Chemical Reduction and Oxidation of Organic Contaminants by Nanoscale Zerovalent Iron

Abstract

This chapter critically reviews the kinetics of NZVI used to reductively and oxidatively transform various kinds of priority organic contaminants including chlorinated ethenes, chlorinated ethanes, chlorinated and aromatic nitro hydrocarbons, chlorinated biphenyls, halogenated bisphenol A, explosives, dyes, and pesticides. All kinds of NZVI, including bare, bimetallic, polymer-modified, and supported NZVI, are reviewed. A total of 102 datasets of laboratory-scale experiments over 20 years (1997–2017) of NZVI research are evaluated to extract state-of-the-art understanding. This chapter elaborates not only reductive transformation pathways of priority organic contaminants but also two factors governing NZVI reactivity: intrinsic properties of the materials and environmental conditions where NZVI particles are applied. These include the particle crystallinity and chemical composition (noble metal), the effect of polymeric surface modification, and the effects of sorptive support, aging effects, pH, anionic and cationic solutes, natural organic matter, aquifer material, contaminant concentration, and the presence of dense nonaqueous phase liquid. Similarly, various factors affecting oxidative degradation of contaminants of concern using NZVI-induced Fenton’s reaction are reviewed.

3.1 Reductive Transformation of Priority Organic Contaminants Using Nanoscale Zerovalent Iron

Nanoscale zerovalent iron (NZVI) is a well-known remediation agent due to its high reactivity to reductively detoxify a variety of contaminants, including metals (Chap. 4) and halogenated organics, which is the focus of this chapter. Table 3.1 provides an intensive review of the reductive transformation of priority organic contaminants, including chlorinated ethene, chlorinated ethane, chlorinated and aromatic nitro hydrocarbon, chlorinated biphenyl, halogenated bisphenol A, explosives, dye, and pesticides, using various kinds of NZVI, including bare, bimetallic, polymer-modified, and supported NZVI. This involves 102 datasets of laboratory-scaled experiments over 20 years (1997–2017) of NZVI research. In addition to presenting a database of organic contaminants, several interesting trends can be determined from Table 3.1, some of which will be discussed later in this chapter.

Table 3.1 Summary of batch experiments for reductive treatment of organic contaminants using various NZVI (1997–2017)

Although the idea of using NZVI for in situ subsurface remediation started in 1997 (Wang and Zhang 1997), bulk ZVI has long been recognized as an electron donor and has been used in the form of iron filings to build a permeable reactive barrier since 1994 (Reynolds et al. 1990; Gillham and O’Hannesin 1994). Undeniably, NZVI is much more reactive than its ZVI counterpart on the mass basis comparison. Nevertheless, it is still debatable if this is mainly due to a much larger specific surface area of NZVI in comparison to iron filings or micron-sized ZVI or if it has something to do with the “nano-effect,” which may result in a greater density of reactive surface sites or surface sites of higher intrinsic reactivity.

The small size of nanoparticles causes an exponential increase in the number of atoms localized at the surface. These atoms are characterized by excess surface energy and are thermodynamically unstable (Ghosh 2015). This results in crystallographic changes including lattice contraction or deformation, the appearance of defects, rearrangements of the surface atoms, or changes in the morphology of nanoparticles (Jiang et al. 2001). This may yield quantum size effects, which cause changes in the Fermi level and band gap, leading to increases in intrinsic reactivity with decreasing particle size (Ghosh 2015).

Nevertheless, Tratnyek’s group revealed that, for tetrachloromethane (or carbon tetrachloride (CT)), no “nano-effect” contributing to the greater intrinsic reactivity of the surface sites or the greater abundance of reactive sites on the surface was observed. This is because, when comparing surface-area normalized rate constants (k_SA) against mass normalized rate constants (k_M) for the reductive degradation of CT by NZVI and bulk ZVI, the results show that NZVI gives larger k_M but similar k_SA values (Nurmi et al. 2005; Tratnyek and Johnson 2006). Nevertheless, when doing the same analysis for other contaminants, including trichloroethylene (TCE), tetrachloroethylene (PCE), and tetrachloroethane (TeCA) (Fig. 3.1), we found that a significant increase in k_SA for NZVI in comparison to ZVN was observed for TeCA and TCE but not for PCE. This suggests a potential “nano-effect,” but it seems to be, at least in part, governed by the interaction between the contaminant and NZVI surface.

Fig. 3.1

Comparison of surface-area normalized rate constants (k_SA) against mass normalized rate constants (k_M) for the degradation of TCE, PCE, and TeCA by NZVI and bulk ZVI . (Data from Johnson et al. (1996), Liu et al. (2005b), Song and Carraway (2005), Amir and Lee (2011), Phenrat et al. (2016), Hepure Technology Inc. (2017), Kim et al. (2017))

Regardless of its particle size and probable nano-effect, the fundamental chemistry of NZVI in an aqueous environment from a reaction viewpoint is the same as ZVI. For environmental remediation purposes, Fe⁰ can be oxidized by halogenated organics (as electron acceptors) as long as such organics have an E_H⁰ greater than −0.447 V. As a result of the electron transfer, in most cases, NZVI transforms such organic contaminants to more environmentally benign byproducts (Eqs. 3.1 and 3.2 using TCE as an example) (Liu et al. 2005a, b). In the meantime, Fe⁰ can also react with water (or H⁺) to produce H₂ gas (Eq. 3.3), which is a competing reaction to the dechlorination reaction and is strongly controlled by the availability of H⁺ (i.e., pH).

$$ {\mathrm{Fe}}^{2+}+2{\mathrm{e}}^{-}\to {\mathrm{Fe}}^0,\kern1em {E_{\mathrm{H}}}^0=-0.447\kern0.5em \mathrm{V} $$

(3.1)

$$ \mathrm{TCE}+n{\mathrm{e}}^{-}+\left(n-3\right){\mathrm{H}}^{+}\to \mathrm{product}+3{\mathrm{Cl}}^{-} $$

(3.2)

$$ {\mathrm{H}}^{+}+{\mathrm{e}}^{-}\to {\mathrm{H}}^{\ast}\to \frac{1}{2}{\mathrm{H}}_2 $$

(3.3)

Although the oxidation of Fe⁰ to Fe²⁺ (Eq. 3.1) is usually assumed, in environmentally relevant applications (i.e., groundwater at a natural pH), the transformation of Fe⁰ to the film of iron oxide, such as magnetite (Fe₃O₄) (Eq. 3.4) and maghemite (Fe₂O₃), is often observed (Liu et al. 2005b; Reinsch et al. 2010). Thermodynamically, the Fe⁰/Fe₃O₄ couple is more favorable than the Fe⁰/Fe²⁺couple at a pH above 6.1. The oxidation of Fe⁰ to Fe₃O₄ provides an additional 2/3 mole of electrons per mole of Fe⁰ oxidized in comparison to the oxidation to Fe²⁺ (Liu et al. 2005b). However, the formation of iron oxide film can passivate NZVI and adversely affect its reactivity and reactive life time (Liu et al. 2005b; Reinsch et al. 2010).

$$ {\mathrm{Fe}}_3{\mathrm{O}}_4+8{\mathrm{H}}^{+}+8{\mathrm{e}}^{-}\to 3{\mathrm{Fe}}^0+4{\mathrm{H}}_2\mathrm{O} $$

(3.4)

As shown in Eqs. 3.1, 3.2, 3.3, and 3.4, electrons from the NZVI oxidation can be used to reduce contaminants or to produce H₂. For bimetallic NZVI (i.e., NZVI modified by noble metals, such as Pd, Pt, and Rh; (Wang and Zhang 1997; Zhang et al. 1998) and highly disordered monometallic NZVI (Liu et al. 2005a); see Sect. 3.1.3.1), H₂ can be activated and used for dechlorination via hydrodechlorination (Eq. 3.5).

$$ {\mathrm{H}}_2\left(\mathrm{g}\right)\leftrightarrow 2{\mathrm{H}}_{\mathrm{ad}}\leftrightarrow 2{\mathrm{H}}_{\left(\mathrm{aq}\right)}^{+}+2{\mathrm{e}}^{-} $$

(3.5)

This explains the database in Table 3.1, in that bimetallic NZVI is typically 10–50 times more reactive than bare NZVI for the same contaminant, considering the enhanced pseudo-first-order reaction rate constant. The cost-effectiveness of in situ remediation using NZVI can be substantially affected by how electrons are utilized. In addition, the ability to utilize H₂ for hydrodechlorination and the characteristics of the iron oxide film formed on the surface of NZVI influence its lifetime, reactivity, and thus treatment efficiency. For NZVI, these phenomena are controlled by (1) intrinsic properties of the materials and (2) environmental conditions where NZVI particles are applied. These factors will be discussed later in this chapter, but we will first discuss reductive degradation pathways of various kinds of contaminants in aqueous environments in Sect. 3.1.2.

3.1.1 Reductive Transformation Pathways

3.1.1.1 Chlorinated Methane

While chlorinated methane, such as tetrachloromethane (or carbon tetrachloride (CT)) and trichloromethane (or chloroform (CF)), are mendable by NZVI, dichloromethane (DCM) shows negligible reactivity toward NZVI, giving the reactivity order of CT > chloroform (CF) >> DCM (Song and Carraway 2006); Table 3.1). Nevertheless, the reductive transformation of CT using NZVI (both Fe^BH and Fe^H2) cannot completely dechlorinate CT to methane. Instead, it produces toxic byproducts including CF (18.4–39.3% of the initial concentration of CT) and dichloromethane (0–2.4% of the initial concentration of CT) as well as unknown byproducts (23.3–42.5% of the initial concentration of CT) presumably methane, carbon monoxide, and formate, while CT retains 15.8–58.2% of its initial concentration after 27 h (see Table 3.1) (Nurmi et al. 2005; Song and Carraway 2006). Because the dechlorination reactions are heterogeneous, the rates depend greatly on the interaction between chlorinated organics and the NZVI surface (iron oxide). For dechlorination of highly chlorinated organics such as CT, CF, and DCM, stepwise electron transfer takes place. The weak sorption of chlorinated methane on the NZVI surface may result in a short residence time and desorption of incompletely dechlorinated compounds. Interestingly, while Fe^BH was reported to perform indirect dechlorination of chlorinated ethene using reactive hydrogen species (Liu et al. 2005a), for chlorinated methane, the dechlorination occurs via a direct electron transfer reduction mechanism, rather than the indirect mechanism (Song and Carraway 2006).

Two possible dechlorination pathways include hydrogenolysis and α elimination. Hydrogenolysis is a two-step electron transfer reaction initiated by direct dissociative electron transfer, in which the first electron addition results in dissociation of the molecule to a trichloromethyl free radical (•CCl₃) and chloride (Mccormick and Adriaens 2004). The free radical can subsequently form CF through the direct abstraction of hydrogen (Fig. 3.2a) or could be further reduced to a trichloromethyl carbanion (:CCl₃), which then forms CF via the rapid addition of H⁺ from the solution (Mccormick and Adriaens 2004; Song and Carraway 2006) (Fig. 3.2a). This is supposed to be a prevalent pathway of CT degradation by NZVI, given that CF is a major byproduct of the degradation. Similarly, hydrogenolysis is supposed to be a major degradation pathway for CF as well, given that DCM is the major byproduct (Fig. 3.2b) (Song and Carraway 2006).

Fig. 3.2

(a) CT degradation pathway forming CF via trichloromethyl radical or carbanion intermediates, (b) CF degradation pathway forming DCM via dichloromethyl radical or carbanion intermediates, and (c) speculated methane (CH₄) formation pathway from the dichlorocarbene intermediate to CH₄ involving multi-electron transfers. (Adapted from Mccormick and Adriaens (2004), Song and Carraway (2006))

On the other hand, CH₄ was also observed along with DCM in the reduction of CT and CF. Since DCM is a rather inert intermediate, the sequential hydrogenolysis of DCM to CH₄ is unlikely. Consequently, methane formation must be caused by a parallel CT degradation pathway, bypassing CF and DCM as an intermediate. This can happen via concerted reductive elimination steps (two-electron transfer (α elimination) and four-electron transfer, respectively) to form dichloromethylene, methylene, and eventually methane (Fig. 3.2c). Notedly, α elimination is a dichloro-elimination process involving a two-electron transfer to the molecule and the elimination of two chlorine atoms. The reaction products are less-saturated aliphatic hydrocarbons and two chloride ions.

Notedly, α elimination is the elimination of chlorine atoms from one carbon atom, while β elimination occurs when chlorine atoms are removed from two different carbons. Reductive β elimination is known to be a preferential pathway for compounds possessing (α, β) pairs of chlorine atoms (Arnold and Roberts 2000), while hydrogenolysis or reductive α elimination is the primary transformation pathway for compounds possessing only α chlorines (Mccormick and Adriaens 2004; Song and Carraway 2005).

3.1.1.2 Chlorinated Ethene and Ethane

The NZVI can completely dechlorinate all chlorinated ethene and ethane (except 1,2-DCA) to non-chlorinated intermediates, such as acetylene, and byproducts including ethene and ethane. Unlike chlorinated methane, the main degradation pathway of chlorinated ethenes via NZVI is dichloro-elimination (β elimination) followed by hydrogenolysis (Fig. 3.3). Similarly, the main degradation pathways of chlorinated ethane, such as HCA, PCA, and 1,1,1,2-TeCA, via NZVI is β elimination followed by hydrogenolysis (Fig. 3.3). Nevertheless, for 1,1,2,2-TeCA and both 1,1,1-TCA and 1,1,2-TCA, dehydrohalogenation becomes equally important, if not more important, than β elimination, especially at high a pH (Song and Carraway 2005).

Fig. 3.3

Dechlorination pathway, ethene and ethane (Tobiszewski and Namieśnik 2012). (Reprinted with permission from Tobiszewski and Namieśnik (2012). Copyright (2012) Springer Nature)

Fundamentally, for chlorinated ethene, β elimination is preferable over hydrogenolysis because thermodynamic reduction potentials for two-electron reduction of chlorinated ethene and ethane via β elimination are more favorable than hydrogenolysis. This is demonstrated by an example of the half reaction of PCE transformation via β elimination (Eq. 3.6) and hydrogenolysis (Eq. 3.7) (Totten and Roberts 2010). From a kinetic viewpoint, the predominance of β elimination over hydrogenolysis is also partially due to the rapidity of intermolecular rearrangements (leading to β elimination) compared to bimolecular collisions with H⁺ (leading to hydrogenolysis) at moderately basic pHs (Arnold and Roberts 2000).

$$ {\mathrm{CCl}}_2={\mathrm{CCl}}_2+2{\mathrm{e}}^{-}\to \mathrm{CCl}\equiv \mathrm{CCl}+2{\mathrm{Cl}}^{-}\kern0.5em {E}_{\mathrm{H}}^0=0.631\kern0.5em \mathrm{V} $$

(3.6)

$$ {\mathrm{CCl}}_2={\mathrm{CCl}}_2+2{\mathrm{e}}^{-}+{\mathrm{H}}^{+}\to {\mathrm{CCl}}_2=\mathrm{CHCl}+{\mathrm{Cl}}^{-}\kern0.5em {E}_{\mathrm{H}}^0=0.592\kern0.5em \mathrm{V} $$

(3.7)

Additionally, Fe^BH was reported to perform indirect dechlorination of chlorinated ethene using reactive hydrogen species (Liu et al. 2005a). This will be discussed in more detail in the next section.

Since the rate-limiting step in the reduction of chlorinated ethene and ethane using ZVI is supposed to be the transfer of a single electron and the formation of an alkyl radical (Johnson et al. 1996; Arnold and Roberts 2000), Scherer et al. (1998) proposed linear free energy relationships (LFERs ) capable of explaining or predicting the rates of dehalogenation by ZVI. They showed that, at the same number of chlorine atoms, dechlorination rate constants of chlorinated ethanes are typically higher than chlorinated ethene. For an internal comparison among chlorinated ethene or among chlorinated ethane, the dechlorination rate constants tend to increase with increasing chlorination of the compounds. As for NZVI, the reactivity of chlorinated ethanes summarized in Table 3.1 reasonably agrees with this trend, in that HCA > PCA > 1,1,1,2-TeCA >1,1,1-TCA > 1,1,2,2-TeCA >1,1,2-TCA > 1,1-DCA. Moreover, the k_SA values of these chlorinated ethanes linearly correlate with the one-electron reduction potential (E1) and the lowest unoccupied molecular orbital (LUMO) energy of chlorinated ethanes used in the LFERs (Song and Carraway 2005). In contrast, chlorinated ethenes do not follow the proposed trend. As shown in Fig. 3.1 and Table 3.1, the k_SA of PCE is lower for TCE for bare Fe^H2 (RNIP) (Liu et al. 2007; Fagerlund et al. 2012). A similar finding was also reported for ZVI for all the series of chlorinated ethene (i.e., k_SA of PCE < TCE < DCEs < VC) (Arnold and Roberts 2000). Based on this finding, Arnold and Roberts (2000) hypothesized that for chlorinated ethene, the transfer of a single electron is not the rate-limiting step but rather the formation of a di-σ-bonded intermediate (Fig. 3.4). This may also explain the findings on NZVI for TCE and PCE.

Fig. 3.4

Hypothesized rate-limiting step of the PCE dechlorination by ZVI (i.e., formation of a di-σ-bonded intermediate). After this step, the reaction may proceed via Steps 3a and 4a to give the reductive β-elimination product dichloroacetylene or via Steps 3b and 4b to give the hydrogenolysis product TCE (Arnold and Roberts 2000). (Adapted with permission from Arnold and Roberts (2000). Copyright (2000) American Chemical Society)

Noticeably, NZVI reduction is known to be incapable of completely treating 1,2-DCA (Maes et al. 2006; Deng and Hu 2007; Su et al. 2012a). The pseudo-first-order rate constants of 1,2-DCA using NZVI at the concentration of 0.4 g/L and at neutral pH were less than 5 × 10⁻⁵/h (Table 3.1) (Song and Carraway 2005). The poor dechlorination efficacy of 1,2-DCA using NZVI may be due to the relatively high C–Cl bond strength (87.2 kcal/mol) (Cioslowski et al. 1997) compared to the higher chlorinated ethanes. For example, the bond strength of HCA and 1,1,1-TCA was 68.83 and 73.6 kcal/mol, respectively, while the k_SA of HCA was four times greater than that of 1,1,1-TCA using Pd/Fe⁰ nanoparticles under the same experimental conditions (Lien and Zhang 2005).

Arnold and Roberts (2000) implied that the slow dechlorination rate of 1,2-DCA by ZVI is explainable by the slow formation of a di-σ-bonded intermediate. Recently, sulfidation of NZVI using dithionite was found to degrade more than 90% of 1,2-DCA, but over the course of a year, pseudo-first-order rate constants ranged from 3.8 × 10⁻³ to 7.8 × 10⁻³/day (Garcia et al. 2016); for more details about NZVI sulfidation, see Chap. 9). Even with this novel NZVI modification, the reductive dechlorination rate constants are still relatively low and may not be practical for source zone treatment of 1,2-DCA. In contrast, the Fenton process has shown its full potential to degrade 1,2-DCA. The NZVI-induced Fenton’s reaction will be discussed more in the latter part of this chapter.

3.1.1.3 Chlorinated and Aromatic Nitro Hydrocarbon

Chlorinated aromatic hydrocarbon , such as chlorophenols, can be detoxified by NZVI to yield a much less hazardous byproduct, such as phenol. Nevertheless, the dechlorination rate constant of pentachlorophenol (PCP) by bare NZVI is relatively low (K_obs = 0.001/h using smectite-templated Fe⁰ containing 13.3 g/L of smectite and 7 mM Fe⁰). Thus, to improve the treatability, a Pd catalyst (0.8 mM Pd⁰) was added to the nanocomposite, yielding the dechlorinate rate constant of 0.487/h. Figure 3.5a illustrates dechlorination pathways of chlorinated phenols by smectite-templated Pd⁰/Fe⁰ (Jia and Wang 2015).

Fig. 3.5

(a) Primary dechlorination pathways of chlorinated phenols by smectite-templated Pd⁰/Fe⁰. Solid arrows refer to the major reaction route, and broken arrows refer to the minor reaction pathway (Jia and Wang 2015) and (b) hydrodechlorination driven by Pd⁰/Fe⁰ nanoparticles (Schrick et al. 2002). (Reprinted with permission from Schrick et al. (2002) and Jia and Wang (2015). Copyright (2015) Elsevier and Copyright (2002) American Chemical Society)

Hydrodechlorination (Eq. 3.5) is the main dechlorination pathway for such a catalytic NZVI. Protons in the aqueous solution acquire electrons from NZVI, forming hydrogen radical on the Pd surfaces (Fig. 3.5b). In this process, the active H atom formed by Pd⁰/Fe⁰ attacks the chlorinated phenols via electrophilic H addition to the π system of the benzene ring (formation of H-aromatic complexes) and then removes the chlorine atom by two-electron transfer. Since this process is promoted by electrophilicity of the chlorinated aromatic hydrocarbon, the more chlorinated the phenols, the lower the dechlorination rate with Pd⁰/Fe⁰ particles because electron-withdrawing groups, such as Cl, that are attached to the aromatic ring reduces the electron density associated with the ring carbons, decreasing the probability of complexation between active H and the corresponding aromatic ring (Jia and Wang 2015). The selectivity of dechlorination of chlorinated phenols is mainly due to charges and steric hindrance governed by the interaction of substituent Cl atoms and individual aromatic carbon.

Figure 3.5a also shows the calculated charge on each carbon of chlorinated phenols in red. Substituent Cl at the ortho-position is preferentially dechlorinated from the phenol ring due to more negative charges associated with the ortho-position C, which is attributed to the influence of the adjacent OH group. The H in the formed H-aromatic ring complexes is prone to interact with more negatively charged carbon, leading to the scission of the CCl bond (Jia and Wang 2015).

Steric hindrance is another important factor governing the dechlorinate rate of chlorinated phenols. For example, for those chlorinated phenols not containing ortho-Cl, Cl located at the meta-position can be dechlorinated radially due to less steric hindrance. Consequently, as summarized in Table 3.1, the dechlorination rate constants of chlorinated phenols follow the order: PCP < 2,3,4,5-TeCP ~ 2,3,4,6-TeCP ~ 2,3,5,6-TeCP <2,3,4-TCP~2,3,6-TCP ~ 2,4,6-TCP < 3,4,5-TCP ~ 2,4,5-TCP ∼ 2,3,5-TCP < 2,6-DCP < 3,5-DCP < 2,3.DCP ∼ 2,4-DCP ∼ 3,4-DCP < 2,5-DCP < 4-CP ∼ 2-CP < 3.CP (Jia and Wang 2015).

Similarly, NZVI and bimetallic NZVI can sequentially dechlorinate chlorobenzenes yielding PeCB, TeCBs, TCBs, and DCBs as byproducts. Unlike chlorophenols, no selectivity of dechlorination product formation was observed for the case of HCB (Zhu et al. 2010). Noticeably, HCB dechlorination is not complete (i.e., no substantial formation of benzene was detected). Interestingly, for HCB dechlorination, Cu appears to be a better catalyst than Pd (Shih et al. 2009; Zhu et al. 2010). Furthermore, nitroaromatic compounds (NACs), such as para-nitrochlorobenzene (p-NCB), are treatable by Ni/Fe⁰ nanoparticles. In addition, Ni plays the role of catalyst, decreasing the activation energy of hydrogenolysis of C–Cl bonds, while Fe⁰ supplies electrons and reduces H⁺ to produce hydrogen gas. Complete transformation (100% removal efficiency) is achieved using Ni/Fe⁰ (2.0% Ni) at the concentration of 6 g/L after 300 min. The p-NCB is first adsorbed by NZVI and then quickly reduced to p-chloroaniline (p-CAN) and eventually to aniline (AN) via hydrogenolysis (Xu et al. 2009).

3.1.1.4 Halogenated Bisphenol a (Flame Retardant)

Bimetallic NZVI can completely dechlorinate halogenated bisphenol A, especially at a slightly acidic pH (pH 5–6). Two dehalogenation mechanisms take place simultaneously for different degrees at different pH ranges. Using debromination of tetrabromobisphenol (TBBPA) by Ni/Fe^H2 nanoparticles as an example, at pH ranging from 3 to 9, sequential dehalogenation via hydrogenolysis takes place as shown in pathway (a) in Fig. 3.6. This hydrogenolysis yields brominated intermediates, such as tri-, di-, and mono-BBPA, as well as a small amount of bisphenol A (BPA; 7.2–7.6%) (Li et al. 2016, 2017). Nevertheless, at a slightly acidic pH (pH 5–6), concerted hydrogenolysis (pathway (b) in Fig. 3.6) is dominant, yielding BPA as a major byproduct (90.7–93.3%) with a small amount of tri-, di, and mono-BBPA (Li et al. 2016, 2017).

Fig. 3.6

Proposed primary debromination pathways of TBBPA by Ni/NZVI at different initial pHs (a) pH 3.0, 5.0, 6.0, 7.0, and 9.0 and (b) pH 5.0 and 6.0 (Li et al. 2016). (Reprinted with permission from Li et al. (2016). Copyright (2016) Elsevier)

3.1.1.5 Pesticides and Dyes

The NZVI can completely transform complex chlorinated pesticides, such as γ-hexachlorocyclohexane (known as lindane) in 24 h at a NZVI concentration of 0.1–0.39 g/L (Elliott et al. 2009). Figure 3.7 summarizes the degradation pathway of lindane, yielding γ-3,4,5,6-tetrachlorocyclohexene (TeCCH) as a major intermediate with benzene as a byproduct. The dihalo-elimination of vicinal chlorides from carbons 1 and 2 of lindane is supposed to be the first and rate-limiting step. Initially, lindane must adsorb onto the NZVI surface. Then, an electron from Fe⁰ is donated to the surface-associated lindane, forming a neutral radical with the release of a chloride ion. Another electron is then transferred from the transient Fe⁺ species to the reacting carbon center of the radical, which then undergoes double bond formation and simultaneous release of chloride from the beta carbon, yielding TeCCH as an intermediate. Noticeably, both chlorine atoms (from carbons 1 and 2 of lindane) occupy axial positions on their respective carbon centers and are oriented antiperiplanar, which maximizes their susceptibility toward reduction. After this initial step, the two subsequent dihalo-elimination steps are believed to occur more rapidly, yielding benzene and chloride as major end products (Elliott et al. 2009).

Fig. 3.7

Dihalo-elimination degradation pathway from lindane to benzene (Elliott et al. 2009); with permission from ASCE). (Reprinted with permission from Elliott et al. (2009). Copyright (2016) American Society of Civil Engineers)

The NZVI can also decolorize dye-contaminated water. Figure 3.8 is an example of the reductive decolorization of methyl orange (MO). It starts with MO adsorption onto the NZVI surface by the formation of the chelate complex of Fe(II) dye. The radicals of H were generated by the NZVI reduction of water or hydrogen ion. Two H radicals were involved in the cleavage of the azo bond with two more electron transfers from NZVI. An electron is transferred from NZVI to the –N=N– bond, causing the breakage of the bond and the combination of a radical of H. In the meantime, another NZVI integrated another two radicals of H to break the N–N bond under the similar process. Eventually, the azo-double bond –N=N– was disconnected from two different amines. The breakage of the azo-double bond makes the visible absorption peaks at 464 nm vanish.

Fig. 3.8

Proposed mechanism for the degradation of MO using Fe⁰ nanoparticles (Chen et al. 2011b). (Reprinted with permission from Chen et al. (2011b). Copyright (2011) Elsevier)

3.1.2 Particle Properties Affecting NZVI Reactivity

Table 3.1 summarizes dechlorination rate constants of various contaminants. Noticeably, the rate constants for the same kind of contaminants by different types of NZVI can be substantially different. Although NZVI has two common characteristics, the small size and high specific surface area, NZVI has a vast variety of physicochemical properties, which can enormously affect dechlorination pathways and kinetics. Since the mass of NZVI needed for remediation and thus the cost of the treatment depend on particle reactivity, reactive lifetime, and particle efficiency, all of which are controlled by NZVI physicochemical properties, it is worth elaborating their effects on NZVI performance in this section.

3.1.2.1 Particle Crystallinity and Chemical Composition (Noble Metal)

The previous section discusses the different degradation pathways of various contaminants. Nevertheless, it also notes that different types of bare NZVI, such as Fe^BH and Fe^H2, have an important contribution to degradation pathways. This is mainly due to the difference in their crystallinity . According to multiple lines of evidence from the transmission electron microscope (TEM), energy-dispersive spectroscopy (EDS), electron energy-loss spectroscopy (EELS), and X-ray diffraction (XRD) (Liu et al. 2005a, b; Nurmi et al. 2005; see also Chap. 2), the morphology of Fe^H2 appears to be a crystallographic angular structure rather than a round, amorphous shape-like Fe^BH. While no significant oxide shell was observed for Fe^BH (Liu et al. 2005b), the iron oxide phases, such as magnetite and maghemite, were observed as the stabilized shell of Fe^H2 (Liu et al. 2005b; Nurmi et al. 2005; Baer et al. 2007; Reinsch et al. 2010). This crystallinity and iron oxide shell critically govern particle reactivity, particle lifetime, and particle efficiency (Liu et al. 2005b; Nurmi et al. 2005).

As shown in Table 3.1, using TCE as an example, the k_SA value using Fe^H2 (0.35 g/L) is 4 × 10⁻⁴ L/(h^.m²), while the k_SA value using Fe^BH is 2 × 10⁻³ L/(h^.m²) at the same experimental condition. This is a fivefold difference in k_SA values (Liu et al. 2005b). The similar trend is also reported for the dechlorination of tetrachloromethane using Fe^BH and Fe^H2 (Nurmi et al. 2005). This is because amorphous Fe^BH can utilize H₂ produced from the H⁺ reduction for TCE dechlorination through a catalytic hydrodechlorination pathway (Liu et al. 2005a). This increases efficiency of electron utilization for Fe^BH because electrons used to produce H₂ can then be used to degrade TCE. This feature is not available in crystalline Fe^H2, making its TCE dichlorination rely on β elimination and hydrogenolysis only. Thus, for Fe^H2, the electron used to produce H₂ is considered unproductive with respect to dechlorination. Nevertheless, due to their amorphous Fe⁰ and the absence of an iron oxide shell, at pH 7, a relatively large fraction of Fe⁰ in Fe^BH interacted with H⁺ and was used to produce H₂, implying that Fe^BH has a poor selectivity of electron utilization for dechlorination (Liu et al. 2005a).

In the same conditions, a relatively small fraction of Fe^H2 was used to produce H₂, implying that Fe^H2 has a good selectivity of electron utilization for dechlorination (Liu et al. 2005b; Liu and Lowry 2006).

In addition to particle crystallinity, different chemical compositions of Fe^BH and Fe^H2 also affect the electron utilization efficiency. Unlike the case of Fe^H2, the formation of iron oxide film around Fe^BH was not observed due to the presence of boron in Fe^BH particles that facilitated the dissolution of the particles and thus regenerated the reactive sites for TCE dechlorination . Consequently, almost all of the Fe⁰ in Fe^BH was available for dechlorination (i.e., high particle efficiency ~92%; (Liu et al. 2005a, b)). On the other hand, the iron oxide shell grew as RNIP (commercially available Fe^H2) was oxidized, making some fraction of Fe⁰ in RNIP unavailable for dechlorination (Liu et al. 2005b). Consequently, particle efficiency went down (only around 52%). However, the disadvantage of Fe^BH particles is their relatively short lifetime in comparison to Fe^H2. For in situ remediation, a long reactive lifetime is preferable for cost-effective treatment of contaminant plumes and to ensure that NZVI particles do not “burn out” prior to reaching contaminated areas (Liu et al. 2005b).

In addition to the effect of boron in Fe^BH on particle dissolution, another example of the influence of chemical composition on NZVI reactivity is the bimetallic NZVI (i.e., NZVI doped with Pd, Pt, Cu, and Ni) to enhance particle reactivity through catalytic pathways (Zhang et al. 1998; Schrick et al. 2002; Zhang 2003). The structure of bimetallic NZVI particles is a reductive Fe⁰ core with the shell of inert noble metals (Fig. 3.5b as an example). Because of the presence of noble metals, bimetallic particles can utilize H₂ for hydrodechlorination (Zhang 2003). As shown in Table 3.1, using chlorinated ethane and ethene as examples, k_SA or k_obs using bimetallic NZVI is around 10–50 times greater than bare NZVI. However, a significant problem with bimetallic NZVI is the observed decrease in reactivity over time due to the deactivation if the thick oxide layer can form and cover the noble metals (Zhu et al. 2006). This deactivation issue is an important obstacle for long-term remediation using bimetallic NZVI.

3.1.2.2 Effect of Polymeric Surface Modification

Polymeric surface modification, as a means to increase NZVI mobility in the subsurface and affinity for specific subsurface contaminants (Chaps. 5 and 6), is essential for in situ remediation using NZVI. However, polymeric surface modification may either enhance or decline NZVI reactivity toward organic contaminants. Since the dechlorination reactions are heterogeneous (the contaminant must contact the particle surface to be degraded), NZVI synthesis in the presence of polymers or polyelectrolytes , such as carboxymethyl cellulose (CMC), guar gum, and polyvinylpyrrolidone (PVP), in a “one pot” manner appears to enhance NZVI reactivity. This surface modification approach yields smaller particles in comparison to the physisorption of polymers or polyelectrolytes onto pre-synthesized NZVI. Conceptually, the Fe²⁺ or Fe3⁺ ions from FeSO₄ or FeCl₃ may be complexed with polyelectrolytes prior to reductive precipitation using NaBH₄. The complexed Fe²⁺ or Fe³⁺ ions behave as nucleation seeding points of NZVI. Consequently, we obtained polymer-modified NZVI with a smaller size that was resistant to aggregation and, thus, more reactive than larger, non-stabilized NZVI. This results in the observed increases in TCE reactivity with polymer-modified Fe-Pd bimetallic nanoparticles at low polyelectrolyte concentrations compared with bare Fe-Pd bimetallic nanoparticles (He and Zhao 2008; Sakulchaicharoen et al. 2010) (Table 3.1).

Nevertheless, the opposite is true for polymeric surface modification of pre-synthesized NZVI, in that polymeric surface modification can decrease the NZVI reactivity via reactive site blocking and mass-transfer resistance. As seen in Fig. 3.9a, the polymeric modified NZVI is covered with a layer of adsorbed macromolecules. According to the Scheutjens and Fleer conceptual model for homopolymer sorption (Scheutjens and Fleer 1979, 1980), charged homopolymers are normally adsorbed onto the surface in the train-loop-tail configuration.

Fig. 3.9

Schematic illustrating site blocking due to adsorbed trains and formation of an extended brush layer (a) of loops and tails (Phenrat et al. 2009b) and (b, c) effect of the Donnan potential in the polyelectrolyte layer on ionic solute distributions (Phenrat et al. 2015). (Reprinted with permission from Phenrat et al. (2009b, 2015). Copyright (2009) American Chemical Society and Copyright (2015) Springer Nature)

Trains are segments of polymer directly attached to the particle surface, which can block NZVI reactive sites, whereas loops and tails form an extended polyelectrolyte brush away from the surface, which can retard the diffusion of chlorinated volatile organic compounds (CVOCs) from the bulk aqueous solution to the NZVI surface. Similarly, for random copolymer, block copolymer, or grafted polymer, trains or anchoring blocks may block electron transfer sites, while a polymer brush may limit the mass transfer or chlorinated organics to the reactive NZVI surface. Several recent studies support this hypothesis. As summarized in Table 3.1, Phenrat et al. (2009a) revealed that, when the surfaces of pre-synthesized NZVIs were modified by the physisorption of polyelectrolytes, the TCE dechlorination rate constant decreased nonlinearly with the increasing adsorbed mass of the polyelectrolytes, with a maximum 24-fold decrease in reactivity. This is due to reactive site blocking and a decrease in the aqueous TCE concentration at the surfaces of the NZVIs due to the partitioning of TCE to the adsorbed polyelectrolytes. A similar finding was also reported by Wang and Zhou (2010) using solvent-responsive, polymer-coated NZVIs to degrade TCE (Wang and Zhou 2010). Noticeably, for the case of “one pot” polymer-modified NZVI, we believe that polymer still blocks the NZVI reactive site and still resists CVOC mass transfer to the NZVI surface, but the two adverse effects are outweighed by the positive effects from the smaller size of NZVI.

Nevertheless, polymeric surface modification makes the NZVI less sensitive to environmental factors in comparison to bare NZVI . As summarized in the next section, non-reducible ionic species, such as Cl⁻, SO₄²⁻, HCO₃⁻, and HPO₄²⁻, decreased the TCE dechlorination rate constant by bare NZVI up to a factor of seven compared with deionized (DI) water at pH 8.9 (Liu et al. 2007). On the other hand, polymeric surface modification reduces the interaction of NZVIs with nontarget groundwater solutes (organic and ionic species). A possible explanation is the effect of the Donnan potential in the polyelectrolyte layer on ionic solute distributions. The Donnan potential (Fig. 3.9b, c) can decrease the concentration of cationic solutes at the surfaces of NZVIs and possibly reduce their blocking effect compared with that for bare NZVIs. A recent study reported that for the first TCE spike, TCE dechlorination rates using polyaspartate (PAP)-modified NZVIs in the actual groundwater samples were 70–85% of the TCE dechlorination rate using PAP-modified NZVIs in DI water (Phenrat et al. 2015), while the TCE dechlorination rates using bare NZVI in the same groundwater samples for the first TCE spike were around ~22% of the TCE dechlorination rate using bare NZVIs in DI water.

Furthermore, over an intermediate period (30 days), in the presence of groundwater solutes, polyelectrolytes, such as PAP, were desorbed from NZVI and thus restored the reactivity of bare NZVIs with TCE. Evidently, the TCE dechlorination rates using PAP-modified NZVIs in the second and third TCE dechlorination cycles (intermediate-term effect) increased substantially (~100% and 200%, respectively, from the rate of the first spike). The desorption of PAP from the surface of NZVIs over time due to salt-induced desorption is hypothesized to restore NZVI reactivity with TCE (Phenrat et al. 2015). This suggests that the modification of the NZVI surface with small charged macromolecules, such as PAP, helps to deliver NZVIs to the subsurface and restores NZVI reactivity over time due to a gradual PAP desorption in groundwater.

3.1.2.3 Sorptive Support

Another means to increase NZVI mobility in the subsurface and the affinity for specific subsurface contaminants is using NZVI supports, such as submicron- or micron-sized porous silica (Zheng et al. 2008) or activated carbon (Mackenzie et al. 2012). These particles serve as carriers for NZVI transport because incorporation of NZVI into porous particles decreases the extent of magnetic attraction among NZVI particles, leading to less agglomeration and increasing their subsurface mobility. Furthermore, these supports are typically sorptive for organic contaminants; the NZVI composite materials behave as sorptive and reactive remediation agents. Figure 3.10 illustrates the dual-phase TCE removal mechanism by Fe^BH entrapped in porous ethyl-functionalized silica (Fe(B)/ethyl-silica system). Noticeably, an immediate sharp decrease of the aqueous TCE concentration to 45% of its original value was observed due to TCE sorption onto the functionalized silica followed by a much slower reaction rate, presumably due to dichlorination. Figure 3.10 also shows the byproduct formation, which supports the sorptive and reactive removal of TCE by the Fe(B)/ethyl-silica system.

Fig. 3.10

TCE removal from solution and gas product evolution rates for Fe(B)/ethyl−silica, where M/M₀ is the fraction of the original TCE remaining, and P/P_f is the ratio of the gas product peak to the gas product peak at the end of 96 h (Zheng et al. 2008). (Reprinted with permission from Zheng et al. (2008). Copyright (2008) American Chemical Society)

Similar behavior is also observed by carbo-iron colloid (CIC), which is Fe^H2 encapsulated in activated carbon. The sorptive and reactive TCE removal mechanism contributes to the higher reactivity and more NZVI utilization efficiency in comparison to bare NZVI without support. Mackenzie et al. (2012) showed that, for bare Fe^H2, dichlorination rate constants (k_obs) decreases with the decrease of the NZVI concentration, while the rate constants (k_obs) for CIC are insensitive to the CIC concentration in the suspension. This is because almost all TCE is adsorbed to the activated carbon of CIC, and since the dechlorination takes place at the surface, the reaction rate is determined by the Fe⁰ content of the carbo-iron and not by the total Fe⁰ concentration in the suspension (Mackenzie et al. 2012).

3.1.2.4 Aging Effect

Aging or longevity is the change of NZVI during immersion in water mainly by reacting with oxygen, water, target contaminants, or naturally occurring subsurface constituents. It can seriously affect both NZVI morphology and reactivity. Because in situ remediation is a moderate- to long-term operation, NZVI aging should be accounted for in determining the required amount of NZVI for contaminant transformation.

Typically, NZVI has a core (Fe⁰) and shell (iron oxide, such as magnetite and maghemite) structure. Aging is a dynamic and complex process conceptually consisting of three processes including (1) breakdown of the existing oxide shell by hydration and auto reduction, (2) oxidation of the freshly exposed underlying Fe⁰ coupled with reduction of reactive solutes, such as oxygen or target contaminants, and (3) subsequent cementation by formation of authigenic mixed-valent Fe(II)-Fe(III) phases (Sarathy et al. 2008). The aging phenomenon is substantially affected by the type of NZVI (Fig. 3.11) (Kim et al. 2012).

Fig. 3.11

Schematic description of the aging procedures of Fe^BH and Fe^H2 nanoparticles (a) from Kim et al. (2012) and (b) from Liu et al. (2015). (Reprinted with permission from Kim et al. (2012) and Liu et al. (2015). Copyright (2012 and 2015) Elsevier)

For Fe^BH, the aging model is described by the outward diffusion of the Fe⁰ core toward the shell. This results in the formation of hollowed-out iron oxide shells, which are further transformed into sheet- and needle-shaped materials (Liu et al. 2015). Various secondary iron oxide mineral phases were formed at different aging times. For example, at 5 days, the main iron oxide shell is magnetite (Fe₃O₄) and maghemite (γ-Fe₂O₃), accompanied by lepidocrocite (γ-FeOOH). For 10-day aging, the ferrihydrite and lepidocrocite were dominant. When aged up to 90 days, the products are mainly γ-FeOOH mixed with small amounts of Fe₃O₄ and γ-Fe₂O₃, as also evident by the formation of the sheet- and needle-shaped materials, the typical morphology of lepidocrocite (Liu et al. 2015). Due to their high reactivity and the outward diffusion of Fe⁰ core, Fe^BH aging can substantially consume Fe⁰, resulting in a significant decrease of Fe^BH reactivity over time (Wang et al. 2010). A similar decrease in reactivity was also reported in Pd/Fe^BH particles in which aging causes both catalyst deactivation (i.e., Pd was completely buried underneath an extensive iron oxide matrix) and Fe⁰ depletion (Yan et al. 2010). The CMC-modified Fe^BH was affected by aging in a similar manner. Nevertheless, the coating of CMC could slow the aging rate of Fe^BH as indicated by the slower drop in Fe⁰ intensity in the XRD pattern (Dong et al. 2016). Moreover, CMC was found to influence the transformation of Fe⁰ and the formation of iron oxide because greater CMC loading in the suspension results in more the lepidocrocite formation as a corrosion product of CMC-modified Fe^BH.

Unlike Fe^BH, aging of Fe^H2 follows a growing shell and shrinking core model with no outward diffusion of Fe⁰ core. The periods of increased, declined, and stabilized reactivity were observed during the aging of Fe^H2. Liu and Lowry (2006) studied reactivity changes with aging of Fe^H2 using TCE as a reactivity probe. They observed a significant initial decrease (for the first 10 days) in TCE reaction rate constants (k_obs,TCE) for Fe^H2 (Fe⁰ = 48%) decreased from 6.2 × 10⁻³ L/(h^.m²) to 1.0 × 10⁻³ L/(h^.m²) after 10 days (Liu and Lowry 2006). They interpreted the initial decrease as the healing of defects in the oxide film formed when the particles were removed from the concentrated suspension and dried. This is followed by a constant or slightly increasing rate constant during which the Fe⁰ content of the particles decreased by ∼40%. Thus, they concluded that the TCE reaction is zero-order with respect to the Fe⁰ content of Fe^H2.

Eventually, reactivity with TCE ceased after 170 days when the Fe⁰ content reached ∼4.6%. A similar trend was observed by Sarathy et al. (2008) who used tetrachloromethane as a reactivity probe. They reported the increase of dechlorination rate constants of tetrachloromethane at the initial state (1–2-day exposure of dried Fe^H2 to DO/DI water) followed by a gradual decrease over 1–6 months of aging (depending on the drying procedure of the Fe^H2). The initial increase in reactivity was explained by depassivation of the particles soon after the first immersion of dried Fe^H2 in water. The depassivation involves breakdown of the iron oxide shell, thereby exposing the underlying Fe⁰ core to the target contaminant. The gradual decrease in dechlorination rate constants after 1–2 days suggests the repassivation due to growth of a new oxide shell or transformation of the oxide shell, which gradually stabilizes the electron transfer from the Fe⁰ core.

3.1.3 Environmental Factors Affecting NZVI Reactivity

Table 3.1 summarizes the batch experiments for reductive treatment of various kinds of organic contaminants using various kinds of NZVI. Most of the studies were conducted in a batch system in DI water at a pH of 8–9, representing an Fe(OH)₂/H₂O or Fe₃O₄/H₂O equilibrium. However, NZVI particles are applied in a subsurface environment, which is far more complex than DI water. The presence of dissolved inorganic and organic species, dense nonaqueous phase liquid (DNAPL), and aquifer materials in the subsurface can physically and chemically affect the performance of NZVI particles. For this reason, this section reviews the effects of these environmental factors on NZVI performance.

3.1.3.1 pH

Groundwater pH typically varies from 6 to 8. The application of NZVI for in situ remediation is normally at only 0.2–0.5 wt %. Thus, because of the high buffer capacity of most soil, the groundwater pH will not change much due to the injection and emplacement of NZVI. Consequently, groundwater pH can substantially affect the electron utilization and lifetime of NZVI (Liu and Lowry 2006). As shown in Eq. (3.3), H⁺ can consume Fe⁰ to produce H₂. This is controlled by pH and competes with electron utilization for dechlorination (Eq. 3.2). Therefore, at low pH, Fe⁰ tends to be utilized for H₂ production rather than dechlorination. For Fe^H2, the H₂ evolution rate constant increased by 27-fold (from 0.008 to 0.22 day⁻¹) due to decreasing pH from 8.9 to 6.5 (Liu and Lowry 2006). However, the TCE dechlorination rate constants were only two times higher (Liu and Lowry 2006). The increase of the TCE dichlorination rate constant is 1.44/h with respect to a one unit decrease of pH.

A similar trend was also observed for Fe^BH for dichlorination of trichloromethane (Song and Carraway 2006) and 1,1,2,2-tetrachloroethane (Song and Carraway 2005) of which the dichlorination rate constants increase by 0.33 and 0.24/h per one unit decrease of pH, respectively. Rapid H₂ evolution and Fe⁰ consumption yield a relatively short lifetime of NZVI. At a pH of 6.5, the reactive lifetime of Fe^H2 was only around 2 weeks (Liu and Lowry 2006). For this reason, both particle lifetime and efficiency decline due to the decrease of pH. The adverse effect of low pH becomes more severe when NZVI particles are applied to treat groundwater plumes in comparison to the source zone treatment because, in plumes, a low concentration of target contaminant is available for reaction with NZVI, and most Fe⁰ will be used for H₂ production. In conclusion, the application of NZVI particles to treat contaminant plumes in aquifers with low pH (~6.5) is not favorable, and additional NZVI injections are expected every few weeks, making it economically unfeasible.

3.1.3.2 Anionic and Cationic Species

Due to geochemical cycles (dissolution and precipitation) of minerals in the subsurface, groundwater normally consists of various anionic species, such as NOO₃⁻, Cl⁻, SO₄²⁻, HCO₃⁻, and HPO₄⁻². Groundwater chemistry can affect NZVI corrosion rate, dechlorination rate, H₂ production, dissolution, and formation of an iron oxide shell. At low concentration (0.2–1 mM), reducible solutes, such as NO₃⁻, did not significantly affect the TCE dechlorination rate. However, at high concentration (~5 mM), NO₃⁻ deactivated Fe^H2 reactivity toward TCE after 3 days, even though Fe⁰ remained in Fe^H2 (Liu et al. 2007). Similarly, NO₃⁻ (7.7 mM) was reported to inhibit dichlorination of hexachlorobenzene using Fe^BH (Su et al. 2012b).

Presumably, two possible hypotheses are that, at this high NO₃⁻ concentration, nitrite may be built up to block reactivity, or nitrate may promote the formation of a passivating and insoluble Fe(III) oxide layer, like maghemite or hematite (Schlicker et al. 2000). However, neither maghemite, hematite, nor goethite was observed using EXAFS (Reinsch et al. 2010), suggesting that the passivating layer may be some other phases or at a too low concentration to be detected using EXAFS. Nevertheless, the passivation by nitrate must happen very rapidly because the Fe^H2 passivated by NO₃⁻ contained Fe⁰, schwertmannite, and magnetite at the same ratio as fresh Fe^H2 as if no Fe⁰ oxidation took place (Reinsch et al. 2010).

In contrast, non-reducible anions, such as Cl⁻, SO₄²⁻, HCO₃⁻, and HPO₄⁻², decreased the TCE dechlorination rate up to a factor of seven in comparison to DI water, and the order of their effect follows their affinity of anion complexation to hydrous ferric oxide (i.e., Cl⁻ < SO₄²⁻ < HCO₃⁻ < HPO₄²⁻) at pH 8.9 (Liu et al. 2007). This implies that the inhibitory effect of these solutes on TCE degradation may be caused by reactive site blocking due to the formation of Fe-anion complexes on the Fe^H2 surface. On the other hand, as for dichlorination of HCB using Fe^BH, HCO₃⁻ did not affect the dichlorination rate, while Cl⁻ and SO₄²⁻ slightly enhanced HCB dechlorination rate constants due to their corrosion promotion (Su et al. 2012b).

Similar to the case of anionic species, cationic species, such as Na⁺, Mg²⁺, Fe²⁺, Cu²⁺, Ni²⁺, Cd²⁺, and Zn²⁺, might be released to the groundwater due to geochemical cycles. Furthermore, these pollutants might be found at high concentrations in DNAPL-contaminated areas as co-contaminants. The NZVI has been demonstrated to be effective for the immobilization of various metals (Ponder et al. 2000; Zhang 2003; Dries et al. 2005; Kanel et al. 2005) through reduction (Eq. 3.8), co-precipitation, and surface complexation (Eqs. 3.9 and 3.10) depending on the relative standard potential E⁰ between NZVI and metals (Li and Zhang 2007; see Chap. 4).

$$ {\mathrm{Me}}^{a+}+{\mathrm{be}}^{-}\to {\mathrm{Me}}^{a-b} $$

(3.8)

$$ \equiv \mathrm{S}-\mathrm{OH}+{\mathrm{Me}}^{2+}\leftrightarrow \equiv \mathrm{S}-{\mathrm{OMe}}^{+}+{\mathrm{H}}^{+} $$

(3.9)

$$ 2\equiv \mathrm{S}-\mathrm{O}\mathrm{H}+{\mathrm{Me}}^{2+}\leftrightarrow {\left(\equiv \mathrm{S}-\mathrm{O}\right)}_2\;\mathrm{Me}+2{\mathrm{H}}^{+} $$

(3.10)

For metal ions such as Zn²⁺ and Cd²⁺, of which E⁰ is very close to or more negative than that of Fe⁰, the main removal mechanism is sorption and surface complexation (Li and Zhang 2007). For metal ions such as Ni²⁺ and Pb²⁺, of which E⁰ is slightly more positive than that of Fe⁰, the removal mechanism is the combination of reduction and sorption (Li and Zhang 2007). For metal ions such as Cu²⁺, Ag⁺, and Hg²⁺, of which E⁰ is greatly more positive than that of Fe⁰, the major removal mechanism is reduction (Li and Zhang 2007). As shown in Eq. (3.8), the reduction of metals consumes electrons (i.e., a competing reaction to dichlorination of chlorinated organics; Eq. 3.2). In addition, the reduction of metals is normally followed by the precipitation of reduced metals on the surface of NZVI. This co-precipitation reaction with surface complexation (Eqs. 3.9, 3.10, and 3.11) can block reactive sites or promote the formation of Fe(III)-metal oxide, a passivating layer, at anodic sites. This decreases NZVI reactivity for dechlorination. For this reason, in general, the presence of metals as co-contaminants adversely affects the dechlorination rate. For example, in the presence of Zn, TCE degradation using ZVI was 2–4 times slower (Dries et al. 2005). Similarly, in the presence of Cr(VI) at a concentration higher than 5 mg/L, the TCE dechlorination rate decreases by a factor of 3–13 (Dries et al. 2005).

Similarly, Fe(II) inhibited the HCB degradation reaction due to passivation layers formed, while Na⁺ and Mg²⁺ did not substantially affect the dichlorination (Su et al. 2012b). However, the presence of noble metals including Cu²⁺, Ni²⁺, Pd²⁺, and Pt²⁺ appeared to enhance the dichlorination rate through the catalytic pathway. The presence of Ni (5–100 mg/L) enhanced the TCE dechlorination due to the catalytic hydrodechlorination by bimetallic Fe⁰/Ni⁰ from the precipitation of Ni⁰ on the surface of ZVI (Dries et al. 2005). Similarly, the presence of Cu²⁺ enhanced dichlorination of HCB (Su et al. 2012b). Furthermore, metal ions including Co²⁺, Cu²⁺, and Ni²⁺ enhanced the dechlorination of 4-chlorobiphenyl (4-ClBP) by NZVI. The dechlorination percentages of 4-ClBP in the presence of 0.1 mmol/L of Co²⁺, Cu²⁺, and Ni²⁺ were 66.1%, 66.0%, and 64.6% in 48 h, and then increased to 67.9%, 71.3%, and 73.5%, after 96 h, respectively (Wang et al. 2011).

3.1.3.3 Natural Organic Matters in Groundwater and Soil

Groundwater naturally contains a significant amount of natural organic matter (NOM) originated from decomposition of animal and plant bodies (Schwarzenbach et al. 2003b). The NOM is a natural charged macromolecule, carrying a net negative charge at a natural pH due to the dissociation of carboxylic groups (Schwarzenbach et al. 2003a). Furthermore, NOM consists of humic and fulvic acids. By operational definition, humic acid is the fraction of NOM that precipitates at pH 2 or lower, while the fulvic acid fraction stays soluble under all pH conditions (Schwarzenbach et al. 2003a). The NOM was found to adsorb various kinds of colloids and nanoparticles (Ramos-Tejada et al. 2003; Hyung et al. 2007). Similarly, carboxylic groups of NOM can specifically adsorb onto the iron oxide surface of Fe^H2 (Fig. 3.12a). The NOM is an anionic polyelectrolyte, which tends to adsorb onto the substrate in a train-loop-tail configuration (Fleer et al. 1998) similar to the polyelectrolyte discussed above.

Fig. 3.12

(a) Adsorption isotherm of HA on RNIP at pH 8.5; the line illustrates the best fit using the Langmuir isotherm. (RNIP is NZVI formed by reduction using H₂). (b) The relationship between standardized TCE dechlorination rate constants using RNIP and HA concentrations in the aqueous phase

There are two different hypotheses of the effects of NOM on ZVI performance. First, NOM can enhance electron transfer and thus ZVI reactivity for pollutant degradation through electron shuttle effects (Tratnyek et al. 2001). Second, adsorbed NOM decreases ZVI reactivity due to reactive site blocking (Tratnyek et al. 2001; Cho and Park 2006; Doong and Lai 2006). Moreover, consisting of the quinone group with standard potential E⁰ of 0.23 V, NOM is hypothesized to transfer electrons from ZVI for the dechlorination of chlorinated ethene (Tratnyek et al. 2001). The enhanced dechlorination due to the presence of humic acid in the ZVI system was observed for PCE dechlorination but not TCE (Cho and Park 2006). In contrast, Tratnyek et al. (2001) reported that TCE degradation kinetics decreased by 21% and 39% in the presence of 20 and 40 mg/L, respectively, of Suwannee River organic matter, presumably due to reactive site blocking.

Figure 3.12b supports and extends the second hypothesis regarding the effect of NOM on TCE dechlorination using NZVI . In the presence of different humic acid concentrations, the TCE dechlorination rates by bare RNIP decreased nonlinearly and exhibited two regions. The TCE dechlorination pathways were not affected, and β elimination remained the dominant dechlorination pathway, yielding acetylene as the reaction intermediate and ethane and ethene as products. Consistent with the Scheutjens-Fleer theory for homopolymer sorption (Fleer et al. 1998), the nonlinear relationship between the dechlorination rate and the surface excess of adsorbed humic acids suggests that adsorbed humic acids decrease reactivity primarily by blocking reactive surface sites at low surface excess where they adsorb relatively flat onto the Fe^H2 surface and by site blocking and decreasing TCE availability at high surface excess where humic acids form an extended layer around the particles. This finding is confirmed by a recent study revealing that the presence of Suwannee River humic acids (SRHA) (10 mg/L) decreased TCE (20 mg/L) dechlorination by Fe^H2 around 23% but did not affect the H₂ production (Chen et al. 2011a). Nevertheless, the mix effect of NOM on NZVI reactivity was also reported. A recent study reported that the presence of humic acid inhibited the reduction of 4-ClBP in the first 4 h, but then significantly accelerated dechlorination by reaching 86.3% in 48 h (Wang et al. 2011).

In addition to the dissolved NOM, the NOM that is adsorbed onto soil or aquifer material can substantially decline dechlorination efficiency. This is an issue of mass transfer of CVOCs from soil to groundwater. The CVOC-sorbed soil may behave as a long-term secondary source, gradually leaching dissolved CVOCs to contaminate the groundwater downgradient. Slow desorption of CVOCs from the soil can result in retardation of reductive detoxification using NZVI. Using TCE as an example, TCE has an arithmetic mean organic carbon partitioning coefficient (k_oc) of 86 (ranging from 18.5 to 150). The soil-water partitioning coefficient (k_d) is the k_oc x fraction of organic carbon in soil (f_oc). This partitioning coefficient can be used to calculate the retardation factor (R) as shown in Eq. 3.11. Subsequently, we can estimate the decrease in TCE dechlorination rate by NZVI in the soil-water system (k_TCE-aq-Soil) with R using Eq. 3.12 in comparison to the TCE dechlorination rate constant using NZVI in water (no sorption onto soil; k_TCE-aq):

$$ R=1+\frac{{k_{\mathrm{d}}}^{\ast }{\rho}_{\mathrm{b}}}{n} $$

(3.11)

$$ {k}_{\mathrm{TCE}\hbox{-} \mathrm{aq}\hbox{-} \mathrm{Soil}}=\frac{k_{\mathrm{TCE}\hbox{-} \mathrm{aq}}}{R} $$

(3.12)

where ρ_b and n are the bulk density of soil and porosity, respectively. Under equilibrium with the partitioning coefficient (k_d) of 1.46 L/kg, R is calculated as 7.28 using Eq. 3.11. Thus, the TCE sorption into soil can decrease the TCE dechlorination rate constant in the soil-water system by 7.28 times in comparison to the system with groundwater alone. This is problematic because, instead of using its reducing power to destroy contaminants, the NZVI reacts with the water to form H₂, which increases the amount of NZVI required for remediation (Liu et al. 2007; Berge and Ramsburg 2010). This is a mass-transfer limitation problem that cannot be solved by modifying NZVI to have greater reactivity, such as by doping with catalysts. Instead, a possible solution is to use NZVI with thermal-enhanced CVOC dissolution or desorption, which will speed up the reaction rate and improve the electron utilization efficiency of the remediation (see Chap. 11).

3.1.3.4 TCE Concentration and the Presence of DNAPL

The NZVI particles are proposed for the remediation of both the DNAPL source zone and groundwater plumes. Therefore, NZVI is subjected to a wide range of CVOC concentration, from low concentration in contaminant plumes to near saturation or even DNAPL in the source zone. The different CVOC concentrations in different zones influence electron utilization, particle efficiency, and the reactive lifetime of NZVI. A recent study (Liu et al. 2007) revealed that changing TCE concentrations from low to medium range (0.03–0.46 mM) insignificantly affected the TCE dechlorination rate using Fe^H2 at pH 7. However, at higher concentrations (1.3 mM) to TCE water saturation (8.4 mM), the TCE dechlorination rate by Fe^H2 decreased by a factor of two, presumably due to reactive site blocking by acetylene, an intermediate.

The higher the TCE concentration, the higher the Fe⁰ utilization efficiency for TCE dechlorination, which is evident from the decrease of H₂ evolution and the shift of byproduct formation toward an unsaturated byproduct (Liu et al. 2007), e.g., acetylene. At pH 7, 40% of Fe⁰ in Fe^H2 was consumed for H₂ production at a TCE concentration of 0.46 mM, while only 7% of the initial Fe⁰ in Fe^H2 was used for H₂ production at a TCE concentration of 8.4 mM (Liu et al. 2007). Evidently, TCE outcompeted Fe⁰ with H⁺. Similarly, acetylene accounted for 86% of the products, while ethene and ethane were 10% and 4%, respectively, at the TCE concentration of 8.4 mM (Liu et al. 2007). Acetylene formation requires fewer electrons than ethene and ethane formation, making TCE dechlorination by Fe^H2 utilize electrons more effectively. The accumulation of acetylene could be the result of TCE saturating reactive sites and blocking acetylene from further transformation.

However, increasing the TCE concentration adversely affects the reactive life time of Fe^H2. The reactive lifetime of Fe^H2 is 10, 40, and 60 days for TCE concentrations of 8.4, 1.3, and 0.46 mM, respectively, at pH 7 buffered by HEPES (Liu et al. 2007). At the DNAPL-water interface without HEPES, TCE dechlorination with Fe^H2 decreased pH to 4–5. The reactive life time of Fe^H2 at DNAPL interface was only 5 days, and no Fe⁰ remained in the particles. The particle efficiency for TCE dechlorination using Fe^H2 at the DNAPL -water interface was only ~15% because the local pH decreased and accelerated H₂ production in a much faster rate than increasing TCE dechlorination (see pH effect). This suggests that Fe^H2 used for source zone treatment will have a relatively short lifetime. Therefore, additional injections for source zone treatment are expected to be more frequent than for plumes.

3.1.3.5 Microorganisms

Various kinds of microorganisms such as sulfate reducers, iron reducers, and methanogens are present under different geochemical conditions in the subsurface. Some microorganisms such as halorespirers can biologically transform chlorinated organics to notorious byproducts under favorable conditions (Harkness et al. 1999; Cupples et al. 2004). Potential synergic effects of abiotic and biotic remediation via stimulated bioremediation using polymer-modified NZVI, leading to the long-term degradation of chlorinated organics, is an active area of research. Chapters 7 and 10 are devoted to this important combined remedy. The effects of microbes on NZVI reactivity and longevity are also discussed in Chap. 10. A reader should consult the microbiological-related materials in Chaps. 7 and 10 prior to designing an in situ remediation using NZVI since, in a real field implementation, interaction between NZVI and microbes is unavoidable and can substantially affect contaminant treatability .

3.2 Oxidative Transformation of Organic Contaminants Using NZVI-Induced Fenton’s Reaction

While being considered a reducing agent, NZVI has also recently gained a lot of interest in the oxidation aspect. In general, NZVI can be used as a reagent in the process of producing oxidative radicals for contaminant degradation. Oxidation approaches of NZVI involve NZVI-induced Fenton’s reaction (Xu and Wang 2011; Choi and Lee 2012; Li and Zhu 2014), NZVI under aeration (Taha and Ibrahim 2014a), and NZVI-induced persulfate system (Al-Shamsi and Thomson 2013; Diao et al. 2016). Nevertheless, the NZVI-induced Fenton’s reaction is the most studied one. The state-of-the-art progress of the NZVI-induced Fenton’s reaction is focused on in this section.

Discovered by Henry J. H. Fenton in 1894, Fenton’s reaction is one of the most common chemical oxidation processes in wastewater treatment and site remediation (Ay and Kargi 2010; Petri et al. 2011; Babuponnusami and Muthukumar 2014; Papoutsakis et al. 2016). Typically, a homogeneous Fenton’s reaction utilizes ferrous sulfate as a catalytic Fe²⁺ source. An acidic condition is required to maintain Fe²⁺ dissolution. However, acidification and the high dose of mobile ferrous/ferric required to obtain an effective treatment are the major drawbacks of conventional Fenton’s reaction (Xu and Wang 2011; Li and Zhu 2014; Yu et al. 2014; Cheng et al. 2015; Usman et al. 2016). For in situ applications, the acidification of the subsurface is probably one of the most challenging problems due to the large buffering capacity at a neutral pH range of the aquifers (Petri et al. 2011). Moreover, acidification might result in the dissolution of metals in the subsurface (Petri et al. 2011). Hence, NZVI-induced Fenton’s reaction has emerged as an alternative approach to avoid these problems.

3.2.1 Mechanism

The mechanism of contaminant degradation using NZVI-induced Fenton’s reaction is primarily based on highly reactive radicals, which are generated from the reaction between Fe⁰ and hydrogen peroxide. The radicals playing the key role in Fenton’s reaction include highly reactive hydroxyl radical (•OH) and hydroperoxyl radical (OOH•). Instead of using acidification to maintain dissolved Fe²⁺ concentration, the NZVI Fenton process is self-catalytic, based on oxidative dissolution of NZVI in the presence of H₂O₂. Interfacial H⁺ is produced at the NZVI surface to provide appropriate local pH, which continuously releases Fe²⁺ for Fenton’s reaction. Babuponnusami and Muthukumar (2012), Chu et al. (2012), and Xu and Wang (2011) indicated the success of NZVI Fenton’s reaction at circumneutral pH 6. Moreover, the NZVI Fenton process is more favorable than the conventional approach using dissolved Fe²⁺ because NZVI can be magnetically recovered and reused for several times (Diya’uddeen et al. 2015).

The critical reactions in a heterogeneous NZVI-induced Fenton’s system are as follows:

$$ {\mathrm{Fe}}^0+{\mathrm{H}}_2{\mathrm{O}}_2+2{\mathrm{H}}^{+}\to {\mathrm{Fe}}^{2+}+2{\mathrm{H}}_2\mathrm{O} $$

(3.13)

$$ {\mathrm{Fe}}^0+{\mathrm{O}}_2+2{\mathrm{H}}^{+}\to {\mathrm{Fe}}^{2+}+{\mathrm{H}}_2{\mathrm{O}}_2 $$

(3.14)

$$ {\mathrm{Fe}}^0+2{\mathrm{H}}^{+}\to {\mathrm{Fe}}^{2+}+{\mathrm{H}}_2 $$

(3.15)

$$ {\mathrm{Fe}}^0+{\mathrm{H}}_2\mathrm{O}\to {\mathrm{Fe}}^{2+}+{\mathrm{OH}}^{-} $$

(3.16)

$$ {\mathrm{Fe}}^{2+}+{\mathrm{H}}_2{\mathrm{O}}_2\to {\mathrm{Fe}}^{3+}+\cdotp \mathrm{OH}+{\mathrm{O}\mathrm{H}}^{-} $$

(3.17)

$$ {\mathrm{Fe}}^{2+}+\cdotp \mathrm{OH}\to {\mathrm{Fe}}^{3+}+{\mathrm{OH}}^{-} $$

(3.18)

$$ {\mathrm{Fe}}^{3+}+{\mathrm{H}}_2{\mathrm{O}}_2\to {\mathrm{Fe}}^{2+}+\mathrm{OOH}\cdotp +{\mathrm{H}}^{+} $$

(3.19)

The NZVI-induced Fenton’s reaction has been reported to effectively degrade a variety of organic contaminants including textile wastewater, pharmaceuticals, halogenated compounds, and other non-halogenated compounds (Table 3.2). The degradation kinetics was found to fit well with the pseudo-first-order reaction. Degradation rates are mainly in the range of 0.01–0.2/min. However, depending on the reaction conditions, the rates can vary from 0.0064/min to 1.79/min. Zhang et al. (2017) reported that the degradation of norfloxacin was at a rate of 0.0064/min at pH 6, while the degradation rate at pH 3 was 0.1/min. Xu and Wang (2011) presented the 100% removal of 4-chloro-3-methyl phenol within 60 min at the degradation rate of 0.35/min at 0.1 g/L of NZVI and 1.79/min at 0.5 g/L of NZVI. In addition, although the NZVI-induced Fenton’s reaction kinetics is pseudo-first-order in most of the cases, some studies suggest that the degradation kinetics is more appropriate for the pseudo-second-order reaction. For instance, Zhou et al. (2015) revealed the pseudo-second-order degradation of 1-alkyl-3-methylimidazolium-bromides at a rate of 0.0415 L/(mM^.min).

Table 3.2 Summary of batch experiments for NZVI-induced Fenton’s reaction of various organic contaminants

In general, NZVI-induced Fenton’s reaction has high reactivity. However, it is a two-stage reaction, in which the first stage usually shows low degradation rates and is followed by a rapid degradation in the second stage. The lag stage is considered an activation process of the surface iron species where iron dissolution occurs on the NZVI surface and participates in the catalytic reaction. Moreover, despite the high reactivity, the degradation by NZVI-induced Fenton’s reaction is usually inhibited quickly ranging from 0.5 to 3 h due to the depletion of H₂O₂. Yu et al. (2014) showed that the textile wastewater removal was inhibited after 40 min gaining 80% efficiency. Cheng et al. (2015) reported the inhibition of pentachlorophenol degradation after an hour with a removal efficiency of 60%. Zha et al. (2014) described that amoxicillin degradation was impeded after 15 min, obtaining approximately 80% removal efficiency. The inhibition even occurs more frequently at a neutral pH (Li and Zhu 2014; Zhou et al. 2015; Wang et al. 2016b). The sequential addition of Fenton’s reagent can be a solution to this problem. Munoz et al. (2014) presented the prolongation of homogenous Fenton reaction to treat sawmill wastewater by the sequential addition of H₂O₂.

3.2.2 Factors Affecting Treatment Efficiency

3.2.2.1 Reagent Dose

The H₂O₂ acts as a scavenger of hydroxyl radical (OH) (Xu and Wang 2011; Li and Zhu 2014; Li et al. 2015; Wang et al. 2016b) as shown in Eqs. 3.20 and 3.21, while ∙OH plays the dominant role in contaminant degradation (Petri et al. 2011). In addition, H₂O₂ can be self-degraded as presented in Eq. 3.22. Therefore, the use of excessive H₂O₂ led to the degradation inefficiency. As a result, sequential H₂O₂ addition, as done in the homogeneous Fenton’s process (Martins et al. 2010; Villa et al. 2010; Munoz et al. 2014), is also a possible solution for sustaining NZVI-induced Fenton’s reaction as to be discussed next in the treatability of 1,2-DCA.

$$ {\mathrm{H}}_2{\mathrm{O}}_2+\cdotp \mathrm{OH}\to {\mathrm{H}\mathrm{O}}_2\cdotp +{\mathrm{H}}_2\mathrm{O} $$

(3.20)

$$ {\mathrm{H}\mathrm{O}}_2\cdotp +\cdotp \mathrm{OH}\to {\mathrm{H}}_2\mathrm{O}+{\mathrm{O}}_2 $$

(3.21)

$$ {\mathrm{H}}_2{\mathrm{O}}_2+{\mathrm{H}}_2{\mathrm{O}}_2\to 2{\mathrm{H}}_2\mathrm{O}+{\mathrm{O}}_2 $$

(3.22)

$$ {\mathrm{Fe}}^{2+}+{\mathrm{H}}_2{\mathrm{O}}_2\to {\mathrm{Fe}}^{3+}+\cdotp \mathrm{OH}+{\mathrm{O}\mathrm{H}}^{-} $$

(3.24)

$$ {\mathrm{Fe}}^{2+}+\cdotp \mathrm{OH}\to {\mathrm{Fe}}^{3+}+{\mathrm{OH}}^{-} $$

(3.25)

The NZVI was the source to generate ∙OH (Li et al. 2015; Wang et al. 2016b) as shown in Eqs. 3.13 and 3.24. Therefore, a higher NZVI concentration provided higher degradation efficiency. However, the excessive iron source has been known as an ∙OH scavenger (Xu and Wang 2011; Wang et al. 2016b), which may prohibit the degradation as shown in Eqs. 3.13 and 3.25. Table 3.2 summarizes the optimal reagent dose and conditions for the treatment of various contaminants.

3.2.2.2 Initial pH

Babuponnusami and Muthukumar (2012) and Xu and Wang (2011) also reported that the heterogeneous NZVI Fenton’s reaction at a neutral pH succeeded in removing phenol and 4-chloro-3-methyl phenol, respectively. Babuponnusami and Muthukumar (2012) reported 65% of phenol removal using NZVI Fenton’s process at a pH 6.2. Chu et al. (2012) showed a decrease of 95% in total phenols and 50% in COD of coking wastewater using iron powder as a Fenton’s catalyst at pH 6.5 and 5.4. Xu and Wang (2011) presented the complete degradation of 4-chloro-3-methyl phenol using an NZVI-induced Fenton’s system at a pH of 6.1. Some examples of optimal pH values for the treatment of different organic contaminants are summarized in Table 3.2.

3.2.3 Treatability of 1,2–DCA

While reductive dechlorination using NZVI is incapable of detoxifying 1,2-DCA as discussed previously, the Fenton process can. Masten and Butler (1986) suggested the success of 1,2-DCA degradation due to free radicals. Noticeably, Vilve et al. (2010) successfully degraded 1,2-DCA at the laboratory scale using a conventional Fenton process. Recently, Le and Phenrat (2018) evaluated the NZVI-induced Fenton process at a neutral pH to degrade 1,2-DCA at a high concentration (2000 mg/L), representing a dissolved 1,2-DCA concentration close to the DNAPL source zone. Approximately 87% of 1,2-DCA was degraded at a neutral pH, with a pseudo-first-order rate constant of 0.98/h using 10 g/L of NZVI and 200 mM of H₂O₂. However, the reaction was prohibited quickly, within 3 h, presumably due to the rapid depletion of H₂O₂. The application of sequential H₂O₂ addition provided a better approach for preventing rapid inhibition via controlling the H₂O₂ concentration in the system to be sufficient but not in excess, thus resulting in the higher degradation efficiency (the pseudo-first-order rate constant of 0.49/h and 99% degradation in 8 h for 10 g/L NZVI and 200 mM but sequential for 25 mM per 30 min). Using NZVI with sequential H₂O₂ addition (25 mM per 30 min) was also successful in degrading 1,2-DCA sorbed onto soil, yielding 99% removal of 1,2-DCA within 16 h at a rate constant of 0.23/h (Fig. 3.13a), around two times slower than in the system without soil, presumably due to rate-limited 1,2-DCA desorption from the soil.

Fig. 3.13

(a) 1,2-DCA degradation in soil-groundwater system and (b) NZVI reuse for three removal cycles (conditions: natural pH, initial 1,2-DCA 2000 mg/L, NZVI 10 g/L, 25 mM H₂O₂ per 30 min)

The NZVI-induced Fenton reaction can be reused for several treatment cycles. Figure 3.13b illustrates the 1,2-DCA degradation kinetics using NZVI-induced Fenton’s reaction in three consecutive cycles. In the first cycle of degradation, 99.9% of 1,2-DCA was degraded for 16 h, obtaining the rate constant of 0.49/h. When the stock 1,2-DCA solution was added into the reactors to restart the treatment cycles, the degradation rate constant declined almost three times to 0.18/h in Cycle 2 and four times to 0.13/h in Cycle 3. Approximately 99.2% and 98.0% of 1,2-DCA were degraded for 24 h and 30 h in Cycles 2 and 3, respectively. This finding suggests that the heterogeneous NZVI Fenton process is a promising approach for in situ treatment. The NZVI can be emplaced in the subsurface close to the DNAPL source zone, while a small amount of H₂O₂ is recirculated to treat 1,2-DCA in both groundwater and soil.