Rescuing Fe0 remediation research from its systemic flaws

The well-accepted assumption of metallic iron (Fe0) acting as electron donor for environmental remediation has created an unstable domain of knowledge for the past 23 years. This assumption is discouraging some outstanding and prospective scientists from correctly interpreting their experimental results. Such a situation is a recipe for long-term decline. The critical situation cannot be solved with simplistic approaches. It is now imperative to develop an understanding to defend the difficulties of this assumption and re-orient Fe0 mediated remediation research as a whole. Correspondence to: Chicgoua Noubactep, Angewandte Geologie, Universität Göttingen, Goldschmidtstraße 3, D-37077, Göttingen, Germany, E-mail: cnoubac@gwdg.de


Introduction
The research on using metallic iron (Fe 0 ) for environmental remediation and water treatment is no more in a golden age. This opportunity can be used to address systemic flaws that have certainly threatened its development, and that were difficult to address some 10 years ago [1][2][3]. A central flaw is the long-held assumption that Fe 0 is a direct reducing agent for the chemical transformation of contaminants [4][5][6][7][8][9]. As a result, the large majority of published works is still considering Fe 0 as an environmental reducing agent [7,8,[10][11][12]. The demonstration that this assumption has been a popular fallacy is now ten years old [2]. In a hyper-competitive atmosphere for research funding, scientific progress is broken as some have received their grants funded possibly for the propagation of misconceptions. Others have seen their grants rejected because they are dancing outside the circle. Thus, promising scientific careers are continued to be misdirected.
Despite such progress, it is remarkable that no bright future is generally expected from Fe 0 filters [39][40][41][42]. Some systems have been almost abandoned [25,43] while even the proven efficient SONO filters [26,28] have not yet reached the expected large-scale distribution. The balance between 'expectations and achievements' are left out from the pure scientific perspective [44,45]. As an example, the very first Fe 0 wall containing only 22% Fe 0 (w/w) [14] is commonly referenced as a success story [10,16]. However, the majority of available Fe 0 walls contain larger Fe 0 proportions [21,22,24]. The rationale for their reported sustainability does not seem to have been discussed yet. Based on these observations and evidences that alternative views have been constantly refuted [8,[46][47][48][49][50] or belittled [51,52]. It is fair to argue that the Fe 0 remediation research has progressed on an unsustainable path [53]. In the light of these findings, an attempt has been made in this communication to describe how this situation arose and proposes/ reiterates some possible remedies. Critical analysis of the published literature revealed that the writing style has been given more attention and weightage than the scientific content.

The reductive transformation concepts
The idea that electrons from Fe 0 mediate contaminant reductive transformations was adopted from a seminal work by Matheson and Tratnyek [4]. Although controversial views were published during 1994 and 1995 [13,71,72] Weber [6] validated and generalized the results of Matheson and Tratnyek [4]. Despite this pseudo-scientific demonstration supported by further reports [5,73,74], O'Hannessin and Gillham [75] acknowledged that the reductive transformation concept was a 'broad consensus'. Recently, Ebelle et al. [76] presented a detailed discussion of Matheson and Tratnyek [4]. It is clearly demonstrated that the conclusions of Matheson and Tratnyek [4] were not supported by own results. The question arises why could such a fallacy be introduced in the scientific literature? The evidence that the whole research community has ignored ancient works on using Fe 0 for safe drinking water provision [31- 35,77,78] is not addressed herein. Using metallic elements (including Al 0 , Fe 0 and Zn 0 ) for water treatment was also common place during the 1980s [79][80][81][82][83].

The propagation of the misconception
The Fe 0 remediation technology was born in North America (Canada, USA) and 'migrated' to Australia, Europe and the rest of the word over the time [18,42,[84][85][86][87][88][89]. The American colleagues were considered the absolute experts of the 'new born'. For example, the Fe 0 remediation technology was introduced in Germany by a literature review [90] presenting the technology as an American innovation. The original literature review was improved two years later to account for novelties mostly based on not readily accessible technical documents [91]. The introduction of Fe 0 walls in Italy and Belgium followed a similar path [42,84,85,87,89,92,93]. The mistakes transported in this trial have largely been identified and discussed [55,70,[94][95][96][97][98][99][100] and will not be duplicated herein. It is sufficient to recall two aspects: (i) no initial critical review of the related scientific literature was performed and (ii) the iron mass balance was fully ignored. Interestingly, a vertical look through the list of authors reveals that many of them are members of editorial boards of authoritative journals. Authors and editors must endorse now the responsibility of the propagated misconception [101]. The next section illustrates the extent of confusion using the Fe 0 /Se system [102].

Lessons from the Fe 0 /Se system
The efficiency of Fe 0 filings to remove Se from agricultural drainage water was demonstrated during the 1980s [80,82,103]. The operating mode of Fe 0 filters for Se removal was elucidated by Anderson [82]. Accordingly, Fe 0 serves as generator of contaminant collectors and Se VI , Se IV and Se 0 are all adsorbed and/or occluded (co-precipitation) in the matrix of iron corrosion products (FeCPs). Fe 0 filters were efficient but not sustainable because iron particles were soon cemented to each other and the initial porosity of the filters was filled by in-situ generated FeCPs (yielding permeability loss). In other words, Harza Fe 0 filters were efficient but not sustainable.
With the advent of Fe 0 reactive walls (permeable reactive barriers), the Fe 0 /Se system was 'rediscovered' and (mostly independently) intensively investigated over the past 17 years [15,40,102,103,[104][105][106][107][108][109][110][111][112][113]. Se VI was tested as redox-sensitive species to be reductively precipitated by Fe 0 . This effort was supported by an impressive number of modern analytical tools, using both bulk and surface methods [15,109,110,114,115]. Related studies tested several types of Fe 0 materials including: (i) iron filings (IF), (ii) sponge iron and (iii) steel wool (SW). Available results can be summarized as follows: (i) Fe 0 reduction can significantly decrease selenium concentrations at laboratory scale, (ii) Fe 0 filters have successfully treated Se in full-scale operations at some sites, (iii) At some sites neither IF (granular Fe 0 ) nor SW (fibrous Fe 0 ) was efficient to meet regulatory requirements, (iv) at other sites SW were successful were IF failed and (v) no clear trend between the residence time in the filters and the efficiency could be established.
The example of the Fe 0 /Se system illustrates the extent of confusion within the Fe 0 research community, despite available powerful analytical tools [38,53]. This communication advocates that the confusion is primarily due to the false assumption that Fe 0 is a reducing agent. It appears that considering Fe 0 as generator of Se collectors solves reported discrepancies and enable the design of (efficient and) sustainable filters.
In fact, what is the long-term kinetics of Fe 0 corrosion? Which Fe 0 materials have the (long-term) ability to produce enough FeCPs for the collection of Se from the polluted water of concern? Answering these questions starts with two points: (i) avoiding pure Fe 0 systems (columns with a zone containing 100 % Fe 0 ) like in the Harza Process and (ii) characterizing the intrinsic reactivity of individual materials [38]. If such a systematic approach is broadly adopted, the modified Harza Process will be sustainable. Moreover, because co-precipitation and size-exclusion are primarily non (or less) selective mechanisms, Fe 0 filters do not only address species with high affinities to FeCPS [59]. This is the rationale for making Fe 0 filters as universal devices for safe drinking water and sanitation [38,69].

The assessment of the peer review system questioned
As it is now established that systematic flaws were governing research on Fe 0 for environmental remediation [35,44,45,53,55,[57][58][59][60][61], there have been some injustice within the research community. In particular, the demands for research money is much larger than the supply. This hyper-competition is supposed to retain just the best scientific workforce for the limited funds available. However, in a context where the expertise of the referees is questioned, it appears that the best scientific workforce is rather punished for its clairvoyance. The great majority of environmental research is conducted by PhD students and postdoctoral fellows. But as a rule, the grants are obtained by more experienced scientists, including professors. In another phrase for two decades (5 generations of PhD student or 10 generations of postdoctoral fellows), experienced scientists working on Fe 0 for environmental remediation have been the 'blind guide of the blind'. Consequently, a growing number of PhDs is graduating within a knowledge system inhibiting their creativity. In this vicious circle, they are performing a contra-productive job without advantage for the taxpayers' investment in their lengthy career of education. The current knowledge system on Fe 0 remediation is a perpetual danger for the integrity of science.

Undermining the scientific knowledge-building process
Science is a collective activity aiming at building a reliable body of knowledge about the physical world. This goal can only be achieved if all scientists are honest and dedicated to the service of their respective nations. In this common project, there is no place for consensus and frauds. Admittedly, some consensus could be temporally met but just in the sake of more clarity. In other words, scientists who are committed to the task must take serious conscious steps to ensure that no falsehoods are introduced in the common knowledge database.
It is difficult to understand why scientists are undermining their own art. Hundreds of articles contain expressions like 'to the best of the author's knowledge' in the introduction to rationalize the investigation of things that are known for more than one century [33][34][35]77,78]. Moreover, several authors have put various arguments forward to maintain that Fe 0 is an environmental reducing agent despite controversial views [8,[46][47][48][49][50][51][52]. This attitude runs the risk of derailing the project of establishing the science of Fe 0 remediation. Scientific knowledge-building should remain the paramount goal, whether it is accompanied by scientific scorekeeping or not.

Missing basic scientific rules
This section analytically describes the Fe 0 /H 2 O system while ignoring primarily, intentionally achievements in the framework of 'Fe 0 remediation'. The reader is then confronted with the dilemma of the 'Fe 0 remediation' research.

An analysis of the Fe 0 /H 2 O system
A piece of reactive Fe 0 immersed in an aqueous solution undergoes spontaneous oxidative dissolution (aqueous iron corrosion) according to Eq. 1 (69,70): Under anoxic conditions, Fe(OH) 2 is polymerized and precipitated as hydroxide which is mostly further transformed to magnetite (Fe 3 O 4 or FeO.Fe 2 O 3 ). A multi-layered oxide scale is formed on the Fe 0 surface and acts as physical barrier for the transport of reactive species, from and to the Fe 0 surface [1][2][3]117,119]. This is the rationale for the universally reported decreased kinetics of iron corrosion by the oxide scale [44,117,118]. It is essential to recall that under natural conditions (no external current), no direct Fe 0 oxidation to Fe III is possible. Moreover, Fe III species are constantly present within anoxic Fe 0 /H 2 O systems and their formation, rationalized by several mechanisms, is not addressed herein [119,120].
Under oxic conditions (presence of oxidizing agents including O 2 ), Eq. 1 is still the main reaction for aqueous iron corrosion [121]. Its reaction kinetics is accelerated by the consumption of Fe II species (LeChatelier Principle) which are further oxidized to less soluble Fe III species (mostly hydroxides and oxides). In another phrase, species reductive transformation in a Fe 0 /H 2 O system results from a chemical reaction between Fe II and H species (indirect reduction) and not from an electrochemical reaction (electrons from Fe 0 , direct reduction). Thus, reaction as per Eq. 2 is impossible under environmental conditions. RCl stands for a halogenated chlorinated hydrocarbon.
Three other important features of the Fe 0 /H 2 O system are: (i) the volumetric expansive nature of iron corrosion [122], (ii) the fact that iron hydroxides precipitate in the presence of foreign species (coprecipitation) [123,124], and (iii) the evidence that iron hydroxides and oxides are good adsorbents for a wide range of dissolved species [29,125,126]. These three features act in synergy for the processes of aqueous contaminant removal in the presence of Fe 0 . In particular, in Fe 0 -based packed beds (Fe 0 filters), volumetric expansion implies that porosity and thus permeability loss occurs with increasing operational duration. The progress of iron corrosion (corrosion rate) depends mainly on the porosity and stability of the oxide scale on Fe 0 [117,118]. Thus, the importance of solution chemistry on the evolution of the oxide scale should be systematically characterized [38].

Separation processes in Fe 0 Filters
Akageneite, amorphous iron hydroxides, goethite, green rusts, lepidocrocite, and magnetite are some common identified solid iron corrosion products from field and laboratory studies using Fe 0 [21,22,[127][128][129][130][131]. These minerals possess little or no permanent surface charge, but variable surface charge is generated through the adsorption of protons to surface hydroxyl groups [20]. For the near neutral pH range, it can be considered that iron oxides/hydroxides exhibit a positively charged surface [132]. This implies that Fe 0 shielded with ion oxides cannot directly influence the transport of dissolved species to its surface. In other words, no matter the chemical reactivity of Fe 0 towards a species (difference between the relative standard electrode potentials), the extent of its effective transformation depends on the species affinity to the oxide scale [59,132,133]: adsorption on the outer surface and diffusion through the film. This basic argument makes on the contaminant removal process makes unstable any discussion not properly considering the oxide scale as a physical barrier [1][2][3]35,44,45,[55][56][57][58][59][60].
As an example, nitrate is considered a nonspecific ion that participate only in electrostatic ion exchange at the oxide surface, and do not undergo chemisorption through inner-sphere attachment [20,134]. This assumption is valid but must be completed by the size of NO 3 and the diffusion potential across the oxide scale (porosity and tortuosity of the oxide scale). In other words, although the adsorption of some species (e.g. oxyanions) on iron oxides has been properly considered, diffusion behavior across the oxide scale has not been thoroughly examined. However, only such a holistic approach could enable the establishment of the science of Fe 0 remediation. Table 1 summarizes the key processes mediating contaminant removal in Fe 0 filters. It is seen that a Fe 0 filter acts as a media filter for the separation of particles, ideally only in-situ generated ones (iron oxides and hydroxides), but also as a chromatographic system for the separation of molecules. Iron oxides and hydroxides are generated in the presence of trace amounts of contaminants. Thus, contaminants are physically enmeshed in the matrix of these iron minerals (coprecipitation). Iron minerals can be regarded as contaminant collectors in Fe 0 filters [31, 32,135]. Lastly, because the precipitated iron minerals progressively fill the pore space within the filter, permeability loss occurs but contaminant removal by size-exclusion is increased [70,[94][95][96][97].
Finally, two mechanisms should be shortly addressed here (precipitation and reduction) for an improved understanding. Both are quantitative at relatively high contaminant concentrations. Water treatment with Fe 0 filters typically implies the removal of contamination at µg/L-level from large volume of waters. Under these conditions, neither precipitation nor reduction could be quantitative [136]. Moreover, the trace amounts of contaminants are immersed in an 'ocean' of iron hydroxides and oxides (oxide scale) such that the opportunity of a quantitative chemical reaction is not given [1][2][3]. For these reason chemical precipitation and reduction in the aqueous phase are not relevant in the context of environmental remediation and safe drinking water provision. In other words, the suitability of Fe 0 for environmental remediation arises from its capacity to generate iron minerals which acts as contaminant collectors [26,82,69,135,136]. This knowledge was used by Tseng et al. [137] to concentrate 60 Co from sea water and by James et al. [138] to remove phosphate from wastewater. As discussed herein, contaminant reduction, when it occurs is mediated by iron corrosion products (e.g. Fe II , H/H 2 , green rusts, magnetite) and is not a relevant removal mechanism for any contaminant [70,95].
The analysis of the Fe 0 /H 2 O system has demonstrated that this popular state-of-the-art knowledge on the operating mode of Fe 0 filters is non-acceptable. In particular, because Eq. 1 is universal, all corrosion products are present in every Fe 0 /H 2 O system. Accordingly, even where reduction is believed to play a key role, adsorbing agents are abundantly available [31,32,82,130,131,139] and have enmeshed some contaminant during their precipitation [1][2][3]56,59,61,62,67,70,95]. Moreover, screening (Table 1) is improved by porosity loss and contaminant removal by size-exclusion is not avoidable. Given that this knowledge is present in the peer-reviewed scientific literature since 2007 [2], it is difficult to understand why the reductive transformation concept is still prevailing [8]. Perhaps changes are more painful to some contemporary scientists than in the days of Carl Sagan. However, the objectivity is the main characteristic of science [148]. A view can be expressed agressively, freshly, harshly or politely in one hand while it should remain objective on the other.

Should scientists learn from politicians?
In 1987, Carl Sagan stated: "In science it often happens that scientists say, 'You know that's a good argument; my position is mistaken,' and then they would actually change their minds and you never hear that old view from them again. They really do it. It doesn't happen as often as it should, because scientists are human, and change is sometimes painful. But it happens every day". The continuation of the citation reads as: "I cannot recall the last time something like that happened in politics or religion". Would researchers on Fe 0 remediation agree with Carl Sagan?
It seems that nowadays, it is the scientist who must learn from politicians. In democratic systems, any politician confronted with any single behaviour contrary to the constitution (the fundamental law of politics) should resign. Accordingly, any scientist confronted with a behaviour contrary to Chemical Thermodynamics (a fundamental law of chemistry) should rectify his understanding. If such an approach was universally adopted, the systematic flaw addressed herein would have not (i) been introduced and (ii) maintained for two decades. The situation is worsen by the evidence that the Fe 0 filtration technology was introduced by distorting the stream of corrosion science and overseeing many other aspects of General Chemistry. In particular, the favoured reductive transformation concept is even not valid in the concentration range of concern (µg/L) [38,53,56,57]. Thus a paradigm shift of the current understanding for Fe 0 filtration technology is very much demanding before the future researchers are dragged deeper into confusion [95].

Conclusions
The Fe 0 remediation research community cannot continue to ignore the warning signs of a knowledge system under great stress. There is even a risk for incipient decline for an already proven efficient technology. It is certain that the tax payer will continue its strong support for a cleaner environment. It is also certain that desirable larger budgets for this aim are possible and defensible. However, for a purposeful use of such budgets, structural flaws within the system should be addressed in a systematic way. There is evidence that the established scientific approaches/methods can correct its vulnerabilities. Some fundamental changes are required to bring the Fe 0 research community back on the highway of iron corrosion science. These changes need to be made in a comprehensive fashion. An immediate paradigm shift is called for considering the worsening of current situation during the last two decades. Widespread engagement is necessary to bring about this paradigm shift, beginning with immediate debate, strong advocacy for change, and action by individual scientists, the funding agencies, academic institutions, and other entities that control and pay for the conduct of science [149].
The envisioned future world of Fe 0 researchers is not more or less talented than the current. It does not need more or less financial support as well. It will just perform better investigations based on the science of aqueous iron corrosion [117,118]. In this perspective, even small laboratories will restore an environment in which talented trainees and scientists can achieve excellent results within some years [38,59,150]. The immediate goal of this communication has been to stimulate the debate on important issues that concern the future of a potentially universal frugal technology, eventually based on a century old knowledge [31, 32,33,34,35,77,78]. This task cannot be left to a self-appointed subset of scientists [56] or to the leaders of funding agencies. Therefore, academic institutions, scientific societies, funding organizations, and other interested parties are encouraged to organize discussions (at regional, national and international levels) with a wide range of relevant constituencies [149]. Some discussions of this type have already begun in countries like Cameroon, Germany, India, Lebanon, Romania and Tanzania [8,55,[150][151][152][153][154][155][156][157][158][159]. However, critical actions are needed on several fronts by many parties to re-orient research on Fe 0 for water treatment. No less than the future credibility of natural science is at stake.