Metallic Iron for Water Remediation: Plenty of Room for Collaboration and Convergence to Advance the Science

Abstract

Scientific collaboration among various geographically scattered research groups on the broad topic of “metallic iron (Fe⁰) for water remediation” has evolved greatly over the past three decades. This collaboration has involved different kinds of research partners, including researchers from the same organization and domestic researchers from non-academic organizations as well as international partners. The present analysis of recent publications by some leading scientists shows that after a decade of frank collaboration in search of ways to improve the efficiency of Fe⁰/H₂O systems, the research community has divided itself into two schools of thought since about 2007. Since then, progress in knowledge has stagnated. The first school maintains that Fe⁰ is a reducing agent for some relevant contaminants. The second school argues that Fe⁰ in-situ generates flocculants (iron hydroxides) for contaminant scavenging and reducing species (e.g., Fe^II, H₂, and Fe₃O₄), but reductive transformation is not a relevant contaminant removal mechanism. The problem encountered in assessing the validity of the views of both schools arises from the quantitative dominance of the supporters of the first school, who mostly ignore the second school in their presentations. The net result is that the various derivations of the original Fe⁰ remediation technology may be collectively flawed by the same mistake. While recognizing that the whole research community strives for the success of a very promising but unestablished technology, annual review articles are suggested as an ingredient for successful collaboration.

1. Introduction

Permeable reactive barriers of metallic iron (Fe⁰ PRBs) are reportedly a well-established technology for groundwater remediation [1,2,3,4]. Since the first four peer-reviewed articles on testing Fe⁰ for groundwater remediation [5,6,7,8], numerous studies have investigated various aspects on the design and installation of efficient Fe⁰ PRBs, including the reaction mechanisms and the long-term hydraulic properties [9,10,11]. Moreover, alternatives to granular Fe⁰ materials were developed, including nano-scale Fe⁰ and sulfidized Fe⁰ [12,13,14,15,16]. The results have been presented in more than 5000 research and review articles [17,18]. The research to date is reported to have followed a conventional collaborative approach [19,20,21,22,23]. In particular, intensive international research collaborations have been documented, as revealed by two recent bibliometric analyses [17,18]. One major aim of research collaboration is boosting the advancement of knowledge while accounting for the increasing professionalization and specialization of science [22]. A closer look behind the publications on the Fe⁰/H₂O system reveals that, despite increased scientific productivity, there have been no frank concerted efforts for the “advancement of knowledge” for the past 15 years.

The first problem of the Fe⁰ PRB technology and the post-1990 Fe⁰ remediation literature is that it was introduced with very little critical literature research [24]. In particular, the old textbook knowledge that granular Fe⁰ generates flocs for water treatment, just like iron salts [25], has not been considered. In one study, it was claimed that a thorough patent research was made, and it revealed some works from the 1970s [26]. This view was directly contradicted in a recent review article by Antia [27] revealing that patents on Fe⁰ for water treatment dating back to the 1850s exist, more than 140 years before the advent of Fe⁰ PRBs. The question then arises of whether or not progress in knowledge can be achieved without a common starting point? Unfortunately, all warnings from investigators suggesting that Fe⁰ might not be a reducing agent under environmental conditions [28,29,30,31,32,33,34] were overlooked by the majority of active researchers. In particular, Lavine et al. [31] used differential pulse polarography to simultaneously monitor the disappearance of nitrobenzene and the appearance of Fe²⁺ in the Fe⁰/H₂O system. Their studies were very “informative” but could not deliver any information about ‘’crucial mechanistic details on the chemical processes controlling the reduction of organic compounds” in the Fe⁰/H₂O system. Jiao et al. [34] and many others have questioned the view that electrons from the metal body (Fe⁰) are of any significance for the reductive transformation of soluble species, including contaminants and O₂ [24,35].

To the authors’ knowledge, only a few research groups have designed experiments to cope with the science of the Fe⁰/H₂O system [36,37,38,39,40,41,42,43,44,45]. In particular, Li et al. [39] have alloyed Fe⁰ to lower the corrosion rate in an effort to have less dissolved iron in the effluent. The three most prominent research groups are the following: (i) Ghauch’s group in Lebanon [46], (ii) Gheju’s group in Romania [47], and (iii) our group in Germany [48]. Interestingly, until 2010/2011, all three research groups reported on the electrochemical nature of contaminant removal in Fe⁰/H₂O systems [49,50,51]. In other words, critics of the reductive transformation model are not enemies of the Fe⁰ technology, but are proponents for the integrity of science [24,50,51,52,53,54]. The aim of the present overview article is to point out the flaws or limitations of the named concept while demonstrating that even leading scientists or proponents contribute to perpetuating the concept. The objective is not to blame any single colleague, but rather to correct past errors and lead early-career researchers of Fe⁰/H₂O systems on the right path. The presentation starts with an explanation of the methodology used (Section 2), followed by a summary of the fundamental flaws (Section 3). Then, an overview on how leading scientists have been ignoring the state-of-the-art knowledge is presented (Section 4). A short conclusion will close the presentation.

2. Methodology

No systematic approach using search terms and combinations thereof was used to identify published scientific literature related to “Fe⁰ for water remediation”. Thus, this paper is not a systematic review, but rather, a critical evaluation of two related aspects that are currently differently appreciated within the research community: (i) permeability loss, and (ii) reactivity loss. These two aspects were well-known in 2007 [9,48,55,56,57] as the major limitations of the Fe⁰ PRBs technology. This critical review is rooted on the premise that the current confusion within the Fe⁰ research community results from the fact that the arguments of Noubactep [48,54] and others [46,58] are still largely overlooked within the research community. For this demonstration, recent publications by researchers from the two research groups that had introduced the technology are evaluated: (i) Paul G. Tratnyek from Oregon (Portland, OR, USA), and (ii) Robert W. Gillham from Waterloo (Ontario, OT, Canada).

While there is agreement on the fact that there is “reactivity loss” and permeability loss in Fe⁰ PRBs, there is a net discrepancy in the appreciation of the origins of these limitations and, thus, on ways to mitigate them. The discrepancy is, in turn, rooted in the significance of the electrochemical nature of aqueous iron corrosion for the process of contaminant removal in Fe⁰/H₂O systems. The majority of active researchers consider that some contaminants can exchange electrons with Fe⁰ (direct reaction, electrochemical mechanism), while a few authors [24,46,47] maintain that all reductive transformations within Fe⁰/H₂O systems are mediated by secondary reducing agents such as Fe^II species, Fe₃O₄, or H₂ (indirect reaction, chemical mechanism). A reader can be irritated by such a statement, but it suffices to read Whitney [59] to realize that the science of aqueous iron corrosion was distorted while introducing the Fe⁰ PRBs technology [24,60].

For the presentation herein, two papers by Dr. Gillham [26,61] and two papers by Dr. Tratnyek [4,14] are considered. From each article, the applied research approach is presented and the soundness of the conclusions is evaluated.

3. Fundamental Flaws in Investigating the Fe⁰/H₂O System

Batch and column experiments are conventionally used to investigate processes relevant for environmental remediation [33,62,63,64,65,66,67,68,69]. However, little guidance is available on: (i) how to perform them, (ii) how they should be set up, and (iii) even how their performance should be evaluated or monitored. The main guidance for investigating Fe⁰/H₂O systems has been the Interstate Technology and Regulatory Council [18,70]. This section covers the design of laboratory experiments to investigate the efficiency of Fe⁰/H₂O systems for water treatment. The presentation is limited to the two major aspects that have avoided or delayed progress in knowledge: (i) the use of vigorously homogenized batch experiments, and (ii) the used of pure Fe⁰ column systems (100% Fe⁰).

Fe⁰ + 2H⁺ ⇒ Fe²⁺ + H₂

(1)

Fe²⁺ + 3H₂O ⇒ FeOOH.H₂O + 3H⁺ + e⁻

(2)

Fe⁰ + RCl + H⁺ ⇒ RH + Fe²⁺ + Cl⁻

(3)

2Fe²⁺ + RCl + H⁺ ⇒ RH + 2Fe³⁺ + Cl⁻

(4)

The system to be investigated is made up of Fe⁰ that is electrochemically corroded It is fine like this by water or H⁺ (Equation (1)). For pH > 4.5, Fe²⁺ is further oxidized, and the resulting Fe³⁺ precipitates in the vicinity of Fe⁰ to form an oxide scale (oxide film). In such a system, no Fe⁰/H₂O interface exists, but rather, there are two interfaces: (i) Fe⁰/oxides, and (ii) oxides/H₂O (Figure 1) [24,35]. Clearly, under environmental conditions, the Fe⁰ surface is permanently shielded by a non-conductive oxide scale acting as: (i) a diffusion barrier for dissolved species, including contaminants and O₂, and (ii) a conduction barrier for electrons from the metal body. In other words, the reductive transformation of a chlorinated hydrocarbon (RCl) by Fe⁰ (Equation (3)) is not possible. Rather, when a reductive transformation is documented, it results from reactions similar to the one depicted in Equation (4) [24,34]. The proponents of the reductive transformation concept fail to explain or totally ignore how the contaminants and O₂ by-pass the oxide scale barrier. The question remains as to why the reductive transformation theory is still prevailing 15 years after it was radically refuted? [48,55]. Some answers are given below while considering the design of batch and column experiments. It should be kept in mind that knowledge of the kinetic behavior of Fe⁰ particles in soils and Fe⁰ filters is very limited [71]. Thus, one major problem in Fe⁰ remediation research is that used materials are not characterized for their (long-term) intrinsic reactivity. In essence, the suitability of all relevant materials (including nano Fe⁰) is based on the electrode potential of the Fe^II/Fe⁰ redox couple: E⁰ = –0.44 V [72,73,74]. Thus, documented differences are due to kinetic factors.

3.1. Batch Experiments

Data to characterize Fe⁰/H₂O systems in batch experiments are conventionally obtained using the bottle-point technique [7,64]. This experimental procedure was borrowed from adsorption studies in which the bottle-point technique is used to determine the adsorption capacity of inert adsorbents. No unified procedure for obtaining adsorption isotherms has been established [62,63,75]. Consequently, each researcher employs a different experimental procedure for the collection of isotherm data. The procedures differ with respect to: (i) buffer application, (ii) the duration of experiments, (iii) Fe⁰ particle size, (iv) Fe⁰ pre-treatment, (v) Fe⁰ type, (vi) shaking type, (vii) shaking intensity and mode, (viii) the volume of the experimental vessels used, (ix) the volume of the added solution, and (x) temperature. As a result, it is very difficult or even impossible to compare independent works from the literature, even for the same contaminants and the same Fe⁰ material and particle size [75]. The fact that achieved results are often modeled using the wrong reaction (Equation (3) instead of Equation (4)) explains why no efficient descriptor could be presented for the characterization of Fe⁰/H₂O systems [76,77].

The worst mistake in investigating Fe⁰/H₂O systems in batch experiments has been the use of vigorously shaken or homogenized experiments [33,78]. In fact, a key feature of the remediation Fe⁰/H₂O system is the stratification of Fe⁰/oxides/H₂O [48,55] and the existence of two different interfaces (Fe⁰/oxides and oxides/H₂O) [35,60]. While vigorously shaking or stirring their experimental vessels, researchers have produced possibly reproducible data with no practical significance [24,60,75]. This is because such vigorously shaken conditions do not occur in real environmental systems, such as permeable reactive barriers. Such vigorous shaking may also putatively interfere with and disrupt the oxide scale, a phenomenon which may not occur under real environmental conditions [48,55,74]. Moreover, such experimental designs are not able to address the question of whether electron transfer can be effective at pits or crevices as maintained by Huang et al. [4]. This is because everything is suspended and the generated Fe²⁺ ions have no opportunity to be further transformed in the vicinity of Fe⁰ (Equation (2)). To the best of the authors’ knowledge, Devlin et al. [79,80] were the first researchers to point out the necessity to keep Fe⁰ fixed in batch vessels. Noubactep et al. [49] went a step further to let the whole remediation process be strictly diffusion-controlled in quiescent batch experiments (no shaking). It is important to mention that the ITRC [18] has not mentioned this aspect at all.

In summary, the bulk of batch experiments performed for the past 30 years were not designed to answer the key research questions, in particular the mechanism of contaminant removal in Fe⁰/H₂O systems. The current corresponding author established these mechanisms 12 years ago [81,82], but it has been largely ignored (Section 4). In particular, the authors identified herein as leading scientists have not even dared to mention that any alternative idea exists [4,14,26,61]. A similar trend was noticed in recent overview articles [3,4,83,84]. Yet, the named alternative concept is rooted on the science of the system, and has been confirmed by historical findings as summarized by Baker [25] in his seminal reference textbook “The quest for pure water”, which has been re-edited a multiple times.

3.2. Column Experiments

Column experiments enable the investigation of the attenuation of specific species within a specific Fe⁰/H₂O system (Fe⁰ filter). The design parameters for Fe⁰ filters include: (i) Fe⁰ size and type, (ii) filter size, (iii) water flow velocity or residence time, (iv) water quality, including pH value, and (v) temperature. By adjusting these five operational parameters, an infinite number of different efficient Fe⁰ filters can be designed. This is equally the reason why published results are difficult to compare to each other. Sarr [85] roots the design of Fe⁰ filters in the stoichiometry of reactions similar to Equation (3). There is already a mistake in such an approach, as water is the sole relevant corroding agent (Equation (1)) [35,59,60]. However, the literature is full of examples of studies comparing the efficiency of Fe⁰/H₂O systems containing various amounts of Fe⁰ [86,87]. For example, O et al. [87] used 1364.2, 1700.6, 1846.4, and 2045.3 g of four different Fe⁰ specimens in an effort to characterize the impact of the initial iron corrosion rate on the long-term performance of Fe⁰ filters. In another example, Bi et al. [86] fixed the “Fe⁰ + sand” mass to 80 g and varied the Fe⁰ mass to investigate the impact of the Fe⁰: sand ratio on the hydraulic conductivity of Fe⁰ filters. The discussion of both series of results is not addressed herein. The question arises of how the efficiency of systems can be compared without reference to the mass balance [33].

Designing Fe⁰ filters starting from the mass balance is possible and appropriate, but the mass balance should be the one of Fe. As Fe⁰ is progressively transformed to its solid corrosion products (FeCPs), a volumetric expansion occurs that is related to the decrease in the density (

Table 1. Structure, formula, and some characteristics of metallic iron (Fe⁰) and its main corrosion products. Data are compiled from Balasubramaniam et al. [92] and Caré et al. [93]. bcc stands for body-centered cubic, and fcc stands for face-centered cubic. d_iron/d_oxide gives an indication of the extent of volumetric expansion, while the percent Fe gives an idea of the decrease in the electronic conductivity of oxides relative to Fe⁰.

Substance	Formula	Structure	Density (d)	dmetal/doxide	MW	Fe	O
		(-)	(g/cm3)	(-)	(g/mol)	(%)	(%)
Iron metal	Fe0	bcc	7.86	1.000	56	100.0	0.0
Wüstite	FeO	fcc	5.67	1.386	72	77.8	22.2
Magnetite	Fe3O4	Cubic	5.18	1.519	232	72.4	27.6
Hematite	α-Fe2O3	Trigonal	5.26	1.494	160	70.0	30.0
Maghemite	γ-Fe2O3	Cubic	4.69	1.676	160	70.0	30.0
Goethite	α-FeOOH	Orthorhombic	4.28	1.836	89	62.9	37.1
Akaganeite	β-FeOOH	Tetragonal	3.55	2.214	89	62.9	37.1
Lepidocrocite	γ-FeOOH	Orthorhombic	4.09	1.922	89	62.9	37.1
	δ-FeOOH	Hexagonal	n.a.	n.a.	89	62.9	37.1
FeOOH (Gel)	FeOOH	Cubic	n.a.	n.a.	89	62.9	37.1
FeII hydroxide	Fe(OH)2	Trigonal	3.40	2.312	90	62.2	37.8
FeIII hydroxide	Fe(OH)3	Trigonal	3.83	2.052	107	52.3	47.7

). The d_oxide/d_iron values in Table 1 suggest that oxides occupy volumes up to 2.3 times larger than the parent metal (Fe⁰). This means that permeability loss is inherent to Fe⁰ filters [88,89]. While rooting the design of Fe⁰ PRBs in the stoichiometry of electrochemical reactions (Equation (3)), it was considered that admixing inert materials such as sand to Fe⁰ represents a dilution of the reactive material (i.e., Fe⁰) with possible negative impacts on the contaminant removal capacity of the resulting Fe⁰ PRBs [86]. However, the expansive nature of aqueous iron corrosion implies that space for accommodating FeCPs must be foreseen and/or the proportion of expansive materials should be decreased. In other words, admixing Fe⁰ with non-expansive aggregates is even a prerequisite for sustainable Fe⁰ filters [88,89]. Clearly, only hybrid Fe⁰/aggregate systems are sustainable (Figure 2) [90]. This evidence is still not clear to many active researchers, mainly because the Fe⁰ PRB technology was demonstrated with a hybrid system (22% Fe⁰ w/w) [91], but the majority of the first-generation Fe⁰ PRBs were built with 100% Fe⁰. Even today, hybrid Fe⁰/aggregate systems are still designed without rationalizing their choice [84]. Such a state of affairs unfortunately implies that the results from most column experiments are not transferable to other systems. but are only valid for the specific experimental setup used [24,68].

Iron oxides have characteristic properties that make them useful for a wide range of applications, including environmental remediation [94,95]. They exist under three different main structures with oxygen ratios varying between 22% for wüstite and 37% for FeOOH (Table 1). The oxides with lower oxygen ratios (FeO, Fe₃O₄, and Fe₂O₃) are semi-conductors, while FeOOH and hydroxides are insulators. For example, hematite and magnetite have significantly different conductive characteristics despite having less than a 3% difference in oxygen content. Hematite has a corundum structure, while magnetite has an inverse spinel structure. Both oxides are n-type semi-conductors [95]. Hematite has a resistivity of over 10⁵ Ω cm, while magnetite has a lower resistivity of approximately 10⁻³ Ω cm. These two examples illustrate how the electrical properties of the oxide scale on Fe⁰ changes with the oxygen content. In essence, electron transfer from the metal (conductor) to FeO and Fe₃O₄ (semi-conductors) is still theoretically possible, but the transfer is already difficult with Fe₂O₃ and is hindered by insulators (FeOOH and hydroxides). This is the reason why: (i) no electron transfer from Fe⁰ to dissolved contaminants is possible, and (ii) batch experiments pertaining to designing Fe⁰ PRBs should be exclusively performed under quiescent conditions [48,55].

In summary, column experiments can provide excellent estimates of all relevant design parameters for field situations [9]. In addition, the obtained results are almost always directly transferable to water treatment at a field scale. The challenge for the future is to systematically test a range of hybrid Fe⁰/aggregate systems to derive a series of applicable designs for water treatment. Such experiments should last for several months, ideally not less than one year [96,97].

4. Disregarding the State-of-the-Art Knowledge

The issue of this communication is that the lack of true collaboration is the major challenge to the Fe⁰ remediation research community. This section illustrates this by presenting five selected recent publications on the removal of methylene blue (MB) in Fe⁰/H₂O systems [98,99,100,101,102]. These works were published more than three years after the MB method was presented in an overview article in the Journal of Environmental Management—Elsevier [103]. The MB method itself was first published a decade ago [104,105]. While the MB method has roots in the historical work of Mitchell et al. [106] and the evidence that iron oxides are poor adsorbents for MB [35], the five referenced papers [98,99,100,101,102] used MB as a model contaminant to characterize some aspects of the Fe⁰/H₂O system. The following section summarizes the studies, and highlights avoidable mistakes by these authors.

Table 2. Summary of some experimental conditions used for the experiments in the five articles. Experiments were conducted in batch using nano-size Fe⁰ (nZVI) modified in different ways (Table 3). “n.s.” stands for “not specified”, MB stands for methylene blue, and [MB] stands for its concentration, while “Nr.” is the number referencing individual articles in Table 3.

Nr.	[MB] (mg/L)	Fe0 Type	Fe0 Dosage (g/L)	pH Value	Duration	Volume (mL)	Mixing	Speed (rpm)	Ref.
1	500–2000	nZVI-LBC	0.5	3–11	5 min	20	Quiescent	0	[102]
2	20–120	nZVI and iron impregnated nanoclay.	0.4–2.8	3–13	20–140 min	50	Agitated	80	[100]
3	50–1000	PAC-mZVI	6	2–10	72 h	100	Stirred	120	[101]
4	10–40	nZVISLP	4–12	2–10	n.s.	100	Stirred	100	[99]
5	10–70	nZVI	2–14	2–12	10–120 min	250	Stirred	100–250	[98]

and

Table 3. Summary of the use of MB and the key findings for each of the five selected articles used to illustrate the evidence that the lack of true collaboration has been the major obstacle to Fe⁰ remediation research progress.

Nr.	MB Use	Key Findings
1	Model contaminant for the investigation of the adsorptive removal capacity of nano zero-valent supported-lemon-derived biochar (nZVI-LBC)	NZVI-LBC composite has an excellent adsorptive affinity for MB, making it a potential candidate for the aqueous removal of cationic contaminants.
2	Model contaminant for the investigation of the adsorptive removal capacity of nano zero-valent iron (nZVI) and iron impregnated nanoclay	Iron impregnated nanoclay comparatively has a better adsorptive removal capacity, making it a more suitable adsorbent for industrial applications of cationic dye removal.
3	Model contaminant to investigate the adsorption mechanism of palm-activated carbon metallic iron (PAC-mZVI) particles	The properties of PAC-mZVI and the structure of MB both have significant effects on the removal of MB.
4	Model contaminant for the investigation of the adsorptive removal characteristics of nano zero-valent iron synthesized from sweet lime pulp waste (nZVISLP)	nZVISLP has a great potential for MB adsorptive removal.
5	Model contaminant to investigate the adsorptive removal capacity of nano zero-valent iron (nZVI)	nZVI is an excellent absorbent for MB aqueous removal.

summarize the experimental conditions of the five articles. It is seen that the MB adsorptive capacities of nano-Fe⁰ and their modifications were investigated in batch studies. Table 3 also presents the key findings of individual studies. The MB initial concentrations varied from 10 to 2000 mg/L, and the Fe⁰ loading varied from 0.5 to 14 g/L. The initial pH value varied from 2.0 to 13.0, and the volume of the working solution varied from 20 to 250 mL. Several mixing speeds and intensities were tested, and the experimental duration, where specified, varied from 5 min to 72 h. The suitability of these operational parameters to meet the intended goal in individual studies has been discussed above (Section 3.1) [78,107]. It suffices to recognize that experimental results obtained under such variable conditions are highly qualitative. Moreover, the five papers have collectively ignored the MB method in their presentation. The fact that quantitative MB removal is reported in all these articles is a proof that co-precipitation is the main removal path in Fe⁰/H₂O systems. The following should be kept in mind: (i) MB adsorbs only very weakly onto iron oxides (FeCPs) [106], and (ii) MB discoloration cannot be achieved by any reductive reaction in the Fe⁰/H₂O system [108].

Table 2 reveals that Wang et al. [101] used the longest experimental duration (72 h or 3 d) and one of the highest mixing speeds (120 rpm) among the five papers. It is certain that Fe⁰ cannot be completely exhausted within 3 days. In other words, it is a mistake to use “adsorption capacity” in discussing the results. There are at least two reasons for this: (i) MB removal by iron oxides is not an adsorptive process, and (ii) no equilibrium state can be achieved in such a dynamic system before total Fe⁰ depletion. Clearly, once immersed in water, Fe⁰ continuously generates FeCPs, which progressively age within the system [104,109,110,111]. The next concern is that the corrosion rate (corrosion kinetics) is not known [97,112,113,114,115].

In summary, the novelty of the five papers is collectively questioned because they failed to consider the state-of-the-art knowledge while designing their experiments and discussing their results. Unfortunately, the five papers, which were selected for having investigated MB removal, are not an exception, but rather, represent the rule within the Fe⁰ remediation research community. The current most holistic understanding of the Fe⁰/H₂O system has been presented by our research group, but is currently ignored by the large majority of active researchers. In other words, an artificial knowledge gap exists, and within this virtual gap, a circular reasoning has already been established for three decades.

5. Scientists Disregard Progress

Using Fe⁰ for in-situ aquifer remediation has undergone rapid development during the past three decades [16,17]. Related applications offer the promise of a more thorough treatment of several classes of contaminants [1,12]. However, there is controversy in the scientific literature about the key issues determining the efficiency and sustainability of Fe⁰ filters (Section 3). A starting researcher, seeking the state-of-the-art knowledge on “remediation using Fe⁰” may identify research groups with the longest expertise [116]. Two such research groups are those of Dr. Paul G. Tratnyek (Oregon/USA) and Dr. Robert W. Gillham (Ontario/Canada). In this section, one research article and one overview article of each group is assessed with regard to the extent to which they have falsified the current state-of-the-art knowledge, which was achieved 11 years ago [81,82]. Interestingly, none of these four articles referenced a single paper out of several of them from Noubactep.

5.1. The Overview Articles

The overview book chapter single-authored by Dr. Gillham [5] retraced the history of “Fe⁰ for groundwater remediation” from the fortuitous observation of Glen Reynolds [117] to about 2007, when Dr. Noubactep first presented strong arguments against the “reductive transformation” concept in the peer-reviewed literature [48,55]. The idea was published before in the grey literature [118,119]. The wording of Dr. Gillham is reproduced here for the sake of clarity:

Glen Reynolds “was investigating the potential for sampling bias caused by sorption of chlorinated organic contaminants to well casings and other materials commonly used in groundwater sampling. From the results of laboratory batch tests, it was clear that contaminants were lost from solution through diffusion into polymer materials such as PVC and Teflon™. Unexpectedly however, losses were also observed from solutions in contact with a variety of metals, though the pattern of loss was inconsistent with a diffusion process. No literature was found that was directly relevant, however reductive dechlorination was proposed as the most likely cause [117]. Though initially observed in 1984, the potential significance of these early results with respect to groundwater remediation was not recognized until 1989. The results were shown to be reproducible, leading to the “laboratory testing and verification” stage of development.”

The mistake made by Dr. Gillham was an incomplete literature review for at least two reasons: (i) it was known since 1903 that, under ambient conditions, Fe⁰ is corroded by water and by water alone [59], and (ii) it is textbook knowledge that Fe⁰ is used in organic synthesis, for example, for the synthesis of aniline (Bechamp Reduction or Bechamp Process, 1854). It is interesting in this regard to state that, while rationalizing nitrobenzene reduction to aniline using Fe⁰ and acid (e.g., HCl), the direct transfer of electrons from Fe⁰ to nitrobenzene has never been convincingly established [120]. On the other hand, while using Fe⁰ in hydrometallurgy for the reductive dissolution of manganese ore (e.g., MnO₂) in acidic solutions, Bafghi et al. [121] also could not established direct electron transfer to MnO₂. Even while using Fe⁰ for wastewater treatment during the 1980s and 1990s, direct electron transfer to contaminants was not established [122,123,124]. In particular, Khudenko [124] used the cementation of Cu²⁺ by Fe⁰ as a tool to induce the reductive “destruction of organics”. Clearly, in the Khudenko process, Fe⁰ is oxidized by Cu²⁺ to produce Fe²⁺ and H_2, which chemically reduce organics. Whitney [59] and Khudenko [124] were not considered while introducing the reductive degradation concept for organics [7,125,126]. This was the beginning of a circular reasoning that cannot be corrected until electron transfer from Fe⁰ is taken out of the equation [24].

The overview article co-authored by Dr. Tratnyek [4] just perpetuated the named mistakes. The article is entitled “Fe(II) redox chemistry in the environment” and comprises a section on “Zero-Valent Iron” wherein the “prominent role of Fe⁰ in the environment” is presented. The following three major mistakes were extracted from Huang et al. [4]:

(i)

Acknowledging the non-stoichiometric nature of decontamination reactions while using pseudo-first-order rate constants (k_obs, k_M, and k_SA) or the corresponding half-lives (t_1/2) to describe the systems.

(ii)

Considering that any Fe⁰/H₂O interface exists, including within pits, crevices, or equivalent defects in the oxide scale on Fe⁰. Direct electron transfer from Fe⁰ to contaminants (electrochemical reduction) supposedly occurs at such sites. Because the formation of oxide scale is spontaneous, it also occurs at defects such that there is always an Fe⁰/oxides interface.

(iii)

Improperly considering the kinetics of iron corrosion, which is non-constant and non-linearly varies with time [90,114,115].

5.2. The Research Articles

The context of Cai et al. [14] is the “development of sulfidated Fe⁰ materials for environmental applications” as recently reviewed by Fan et al. [127]. This application is regarded as the convergence of two developments over the past three decades: (i) the diversification of the use of Fe⁰/H₂O systems for water treatment, and (ii) the optimization of the role of Fe_xS_y minerals in contaminant transformation processes [14,127]. It is certain that sulfidation of Fe⁰ increases the number of galvanic cells (Fe-S) in the metal body or at its surface and, thus, enhances the reactivity of Fe⁰. However, enhancing electron transfer within the metal body does not change the non-conductive nature of the oxide scale in the vicinity of Fe⁰. This fact was revealed to the research community as early as 2009 [128], but has been largely overlooked in the majority of published studies. In fact, according to Google Scholar, Noubactep [128] has been referenced only 33 times during the past 12 years. That is less than three times per year. The specific aims of Cai et al. [14] are not addressed herein, as they fall under the shortcoming of batch experiments, as discussed in Section 3. On the other hand, these authors could have investigated parallel systems for MB discoloration (MB method) to improve the discussion of their results (Section 4).

Jeen et al. [61] characterized the degradation of two chlorofluorocarbons (CFC11 and CFC113) by granular Fe⁰, Fe⁰/Ni⁰, and Fe⁰/Pd⁰ in flow-through column tests. The results revealed rapid and better degradation efficiencies in bimetallic systems (Fe⁰/Ni⁰ and Fe⁰/Pd⁰) relative to Fe⁰ for both compounds. However, for CFC113 the average pseudo-first-order rate constants were only about two to three times greater than for Fe⁰. These smaller-than-expected differences in degradation rate constants between Fe⁰ and its bimetallic counterparts was a clear hint that the underlying reaction mechanism may be unstable. Instead of checking this alternative explanation, Jeen et al. [61] suggested very complex explanations to cope with the reductive degradation concept. The approach of Jeen et al. [61] is rather characteristic within the Fe⁰ remediation research community.

6. Discussion

Fe⁰ PRBs have been used for groundwater remediation since the early 1990s [26]. It is a tangible fact that Fe⁰ oxidation by groundwater releases ferrous iron species (Fe^II) and hydrogen (H₂) and typically increases the pH value (i.e., H⁺ consumption) (Equation (1)) while lowering the oxidation/reduction potential (creating reducing conditions). The generated Fe^II species are further transformed to various Fe^II/Fe^III oxides and hydroxides, which are excellent contaminant scavengers. These modifications cause the reductive degradation of some parent chemicals (including contaminants), but ideally cause the removal of all species (contaminants and daughter products) by adsorption and co-precipitation. The contaminant scavengers account for the removal of non-reducible contaminants such as methylene blue (MB) and Zn²⁺. In contrast, proponents of the reductive transformation concept fail to explain how non-reducible contaminants are removed in Fe⁰/H₂O systems or consider that only such contaminants are removed by adsorption and co-precipitation. To date, the evaluation of the performance of Fe⁰ PRBs at field sites has focused on monitoring the following: (i) compliance with groundwater standards, (ii) permeability alterations, (iii) identifying the nature of mineral precipitates, (iv) the extent of Fe⁰ exhaustion, and (v) potential biologically mediated reactions [129,130,131,132]. The discussion of Fe⁰ longevity has been a function of amount and size, but has not yet considered the Fe⁰ ratio within the reactive zone. Unfortunately, without considering the Fe⁰ ratio no convincing discussion of the biological activity can be given [97,131,132,133,134,135]. Ideally, the reactivity and size of the admixing aggregates (e.g., FeS₂, MnO₂, and sand) should also be considered. In other words, existing guidance documents for the design of Fe⁰ PRBs (e.g., ref. [18]) have to be amended.

Apart from the MB method, no other quantification method to account for the extent of Fe⁰ “passivation” on the efficiency of Fe⁰ filters has been presented. On the other hand, the permeability of a Fe⁰ PRB depends both on Fe⁰ reactivity and the amount of in-situ generated iron minerals (FeCPs). However, the law of generation of these FeCPs is not known and its extent depends on the (residual) porosity and the pore size distribution [114,115]. Virtually, the porosity increases if the following three conditions are observed: (i) there is no inflow of suspended particles, (ii) there is no significant accumulation of removed contaminants, and (iii) there is no mineral precipitation. This idealized case corresponds to the situation where contaminants are present in trace amounts. However, at pH values larger than 4.5, mineral precipitation is certain and quantitative. Moreover, the density of individual FeCPs (d < 5 g/cm³) is lower compared to that of Fe⁰ (d = 7.86 g/cm³), meaning that each FeCP occupies a larger volume than Fe⁰ in the porous media. This implies that porosity decrease is the rule for field Fe⁰ PRBs, regardless of the material’s intrinsic reactivity. Consequently, to understand the mechanisms underlying porosity and permeability changes occurring in PRBs, the reasoning should start with the extent of Fe⁰ depletion. The problem with the available models is that they have considered iron corrosion by water as a side reaction, thereby neglecting the main pore filling factor. It is, thus, not surprising that no conclusive discussion was presented [89,97]. In one case, a time-dependent increase in the permeability of Fe⁰ PRBs was predicted [136].

Another possible reason accounting for the continued propagation of the reductive transformation concept despite its established flaws [81,82,137,138,139] is the bandwagon or ‘Matthew’ effect. The bandwagon or Matthew effect implies that a concept (even a flawed one) that has been widely publicized and cited may attract even more subsequent research attention compared to other competing theories. Overall, this creates bias in research with respect to research focus and the allocation of scarce resources, including research grants [140]. This is particularly important for early career researchers who may be desperate to be published or secure research grants. The Matthew effect, its impacts on research, and possible solutions have been discussed extensively in earth sciences [141,142] and environmental sciences [143,144]. The Fe⁰/H₂O remediation community need to develop active measures to guard against such phenomenon and root the field in the fundamentals of aqueous iron corrosion [24,145,146,147].

7. Questioning the Reputation of Journals

There is a current discussion on the meaning of journal ranking for what can be called the survival of science [140,148]. Reputed journals advertise with their rejection quotes [148]. Some journals publish submitted manuscripts only if the comments of at least 66% of the invited reviews are favorable. From the standpoint of the Fe⁰ literature, it is not exaggerated to assume that part of the rejected manuscripts, constituting up to 33%, could be scientifically sound. This is because, as in any field it is difficult, if not even impossible, to publish a paper with an alternative theory contrary to what is believed to be correct by the majority [140,148]. This is the case, even when the majority may believe in a flawed concept [149,150,151,152]. In fact, as discussed in Section 3, in the case of Fe⁰/H₂O systems, it has been very difficult to publish articles moving “against the tide” that considers Fe⁰ as a reducing agent [24]. The fundamental reason why Fe⁰ is suitable for environmental remediation is two-fold: (i) Fe^II and Fe^III species are stable under environmental conditions, and (ii) the structures of Fe⁰, Fe^II, and Fe^III precipitates (hydroxides and oxides) are considerably different (Table 1) [72,73,153,154]. In fact, Al⁰ (E⁰ = –1.676 V) and Zn⁰ (E⁰ = –0.762 V) are by far stronger reducing agents than Fe⁰ (E⁰ = –0.440 V) and have been reported to be initially more reactive for contaminant reductive transformations [8,28]. However, as corrosion proceeds, Al⁰ and Zn⁰ are progressively covered by impervious layers of hydroxide (Al(OH)₃) or oxide (ZnO), respectively. Conversely, Fe⁰ is just shielded by a porous layer of mixed oxides and hydroxides that are non-protective [153,154]. This knowledge was presented as a short communication in the peer-reviewed literature as early as 2010 [154], but has been largely ignored, with just 76 citations in more than 12 years (Google Scholar).

The Fe⁰/H₂O system is certainly a very complex one whose profound understanding requires knowledge from various branches of sciences, including chemistry, electrochemistry, materials science, mechanical engineering, and microbiology [155]. The first four peer-reviewed journal articles on Fe⁰/H₂O systems [5,6,7,8] appeared in reputed journals: Chemosphere [6,8], Environmental Science & Technology [7], and Ground Water [5]. Over the years, both good articles based on the fundamental science of aqueous iron corrosion [47,156] and poor articles perpetuating a flawed concept considering Fe⁰ as a reducing agent [12,14] were published in all relevant journals. This state of affairs questions the relevance of journal impact factors in evaluating scientists and grant proposals. In fact, a number of cases have been reported in recent years where the impact factors and citations of reputable journals have been manipulated by some members of the journal editorial teams [157,158,159,160]. It seems the right time to consider that each article has its own impact, and should be assessed based on its merit.

8. Conclusions

An increasing amount of attention is directed at documenting and explaining the association between Fe⁰ oxidative dissolution (iron corrosion) and contaminant removal in Fe⁰/H₂O systems, including Fe⁰ PRBs. This collective search for the rationale for efficient Fe⁰-based systems has revealed the key information about the sustainability of Fe⁰ PRBs. However, the proper interpretation of the available data is biased by the initial consideration that Fe⁰ is an environmental reducing agent. The biogeochemical complexity of the Fe⁰/H₂O system is obvious, but the fact that Fe⁰ reactivity declines with the service life of Fe⁰-based systems (reactivity loss) is also well-known. Therefore, prominent theoretical considerations should be revisited in an effort to determine the explanatory value of standard hypotheses and explore alternative explanations for tangible experimental and field observations. Furthermore, the theoretical perspectives presented herein deserve greater attention because they have already enabled the explanation of many reported controversies. In particular, the current three-fold uncertainty in designing Fe⁰ PRBs (mechanisms, permeability, and reactivity) is reduced to just one, since mechanisms and reactivity are known. Thus, only changes in the permeability are uncertain, but they are coupled to the following: (i) Fe⁰/aggregate ratio, (ii) Fe⁰ amount, (iii) Fe⁰ reactivity, (iv) Fe⁰ size, and (v) initial porosity. Conventional methodological tools to evaluate the performance of Fe⁰ filtration systems are limited in their efficacy, but at the same time, the limitations present good opportunities to explore new tools for progress in this line of research. Deciphering the sources of permeability loss has the potential to establish Fe⁰ filtration systems as a reference technology for water treatment at several scales. All that is needed is true collaborative research according to the rules of good science. The minimum requirement is the consideration of the state-of-the-art knowledge while presenting one’s own achievements or contributions.

The highly interdisciplinary nature of research on the Fe⁰/H₂O system and the abundance of related publications suggest that annual review articles in relevant journals (including Water/MDPI) can be a tool to offer early-career scientists the best assistance. Currently, it is very difficult to decide where to start, despite the large availability of review articles. Such an annual review can be written as a scoping communication focusing on the papers published within the calendar year (January/December).