contain aero-anaerobic and anaerobic bacteria that use, in the absence of oxygen, either fermentation processes or anaerobic respiration to reduce organic and inorganic electron acceptors other than oxygen such as NO3−, Fe(III), Mn(IV), SO42−, fumarate, CO₂. Iron-reducing bacteria (IRB) able to reduce ferric iron from soluble, e.g., organo-ferric complexes, or solid compounds, e.g., oxides or (oxyhydr)oxides, are present in large populations and communities not only in hydromorphic soils (Redoxisols, Reductisols), but in well-drained and non permanently waterlogged soils as well. These iron-reducing bacteria use directly ferric iron as main electron acceptor, e.g., Shewanella sp., Geobacter sp., or use it simultaneously along to fermentation, e.g., Bacillus sp., Clostridium sp. Their activities depend on physical (e.g., mineral particle size), chemical and crystallochemical (e.g., metal substitution in the structure), biological (e.g., nutritional and energetic requirements, contact between bacteria and mineral) parameters. Soil organic matter and ferric iron bioavailability are among the main environmental parameters. Such bacteria are responsible for weathering processes of ferric (oxyhydr)oxides and release of soluble ferrous iron and associated trace metals (e.g., Mn, Ni, Co, Cr). They are also involved for redistributing these elements in the geochemical soil compartments.

Natural occurrence of GR has been observed in Reductisols (Gleysols) and Redoxisols (pseudogley), and was well characterized in a Redoxisol of the forest of Fougères (France). It was named fougerite. The Fougères Redoxisol contains, in the horizons where fougerite was discovered, large fermentative and non-fermentative IRB communities, which present efficient iron-reducing activities, using soil organic matter as carbon and energy sources. The fermentative IRB communities appear larger than the non-fermentative and have a high potential activity in the soil horizons where GR formed. However, their involvement in GR formation is not yet defined. The significant relation between bacterial iron-reducing activity and GR formation, has been well established in laboratory experiments with a non-fermentative bacterium Shewanella putrefaciens, used as a reference IRB cultivated in the presence of methanoate as electron donor and of lepidocrocite as source of iron and electron acceptor. By bacterial reduction of iron, the concomitant productions of Fe(II) and HCO3− are responsible for the formation of hydroxycarbonate green rust of type 1.

Iron-reducing bacteria appear as the main factor involved in GR formation. However, environmental parameters such as the availability of carbon energy sources and of ferric iron, dynamic and activity of bacterial populations have to be defined more accurately in field conditions to progress in the knowledge of the processes and of their impact and possible application in the cycling of iron and associated elements in soil functioning.