Introduction

Rampant anthropogenic activity has detrimentally increased the pollution levels in all of the earth’s natural reservoirs [e.g. 1, 2]. Because sediments represent a sink for these pollutants, sediment pollution in marine, estuarine, lacustrine and fluvial systems is increasingly being recognised as an emerging global concern [3–5]. Accumulation of toxic metals, metalloids and persistent organic contaminants in aquatic sediments poses a threat to benthic organisms and can further impact on other aquatic organisms if contaminants are released into the aqueous phase either through sediment re-suspension events (both natural (e.g. slumping, bioturbation) and induced (e.g. dredging activity)) or through chemical variability in the sediment environment (e.g. redox changes, pH changes). Trace metals are particularly important in this regard, as their persistence, toxicology and propensity to bio-accumulate in higher trophic levels are concerning for public health [e.g. 6, 7].

Iron is the most abundant transition metal in the Earth’s crust, yet it is not commonly reported as a soil or sediment pollutant [8, 9]. An exception lies in the rice industry, where low oxygen conditions enhance Fe(II) concentrations and can result in Fe toxicity, which detrimentally affects rice yields [10, 11]. Under oxic conditions, however, the bioavailability and toxicity of iron is limited by its exceedingly low solubility [12] which results in the precipitation and subsequent stability of iron (oxy-hydr)oxide minerals (e.g. ferrihydrite, goethite, hematite). These ubiquitous mineral phases are redox reactive and are well known for their propensity for adsorbing various inorganic and organic pollutants [e.g. 13–17]. They have thus attracted much attention as a viable mineral amendment option and have successfully been applied to remediation of polluted soils [see reviews 18–20] and sediments [21–23]. Typically, these amendments can be applied as in situ soil treatments whereas sediment treatments require either prior dredging [24] or the development of a confined aquatic disposal cell [25]. Qian and co-workers [23] have shown that in situ ferrihydrite additions are able to immobilise various trace metal contaminants in sediments whilst simultaneously decreasing their bioavailability and toxicity. Hematite and zero-valent iron additions have also been shown to reduce trace metal toxicity in sediments [26] and pilot-scale work on dredged sediment samples has shown that these mineral phases have significant ability to immobilise As, Cd, Cu, Mo, Ni and Zn when applied at a 5 % amendment ratio [24].

Because of the importance of Fe and Fe minerals to contaminant cycling and stabilisation, this work assesses the status of knowledge of these interactions in contaminated sediment systems. Although high iron loading can result in Fe toxicity, these cases are relatively isolated and discussions on Fe as a pollutant thus fall outside of the scope of this review. The role of iron mineral amendments for remediation of polluted sediments is evaluated, and the recent advances in Fe-based amendment strategies (particularly nano-particulate and zero-valent Fe amendment strategies) are critically reviewed.

Iron Mineralogy and Cycling in Sediments

Iron represents the fourth most abundant element and the most abundant transition metal in the earth’s continental crust. Much of this iron originates from Fe(II) bound in the lattice structures of ferro-magnesian silicate minerals (e.g. olivine, pyroxene, biotite and amphiboles) and in magnetite, a mixed-valence Fe oxide. Weathering and oxidation under surface environmental conditions cause the low temperature geochemical transformation of these minerals into secondary oxides and clay minerals [e.g. 12]. Under these oxic conditions, ferric iron (Fe(III)) is the thermodynamically favourable oxidation state and predominantly forms highly insoluble ferric oxide and ferric oxy-hydroxide mineral phases. These can exist either as discreet crystals (including nano particles) or as coatings on other mineral phases, and due to their ubiquity, redox activity and amphoteric surface reactivity, they are known to have a strong control on the chemical properties of soils and sediment.

Depending on the climatic conditions, the most common soil and sediment Fe (oxy-hydr)oxides are goethite (α-FeOOH), hematite (Fe2O3) and amorphous and poorly crystalline phases (including ferrihydrite and green rusts). Lepidocrocite (γ-FeOOH) is less common and is typically found in sub-micron scale association with goethite [27], and akaganeite (β-FeOOH) is confined to unique chloride-rich environments (e.g. acid sulphate wetland sediment [28] and acid mine drainage sediment [29]). The adsorption of contaminants to these Fe (oxy-hydr)oxides is governed by the ambient pH conditions and by the reactivity, structure and chemistry of the hydroxyl groups on the mineral surfaces [30]. Contaminant adsorption to Fe mineral surfaces can occur through inner-sphere and outer-sphere complexation reactions, ligand exchange reactions and through the formation of ternary complexes (Eqs. 1–7; after [30]). Spectroscopic investigations have shown that Fe (oxy-hydr)oxide mineralogy can exert a control of the nature of these bonding interactions, thereby impacting the stability and mobility of the complexed contaminants (Fig. 1). For example, spectroscopic techniques have shown that Se(VI) forms outer-sphere complex with goethite [32] whereas it coordinates with ferrihydrite via inner-sphere bidentate bonding ([33]; additional examples in [31] and in Fig. 1).

Fig. 1
figure 1

Comparison between selected bonding interactions at a goethite surface (a) and a ferrihydrite surface (b) highlighting the effect of mineralogical controls on contaminant sorption. Se(VI) forms outer-sphere complexes with goethite and inner-sphere complexes with ferrihydrite. Sr(II) forms inner-sphere bidentate bonds with goethite but outer-sphere bonds with ferrihydrite. As(V) forms inner-sphere bidentate and minor monodentate bonds with goethite, and inner-sphere bidentate bonds with ferrihydrite [31, and references therein]. Co-precipitation and occlusion is another mechanism for contaminant sequestration (e.g. Ni(II) in ferrihydrite structure). Figures are orientated facing the 111 plane

Full size image
figure a

The iron (oxy-hydr)oxides are stable in the oxic surface layers of the sediment (upper few mm to cm, depending on sediment grain size distribution); however, as minerals and their associated contaminants are buried and undergo early diagenesis, their stability and speciation change (Fig. 2). Ageing reactions lead to the transformation of amorphous and poorly crystalline phases (e.g. ferrihydrite, schwertmannite) into more crystalline minerals (e.g. goethite, hematite), whereas reductive biogeochemical reactions (notably organic matter degradation) cause a rapid decrease in the redox potential with depth thereby affecting mineral stability. In the sub-oxic zone, Fe(III) oxides act as electron acceptors and, following reductive dissolution pathways, Fe(II) and adsorbed and co-precipitated contaminants are released into solution. Mobilised Fe(II)aq can migrate to the redox front to further interact with pollutants as a reducing agent. In anoxic sediment, oxygen concentrations are diminished, Fe(III) oxides have largely been reduced and sulphate is an important electron acceptor. Under strongly anoxic (i.e. sulphidic) conditions, sulphide mineral solubility and stability have a strong control on the partitioning of inorganic pollutants between dissolved complexes and solid mineral phases. Iron sulphide minerals include mackinawite (tetragonal FeS) which typically precipitates as a nano-particulate phase; greigite (Fe3S4, inverse spinel structure) which can form from the solid state transformation of mackinawite; and pyrite (FeS2) which is typically micro-particulate due to its rapid crystal growth and is the most common iron sulphide in sediments [36, 37].

Fig. 2
figure 2

Interactions between Fe mineral phases and inorganic and organic contaminants and the transformations and variations that occur due to sediment biogeochemistry. (1) Ageing transformations from poorly crystalline phases (e.g. ferrihydrite, schwertmannite) to more crystalline phases (e.g. goethite, hematite) typically lead to an expulsion of adsorbed contaminants due to decreases in mineral surface area. However, divalent cations associated with the more crystalline phases typically become more recalcitrant [34]. (2) Reductive dissolution of Fe oxide minerals leads to the release of adsorbed and co-precipitated contaminants, with a concomitant release of reduced Fe(II) into solution. (3) Disturbances to the sediment structure during diagenesis (e.g. bioturbation, dredging and sediment re-suspension) can lead to speciation changes (due to infiltrating O2) and can also lead to release of contaminants into the water column, thus posing a threat to aquatic organisms. (4) The released flux of Fe(II) can act as an electron donor during coupled redox transformation of contaminant phases. This is the premise on which the Fe(0) and Fe(II) amendment strategies are based. (5) The reduced Fe(II) flux can also re-precipitate at the redox boundary to form poorly crystalline phases (including mixed metal species such as FeAsO4-H2O) which can incorporate various trace metal contaminants. (6) Under highly anaerobic conditions, sulphur speciation has the strongest control on contaminant chemistry. Pyrite readily precipitates and can incorporate trace metals such as Co and Ni. Effect of Eh on redox couples is after [35]

Full size image

Iron Mineral Interactions with Sediment Pollutants

Because iron mineral stability is strongly dependent on the ambient redox properties of the sediment, large differences in the Fe mineral interactions with various inorganic and organic contaminants can be expected between oxic, sub-oxic and anoxic sediment horizons. The largest volumes of work have been conducted on Fe oxides which are stable in the oxic horizons and which are well known for their adsorption capacity for metals and metalloids in soils and in experimental systems. Because of the complexities associated with sampling and analysis [38], a far smaller volume of work has been conducted on the pollutant-mineral interactions in reduced and in reducing sediment horizons, where Fe carbonates, phosphates and sulphides are the dominant Fe minerals present.

Contaminant-Mineral Interactions in Oxic Sediments

Arsenic

Arsenic exists in soils and sediments predominantly as inorganic arsenate (As(V)) and arsenite (As(III)). It is discussed here separately from toxic trace metals because of its classification as a metalloid because it is characterised by a strong and well-studied association with Fe oxides and because As toxicity has been highlighted as a major and currently occurring human health concern (e.g. Bangladeshi drinking water [39]). Arsenite is more toxic and more mobile than As(V) [40], and both forms can be more toxic than many of the organoarsenical moieties [41]. Arsenic toxicity can be mitigated by immobilisation through adsorption to or co-precipitation with different Fe (oxy-hydr)oxide mineral phases [15]. Although both forms show a strong geochemical association with Fe, the oxidised form (As(V)) is more strongly bound to Fe (oxy-hydr)oxide surfaces than As(III). This is achieved through the formation of bidentate inner-sphere bonding coordination of As(V) at the oxide surface [15, 42], although monodentate bonding has also been reported, typically when the pH is greater than eight [42].

Aside from the effects of As speciation and pH, other chemical and mineralogical factors can affect the stability and extent of As adsorption. For example, among the different Fe (oxy-hydr)oxides, ferrihydrite is known to have the highest sorption capacity for As due to its high surface area [17, 43], and Al substitution for Fe(III) in the various oxide mineral structures has been shown to decrease As adsorption [44]. Arsenic adsorption can further be decreased by the presence of Fe-binding ligands (e.g. dissolved organic matter [45] and inorganic ligands such as phosphates [46]) which compete for surface bonding sites on Fe (oxy-hydr)oxide surfaces. In contrast, the presence of Zn can significantly increase the adsorption of As by goethite through the formation of Zn-AsO4 surface complexes [47].

Trace Metal Pollutants

Unlike organic pollutants which can be degraded, trace metals (e.g. Cd, Cr, Cu, Ni, Pb, Sn, Zn) remain persistent in sediment systems, where they can pose severe toxicity risks to benthic and aquatic organisms [20]. These trace elements must thus either be removed from the system (relocation and/or ex situ remediation) or their toxicity must be mitigated through changing their speciation (e.g. Cr(VI) to Cr(III)) or through decreasing their mobility (i.e. through geochemical stabilisation with sediment constituents such as Fe (oxy-hydr)oxide phases). Relative to As, trace metals are not as strongly adsorbed to Fe mineral surfaces and the stability of these various adsorption complexes is often more strongly dependent on parameters (especially pH) other than the Fe mineralogy and concentration [48]. For example, Cd(II) is known to form edge-sharing and corner-sharing inner-sphere complexes with goethite and ferrihydrite [16], yet in marine sediments with high Cl− activity, the formation of mobile Cd-chloride complexes inhibits surface adsorption to sediment minerals [49].

Chromium exists in sediment horizons in either its trivalent (Cr(III)) or more toxic hexavalent (Cr(VI)) form [e.g. 50], and both species have been shown to interact with Fe minerals [51]. Depending on Fe (oxy-hydr)oxide mineralogy, bonding interactions range from inner-sphere monodentate and bidentate coordination to epitaxial and non-epitaxial precipitation on oxide surfaces (see Table 1 in [31]). Notably, Cr(VI) toxicity can be mitigated by reductive precipitation at Fe(II) structural sites on magnetite surfaces [52]. Ferric (oxy-hydr)oxide surfaces can also play a strong role in catalysing the reduction of Cr(VI) by Fe(II) resulting in the surface precipitation of Cr(OH)3 [53]. Finally, adsorption of Cr(III) to Fe (oxy-hydr)oxide surfaces can inhibit its Mn-oxide-mediated oxidation to toxic Cr(VI) [54].

Adsorption of divalent metal ions (e.g. Cu(II), Pb(II), Zn(II)) to Fe (oxy-hydr)oxide surfaces is most strongly controlled by ambient sediment pH. Because the point of zero charge (PZC) for these minerals occurs at circumneutral pH [12], their surfaces become increasingly positive as pH decreases, thereby decreasing the likelihood of interaction with cationic metals. Fe (oxy-hydr)oxides are generally regarded as being a good sink for Cu [20] and have been shown to form inner-sphere complexes with goethite and poorly crystalline Fe oxy-hydroxides (see Table 1 in [31]). Adsorption of Cu has also been shown to block the complexation of other divalent contaminants (i.e. Zn) through competition for sorption sites on Fe (oxy-hydr)oxide surfaces [55]. Zn also forms inner-sphere complexes with Fe (oxy-hydr)oxides and the exact bonding configuration is strongly dependent on Fe mineralogy and pH [56, 57]. Similarly, Pb bonding interaction is partially controlled by mineralogy, forming bidentate edge-sharing complexes with goethite [58] and multinuclear complexes with ferrihydrite [59]. In spite of the formation of these inner-sphere bonds to Fe (oxy-hydr)oxides, Pb shows a much stronger bonding affinity for Mn oxides which are thus up to 40 times more efficient in stabilising this toxic metal [60, 61].

Organic Pollutants

Iron oxides are known to have strong geochemical interactions and associations with organic matter in soils [62] and in deposited and suspended sediments [63, 64]. The interactions between Fe minerals and organic contaminants (e.g. polycyclic aromatic hydrocarbons, polychlorinated biphenols, antibiotics etc.) are complicated by the wide range in organic contaminant structure and functional group chemistry. However, multiple studies have shown that Fe mineral phases can adsorb organic contaminants thereby stabilising their mobility in soil and sediment systems [e.g. 13, 14, 65]. For antibiotic contaminant phases, the bonding interaction is attributed to chemical complexation at the Fe oxide surfaces [14, 65] and can be inhibited by the presence of competing ions (e.g. P, Ca [14]). However, the formation of these surface complexes may not always lead to contaminant immobilisation, as the complexation can lead to ligand-promoted dissolution of Fe mineral phases [66]. Fe minerals can also play a catalytic role in organic pollutant degradation through oxidative Fenton-like reaction pathways [67, 68].

Contaminant-Mineral Interactions in Sub-Oxic and Anoxic Sediments

At depths below the upper-most oxic sediment layer, the sediment redox potential decreases, causing a decrease in the stability of Fe(III) mineral phases with a concomitant increase in the stability of porewater Fe(II) and reduced secondary minerals (e.g. vivianite (Fe3(PO4)2; siderite (FeCO3) and Fe(II) sulphides). This reduction of the Fe(III) (oxy-hydr)oxide minerals and production of Fe(II) gives rise to a number of dynamic chemical processes that can have impact on the speciation and mobility of sediment pollutants (Fig. 2). Importantly, the reductive dissolution of Fe (oxy-hydr)oxide minerals leads to the release of adsorbed and co-precipitated pollutants into the pore waters, potentially representing a significant source of pollutants into the overlying water column [69]. For example, the reducing conditions associated with flooded rice paddies results in the release of high concentrations of As(III) to the soil porewater [70, 71]. The released flux of Fe(II) can affect contaminant fate in two ways: direct interaction with the contaminants through coupled redox transformation [35] or in situ re-oxidation to precipitate oxide phases with co-precipitated contaminants [72].

Interaction between reduced Fe and pollutants has been well studied, and a range of remediation techniques utilises this principle to stabilise inorganic pollutants and degrade organic pollutants through oxidation by Fe(II) (e.g. FeSO4 and Fe(0) amendment technologies (Table 1; [18]). For example, Fe(II) can facilitate the reduction of Cr(VI) to its less toxic Cr(III) form [135], and Cu(II) has been shown to reduce to Cu(I) [136] and to native copper [137]. Fe(II) is not significant in reducing As(V) to the more toxic As(III) form [138]; however, the reverse reaction can be catalysed under oxic and anoxic conditions when As(III) is oxidised by enhanced Fe(II) reactivity at a Fe(III) (oxy-hydr)oxide mineral surface [139, 140].

Table 1 Non-exhaustive list of laboratory and field experiments involving Fe materials interacting with common soil and sediment contaminants
Full size table

The Fe(II) flux can also re-oxidise to precipitate in situ a variety of poorly crystalline mineral phases (e.g. schwertmannite (Fe8O8(OH)6(SO4).nH2O), green-rust (mixed-valence amorphous phase) and ferrihydrite) depending on the ambient solution properties (e.g. pH, Eh, Fe and contaminant concentrations). Ferrihydrite is known to incorporate As, Cd, Cu, Pb and Zn during precipitation [e.g. 141] with many of the divalent cations becoming less extractable upon ageing to goethite [34]. Similarly, ageing of schwertmannite to goethite leads to increased Cr incorporation in the mineral structure, whereas Cu, Ni and Zn concentrations have been found to decrease with ageing [142]. In sediment systems with high As concentrations, a number of discreet co-precipitated mineral phases are known to form (e.g. FeAsO4-H2O, FeAsO4.2H2O and Fe3(AsO4)2), many of which are more stable and less soluble than the complexes that form when As is merely adsorbed to Fe (oxy-hydr)oxide surfaces [34, 36].

Under increasingly anoxic conditions, sulphate reduction proceeds thus causing the sediments to become sulphidic and leading to the precipitation of various sulphide mineral phases. The precipitation of metal sulphides in anoxic sediments is controlled by the speed of the metals’ water exchange reaction kinetics: Cd, Cu, Pb and Zn sulphides typically precipitate before the formation of pyrite (FeS2) whereas Co and Ni have slower reaction kinetics and can be incorporated into the pyrite structure [143, 144]. Mackinawite and greigite are also known to incorporate variable amounts of Cu, Co and Ni [36, 145]. Reduced As is incorporated into the pyrite structure, whereas reduced Cr is not chalcophyllic and therefore does not associate with precipitating pyrite [143]. Pyrite shows relatively slow dissolution kinetics and thus represents a relatively good sink for co-precipitated and occluded contaminant phases. Furthermore, the pyrite mineral surface can also play a role in contaminant transformation [146] and sorption [147] processes.

Iron-Based Remediation of Contaminated Sediments

Based on the chemical and mineralogical interactions of natural Fe with contaminant phases, a range of techniques has become available in the last two decades to treat contaminated sediments and soils using iron amendment strategies (Table 1). These techniques exploit either the adsorption or reduction properties of Fe [148–150] and can be applied as in situ [23] or ex situ [24] remediation strategies. The efficiency of either pathway is controlled by factors affecting the biogeochemical cycling of iron [151].

Sorption-based technologies typically utilise the known associations and chemical bonding interactions between mineral surfaces and toxic metals and metalloids, as described in ‘Iron Mineral Interactions with Sediment Pollutants’ section. These remediation techniques are thus used as in situ amendments to immobilise toxic metal(loid)s and to reduce their bioavailability in contaminated sediments [e.g. 19], and their efficiency is partially controlled by the mineralogy of the Fe mineral amendment. Due to their high specific surface areas, amorphous Fe (oxy-hydr)oxides are the most effective amendment when considering sorption capacity, yet their long-term stability may be compromised by contaminant release during ageing and/or recrystallization (e.g. As, Cu, Zn, Ni lost to solution during the schwertmannite to goethite transformation [142, 152]. However, dynamic recrystallization of Fe minerals (e.g. during ageing) is known to have complex effects on trace metal redistribution [153] and can represent an important trapping mechanism (e.g. for As [70]). Overall, when considering long-term immobilisation, the more crystalline Fe oxides are typically considered to be more effective, as they have shown better long-term retention of trace metals in soil systems [19, 34]. Mechanistically, the trace metal immobilisation can be achieved by inner-sphere or outer-sphere surface complexation reactions (Fig. 1), with inner-sphere coordination regarded as being more stable (Eqs. 1, 2, 4, 5). For example, the inner-sphere complexation reactions of Ni(II) and Cr(VI) with Fe mineral surfaces can be expressed as:

$$ \equiv \mathrm{Fe}\mathrm{O}\mathrm{H} + {\mathrm{Ni}}^{2+}\leftarrow \to \equiv \mathrm{F}\mathrm{e}\mathrm{O}-{\mathrm{Ni}}^{+}+{\mathrm{H}}^{+} $$
(8)
$$ \equiv \mathrm{Fe}\mathrm{OH} + {\mathrm{Cr}}_2{{\mathrm{O}}_7}^{2\hbox{-}}\leftarrow \to \equiv \mathrm{F}\mathrm{e}-{\mathrm{Cr}}_2{{\mathrm{O}}_7}^{\hbox{-} } + {\mathrm{O}\mathrm{H}}^{-} $$
(9)

In chemical treatment methods, iron is generally used as a reductant for its ability to act as an electron donor (Fig. 3). From our review of literature, it is clear that the chemical treatment methods are primarily used for the treatment of contaminants in the porewater solution and limited research is available for their use for the treatment of soils and sediments [154]. Most commonly, the Fenton Reaction or a similar pathway is used (Fig. 3) where Fe(0) or Fe(II) are used to produce non-selective highly reactive oxidants for the breakdown of persistent organic compounds, such as PCBs, pesticides, fuels and explosives [148, 155, 156], and to change the speciation of redox reactive inorganic pollutants (e.g. Cr, Cu). Nano-zero-valent iron (Fe(0)) is the most commonly used iron phase in these treatments because of its low standard electron potential (Eo = −0.44 V; [155]), which is lower than many trace metals and organic compounds [157]. Ferrous iron amendments including phosphates [158, 159], sulphates [73, 160] and sulphides [150] are also applicable to the reductive treatment of industrial wastes. In addition to their reductive capabilities, the introduction of reduced Fe phases to contaminated sediment porewater systems can lead to the in situ precipitation of Fe sulphides (under anoxic conditions) which can sequester Co, Ni and other trace divalent trace metal cations [143, 144]. Similarly, under oxic conditions, the in situ precipitation of Fe oxy-hydroxides can strongly stabilise contaminants through co-precipitation pathways. For example, iron sulphate amendments are particularly effective in this regard and can have better metal retention efficiency than Fe oxide amendments when compared in soil systems [81]. However, such Fe(II) amendments may require additional buffering capacity (e.g. lime additions) to counteract the acidity released during Fe oxy-hydroxide formation (Eq. 10), as some trace metals (e.g. Cu, Zn) show enhanced mobility with decreasing pH [160].

Fig. 3
figure 3

Example of Fenton-like reaction catalysed by Fe(0) amendment

Full size image
$$ 4{\mathrm{FeSO}}_4+{\mathrm{O}}_2+6{\mathrm{H}}_2\mathrm{O}\to 4\mathrm{FeOOH}+4{{\mathrm{SO}}_4}^{2-}+8{\mathrm{H}}^{+} $$
(10)

Recent Advances in Fe-Based Remediation

A review of the recently published literature (Scopus and Web of Science) reveals that the most notable developmental work occurring in the last 5 years has been in the field of iron nanoparticle (particularly zero-valent iron nanoparticle) remediation technologies. Over 500 related articles have been published within the last 3 years alone. Thus, despite the historical focus on the use of bulk iron minerals for sediment amendment, we focus here on the recent developments within the field of Fe nanoparticle amendments for improving the efficiency of waste remediation.

Although there is a lack of consensus on the definition of a nanoparticle, generally, any particle with one dimension less than 100 nm is regarded as being nano-particulate [161]. Iron nanoparticles may occur naturally or are produced as by-products of other reactions. However, for remediation purposes, engineered particles are preferred as their specific properties and composition can be controlled, and their efficiency can thus be optimised [e.g. 162–164]. Depending on the properties desired, these engineered Fe nanoparticles can be produced by a variety of physical, chemical and biological methods, with co-precipitation, hydrothermal and microemulsion chemical techniques being most prevalent [165, 166]. The choice of preparation technique can have a strong control on the resultant nanoparticle size distribution, stability, magnetism, reactivity and surface chemistry, and these properties can be further modified by the application of surface coatings [165]. A diverse array of engineered Fe nanoparticles has thus been developed, and although zero-valent Fe nanoparticles (and modifications thereof) are most common, other Fe minerals have also been synthesised at the nano-scale and applied to contaminant remediation studies. For example, Lin and co-workers [167] synthesised magnetic γ-FeOOH nanoparticles that adsorb As(III) and As(V) over a wide pH range (3–11) and cellulose-stabilised FeS nanoparticles have been shown to reduce Hg leachability by 99 % [150]. Magnetite [108, 168], hematite [169], maghemite [170], hydrous iron oxide [171] and Fe phosphate nanoparticles [158, 159] have also all been developed and tested for their reactivity towards various contaminant phases.

Figure 4 highlights a number of factors that render these different Fe nano-particulate minerals more effective at contaminant breakdown and adsorption, relative to their respective bulk counterparts. In particular, Fe nanoparticles are characterised by a large surface area to particle size ratio (Fig. 4a), where the difference between available surface area in micro- (10 m2 kg−1) and nano- (30,000 m2 kg−1) particles is shown to be substantive [172]. Laboratory studies conducted by Waychunas and co-workers [31] have shown that the absolute uptake of bivalent metals follows the expected trend, where adsorption was highest in nanoparticles with the smallest size and largest surface area (∼300 m2 g−1). The reactivity of Fe nanoparticle surfaces is also much greater than observed for bulk Fe mineral surfaces (Fig. 4b). This is reflected in mathematical modelling, where Noubactep et al. [173] have shown that smaller sized zero-valent iron particles release a greater number of electrons and follow faster release kinetics (Fig. 4c) than larger sized particles, and this could enhance their effectiveness and efficiency during reductive treatment of affected sediments.

Fig. 4
figure 4

Characteristics and reactivity of iron particles as a function of particle size. Change in specific surface area (based on average densities of Fe(0) and Fe3O4) (a), reactivity (b) and kinetic electron release (c) suggests improved capacity of nano-sized Fe minerals to interact with contaminants in sediments. Figure modified after [151, 172, 173]

Full size image

This enhanced reactivity and chemical instability can however also be a hindrance to the use of iron nanoparticles, as it affects their mobility and results in untargeted reactivity [174–177]. To overcome these issues and to increase Fe nanoparticle effectiveness, scientists modify Fe nanoparticles using other metallic elements (e.g. bimetallic nanoparticles) or by altering their surface characteristics, typically through the addition of a coating media. The performance of Fe(0) nanoparticles has been improved by synthesising bimetallic Fe nanoparticles, which typically exhibit better catalytic abilities in breaking down various pollutants [163]. Bimetallic nanoparticles have been synthesised with metals that have a higher redox potential than Fe, including Ag [178], Cu [179], Ni [104, 121] and Pd [114, 119, 148]. By acting as electron transfer media, these metals prevent the self-inhibitory effects associated with oxy-hydroxide precipitation on monometallic Fe(0) nanoparticle surfaces (Hu et al. ref 27-in Liu). Surface modification and membrane coatings can further improve the effectiveness of Fe nanoparticles, particularly in soil and sediment systems where they readily aggregate and where their reactivity and adsorptive behaviour can be misdirected towards the non-pollutant moieties in the soil/sediment system. Nanoparticle dispersion (and hydrophobicity), mobility and stability have been enhanced by different organic surface modifiers including starch [102, 180], cellulose [181], oils [182] and various other anionic polymers and non-ionic surfactants [177, 183–185]. Silicon-based coatings (e.g. tetraethyl orthosilicate) can also enhance nanoparticle dispersion, and commercial applications of this technology are now available [186]. Besides increasing particle mobility, surface coatings can increase the specificity of Fe nanoparticle reactivity, particularly when functional groups with selective chelating ability are used [187, 188].

In summary, recent developments in Fe-based remediation of polluted sediments and soils have drawn on a fundamental understanding of the adsorptive and redox reactive properties of Fe and Fe minerals (‘Iron Mineral Interactions with Sediment Pollutants’ section). The advent of nanotechnology has enabled the synthesis and study of nano-sized Fe materials, which have been shown to have higher specific surface area and reactivity than their bulk counterparts (Fig. 4). As such, most developmental work has focused on the synthesis of novel Fe-based nanotechnologies for contaminant remediation. In particular, bimetallic particles and various surface modifications have been used to ensure that particles are easily mobilised in sediments (i.e. they do not aggregate easily or rapidly [164, 189, 190]); that the particles contain enough zero-valent Fe content and surface area for optimised reaction or targeted interaction with contaminants [150, 184]; and that the stabiliser characteristics do not negatively affect mobility and reactivity [157, 191, 192]. However, many of these technologies have yet to be tested in natural soil and sediment systems, where their impacts on biodegradation pathways [151] remain to be evaluated. Future work should further focus on field testing and on establishing the long-term stability and ageing of the amendment products, particularly if the amendments are to be applied as in situ treatments [e.g. 23] to sediments and soils with dynamic redox conditions.