A Novel and Facile Method to Characterize the Suitability of Metallic Iron for Water Treatment

: Metallic iron (Fe 0 ) materials have been industrially used for water treatment since the 1850s. There are still many fundamental challenges in a ﬀ ordably and reliably characterizing the Fe 0 intrinsic reactivity. From the available methods, the one using Fe 0 dissolution in ethylenediaminetetraacetic acid (EDTA—2 mM) was demonstrated the most applicable as it uses only four a ﬀ ordable chemicals: Ascorbic acid, an ascorbate salt, EDTA and 1,10-Phenanthroline (Phen). A careful look at these chemicals reveals that EDTA and Phen are complexing agents for dissolved iron species. Fe 3 -EDTA is very stable and di ﬃ cult to destabilize; ascorbic acid is one of the few appropriate reducing agents, therefore. On the other hand, the Fe 2 -Phen complex is so stable that oxidation by dissolved O 2 is not possible. This article positively tests Fe 0 (0.1 g) dissolution in 2 mM Phen (50 mL) as a characterization tool for the intrinsic reactivity, using 9 commercial steel wool (Fe 0 SW) specimens as probe materials. The results are compared with those obtained by the EDTA method. The apparent iron dissolution rate in EDTA (k EDTA ) and in Phen (k Phen ) were such that 0.53 ≤ k EDTA ( µ g h − 1 ) ≤ 4.81 and 0.07 ≤ k Phen ( µ g h − 1 ) ≤ 1.30. Higher k EDTA values, relative to k Phen , are a reﬂection of disturbing Fe 3 species originating from Fe 2 oxidation by dissolved O 2 and dissolution of iron corrosion products. It appears that the Phen method considers only the forward dissolution of Fe 0 . The Phen method is reliable and represents the most a ﬀ ordable approach for characterizing the suitability of Fe 0 for water treatment.


Introduction
Inadequate access to safe drinking water (and sanitation) is the most pervasive problem currently afflicting the developing world [1][2][3][4][5][6]. Moreover, water scarcity is expected to grow worse in the coming decades, even in regions currently considered water-rich [1,7]. Therefore, new robust, efficient, affordable, and applicable water treatment technologies are urgently needed [1,7,8]. Ideally, such innovative technologies should be simple to operate and maintain, be able to function without electricity and be based on local resources and skills [2,4,5]. Filtration on metallic iron (Fe 0 ) beds has been discussed in the literature as such an applicable technology for universal access to safe drinking water to low-income communities [9,10].
The purpose of this study was to develop an experimental method to characterize the intrinsic reactivity of Fe 0 SW using their initial dissolution in 1,10-Phenanthroline (ortho-Phenanthroline or Phen). Nine (9) commercial Fe 0 SW were used. The selection of Phen was justified by the extreme stability of the Fe 2 -Phen complex [53,55]: (1) avoiding any further oxidation of Fe 2 from Fe 0 dissolution and (2) non-addressing any Fe 3 species already available in the system [52]. Parallel experiments using EDTA were performed and the results are comparatively discussed.

Fe 0 Materials
A total of nine (9) locally available commercial steel wool (Fe 0 SW) specimens were purchased from East Africa Steel Wool (Ind.) Limited in Dar Es Salaam (Tanzania) and used. The tested 9 Fe 0 SW covered all grades of steel wools. Table 1 summarizes the main characteristics of the materials together with their elemental composition. Table 1. Major characteristics of tested metallic iron (Fe 0 ) steel wool (SW) specimens. The elemental composition (%) was determined in this study. SW1 to SW8 originated from China and SW9 from Tanzania. SW1 to SW8 were brand new, direct from the manufacturer, while SW9 was bought on the market in Tengeru (Tanzania). SW9 was thus air oxidized to some extent. The n.d. stands for not detected. The procedure presented by [57] was used. A solution containing HNO 3 (70%), HCl (35%) and de-ionized water was prepared at a 1:1:1 ratio by volume. Analytes were prepared by adding 12 mL of the mixed acid solution to 0.5 g of Fe 0 SW in a beaker. The beaker was covered with a watch glass, and after 10 min pre-digestion at room temperature (23 ± 2 • C), the solution was gradually heated on a hotplate for approximately 30 min to 200 • C. After cooling for about 15 min, the sample digest was made up to 100 mL with de-ionized water, and an aliquot was transferred to a 50 mL auto-sampler vial for analysis. Then, 1 mL of the sample was diluted with de-ionized water to a mark in a 250 mL volumetric flask prior to Inductively Coupled Plasma (ICP) analysis.

Iron Dissolution
Iron dissolution studies were accomplished by using 0.10 g of each Fe 0 SW in 50 mL of the complexing agent (2 mM Phen, 6 mM Phen, and 2 mM EDTA) for up to 5 days (120 h). The experiments were performed with the conventional quiescent experimental protocol for the EDTA test [45]. Tested SW specimens were cut into 2 to 3 cm before weighing and submerged in 50 mL graduated plastic cell-star tubed (narrow beakers) containing EDTA or O-Phen. All experiments were performed in triplicates, protected from direct sunlight.

Analytical Method
The aqueous iron concentration was determined with a Rayleigh UV/VIS Spectrophotometer (Beijing Beifen-Ruili Analytical Instruments (group) Co., Ltd., Beijing, China), at a wavelength of 510 nm using a 1.0 cm cuvette. The instrument was calibrated for iron concentration ≤10 mg L −1 . An ascorbate buffer was used to reduce Fe 3 -EDTA before adding Phen for Fe 2 complexation before Fe determination. Samples from the Phen experiments were treated in the same way as the same calibration curve was used [21].

Expression of Results
The basic idea of using complexing agents to characterize the intrinsic reactivity of Fe 0 materials is that iron dissolution (from Fe 0 and FeCPs) is initially a linear function of the time [45]. Thus, for a certain time frame (t 1 > t 0 ) after the initiation of the experiment (t 0 ) the total iron concentration ([Fe] t ) as defined in Equation (1) is a linear function (of time). [ Therefore, the goal of this study was to identify the time frame for the linearity of Equation (1). The regression coefficients (a and b) are characteristic for each Fe 0 specimen. In fact, "a" is the rate of Fe dissolution from Fe 0 while "b" gives an estimation of the amount of FeCPs on the material or the fraction thereof that is dissolved by the used complexing agent (here EDTA and Phen). For the EDTA and Phen dissolution experiments, the corresponding values are termed as k EDTA and k Phen , respectively. Herein, k EDTA and k Phen values were obtained from Microsoft Excel 2017 (Microsoft, Redmond, WA, USA).

Results and Discussion
The results of elemental analysis are presented in Table 1. It is seen that all 9 specimens are made up of at least 98.4 % Fe. Accordingly, differences in intrinsic reactivity are due to the manufacturing process (not accessible) and to the distribution of the remaining 1.4% alloying elements. Relevant elements are Co, Cr, Cu, Ni, and Pb (Table 1). For both Grade 00 specimens (very fine: SW3 and SW9) differences can also be attributed to the surface oxidation state.

Appropriateness of the Experimental Approach
In the presence of Phen, Fe 2+ is complexed to form very stable Fe-Phen (Equation (3)) which blocks the Fe 2+ oxidation by dissolved oxygen (Equation (4)). In the presence of EDTA, Fe 2+ oxidation to Fe 3+ (Equation (4)) is rather accelerated because (Fe-EDTA) 3+ is more stable than (Fe-EDTA) 2+ [53,55,58]. Resulting Fe 3+ is then complexed by EDTA (Equation (5)). It should be recalled that (Fe-EDTA) 3+ is not destabilized by many inorganic reducing agents, including hydroxylamine. Hydroxylamine is used in the standard spectrophotometric method to reduce Fe 3 species prior to the formation of the characteristic orange complex of Fe 2 -Phen [59]. Lastly, amorphous Fe 3 corrosion products are (at least partly) dissolved in EDTA (Equation (6)). Amorphous Fe 2 corrosion products, if available, would dissolve both in EDTA and Phen [52].
The discussion until now clearly shows that the Phen method was (1) not sensible to dissolved oxygen and (2) less sensible to the presence of atmospheric corrosion products on Fe 0 than the EDTA method. Thus, the Phen method was free from an inherent limitation of the EDTA method making its application for characterizing (1) very reactive materials and (2) fine Fe 0 specimens containing large amounts of atmospheric corrosion products [20,51]. Hildebrandt [51] recently presented an original approach to address the second challenge. This consisted of reducing the mass of Fe 0 SW from 0.1 g to 0.01 g in 50 mL of the EDTA solution. Even then, a linear relationship between the elapsed time and the dissolved iron concentration (Equation (1)) was difficult to obtain In the presence of Phen, Fe 2+ is complexed to form very stable Fe-Phen (Equation (3)) which blocks the Fe 2+ oxidation by dissolved oxygen (Equation (4)). In the presence of EDTA, Fe 2+ oxidation to Fe 3+ (Equation (4)) is rather accelerated because (Fe-EDTA) 3+ is more stable than (Fe-EDTA) 2+ [53,55,58]. Resulting Fe 3+ is then complexed by EDTA (Equation (5)). It should be recalled that (Fe-EDTA) 3+ is not destabilized by many inorganic reducing agents, including hydroxylamine. Hydroxylamine is used in the standard spectrophotometric method to reduce Fe 3 species prior to the formation of the characteristic orange complex of Fe 2 -Phen [59]. Lastly, amorphous Fe 3 corrosion products are (at least partly) dissolved in EDTA (Equation (6)). Amorphous Fe 2 corrosion products, if available, would dissolve both in EDTA and Phen [52].
The discussion until now clearly shows that the Phen method was (1) not sensible to dissolved oxygen and (2) less sensible to the presence of atmospheric corrosion products on Fe 0 than the EDTA method. Thus, the Phen method was free from an inherent limitation of the EDTA method making its application for characterizing (1) very reactive materials and (2) fine Fe 0 specimens containing large amounts of atmospheric corrosion products [20,51]. Hildebrandt [51] recently presented an original approach to address the second challenge. This consisted of reducing the mass of Fe 0 SW from 0.1 g to 0.01 g in 50 mL of the EDTA solution. Even then, a linear relationship between the elapsed time and the dissolved iron concentration (Equation (1)) was difficult to obtain [51]. The perfect linearity observed for Phen in Figure 1 confirms that using the Phen method practically only Fe 0 oxidative dissolution was characterized [60]. Hu et al. [52] recently presented the first application of the Phen test for granular Fe 0 materials. Figure 2 compares changes of iron concentration as a function of time for the 9 tested Fe 0 SW in 2 mM Phen for up to 120 h. For all materials, Fe values increased progressively with the experimental duration, even after 120 h of contact. However, a linear increase was observed only during the first 24 h (1 day). In the Fe 0 /Phen/H 2 O system, there is no (porous) oxide scale which could justify slow mass transfer into the bulk solution. The solution was also not saturated (Section 3.1). Hence, at longer experimental times, the data indicated the occurrence of a chemical reaction that was slow and specific to each Fe 0 SW specimen. This is the first time that the differential kinetics of Fe 0 oxidative dissolution was clear evidence in short-term quiescent batch experiments.

Evidencing the Non-Linear Kinetic of Iron Corrosion
Water 2019, 11, x 6 of 14 [51]. The perfect linearity observed for Phen in Figure 1 confirms that using the Phen method practically only Fe 0 oxidative dissolution was characterized [60]. Hu et al. [52] recently presented the first application of the Phen test for granular Fe 0 materials. Figure 2 compares changes of iron concentration as a function of time for the 9 tested Fe 0 SW in 2 mM Phen for up to 120 hours. For all materials, Fe values increased progressively with the experimental duration, even after 120 hours of contact. However, a linear increase was observed only during the first 24 hours (1 day). In the Fe 0 /Phen/H2O system, there is no (porous) oxide scale which could justify slow mass transfer into the bulk solution. The solution was also not saturated (Section 3.1.). Hence, at longer experimental times, the data indicated the occurrence of a chemical reaction that was slow and specific to each Fe 0 SW specimen. This is the first time that the differential kinetics of Fe 0 oxidative dissolution was clear evidence in short-term quiescent batch experiments. Compared to granular Fe 0 used in EDTA tests [20], filamentous Fe 0 SW was more reactive, but the same mass (0.1 g) could not achieve iron saturation (112 mg L −1 ). It is recalled that the EDTA test was stopped after 4 days because saturation was approached for more reactive granular samples [45]. As discussed above, dissolution of atmospheric corrosion products also occurred in EDTA. The data in Figure 2 show that the most reactive Fe 0 SW used herein could not dissolve 30 mg L −1 of iron in 2 mM Phen after 120 days. The reason has been that only the slow forward dissolution of Fe 0 Compared to granular Fe 0 used in EDTA tests [20], filamentous Fe 0 SW was more reactive, but the same mass (0.1 g) could not achieve iron saturation (112 mg L −1 ). It is recalled that the EDTA test was stopped after 4 days because saturation was approached for more reactive granular samples [45]. As discussed above, dissolution of atmospheric corrosion products also occurred in EDTA. The data in Figure 2 show that the most reactive Fe 0 SW used herein could not dissolve 30 mg L −1 of iron in 2 mM Phen after 120 days. The reason has been that only the slow forward dissolution of Fe 0 was addressed. This evidence makes the Phen test more reliable that the EDTA test. In addition to using fewer chemicals than the EDTA test, the Phen test was also more affordable [52,60].

Evidencing the Non-Linear Kinetic of Iron Corrosion
In a purely analytical perspective, it is recalled that samples from the Phen test, despite already depicting the orange coloration of Fe-Phen complexes, should be treated in the same way like the calibration standards. In this study ascorbic acid was used because of parallel experiments with EDTA [21]. In routine Phen tests, all other reducing agents (e.g., hydroxylamine) can be used, but the authors recommend ascorbic acid for its less toxicity and even its better affordability [59].  (Figure 3b), no such variability was observed as the Phen test is neither sensible to Fe 3 species nor to dissolved O 2 (Section 3.1). Calculations showed that 2 to 16 times more iron was dissolved in the EDTA systems than in the corresponding Phen systems. Extending the experimental duration to 120 h (5 days) revealed differential behaviors for individual Fe 0 SW specimens (Section 3.2, Figure 2). The general trend is that iron dissolution from the coarsest material (SW8) was progressive through the end of the experiments (Figure 2). This trend was also observed for SW9, which was the most reactive by virtue of its "pre-oxidized" surface state. In a purely analytical perspective, it is recalled that samples from the Phen test, despite already depicting the orange coloration of Fe-Phen complexes, should be treated in the same way like the calibration standards. In this study ascorbic acid was used because of parallel experiments with EDTA [21]. In routine Phen tests, all other reducing agents (e.g., hydroxylamine) can be used, but the authors recommend ascorbic acid for its less toxicity and even its better affordability [59].  Figure 3a depicts a large variability of the standard deviation (results not shown) which results from various amounts of atmospheric corrosion products initially available on the SW specimens and the evidence that dissolved O2 accelerates iron dissolution. For the experiments with Phen (Figure 3b), no such variability was observed as the Phen test is neither sensible to Fe 3 species nor to dissolved O2 (Section 3.1.). Calculations showed that 2 to 16 times more iron was dissolved in the EDTA systems than in the corresponding Phen systems. Extending the experimental duration to 120 hours (5 days) revealed differential behaviors for individual Fe 0 SW specimens (Section 3.2., Figure  2). The general trend is that iron dissolution from the coarsest material (SW8) was progressive through the end of the experiments (Figure 2). This trend was also observed for SW9, which was the most reactive by virtue of its "pre-oxidized" surface state.   Table 2. Table 2 summarizes the k values for both tests for 5 experimental points (N = 5) and an experimental duration of 24 hours, corresponding to linearity. The 'a' and R 2 values are also given. The R 2 values for EDTA vary between 0.534 for SW2 and 0.994 for SW3. For Phen all R 2 values were larger than 0.94, confirming the better suitability of the Phen test as already discussed (see also [52]). As concerning the 'b' values, it is seen that for the EDTA test, four from nine values are negative and large in absolute values (b > 1.03 mg). For the Phen test, only three values are positive but all low (0.055, 0.312, and 1.100 mg), while the 6 negative numbers are very low, (< 0.6 mg). The negative 'b' values from the Phen test were due to the very slow kinetics for iron dissolution from less reactive samples (Figure 3b) while negative value from EDTA test was attributed to dissolved O2 and atmospheric corrosion products [20,21,52]. The most important parameters from these tests are the k values.  Table 2.  Table 2 summarizes the k values for both tests for 5 experimental points (N = 5) and an experimental duration of 24 h, corresponding to linearity. The 'a' and R 2 values are also given. The R 2 values for EDTA vary between 0.534 for SW2 and 0.994 for SW3. For Phen all R 2 values were larger than 0.94, confirming the better suitability of the Phen test as already discussed (see also [52]). As concerning the 'b' values, it is seen that for the EDTA test, four from nine values are negative and large in absolute values (b > 1.03 mg). For the Phen test, only three values are positive but all low (0.055, 0.312, and 1.100 mg), while the 6 negative numbers are very low, (< 0.6 mg). The negative 'b' values from the Phen test were due to the very slow kinetics for iron dissolution from less reactive samples (Figure 3b) while negative value from EDTA test was attributed to dissolved O 2 and atmospheric corrosion products [20,21,52]. The most important parameters from these tests are the k values.

Comparing kEDTA and kPhen Values
The k values from Table 2 enable a better classification of the order of reactivity of the tested SW specimens. In 2 mM EDTA and Phen, the increasing order of reactivity are: In both cases, SW9 was the most reactive, while SW1 and SW6 were the least reactive (Figures 2  and 3). While the results from the Phen test are more reliable (forward iron dissolution), more research is needed to ascertain whether in the long term the specimens depicting very similar reactivity herein could be differentiated. One curious observation herein is that it is rather the coarsest sample (SW8) which exhibited the largest k Phen value from the brand new materials. Based on the larger surface area depicted by finer materials, the opposite would have been expected. It is recalled that SW9 was from a different origin and had a pre-oxidized surface.
Certainly, the dynamic process of iron dissolution from concentric layers of Fe 0 atoms making up the SW filaments cannot be linear [61,62]. Effects similar to "case hardening" for food and wood drying will lead to "surface hardened layers" [63][64][65] leading to differential kinetics/extents of Fe dissolution for different particle size ranges. In other words, the extent of restricted corrosion rates through resulting surface hardened layers will be different for different grades of Fe 0 SW. The results of this research suggest that "case hardening" implies that Fe 0 is easier to dissolve from coarser SW than from finer. However, such a conclusion cannot be drawn without considering the manufacturing and storage processes of the different specimens [20].
Comparing the k Phen values having in mind the elemental composition (Table 1) would suggest, for SW8 and SW9, that the Ni content is essential for the Fe 0 reactivity. Both samples practically only differ in their Ni content: 0.10 for SW8 and 0.03 for SW9. However, such a key conclusion cannot be drawn based on one example. It has to be remembered that the goal of this research is to screen and recommend Fe 0 specimens for further research. On the other hand, designing an efficient Fe 0 -based water treatment includes selecting the best material to solve a site-specific problem. This might imply purposefully selecting a low reactive material, able to generate satisfactory amounts of corrosion products to achieve water treatment standards for a long time. Figure 3b shows, interestingly, that iron dissolution from most SW specimens did not take place immediately upon addition of the Phen solution. Iron quantitative dissolution started after a lag time of up to 6 h. Such a lag time was already documented in the process of nitrate removal in Fe 0 /H 2 O systems [66,67]. The results herein justified why this was not a typical/general observation. A researcher using SW9 would not necessarily observe a lag time because iron dissolution is comparatively rapid.

Discussion
The need for applicable methods to (1) assesses the intrinsic reactivity of Fe 0 materials and (2) assure their quality control is inherent to the technology and has been recognized in the very early stage of its recent development (post 1990). Summarizing, the experience of the first decade of research [43] recognized difficulties in correlating parameters that had been individually demonstrated to significantly impact the efficiency of remediation systems using Fe 0 . Relevant parameters include chemical composition, cooling history, feedstock's processing, kiln firing atmosphere, specific surface area, surface morphology, and type and amount of oxide coatings. The need to further explore these relationships in future research was expressed. However, this effort was not really continued and the Fe 0 reactivity was mainly characterized using selected Fe 0 specimens and various kinds of pollutants (e.g., Cr 6 , methylene blue, nitrate, trichloroethylene) [44,47]. The reactivity of the materials was then reflected by the rate constants of their degradation. However, possible probing reactants are inherently different in their chemistry and rate constants vary by several orders of magnitude, making achieved results of limited guidance in selecting Fe 0 for the practice [47,49]. Some few pollutant-independent characterization methods have been presented and are summarized below in the chronological order of their publication in the peer-reviewed literature. They are briefly commented.
Reardon [42] presented H 2 evolution at neutral pH values as a tool to characterize the reactivity of iron. Probably due to the large Fe 0 amount needed to quantitatively produce H 2 and the long duration of the experiments, this method has not been considered [21]. Noubactep [68] presented H 2 evolution in HCl (initial pH = 2) as a more rapid mean for the same purpose. This tool was completely ignored, perhaps because of the clear different mechanism of H 2 evolution at pH < 4.0 (compared to the field situation pH > 6.0). Noubactep et al. [45] presented the EDTA method which was independently introduced two years later by Pierce et al. [50]. The EDTA method has been continuously used and revisited by Noubactep and collaborators as summarized in [21,52]. Kim et al. [47] introduced a method to measure the 'reductive power' of Fe 0 . A tri-iodide/iodide system (I 3 − /I − , iodine method) was used for the oxidative dissolution of Fe 0 specimens and the rate constant of I 3 − reduction used to characterize the Fe 0 reactivity. The iodine method has not considered all previous methods, perhaps because it uses a redox reaction. Li et al. [49] introduced a method based on the Fe 3 /Fe 0 redox reaction (ferric method), which consists of adding a concentrated Fe 3 solution (up to 3000 mg L −1 ) to a Fe 0 specimen and quantifying the increase in iron concentration after 4 h. The ferric method seems very simple but ferric iron must be available, and strong dilution must precede iron determination, up to factor 100. Lastly, the Phen method presented in this study uses only one chemical, that is also used in all other methods if the iron is to be determined by spectrophotometry using the 1,10 Phenanthroline method [60]. It is thus not only the simplest but also the most affordable. The Phen test has successfully characterized the intrinsic reactivity of steel wool specimens while using the same mass of material (0.1 g) used for the EDTA test. It is obvious that this test will work for coarser materials. In the meantime, Hu et al. [52] have already validated the Phen test using eight commercial granular materials. The major merit of the Phen test is that it uses the same mass of material (0.1 g) regardless of the size and aggregation state (e.g., filamentous, granular). The well-established common process implying water decontamination by Fe 0 materials is their oxidative dissolution (corrosion). Whether the contaminant is then adsorbed or chemically transformed (oxidized or reduced), oxidation of Fe 0 to Fe 2 (Equation (2)) is the first step [10,30]. Precisely, this process is characterized by the Phen method and is neither disturbed by dissolved O 2 nor atmospheric FeCPs. Therefore, k Phen values provide a reliable and confident parameter to characterize reactive Fe 0 materials. Following the roadmap of [18], it is time to build a significant k Phen database for available Fe 0 materials to prepare their systematic testing in long-term column experiments. Once a significant number of materials differing in their k Phen values are tested in such long-term experiments, it will be possible to reliably recommend selected ones for the site-specific design of field Fe 0 /H 2 O systems. In other words, the data provided herein are a clear starting point for the design of future laboratory, pilot, and field-scale studies for the design of the next generation Fe 0 remediation systems to achieve the United Nations Sustainable Development Goals (UN SDGs) [18]. Further work should also use other affordable approaches like the methylene blue method (MB method) [69] to quickly access the impact of various parameters of the system efficiency. The MB method takes advantage of the differential affinity of MB toward sand and iron oxides (FeCPs). Thus, a system containing sand and Fe 0 discolors always lesser MB the more sand is in situ coated by FeCPs.
The results presented herein also clarify the influence of complexing agents (including EDTA and Phen) on the efficiency of Fe 0 /H 2 O systems. Complexing agents are often used to avoid the precipitation of Fe(OH) 3 and, thus, to clarify the mechanisms of contaminant removal in Fe 0 /H 2 O systems [58,70,71]. Despite 25 years of intensive research, the reaction mechanism is still controversially discussed [72][73][74][75]. The present study confirms that both EDTA and Phen avoid iron precipitation before the solution is saturated (e.g., [Fe] > 112 mg L −1 for 2 mM EDTA). The results also suggest that (1) contaminant reduction should be quantitative in the presence of both EDTA and Phen because Fe 0 dissolution is accelerated or (2) any observed contaminant reduction in their presence results from Fe 2 oxidation (electrons from Fe 2 ). Case (2) suggests that contaminant reduction is only observed in the presence of EDTA as Phen blocs Fe 2 oxidation. Exactly, this was observed by [58] investigating aqueous Cr 6 removal as influenced by EDTA, NaF, and Phen. Indirectly, [58] have demonstrated that aqueous Cr 6 is removed only when Fe precipitates. Their observation is thus consistent with the concept that adsorption and co-precipitation are the fundamental mechanisms of contaminant removal in Fe 0 /H 2 O systems [4,10,[73][74][75][76].

Conclusions
Iron dissolution in 2 mM 1,10-Phenanthroline (Phen test) provides an accurate and reliable method to characterize the intrinsic reactivity of Fe 0 materials for water treatment. The Phen test uses only one reagent, is environmentally friendly, and takes just 16 h. Here, it is conducted with a conventional spectrophotometer in a low-equipped laboratory, and it is thus universally applicable. The Phen test can be further developed to be used to assess the long-term kinetics of iron corrosion to help design sustainable filtration systems. This simple method is ready to serve as a universal testing procedure for Fe 0 suppliers and is a standard candidate procedure for testing Fe 0 materials for environmental applications.