Metallic Iron Filters for Universal Access to Safe Drinking Water

Chicgoua Noubactep1

Angelika Sch�ner2

Paul Woafo3

1Angewandte Geologie, Universit�tG�ttingen, G�ttingen, Germany.

2Institut f�r Geowissenschaften,Ingenieurgeologie, Martin-Luther-Universit�t Halle, Halle, Germany.

3Laboratory of Modeling and Simulation inEngineering and Biological Physics,Faculty of Science, University ofYaounde I, Yaounde, Cameroon.

Review

Metallic Iron Filters for Universal Access to SafeDrinking Water

The availability of sustainable safe drinking water is one of Millennium DevelopmentGoals (MDGs). The world is on schedule to meet the MDG to ,halve by 2015 the propor-tion of people without sustainable access to safe drinking water in 2000’. However,present technologies may still leave more than 600 million people without access tosafe water in 2015. The objective of the present article is to present a concept for uni-versal water filters primarily made of metallic iron (Fe0) and sand. The concept of Fe0/sand filters is based on a combination of: (i) recent developments in slow sand filtra-tion and (ii) recent progress in understanding the process of contaminant removal inFe0/H2O systems. The filters should be made up of more than 60% sand and up to 40%Fe0. The actual Fe0 proportion will depend on its intrinsic reactivity. The most impor-tant question to be answered regards the selection of the material to be used. Thedesign of the filter can be derived from existing filters. It appears that Fe0/H2O basedfilters could be a technology with worldwide applicability.

Keywords: Contaminant removal; Point of use; Safe water; Water filtration; Zerovalent iron;

Received: May 28, 2009; revised: August 05, 2009; accepted: September 04, 2009

DOI: 10.1002/clen.200900114

1 Introduction

Natural, readily available waters (shallow groundwater, surfacewater, water from boreholes and springs) are the main sources fordrinking water production. These waters may become contami-nated with organic/inorganic chemical pollutants and pathogenicmicroorganisms (bacteria, fungi, protozoa, viruses) of various ori-gins: (i) natural (geogenic) hydrogeochemical processes, (ii) artificialrecharge with wastewater or water from septic tanks or leaking sew-age pipes, (iii) discharge of wastewater and/or manure run off fromagricultural land.

To produce safe drinking water from natural waters, chemicalpollutants and pathogens need to be removed. One effective naturalway is the passage of surface water through soil, as is the case inbank filtration [1–4]. The effectiveness of bank filtration to producesafe drinking water from natural waters depends on four major fac-tors: (i) the extent of water pollution, (ii) the nature of the soil, (iii)the thickness of the soil or the travel distance within the soil, and(iv) the water flow rate or travel time. One can learn from natureand produce drinking water on a small scale. For this purpose, itcan be assumed that for a sufficient amount of any relevant treat-ment material in a filter, an adequate water flow rate will yield sat-

isfactory contaminant mitigation. The reactive material can be (i) anatural material, (ii) a synthetic material or composite, or (iii) a mix-ture of materials. The amount of material determines the thicknessof reactive layer in the filter. The water flow rate determines traveltime.

During the past three decades, treating water at the householdlevel has been shown to be one of the most efficient means of pre-venting waterborne diseases worldwide [5–10]. The following are sixexamples illustrating this: the use of (i) commercial filtration devi-ces (e. g., Brita), used mostly in developed countries [11–13], (ii)ceramic pot filters or Kosim filters in Ghana [14], (iii) 3-Kolshi filtersand SONO filters in Bangladesh [6, 15, 16], (iv) KanchanTM arsenic fil-ters in Nepal [17, 18], (v) Danvor plastic biosand filters [19], and (vi)Potters for Peace Filtron ceramic filters [19, 20]. Promoting house-hold water treatment and safe storage helps populations to activelytake charge of their own water security. Therefore, providing peoplewith the knowledge and affordable tools to treat their own drinkingwater at the point of use is a noble objective.

The present study presents the concept of using metallic iron (Fe0)as a universal filter material for water treatment at the point of use(POU) in general and at the household level in particular. The secon-dary objective is to demonstrate the suitability of Fe0 for water treat-ment in remote locations of developing countries. In these regions,available water could be chemically safe (no geogenic or anthropo-genic contamination). In this case, microbial contamination due tolack of sanitation and improper hygiene is the sole source of watercontamination. Some relevant water pathogens and related diseasesare summarized in Tab. 1 [21]. The presentation is based on state ofthe art knowledge of the mechanisms of contaminant removal inthe presence of Fe0 (Fe0/H2O systems) which will be presented first.

Correspondence: Dr. C. Noubactep, Angewandte Geologie, Universit�tG�ttingen, Goldschmidtstr. 3, D-37077 G�ttingen, Germany.E-mail: [email protected]

Abbreviations: AMD, Acid mine drainage; BSF, Biosand filter; CIM, Com-posite iron matrix; EDTA, Ethylenediaminetetraacetate; KAF, KanchanTM

arsenic filters; MDG, Millennium Development Goal; NGO, Non-govern-mental organization; POU, Point of use; ZVI, Zerovalent Iron

i 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.clean-journal.com

930 Clean 2009, 37 (12), 930 –937

Clean 2009, 37 (12), 930 –937 Metallic Iron Filters for Universal Access to Safe Drinking Water 931

2 Discussion

2.1 Mechanisms of Contaminant Removal inFe0/H2O Systems

The passive remediation of water pollution by using Fe0/H2O systemshas been developed for about 20 years [22–24]. Applications of theseinnovative systems have included (i) groundwater remediation, (ii)drinking water treatment, and (iii) wastewater treatment. Success-ful quantitative removal of metals (e. g., Cd, Co, Cr, Cu, Pb, U, Zn),non metals (e. g., As, Mo, Se, Sn), anions (e. g., AsO4

3–, F–, MoO42–, NO3

–,PO4

3–, SeO42–, SO4

2–), organic dyes, organic compounds (e. g., benzene,chlorinated solvents, phenol, pesticides, toluene), bacteria, sus-pended solids, and viruses has been reported [16, 25–30]. Almost allstudies dealing with pollutant removal were limited to proving theviability of Fe0/H2O systems for a few target pollutants and were notincorporated within a broad based understanding of Fe0 remedia-tion technology.

It is interesting to note that another technology, electrocoagula-tion [31–33], using Fe0 could exhibit similar efficiency for a largespectrum of contaminants. Iron electrocoagulation can be consid-ered an electrochemically driven accelerated corrosion process. Inboth cases (passive Fe0/H2O and iron electrocoagulation), pollutantsare removed from water by a multitude of mechanisms operatingsynergistically.

Likely removal mechanisms include: (i) adsorption onto Fe0 andFe oxides/hydroxides, (ii) coprecipitation with Fe oxides (or copreci-pitation on the substrate), (iii) direct reduction by Fe0, (iv) indirectreduction by FeII or H/H2, (v) indirect oxidation by in situ generatedradicals (HO)9. The exact sequence of these reactions depends on thesystem. Because redox insensitive pollutants have been quantita-tively removed in passive Fe0/H2O systems, any individual redoxprocess could not be the fundamental removal mechanism as origi-nally assumed [23, 34]. Therefore, the fundamental mechanisms ofcontaminant removal in Fe0/H2O systems are adsorption (includingsurface complexation) and coprecipitation [35, 36].

Certainly, for some contaminants, such as CrVI which are reduci-ble even in the aqueous phase by FeII (under local anoxic conditions),quantitative reduction may precede adsorption and coprecipita-

tion. However, considering the nature of the Fe0/H2O system, thereis no reason to consider quantitative reductive transformation a pri-ori for some species. Moreover, a Fe0/H2O system should be consid-ered as a zone of precipitating iron oxides/hydroxides [37]. Duringthis process expansive amorphous iron hydroxides are generatedand further transformed by dehydration into more crystallineoxides (see Tab. 2) [38, 39].

2.2 Diversity in the Reactivity of Iron Oxides inFe0/H2O Systems

In order to better understand the role of metallic iron in Fe0/sandfilters, the dynamic nature of iron corrosion in Fe0/H2O systems isdiscussed. For this purpose, synthetic bulk oxides and oxide filmson Fe0 should be compared in terms of reactivity towards contami-nant removal. Synthetic oxides and in situ formed oxide films arefundamentally different in that the latter are reactive systems( ,reactive oxides’) by virtue of the continual generation and annihi-lation of point defects at the interfaces. Furthermore, oxide films(passive layers), unlike synthetic oxides, continually grow into themetal at the Fe0/layer interface while being simultaneouslydestroyed by dissolution or restructuring at the layer/H2O interface[40, 41]. Accordingly, synthetic oxides are strictly

,

coatings’ andhence, with respect to the processes that influence Fe0 reactivity,they are

,

dead’ [40]. Synthetic oxides can only act as contaminantadsorbents. Although synthetic oxides may be capable of simulatingthe properties of the deposited outer layers (at the layer/H2O inter-face), it is difficult to see how they can simulate the barrier layer (atthe Fe0/layer interface,

,

reactive oxides’), whose defect concentra-tion is normally far in excess of that which can be obtained in bulkoxides.

,

Reactive oxides’ are better adsorbents and may coprecipi-tate contaminants during their formation and transformation to

,dead oxides’ [25, 35, 36, 42, 43]. ,Dead oxides’ are good adsorbentsof limited and selective affinity to some contaminants. The conse-quence of this analysis is that Fe0 can be regarded as a permanentsource of highly reactive hydroxides ( ,reactive oxides’) in a treat-ment system (Statement 1). Statement 1 is the major argument onwhich the concept of Fe0/sand filters for universal access to safedrinking water is built.

A schematic illustration of the time dependant evolution of ironcorrosion products is given in Fig. 1. The cross section of a sphericalFe0 material experiencing uniform corrosion is represented. It isassumed that the initial material (Fe0 in Fig. 2) is progressively cov-ered by concentric layers of iron oxides from which the three mostinternal layers (see d1, d2, d3, Fig. 2) are

,

reactive’ (see above) and allthe external layers (d4) are comparative in their reactivity to syn-thetic bulk oxides (

,

coatings’). While this assumption is somewhatarbitrary, d1 can be regarded as a layer of amorphous ferrous ironhydroxides (Fe(OH)2), d2 a layer of mixed amorphous iron hydroxides(Fe(OH)2 and Fe(OH)3), d3 a layer of mixed amorphous iron hydrox-ides/oxides (green rust, magnetite) and d4 a layer of aged corrosionproducts (crystallized iron oxides).

The schematic representations in Fig. 1 and 2 have intentionallyneglected the expansive nature of the process of iron oxide/hydrox-ide production form metallic iron [39, 44]. In fact, the theoreticalratio (a1 = Voxide/VFe) between the volume of corrosion products andthe volume of iron in the metallic structure varies between 2.0 forFe3O4 and 6.4 for Fe(OH)3 . 3 H2O (see Tab. 2). This volume increase isthe principal cause of the expansion and ultimately the loss ofhydraulic conductivity (permeability loss) for Fe0/H2O remediation


Table 1. Some relevant water pathogens and related diseases. Compiledfrom Hoffman et al. [21].

Microbes Examples Diseases

Bacteria Vibrio cholerae CholeraShigella sonnei Bloody diarrheaSalmonella enterica Gastrointestinal illnessHelicobacter pylori Chronic ulcers and cancerEscherichia coli Bloody diarrhea, kidney failureArcobacter butzleri Acute gastrointestinal illness

Viruses Astro- and Caliciviruses Gastrointestinal illnessHepatitis A/E virus Acute liver disease (hepatitis)Rotavirus Severe diarrhea

Protozoa Blastocystis hominis Diarrhea and abdominal painEntamoeba histolytica Gastrointestinal illnessNaegleria fowleri Amebic meningoencephalitis

Fungi Aspergillus fumigatus Respiratory illnessFusarium solani Skin-related infections

932 C. Noubactep et al. Clean 2009, 37 (12), 930 –937

systems. Clearly, the end of service of Fe0/sand filters is not (orshould not be) dictated by Fe0 depletion but rather by loss ofhydraulic conductivity. The practical expansion of corrosion prod-ucts can be discussed from a proper kinetic law which correlates theradius loss of Fe0 and the corresponding volume of accumulatediron oxides [39]. In this work, mixing inert sand and Fe0 is a practicaltool to extent filter service life. It is expected that for any appropri-

ate Fe0 material, an optimal weight ratio sand/Fe0 should exist forsatisfactory water treatment in the medium for long term (e. g., F12 months).

2.3 Water Treatment by Sand Filtration

Promising strategies for providing people with access to safe waterare available, but may not be suitable or affordable for areas of lowpopulation density with little perspective for economic growth,e. g., rural areas, [45, 46]. It may further prove challenging to guaran-tee the quality of drinking water in these populations. Methods thatallow the treatment of water at the place where it is consumed (POUmethods) may provide a low cost, promising, easy and flexible solu-tion for increasing drinking water quality in much of the popula-tion who need it. Slow sand filtration is one of these methods.

2.4 Basic Sand Filtration

The earliest form of water treatment was slow sand filtration. Slowsand filters were developed in the 1820’s in Scotland by RobertThom and in England by James Simpson. They became successfullyestablished in Europe by the end of the 19th century [47–50]. Thistechnology used sand filter beds through which water was slowlytrickled. The natural formation of a biological layer (biofilm, widely


Table 2. Some relevant characteristics of metallic iron and its main corrosion products. a is the molecular weight of iron to the molecular weight of thecorrosion products. a1 is the ratio of volume of expansive corrosion products to the volume of iron in the metallic structure. Compiled from Noubactep[38] and Balasubramaniam et al. [39].

Name Formula Structure Density(kg/m3)

a a1

Iron a-Fe bcc 7800 – –Hematite 1/2 Fe2O3 Trigonal 5260 0.699 2.12Magnetite 1/3 Fe3O4 Cubic 5180 0.724 2.08Goethite a-FeOOH Orthorhombic 4260 0.629 2.91Akageneite b-FeOOH Tetragonal 3560 0.629 3.48Lepidocrite c-FeOOH Orthorhombic 4090 0.629 3.03

Fe(OH)2 Trigonal 3400 0.622 3.75Fe(OH)3 n.a.a n.a.a 0.523 4.20Fe(OH)3 N 3 H2O n.a.a n.a.a 0.347 6.40

a n.a. = not available

Figure 1: Time dependant evolution of Fe0 and Fe corrosion products from a spherical material assuming uniform corrosion. Fe0 corrosion continuouslyproduces concentric layers of iron hydroxides which are transformed to iron oxides. It is assumed (see text) that the three layers next to the Fe0 surfaceare reactive (d1, d2, d3) while the outer layers are non reactive (d4).

Figure 2: Cross sectional diagram of a uniformly corroding Fe0 particle att > t2 (see Fig. 1). d1, d2 and d3 are supposedly the reactive or transforminglayers in which contaminants may be adsorbed, coprecipitated or chemi-cally transformed. d4 is the growing layer of aged corrosion products onwhich contaminants could adsorb. d4 is less porous than the inner layersbut may be fissured.


termed as Schmutzdecke) and the filtering action of sand removesbacteria, silt and chemical pollutants. The treatment efficiency isaffected partly by physical straining, but more importantly by bio-logical action within the biofilm that forms on top of the sand [51].A household scale intermittent slow sand filter (the Davnor Biosandfilter or Davnor BSF) developed by Manz has been successfully testedby several governments, research and health institutions and NGOsin Bangladesh, Brazil, Canada, Haiti, Nicaragua, Vietnam and othercountries [52, 53]. Lantagne et al. [5] summarized the drawbacks ofthe BSF as follows: (i) low rate of virus inactivation, (ii) lack of resid-ual protection and removal of less than 100 percent of the bacteria,which leads to recontamination, (iii) current lack of studies provinghealth impact, and (iv) difficulty in transport and high initial cost,which make scalability more challenging. Accordingly, the majorproblem with BSF is that complete pathogen removal cannot beguaranteed. Therefore, if water can be successfully freed frompathogens within a sand filter, this technology can be used withmore confidence for safe drinking water at a household level. Thepresentation on the mechanisms of contaminant removal in Fe0/H2O systems suggests that coupling reactive Fe0 to sand could solvethe problem (Statement 1). This work has already been done but Fe0

aimed at eliminating arsenic as presented below.

2.5 Improved Sand Filtration

Recently, the Arsenic Biosand Filter (KanchanTM Arsenic Filter, KAF)was developed and distributed in Nepal by Ngai et al. [17, 18]. TheKAF is built on the platform of a slow sand filter, modified toinclude/increase arsenic removal capability. The KAF combines theconcepts of slow sand filtration and an intermittent householdscale system with the innovation of a diffuser basin containing ironnails for arsenic removal. In the KAF, arsenic is certainly quantita-tively removed by adsorption onto or by coprecipitation with ironoxyhydroxides from rusting iron nails. Pathogens (e. g., bacteria)have been reported to be removed mostly by physical straining pro-vided by the fine sand layer, by attachment to previously removedparticles and, to a lesser degree, by biological predation occurringin the top few centimeters of the sand. The KAF was demonstratedsuccessful for simultaneous arsenic and pathogen removal and isconsidered as the best among all household arsenic filters availablein Nepal [18]. Chiew et al. [54] have recently published primaryresults on attempts to extend the concept of KAF filtration in Cam-bodia (see Fig. 3).

Recent developments in understanding the process of contami-nant removal in Fe0/H2O systems demonstrates that pathogens arecertainly removed/inactivated by coprecipitation with iron oxy-hydroxides from rusting Fe0. Therefore, a further development ofthe KanchanTM Arsenic Filter is to disseminate Fe0 in of the pathogenremoval unit of the filter (BioSand filter, see Fig. 2) or to replace thearsenic removal unit by a unit of water saturated Fe0/sand layer. Thesecond possibility is discussed below. Such a filter will remove allpossible contaminants (including arsenic and bacteria) from theaqueous solution.

2.6 Fe0/Sand Filters

The works of Dr. Manz yielded an efficient intermittent householdscale slow sand filtration system for

,

safe’ drinking water [50]. Ngaiet al. [18] further developed this system by adding a small amount of

metallic iron to optimize arsenic (and pathogen) removal. As a rule,the nature and extent of water contamination in rural areas ofdeveloping countries is unknown [46]. However, despite the variabil-ity of the influents, the system outputs must be satisfactory. Basedon state of the art knowledge of the mechanism of contaminantremoval by Fe0/H2O systems, the use of metallic iron as a universalmaterial for safe drinking water has been suggested [55, 56]. Theworks of Ngai et al. [18] suggested that the amount of Fe0 in the filteris not necessarily high (e. g., 10% weight). Furthermore, since chem-ical contaminants and pathogens are certainly coprecipitated withprecipitating iron corrosion products, the flow rate of contami-nated waters through the Fe0/sand filter can be enhanced to reachthe daily demand of individual families within tens of minutes.That is the idea behind Fe0/sand filters.

2.7 Innovation in Fe0/Sand Filters

People familiar with the problem of arsenic contamination in Ban-gladesh and Nepal are aware that an effective Fe0/sand filtration sys-tem (3-Kolshi system) was abandoned because of loss of porosity inthe system. A close look at the 3-Kolshi filtration system reveals thatthe top Kolshi was filled with 3 kg of cast iron and 2 kg of sand ontop of the iron turnings [15]. In other words the Kolshi with Fe0

(here cast iron) contains a zone of 100% Fe0 overlying the sand layer.The 3-Kolshi filtration system was replaced by the SONO filter. Here,the primary active material is a porous composite iron matrix (CIM),a mass made of cast iron turnings through a proprietary process to


Figure 3: Design of the filter tested for arsenic and pathogen removal inCambodia by Chiew et al. [54]. The pathogen removal unit is a conven-tional BioSand filter in which suspended solid are removed.


maintain active CIM integrity for years [6, 16]. Again, a 100% layer ofcomposite is sandwiched between two layers of (coarse) sand. Thereported loss of porosity is necessarily coupled to an overly high pro-portion of Fe0 in the reactive zone (which is actually 100%). Giventhe relatively low concentration of arsenic in contaminated waters,this huge amount of Fe0 (and possibly composite) is obviouslyunnecessary. As discussed above, Fe0 depletion should not occur.However, even the KAF of Ngai et al. [18] contains a 100% layer ofiron nails on top of fine sand, coarse sand and gravel. The relativeamount of iron nails in KAF was small compared to that of reactivematerials in 3-Kolshi and SONO filtration systems and the KAF areefficient for several years [17]. Therefore, reducing the proportion ofFe0 in the system is a sensible modification to maintain filter perme-ability for a long time.

The real novelty with the proposed Fe0/sand filters is that no 100%Fe0 will be available. Rather, the Fe0 reactive layer will be a mixtureof at least 60% sand, gravel or porous volcanic rocks and up to 40%Fe0. The actual proportion of Fe0 will depend on its intrinsic reactiv-ity and particle size. The advantage of porous volcanic rocks [57, 58]is that generated iron hydroxides may fill their porous structureextending service life (or retarding loss of hydraulic conductivity).On the other hand, Fe0 materials could be produced locally [59–62]or selected from available iron products (including productionwastes and by products) such as nails rivets, nuts, bolts, barbed wire,packing wire, chicken wire mess, fencing wire, steel wool, construc-tion materials and other Fe0 products. Using available Fe0 materialswill protect filter users from the market law as increased demand isonly indirectly coupled to the original use of the material. For exam-ple, it cannot be expected that the price of packing wire wouldincrease just because it is used for water treatment. Anotherapproach will consist of testing the potential Fe0 materials in indus-trialized countries (including China), that are likely to be exportedto developing countries and build a database of suitable materials.

2.8 Potential Beneficiaries of Fe0/Sand Filters

It is certain that Fe0/sand filters would accelerate the health gainsassociated with improved drinking water until the goal of universalaccess to piped, treated water is achieved [5]. Even after this hypo-thetical goal is achieved, the world will still have to face critical sit-uations (e. g., accidental contamination, earthquakes, epidemicplagues, wars, tsunamis), in which available water should be rapidlytreated. These critical situations are managed worldwide mostly byarmies and NGOs (including the Red Cross). Therefore, the Fe0/sandfilters are not only suitable for developing countries but for thewhole world. Recently, uranium contamination was discovered indrinking water production wells in Barlissen in Lower Saxony, Ger-many and the precautionary measure of putting the wells out ofservice was taken [63]. Barlissen is a typical situation where Fe0/sandfilters could help in industrialized countries.

On the other hand, working on Fe0/sand filters will give research-ers from developing countries the opportunity to solve a crucialproblem with local solutions. The technology is low cost and goodresults can be achieved with simple equipment. Furthermore,experiments do not involve any stringent reaction conditions norexpensive devices, as the major experiments are to be conductedunder atmospheric conditions. It is expected that research groupsworking on Fe0/sand filters will achieve results which are helpfulfor the further development of iron reactive walls. In fact, Fe0/sandfilters can be regarded as a sort of rapid small scale column test

which could help to bridge the gap between short term studies inthe lab (few weeks) and field Fe0 reactive barriers (some two decades).In fact, testing Fe0/sand filters offers a unique opportunity to testthe reactivity of the same Fe0 material at several sites with naturalwaters of various characteristics. Clearly, a South/North symbiosiscan be expected: a technology developed in the North tested andused in a modified version in the South could contribute to the fur-ther development of the original technology (not only for theNorth).

2.9 Ways to Affordable Fe0/Sand Filters

The concept of Fe0/sand filters is based on a scientific understandingof the complex chemical and physical processes involved in anevolving technology (iron remediation technology) that has beingsuccessfully applied for almost 20 years. Provided that a relevantreactive Fe0 is used, the effectiveness of Fe0/sand filters does not needto be demonstrated, except some technology verification in the field(monitoring). The sole tasks are: (i) selecting and processing theappropriate Fe0 materials, (ii) designing the filters, and (iii) avoidingthe use of Fe0/sand filters for waters of pH f 5. In fact, Fe0/sand filtersare based on the anodic dissolution of iron in neutral and close toneutral aqueous systems. In this pH range, primary iron dissolutionis followed by a continuous build up and transformation of a corro-sion product layer in the vicinity of Fe0 [41, 64, 65].

2.10 Material Selection

Practically any available Fe0 material (mainly low alloyed steel andcast iron) is theoretically applicable for water treatment. Ideally,any newly obtained Fe0 material should be characterized and testedfor water treatment capacity by a standard method. In using acti-vated carbon for wastewater treatment, for example, it is generallyaccepted that a good decolorizing carbon should fulfill at least 200mg/g removal capacity for methylene blue in batch experiments[66]. Until recently, there was no experimental parameter to charac-terize the intrinsic reactivity of Fe0 materials [67–69]. Noubactep[68, 69] introduced a parameter, kEDTA, which could enable purpose-ful material selection. kEDTA is the slope of the line of the timedependent oxidative dissolution of ions from a given Fe0 material ina 2 mM EDTA solution. kEDTA is determined in batch experimentsand characterizes material intrinsic reactivity under any givenexperimental conditions [70]. For example, materials used in Fe0/sand filters should exhibit a kEDTA value above a critical value. Thiscritical value is yet to be determined. The kEDTA value of the iron nailsused by Ngai et al. [18] could be used as guide to select the Fe0 load-ing of the filter. For a less reactive material, a larger Fe0 loading thanin the KanchanTM Arsenic Filter would be applied and lesser Fe0 load-ings are needed for more reactive materials. The long term reactiv-ity of available Fe0 materials has to be tested as well. Initially, ironnails from the KAF filters or iron composites from the SONO filterscan be used as starting point.

2.11 Filter Designs

Basically, Fe0/sand filters are slow sand filters for intermittent use[49, 50]. Almost 130 years of experience is available for this type ofsystem. The most recent developments are those of KanchanTM

Arsenic Filter [18], SONO filters [6, 16] and Kosim filters [14]. These



devices can be modified to use a Fe0/sand reactive layer as describedabove. Alternatively, new designs may be conceived based on localspecifications [70]. A cross-section of the filter design used by Chiewet al. [54] is shown in Fig. 3. Arsenic removal and bacterial removalunits are illustrated. A potential shortcoming of this design is thatthe arsenic removal unit experiences wet and dry periods. In animproved design, the height of the outlet pipe should enableimmersion of the Fe0 layer.

In all cases the filters should be constructed from local materialsand the reactive layer should be embedded in the sand layers. Avail-able Fe0 materials (e. g., iron nails, steel wool) should be tested andapproved at least in the early stage of technology implementation.The Fe0/sand filter should be manufactured locally by trained work-ers regardless of their scholarly education. To sustain Fe0 reactivity,intermittent additions (e. g., once a month) of boiled water [16] orlemon juice [71] are envisaged.

While Hussam and Munir [16] used boiled water to kill pathogens,the present review demonstrated that upon purposeful dimension-ing, pathogens will be sequestrated by iron corrosion products.Therefore, boiled water should increase/sustain Fe0 reactivity by ele-vating the temperature. Intermittent temperature elevation cer-tainly disturbs the process of iron passivation, and thus sustains Fe0

reactivity as reported by Hussam and Munir [16].The rational for the use of citric acid is the well known increase in

iron corrosion with decreasing pH value [65, 66]. Cornejo et al. [71]tested three commercially available lemon species as citric sources.Commercial lemon species may contain undesirable preservativesand little fruit. Even though these undesirable contents are nontoxic, these additives should be regarded as contaminants and willbe removed in the filtering system as well. In rural areas naturallemon juice should be at least seasonally available at low cost orcheaper than commercial lemon juices. While testing the ability ofcitric acid (pKa1 = 3.15, pKa2 = 4.77, pKa3 = 6.40) to sustain reactivity,care should be taken to keep the pH > 4.5 to avoid dissolved Fe in thefiltered water. Again, this technology is not applicable for the treat-ment of waters of initial pH a 5. Such waters, e. g., from acid minedrainage (AMD), occur worldwide at abandoned and active miningsites [72].

3 Conclusions

The fact that field Fe0/H2O systems have successfully removed vari-ous contaminants from polluted water is a testament to their poten-tial, which is yet to be fully realized. This paper presents a way forthis potential to be realized at a household level by simple water fil-tration through a Fe0/sand column. The Fe0/sand filter is an innova-tion combining two proven water treatment techniques: (i) adsorp-tion and coprecipitation in Fe0/H2O systems, and (ii) size exclusionby slow sand filtration. There is no doubt that proper material selec-tion and design of the filtration system will contribute to achieving,and even exceeding, the Millennium Development Goal. The origi-nal merit of this concept is that it allows researchers from develop-ing countries to actively work on a crucial problem while resultswill be useful for the developed world as well. Intensive laboratoryand field research is needed to develop Fe0/sand filters. The effortsshould be accompanied by numerical modeling. This aspect isunder investigation in the research group of one of the authors (P.Woafo) using mathematical equations describing the spatio-tempo-ral evolution of pollutant concentration and porosity in the filters

complemented by experiments. On the other hand, reportedly effi-cient filters (e. g., Danvor BSF, Filtron ceramic filter, Kosim filter)can be amended with Fe0 to enhance pathogen inactivation.

Acknowledgments

Thoughtful comments provided by Prof. Abul Hussam (Center forClean Water and Sustainable Technologies, George Mason Univer-sity, USA) on the draft manuscript are gratefully acknowledged.Sven Hellbach (student research assistant) is acknowledged for tech-nical assistance. The manuscript was improved by the insightfulcomments of anonymous reviewers from Clean: Soil, Air, Water.

The authors have declared no conflict of interest.

References

[1] J. F. Schijven, S. M. Hassanizadeh, Removal of Viruses by Soil Passage:Overview of Modeling, Processes, and Parameters, Crit. Rev. Env. Sci.Technol. 2000, 30 (1), 49.

[2] K. M. Hiscock, T. Grischek, Attenuation of Groundwater Pollution byBank Filtration, J. Hydrol. 2002, 266 (3 – 4), 139.

[3] S. Gr�nheid, G. Amy, M. Jekel, Removal of Bulk Dissolved OrganicCarbon (DOC) and Trace Organic Compounds by Bank Filtration andArtificial Recharge, Water Res. 2005, 39 (14), 3219.

[4] W. J. Weiss et al., Riverbank Filtration for Control of Microorgan-isms: Results from Field Monitoring, Water Res. 2005, 39 (10), 1990.

[5] D. S. Lantagne, R. Quick, E. D. Mintz, Household Water Treatmentand Safe Storage Options in Developing Countries: A Review of Cur-rent Implementation Practices, in Water Stories: Expanding Opportuni-ties in Small-scale Water and Sanitation Projects (Eds.: M. Parker, A. Wil-liams, C. Youngblood, R. Weisshaar), Woodrow Wilson InternationalCenter for Scholars, Washington, DC 2006, pp. 17 – 38.

[6] S. Ahamed, A. K. M. Munir, A. Hussam, Groundwater ArsenicRemoval Technologies Based on Sorbents: Field Applications andSustainability, in Handbook of Water Quality and Water Purity (Ed.: S.Ahuja), Academic Press, London 2009.

[7] M. D. Sobsey et al., Point of Use Household Drinking Water Filtra-tion: A Practical, Effective Solution for Providing Sustained Access toSafe Drinking Water in the Developing World, Environ. Sci. Technol.2008, 42 (12), 4261.

[8] T. Clasen et al., Comment on

,

Household Water Treatment in PoorPopulations: Is There Enough Evidence for Scaling up Now?’ Environ.Sci. Technol. 2009, 43 (14), 5542.

[9] W.-P. Schmidt, S. Cairncross, Household Water Treatment in PoorPopulations: Is There Enough Evidence for Scaling Up Now? Environ.Sci. Technol. 2009, 43 (4), 986.

[10] W.-P. Schmidt, S. Cairncross, Response to Comment on ,HouseholdWater Treatment in Poor Populations: Is There Enough Evidence forScaling up Now?’ Environ. Sci. Technol. 2009, 43 (14), 5545.

[11] F. D. Daschner, Holy Water, Tap Water, Mineral Water or Water Fil-ters? J. Hosp. Infect. 1997, 35 (1), 71.

[12] B. L. Gulson et al., The Efficiency of Removal of Lead and other Ele-ments from Domestic Drinking Waters Using a Bench-top Water Fil-ter System, Sci. Total Environ. 1997, 196 (3), 205.

[13] S. Levesque et al., Effects of Indoor Drinking Water Handling on Tri-halomethanes and Haloacetic Acids, Water Res. 2006, 40 (15), 2921.

[14] S. M. Johnson, R. L. Peletz, S. Murcott, Results from HouseholdCeramic Filter Evaluation in Northern Ghana, Proceedings of the 33rdWEDC International Conference, Accra, Ghana 2008, p.463.

[15] A. K. M. Munir et al., Evaluation of Performance of Sono 3-Kolshi Fil-ter for Arsenic Removal from Groundwater Using Zero Valent IronThrough Laboratory and Field Studies, Proceedings of the International



Workshop on Technology for Arsenic Removal from Drinking Water, Bangla-desh University of Engineering and Technology, Dhaka and UnitedNations University, Tokyo 2001, pp. 171 – 189.

[16] A. Hussam, A. K. M. Munir, A Simple and Effective Arsenic FilterBased on Composite Iron Matrix: Development and DeploymentStudies for Groundwater of Bangladesh, J. Environ. Sci. Health, Part A:Toxic/Hazard. Subst. 2007, 42 (12), 1869.

[17] T. K. K. Ngai et al., Development and Dissemination of KanchanTMArsenic Filter in Rural Nepal, Water Sci. Technol.: Water Supply 2006, 6(3), 137 – 146.

[18] T. K. K. Ngai et al., Design for Sustainable Development-householdDrinking Water Filter for Arsenic and Pathogen Treatment in Nepal,J. Environ. Sci. Health, Part A: Toxic/Hazard. Subst. 2007, 42 (12), 1879.

[19] J. Baumgartner, S. Murcott, M. Ezzati, Reconsidering

,

AppropriateTechnology’: The Effects of Operating Conditions on the BacterialRemoval Performance of Two Household Drinking-water Filter Sys-tems, Environ. Res. Lett. 2007, 2, 024003.

[20] D. Van Halem et al., Assessing the Sustainability of the Silver-impreg-nated Ceramic Pot Filter for Low-cost Household Drinking WaterTreatment, Phys. Chem. Earth 2009, 34 (1 – 2), 36.

[21] R. Hoffman, M. M. Marshall, M. C. Gibson, P. A. Rochelle, PrioritizingPathogens for Potential Future Regulation in Drinking Water, Envi-ron. Sci. Technol. 2009, 43 (14), 5165.

[22] R. W. Gillham, S. F. O'Hannesin, Enhanced Degradation of Halogen-ated Aliphatics by Zero-valent Iron, Ground Water 1994, 32 (6), 958.

[23] L. J. Matheson, P. G. Tratnyek, Reductive Dehalogenation of Chlori-nated Methanes by Iron Metal, Environ. Sci. Technol. 1994, 28 (12), 2045.

[24] S. F. O’Hannesin, R. W. Gillham, Long-term Performance of an InSitu ,Iron Wall’ for Remediation of VOCs, Ground Water 1998, 36 (1),164.

[25] S. L. S. Stipp et al., Behavior of Fe-oxides Relevant to ContaminantUptake in the Environment, Chem. Geol. 2002, 190 (1 – 4), 321.

[26] Y. You, J. Han, P. C. Chiu, Y. Jin, Removal and Inactivation of Water-borne Viruses Using Zerovalent Iron, Environ. Sci. Technol. 2005, 39 (23),9263.

[27] A. D. Henderson, A. H. Demond, Long-term Performance of Zero-val-ent Iron Permeable Reactive Barriers: A Critical Review, Environ. Eng.Sci. 2007, 24 (4), 401.

[28] A. B. Cundy, L. Hopkinson, R. L. D. Whitby, Use of Iron-based Technol-ogies in Contaminated Land and Groundwater Remediation: AReview, Sci. Total Environ. 2008, 400 (1 – 3), 42.

[29] R. Thiruvenkatachari, S. Vigneswaran, R. Naidu, Permeable ReactiveBarrier for Groundwater Remediation, J. Ind. Eng. Chem. 2008, 14, 145.

[30] R. L. Johnson, R. B. Thoms, R. O. B. Johnson, T. Krug, Field Evidencefor Flow Reduction through a Zero-valent Iron Permeable ReactiveBarrier, Ground Water Monit. Remed. 2008, 28 (3), 47.

[31] M. Y. A. Mollah, R. Schennach, J. R. Parga, D. L. Cocke, Electrocoagula-tion (EC)-Science and Applications, J. Hazard. Mater. 2001, 84 (1), 29.

[32] D. Lakshmanan, D. A. Clifford, G. Samanta, Ferrous and Ferric IonGeneration During Iron Electrocoagulation, Environ. Sci. Technol.2009, 43 (10), 3853.

[33] C. H. A. Moreno et al., Electrochemical Reactions for Electrocoagula-tion Using Iron Electrodes, Ind. Eng. Chem. Res. 2009, 48 (4), 2275.

[34] E. J. Weber, Iron-mediated Reductive Transformations: Investigationof Reaction Mechanism, Environ. Sci. Technol. 1996, 30 (2), 716.

[35] C. Noubactep, Processes of Contaminant Removal in ,Fe0-H2O’ Sys-tems Revisited. The Importance of Co-precipitation, Open Environ. J.2007, 1 (1), 9.

[36] C. Noubactep, A Critical Review on the Mechanism of ContaminantRemoval in Fe0-H2O Systems, Environ. Technol. 2008, 29 (8), 909.

[37] C. Noubactep, An Analysis of the Evolution of Reactive Species in Fe0/H2O Systems, J. Hazard. Mater. 2009, 168 (2 – 3), 1626.

[38] R. Balasubramaniam, A. V. R. Kumar, P. Dillmann, Characterizationof Rust on Ancient Indian Iron, Curr. Sci. 2003, 85 (11), 1546.

[39] S. Car�, Q. T. Nguyen, V. L'Hostis, Y. Berthaud, Mechanical Propertiesof the Rust Layer Induced by Impressed Current Method in Rein-forced Mortar, Cement Concr. Res. 2008, 38 (8 – 9), 1079.

[40] E. Sikora, D. D. Macdonald, The Passivity of Iron in the Presence ofEthylenediaminetetraacetic Acid I. General Electrochemical Behav-ior, J. Electrochem. Soc. 2000, 147 (11), 4087.

[41] P. Schmucki, From Bacon to Barriers: A Review on the Passivity ofMetals and Alloys, J. Solid State Electrochem. 2002, 6 (3), 145.

[42] R. J. Crawford, I. H. Harding, D. E. Mainwaring, Adsorption andCoprecipitation of Single Heavy Metal Ions onto the HydratedOxides of Iron and Chromium, Langmuir 1993, 9 (11), 3050.

[43] R. J. Crawford, I. H. Harding, D. E. Mainwaring, Adsorption andCoprecipitation of Multiple Heavy Metal Ions onto the HydratedOxides of Iron and Chromium, Langmuir 1993, 9 (11), 3057.

[44] K. Bhargava, A. K. Ghosh, Y. Mori, S. Ramanujam, Modeling of Timeto Corrosion Induced Cover Cracking in Reinforced Concrete Struc-tures, Cement Concr. Res. 2005, 35, 2203.

[45] M. Hobbins, Home-based Drinking Water Purification through Sun-light: From Effectiveness to Health Effectiveness, Ph.D. Thesis, Univer-sity of Basel 2004.

[46] M. N. B. Momba, C. L. Obi, P. Thompson, Survey of Disinfection Effi-ciency of Small Drinking Water Treatment Plants: Challenges FacingSmall Water Treatment Plants in South Africa, Water SA 2009, 35 (4),485.

[47] P. F. Schuler, M. M. Ghosh, P. Gopalan, Slow Sand and DiatomaceousEarth Filtration of Cysts and Other Particulates, Water Res. 1991, 25(8), 995.

[48] L. C. Campos, M. F. J. Su, N. J. D. Graham, S. R. Smith, Biomass Devel-opment in Slow Sand Filters, Water Res. 2002, 36 (18), 4543.

[49] G. Logsdon, R. Kohne, S. Abel, S. LaBlonde, Slow Sand Filtration forSmall Water Systems, J. Environ. Eng. Sci. 2002, 1 (5), 339.

[50] D. H. Manz, P. Eng, New Horizons for Slow Sand Filtration, Proceed-ings of the 11th Canadian National Conference and 2nd Policy Forum onDrinking Water and the Biennial Conference of the Federal-Provincial-Territo-rial Committee on Drinking Water, Calgary, Canada 2004, pp. 682 – 692.

[51] G. Palmateer et al., Toxicant and Parasite Challenge of the ManzIntermittent Slow Sand Filter, Environ. Toxicol. 1999, 14 (2), 217.

[52] W. Duke, A. Mazumder, R. Nordin, D. Baker, The Use and Perform-ance of Biosand Filters in the Artibonite Valley of Haiti: A Field Studyof 107 Households, Rural Rem. Health 2006, 6 (3), 530.

[53] M. A. Elliott et al., Reductions of E. coli, Echovirus Type 12 and Bacter-iophages in an Intermittently Operated Household-scale Slow SandFilter, Water Res. 2008, 42 (10 – 11), 2662.

[54] H. Chiew et al., Effect of Groundwater Iron and Phosphate on theEfficacy of Arsenic Removal by Iron-amended Biosand Filters, Envi-ron. Sci. Technol. 2009, 43 (16), 6295 – 6300.

[55] C. Noubactep, Besseres Trinkwasser an jeder Stelle in l�ndlichenGebieten Afrikas, Afrika Wissenschaft 2008, 1 (3), 36.

[56] C. Noubactep, P. Woafo, Elemental Iron (Fe0) for Better DrinkingWater in Rural Areas of Developing Countries, in Uranium in the Envi-ronment (Eds.: B. J. Merkel, A. Hasche-Berger), Springer, Berlin 2008,pp. 121 – 130.

[57] H. M. N. Wright, K. V. Cashman, E. H. Gottesfeld, J. J. Roberts, PoreStructure of Volcanic Clasts: Measurements of Permeability andElectrical Conductivity, Earth Planet. Sci. Lett. 2009, 280 (1 – 4), 93.

[58] M. Kitis et al., Adsorption of Natural Organic Matter from Waters byIron Coated Pumice, Chemosphere 2007, 66 (1), 130.

[59] N. David, R. Heimann, D. Killick, M. Wayman, Between Bloomeryand Blast Furnace: Mafa Iron-smelting Technology in North Came-roon, Afr. Archaeol. Rev. 1989, 7, 183 – 208.

[60] S. T. Childs, Style, Technology, and Iron Smelting Furnaces in Bantu-speaking Africa, J. Anthropol. Archaeol. 1991, 10 (4), 332.

[61] A. Verlinden, B. Dayot, A Comparison between Indigenous Environ-mental Knowledge and a Conventional Vegetation Analysis in NorthCentral Namibia, J. Arid Environ. 2005, 62 (1), 143.



[62] S. Chirikure, Metals in Society: Iron Production and its Position inIron Age Communities of Southern Africa, J. Soc. Archaeol. 2007, 7 (1),72.

[63] A. Fuhrmann, Uran in Trinkwasser: Der Brunnen Barlissen wurdevom Wasserverband Leine-S�d bis auf Weiteres stillgelegt, G�ttingerTageblatt, 13 September 2008.

[64] A. Y. Aleksanyan, A. N. Podobaev, I. I. Reformatskaya, Steady-stateAnodic Dissolution of Iron in Neutral and Close-to-neutral Media,Protect. Met. 2007, 43 (1), 66.

[65] S. Nesic, Key Issues Related to Modeling of Internal Corrosion of Oiland Gas Pipelines: A Review, Corros. Sci. 2007, 49, 4308.

[66] A. A. Attia, B. S. Girgis, N. A. Fathy, Removal of Methylene Blue byCarbons Derived from Peach Stones by H3PO4 Activation: Batch andColumn Studies, Dyes Pigm. 2008, 76, 282.

[67] C. Noubactep et al., Testing the Suitability of Zerovalent Iron Materi-als for Reactive Walls, Environ. Chem. 2005, 2 (1), 71.

[68] C. Noubactep, Characterizing the Effects of Shaking Intensity on theKinetics of Metallic Iron Dissolution in EDTA, J. Hazard. Mater. 2009,170 (2 – 3), 1149 – 1155.

[69] C. Noubactep et al., Exploring the influence of operational parame-ters on the reactivity of elemental iron materials, J. Hazard. Mater.2009, 172 (2 – 3), 943 – 951.

[70] R. J. Frankel, Design, Construction and Operation of a New FilterApproach for Treatment of Surface Waters in Southeast Asia, J.Hydrol. 1981, 51 (1 – 4), 319.

[71] L. Cornejo et al., In Field Arsenic Removal from Natural Water byZero-valent Iron Assisted by Solar Radiation, Environ. Pollut. 2008, 156(3), 827.

[72] H. Tutu, T. S. McCarthy, E. Cukrowska, The Chemical Characteristicsof Acid Mine Drainage with Particular Reference to Sources, Distri-bution and Remediation: The Witwatersrand Basin, South Africa asa Case Study, Appl. Geochem. 2008, 23 (12), 3666.


Metallic Iron Filters for Universal Access to Safe Drinking Water

Documents

Transcript of Metallic Iron Filters for Universal Access to Safe Drinking Water