Dimensioning metallic iron beds for efficient contaminant removal 1 
2 
3 
4 
5 
6 
7 
8 
9 
10 
11 
12 
13 
14 
15 
16 
17 
18 
19 
20 
21 
22 
23 
24 
25 
26 
Noubactep C.(a,c), Caré S.(b) 
(a) Angewandte Geologie, Universität Göttingen, Goldschmidtstraße 3, D - 37077 Göttingen, Germany. 
(b) Université Paris-Est, Laboratoire Navier, Ecole des Ponts - ParisTech, LCPC, CNRS, 2 allée Kepler, 77420 
Champs sur Marne, France. 
(c) Kultur und Nachhaltige Entwicklung CDD e.V., Postfach 1502, D - 37005 Göttingen, Germany. 
e-mail: cnoubac@gwdg.de; Tel. +49 551 39 3191, Fax: +49 551 399379 
Abstract 
Remediation of contaminated groundwater is an expensive and lengthy process. Permeable 
reactive barrier of metallic iron (Fe0 PRB) is one of the leading technologies for groundwater 
remediation. One of the primary challenges for the Fe0 PRB technology is to appropriately 
size the reactive barrier (length, width, Fe0 proportion and nature of additive materials) to 
enable sufficient residence time for effective remediation. The size of a given Fe0 PRB 
depends mostly on accurate characterization of: (i) reaction mechanisms, and (ii) site-specific 
hydrogeologic parameters. Accordingly, the recent revision of the fundamental mechanisms 
of contaminant removal in Fe0/H2O systems requires the revision of the Fe0 PRB 
dimensioning strategy. Contaminants are basically removed by adsorption, co-precipitation 
and size exclusion in the entire Fe0 bed and not by chemical reduction at a moving reaction 
front. Principle calculations and analysis of data from all fields using water filtration on Fe0 
bed demonstrated that: (i) mixing Fe0 and inert additives is a prerequisite for sustainability, 
(ii) used Fe0 amounts must represent 30 to 60 vol-% of the mixture, and (iii) Fe0 beds are 
deep-bed filtration systems. The major output of this study is that thicker barriers are needed 
for long service life. Fe0 filters for save drinking water production should use several filters in 
series to achieve the treatment goal. In all cases proper material selection is an essential issue. 
Keywords: Drinking water, Inert materials, Iron filter, Multi-filtration, Zerovalent iron. 
 1
1 Introduction 26 
27 
28 
29 
30 
31 
32 
33 
34 
35 
36 
37 
38 
39 
40 
41 
42 
43 
44 
45 
46 
47 
48 
49 
50 
Packed beds with metallic iron (Fe0) are currently used as contaminant mitigating agent in 
several contexts including groundwater remediation, wastewater treatment and drinking water 
production [1-4]. Fe0-based materials are used in particular (i) as reducing agents in 
permeable reactive walls [5-8], and (ii) as reagents to assist biofiltration in household filters 
[3,9,10]. The fundamental process responsible for contaminant removal in both contexts is 
necessarily the oxidative dissolution of Fe0 (iron corrosion) which may be coupled with 
contaminant reduction (reactive walls) or the subsequent precipitation of iron hydroxides 
which may be coupled with contaminant adsorption and co-precipitation (household filters). 
Adsorption, co-precipitation and chemical transformations (oxidation and/or reduction) are 
not mutually exclusive [11-13]. It is obvious that in household filters and reactive walls, a 
synergy between these three processes is responsible for expected and observed 
decontamination. Moreover, these processes proceed in the inter-granular porosity of the 
packed beds which are made up of Fe0 (100 %) or a mixture of Fe0 and an inert material (e.g. 
gravel, pumice, sand) [2,14]. Because of the volumetric expansive nature of the process of 
iron corrosion [15], the porosity of the filtrating systems certainly decreases with increasing 
service life, possibly yielding complete permeability loss system (filter clogging) [16,17]. The 
filling of the pore volume by corrosion products is necessarily coupled with improved size 
exclusion capacity. Therefore, a fourth mechanism for decontamination in packed Fe0 beds is 
identified. 
The very recent concept that adsorption, co-precipitation and size exclusion are the 
fundamental mechanisms of aqueous decontamination in Fe0 packed beds [13] is yet to be 
discussed in the scientific literature. The two main objectives of this communication are (i) to 
give some arguments supporting the new concept, and (ii) to enumerate some consequences 
for the further development of the iron filtration technology. In this effort a particular 
 2
attention is paid to filter dimensioning or bed sizing. For the sake of clarity, the presentation 
will start with the short history of Fe
51 
52 
53 
54 
55 
56 
57 
58 
59 
60 
61 
62 
63 
64 
65 
66 
67 
68 
69 
70 
71 
72 
73 
74 
75 
76 
0 for reactive walls and household filters. 
2 Metallic iron for reactive walls 
The Fe0 reactive wall technology is one aspect of the materialization of the original idea of 
McMurty and Elton [18] that a passive design “using natural groundwater flow and a 
treatment media” can “capture or treat the contaminants without the need for regeneration or 
replacement”. With the publication of this innovative concept in August 1985, an ongoing 
effort for efficient reactive materials for permeable reactive barriers started. In 1990, Gillham 
and his colleagues fortuitously found that corroding Fe0 (reductively) eliminated aqueous 
trichloroethylene [19]. This discovery was the starting point of remediation with elemental 
metals. Elemental metals (e.g. Al0, Fe0, Zn0 and bimetallics) are now recognized as competent 
alternatives for remediation of groundwater that is contaminated with reducible substances 
[8,20]. 
Currently, however, Fe0 has exceeded all expectations because non-reducible substances have 
been quantitatively removed as well. For example, aqueous ZnII which is thermodynamically 
non-reducible by Fe0 has been efficiently removed [21]. The results of Morrison et al. [21] 
attested the synergic effects of removal mechanisms as the investigated systems also 
contained MoVI and UVI. MoVI and UVI could be reduced to less soluble species. Furthermore, 
MoVI known for its poor adsorptive capability onto iron oxides at pH > 5 [22,23] was 
quantitatively removed, suggesting that improved size exclusion might had been effective. 
The concept that contaminants are fundamentally removed by adsorption and co-precipitation 
is consistent with many experimental observations which remained non-elucidated by the 
reductive transformation concept [11,12]. Although researchers are continuing to maintain the 
validity of the latter concept [24-26], the new concept was validated [27,28] and has been 
independently verified [29,30]. As a matter of course the concept of adsorption/co-
precipitation (and size exclusion for packed bed) should have been challenged by researchers 
 3
working on remediation in Fe0/H2O systems. The motivation of using Fe0 at household level 
corroborates the validity of the adsorption/co-precipitation/size exclusion concept.  
77 
78 
79 
80 
81 
82 
83 
84 
85 
86 
87 
88 
89 
90 
91 
92 
93 
94 
95 
96 
97 
98 
99 
100 
101 
3 Metallic iron for household filters 
While using slow sand filtration for water treatment in rural Bangladesh, it was observed that 
the filter efficiency for arsenic removal depends on the iron content of natural waters. Arsenic 
was readily removed from Fe-rich natural waters. Accordingly, Fe0 is used “to provide a 
constant input of iron (soluble or surface precipitate) for groundwater low in soluble iron” 
[32]. The very efficient resulting filter for As removal was the 3-Kolsi filter [10,33]. A typical 
3-Kolsi filter contained a layer of about 3 kg Fe0 (100 % Fe0). However, the 3-Kolsi filter was 
not sustainable as it clogged after some 8 weeks of operation [3,34]. 
The remarkable efficiency of 3-Kolsi filters has prompted researchers to further develop the 
system for improved sustainability [9,10,31,34-37]. The best product is the SONO arsenic 
filter in which the 100 % Fe0 layer is replaced by a proprietary porous Fe0-based material 
(termed as Composite Iron Material - CIM) [32,33]. The two most important features of CIM 
are: (i) its porosity and (ii) its low content of Fe0. In consequence, two opposite effects may 
be observed: (i) the porous structure of the CIM induces a larger reactive surface compared to 
non-porous Fe0 particle (or compact Fe0); the internal porosity could be regarded as magazine 
for in-situ generated iron corrosion products and (ii) less initial Fe0 is used compared to 
compact Fe0 particle. The former effect (larger reactive surface) is well-documented as tool to 
improve Fe0 efficiency and is the rationale for using nano-scale Fe0 for water treatment [38]. 
The latter effect (less initial Fe0) could not improve Fe0 efficiency in term of Fe0 reactivity but 
is known as tool to delay or avoid porosity loss [39-41]. of the filter system, but not the 
second. These observations suggest that the 100 % Fe0 layer in the 3-Kolsi filter was the 
major reason for its too short service life. Leupin and Hug [35] have considerably reduced the 
proportion of Fe0 (1.5 g Fe0 for 60 g sand). More recently, Gottinger [42] demonstrated in a 
 4
pilot study that a volumetric mixture Fe0:sand of 30:70 was very efficient for water treatment 
at a small community level. 
102 
103 
104 
105 
106 
107 
108 
109 
110 
111 
112 
113 
114 
115 
116 
117 
118 
119 
120 
121 
122 
123 
124 
125 
126 
127 
It is important to notice that household Fe0 filters primarily treat water of unknown 
composition. Design efforts are focused on keeping filter permeability. Available filters were 
designed for As removal but SONO filters have efficiently removed several other chemicals 
and pathogens [32,43,44]. It is obvious that Fe0-based filters regarded as “Fe0 assisted sand 
filtration” are not designed to chemically reduce any contaminant. Even arsenics for which 
the majority of household filters were designed is removed by adsorption, co-precipitation 
and size exclusion, although AsV reduction to AsIII and As0 is thermodynamically favorable 
([3,45] and references therein). 
A typical SONO filter contains 5 to 10 kg of porous CIM (CIM: 92 - 94 % Fe, 4 - 5 % C, 1 - 2 
% SiO2, 1 - 2 % Mn, 1 - 2 % S,P) and may function for up to 11 years while filtering waters 
containing up to 1000 μg As/L. It is important to notice that only a fraction of the 92 - 94 % 
Fe in SONO filters is in metallic form (Fe0) and could undergo volumetric expansion. 
Therefore, learning from SONO filters to design efficient Fe0 beds consists in reducing the 
proportion of Fe0 and create place for in-situ generated iron corrosion products. Prior to 
discuss an efficient designing tool, an overview of current design options to limit the impact 
of fouling in Fe0 PRB will be given. 
4 Current design approach to limit Fe0 PRB clogging 
Fe0 PRBs are currently believed to create redox conditions for contaminant degradation or 
immobilization [2]. Accordingly, the precipitation of iron corrosion products and other 
secondary minerals is regarded as perturbing side effect yielding reactivity and porosity loss 
[2,14,16,17]. Accordingly, the design of a PRB requires profound knowledge of local water 
flow velocity (residence time), aquifer porosity, influent contaminant concentration. 
Additionally, the contaminant degradation rate by used Fe0 is usually estimated in laboratory 
and pilot studies and used to size the PRB. Sizing aspects include the amount of Fe0 to be 
 5
used and the thickness of the bed (filter or wall). The first problem with this approach is that 
used Fe
128 
129 
130 
131 
132 
133 
134 
135 
136 
137 
138 
139 
140 
141 
142 
143 
144 
145 
146 
147 
148 
149 
150 
151 
152 
0 media can not be each other compared in reactivity as there is no standard procedure 
to this end [46]. 
Recently, Li and Benson [2] identified and discussed five relevant strategies to limit the 
clogging of Fe0 PRBs: (i) pea gravel equalization zones up gradient and down gradient of the 
reactive zone to equalize flows (strategy 1), (ii) placement of a sacrificial pre-treatment zone 
upstream of the reactive medium (strategy 2), (iii) pH adjustment (strategy 3), (iv) use of 
larger Fe0 particles (strategy 4), and (v) periodic mixing of the Fe0 to break up and redistribute 
secondary minerals (strategy 5). 
In the light of the concept that contaminants are basically removed by adsorption, co-
precipitation and size exclusion, the following comments can be made on the five strategies. 
Strategy 1 necessarily has a double function as quantitative contaminant removal may occur 
in the equalization zones. The same remark is valid for strategy 2 as this study shows that 
reactive zones with 100 % Fe0 are not sustainable. Strategy 3 is recommended because iron 
corrosion is sustained by FeS2 dissolution (or H+ production). Accordingly, FeS2 should be 
regarded as useful reactive additive (Fe0/FeS2 system or Fe0/FeS2/sand system). Hereby, care 
should be taken that the added proportion of FeS2 don’t induce a pH shift below a value of 
5.5. In fact, if the final pH < 5.5 the Fe solubility is increased and the effluent may exhibit too 
high Fe concentration. On the other hand, if dissolve Fe is transported away from the reactive 
zone, the bed porosity will increase and the filtration efficiency will decrease. Another 
positively tested reactive additive is MnO2 [29,47,48]. MnO2 reductive dissolution is driven 
by FeII from Fe0 oxidation, sustaining Fe0 dissolution is beneficial for the decontamination 
process. Strategy 4 will be effective only at certain sites depending on the extent of 
contamination. In fact, larger size Fe0 means larger pore space and poorer size exclusion. 
Finally, Strategy 5 can be rendered superfluous by a proper bed design. 
 6
The approach based on the concept that contaminants are removed by adsorption, co-
precipitation and size exclusion has the advantage that only iron corrosion with site-specific 
water or relevant model water has to be characterized for proper barrier design. Accordingly, 
an aggressive groundwater will rapidly corrode iron, rendering a thin wall satisfactorily. For 
less aggressive waters a thicker wall is necessary to enable completed contaminant removal 
by multi-filtration (see discussion section). The same systematic can be applied to Fe
153 
154 
155 
156 
157 
158 
159 
160 
161 
162 
163 
164 
165 
166 
167 
168 
169 
170 
171 
172 
173 
174 
175 
176 
177 
178 
0 media 
of various reactivity. The less reactive a material in a groundwater, the thicker the reactive 
barrier. Therefore, the selection of the most appropriate Fe0 material at each site is a key issue 
for wall or generally bed efficiency. The next section is focused on better designing Fe0 beds. 
5 Designing Fe0 beds 
The presentation above suggests that Fe0 bed design must be based on the available pore 
volume for volumetric expansion of corroding iron. Accordingly, for a given bed size 
replacing a portion of reactive iron by an inert material is the first tool to extend filter service 
life. The very first additive material in this regard is a non-porous material as quartz (0 % 
porosity). The next step could consist in partly or totally replacing quartz by porous materials 
like sandstone (up to 40 % porosity) or pumice (up to 90 % porosity). In each case a critical 
Fe0:additive ratio must exist for which bed porosity is lost upon Fe0 depletion as illustrated 
below. 
5.1 Sustaining Fe0 bed reactivity by addition of inert materials: Bed design 
A random packed Fe0 bed of identical spheres is considered. The initial bed porosity Φ0 has a 
fundamental value of 36 % [49]. In other words, regardless from the actual dimension of the 
bed, 64 % of the bed volume V is filled by dense Fe0 and 36 % is available as inter-granular 
pore space for corrosion products. It can be noticed that the compactness C of the granular 
medium is C=1-Φ0=Vinitial Fe / V=0.64 where Vinitial Fe is the initial volume of iron. If a 
volumetric fraction of Fe0 is replaced by non-porous quartz (with the same particle diameter), 
36 % of the bed volume is still available for corrosion products but more Fe0 will corrode 
 7
before the bed porosity decreases to zero (Fig. 1). Calculations could enable the identification 
of critical Fe
179 
180 
181 
182 
183 
184 
185 
186 
187 
188 
189 
190 
191 
192 
193 
194 
195 
196 
197 
198 
199 
200 
201 
202 
0:additive ratios. Two hypothetical examples will be used here for illustration: (i) 
a rectangular reactive wall, and (ii) a cylindrical household filter. 
The dimensions of the demonstration reactive wall in Borden (Ontario, Canada) are used for 
the hypothetical reactive wall [5]. The dimensions of the wall were 5.5 x 1.6 x 2.2 m (l x w x 
h), giving a volume V = 19.36 m3. For the hypothetical cylindrical household filter the 
dimensions of field columns used by Westerhoff and James [49] are adopted. The columns 
had a total capacity or volume V = 4.022.10-3 m3 (4.022 L): diameter 7.5 cm and height 91 
cm. 
The filling of the bed porosity  by iron corrosion products can be estimated from a simplified 
modeling (Fig. 1) based on the following assumptions:  
(i) uniform corrosion: the diameter reduction of the particle is the same for all the Fe0 
particles,  
(ii) iron corrosion products are fluid enough to progressively fill available pore space.  
Assuming that the coefficient of volumetric expansion (η) of the iron corrosion products is: 
η = Voxides/VFe     (1) 
where Voxides is the volume of the iron corrosion products and VFe the volume of parent Fe0. 
The surplus volume of the iron corrosion products contributing to porosity loss is V’oxides. Per 
definition V’oxides is the difference between the volume Voxides of iron corrosion products and 
the volume VFe of parent Fe0. V’oxides is given by Eq. 2:  
V’oxides = (η - 1) * VFe     (2) 
Assuming that the bed is clogged when the volume V’oxides is equal to the initial inter granular 
voids (Φ0.V), the volume V Fe, clogging of the consumed iron leading to clogging of the bed is 
then estimated by: 
1
V.=V 0ginglogc,Fe −η
Φ .      (3) 203 
 8
In this case (Eq.3), the volume V Fe, clogging of the consumed iron is inferior to the initial 
volume of dense Fe
204 
205 
206 
207 
208 
0. It means that clogging appears before depletion of Fe. It can be noticed 
that, in some cases, the initial volume of iron may be too low so that there is no clogging and 
the bed porosity is not completely filled by iron corrosion products.  
The residual porosity Φr defined by Φr = V residual voids / V is evaluated by Eq. 4:  
V
V).1( consumedFe0r −η−Φ=Φ     (4) 209 
210 
211 
212 
where VconsumedFe is the volume of Fe which is consumed. When the clogging appears before 
depletion of Fe0, the volume VconsumedFe is given by the equation 3 and the residual porosity is 
equal to Φr=0. When there is no clogging, the volume VconsumedFe is equal to the initial volume 
of Fe and there is residual porosity ( 0r ≠Φ ).  213 
214 
215 
216 
217 
218 
219 
220 
221 
222 
223 
224 
225 
226 
227 
228 
These calculations allow the evaluation of the efficiency of the bed (reactive wall or filter) 
related to the possible clogging. Two cases are discussed in the following. 
5.2 Case of a 100 % Fe0 bed 
Considering that the density of Fe0 is 7,800 kg/m3, the 12.4 m3 (64 % of the total volume) 
available in the hypothetical reactive wall (Tab. 1) can be filled by 96,645 kg of Fe0. The 
calculations in Tab. 2 demonstrated that from this Fe0 amount only a maximum of 50,336 kg 
can be oxidized to yield porosity loss (no residual porosity, Φr=0). The weight proportion of 
consumed Fe0 ranges between 10.4 % and 52.1 % when the main corrosion products are 
Fe(OH)3.3H2O (ferrihyrite) or Fe2O3 (hematite) respectively, showing that 100 % Fe0 reactive 
walls are pure material wastage. The calculations for the hypothetical household filter 
demonstrated that only 2.1 to 10.5 kg of Fe0 will be consumed corresponding to the same 
weight percent like for the hypothetical reactive wall. 
Ideally, when Fe0 is mixed with quartz, a bed containing more than 52.1 wt-% Fe0 of the mass 
of Fe0 necessary to have a 100 % Fe0 bed should not be constructed because bed clogging will 
happen and excess Fe0 will not react (material wastage). The actual Fe0 proportion will 
 9
229 
230 
231 
232 
233 
234 
235 
236 
237 
238 
239 
240 
241 
242 
243 
244 
245 
246 
247 
248 
249 
250 
251 
252 
253 
254 
depend on its intrinsic reactivity and the kinetics of iron oxidative dissolution. Kinetics 
aspects are not considered in this study. 
5.3 Case of a volumetric Fe0:quartz ratio of 50:50 
The calculations above suggests that only about 10.4 to 52.1 wt-% Fe0 is necessary to fill the 
pore space of a 100 % Fe0 filter regardless from the bed dimensions. In this section, the 
calculations are made for a volumetric Fe0:quartz ratio of 50:50. To calculate the 
corresponding weight ratio, one should use the particle size and the densities. However, 
because the same beds (wall and filter) are used, the bed volume occupied by 50 vol-% is 
necessarily one half of the value used in the pure Fe0 bed: (i) 6.20 m3 or 48,322 kg for the 
wall and (ii) 1.3*10-3 m3 (1.3 L) or 10.4 kg for the filter. It is evident that the Fe0 masses 
consumed to yield bed clogging are the same as in the 100 % Fe0 case. The percent 
consumption is then higher (more iron is consumed to obtain the same volume of iron 
corrosion products at Fe0 depletion, Fig. 1 and varies from 21 % for Fe(OH)3.3H2O to 100 % 
for Fe2O3 and Fe3O4. 
The bed containing 50 vol-% Fe0 is necessarily clogged at Fe0 depletion; no residual porosity 
(Φr=0). However, an ideal treatment system should keep a certain residual porosity. This is 
particularly important for subsurface reactive barriers. To warrant a residual porosity (Φr ≠ 0) 
while using a constant Fe0 amount in the bed, it appears that thicker beds have to be 
considered. For example the amount of additive material can be increased such that the 
resulting volumetric proportion of Fe0 is 35 %. Another tool to sustain Fe0 reactivity is to use 
porous additive instead of non-porous quartz. In this way, the total volume for the storage of 
in-situ generated iron corrosion products is increased and the residual bed porosity at Fe0 
depletion is warranted as will be illustrated in the next section. 
5.4 Lengthening Fe0 bed service life by porous additives 
When quartz particles from section 5.3 are replaced by VPP of porous particles (with VPP = V - 
Vinitial Fe), the available porosity Φ0’ for iron corrosion products is increased according to:  
 10
255 Φ0’ = Φ0 + ϕpp . f pp     (5) 
256 where ϕpp  (-) is the critical porosity of the porous particles; 
257 
258 
259 
260 
261 
262 
263 
264 
265 
266 
267 
268 
269 
270 
271 
272 
273 
274 
275 
276 
277 
278 
279 
 f pp  (-) is the porous particle volume fraction (here f pp =VPP/V).  
The volume VFe of the consumed iron leading to clogging of the bed (Eq. 3) or the residual 
porosity Φr (Eq. 4) can be obtained by replacing Φ0 by Φ0’. The calculations in Tab. 3 show 
that it is possible to increase the efficiency of the filtration system. More iron may be 
consumed and transformed into iron corrosion products before clogging. In two cases (Fe2O3 
and Fe3O4), a residual bed porosity is available at Fe depletion.  
Figure 2 shows that replacing quartz by sandstone or more porous (or less dense) materials 
could further extend Fe0 bed service life. This conclusion is justified by the fact that heavier 
materials are less porous. However, the most important feature from Fig. 2 is that weight-
based and volumetric ratios are not linearly dependent. Therefore, the description of any 
experimental design should comprise data on Fe0 and additives (form, density, porosity, size) 
and filter dimensions together with the volumetric proportion of Fe0. This procedure will 
enable or ease comparability of published results. 
5.5 Discussion 
The calculations above have shown that in a 100 % Fe0 bed, system clogging will occur when 
only about 52 wt-% of used Fe0 is consumed. In a 50 % Fe0 bed material depletion (100 % 
consumption) is only possible if the corrosion products are Fe3O4 and Fe2O3 (no residual 
porosity). By replacing quartz by sandstone, a residual porosity Φr about 12 % is obtained 
when the corrosion products are Fe3O4 and Fe2O3. But even in these cases, crystalline Fe3O4 
and Fe2O3 are the final stages of transformations which go through several more volumetric 
amorphous stages (e.g. Fe(OH)2, FeOOH). Accordingly, a volumetric ratio 50:50 should be 
regarded as the highest proportion of Fe0 for long-term efficiency of Fe0 beds. In the literature 
however, a 50:50 weight ratio is usually used based on a pragmatic approach [50]. The 
 11
volumetric 50:50 ratio for the Fe0:quartz mixture (quartz: 2.6 kg/m3) corresponds to a 
Fe
280 
281 
282 
283 
284 
285 
286 
287 
288 
289 
290 
291 
292 
293 
294 
295 
296 
297 
298 
299 
300 
301 
302 
303 
304 
305 
0:quartz weight ratio of 75:25. The suitability of the volumetric ratio in this context arises 
from the fact that the expansive nature of iron corrosion is to be considered. Finally, a 
consideration of the conditions used by O’Hannesin and Gillham [5] and Westerhoff and 
James [50] is made. 
O’Hannesin and Gillham [5] used only 22 wt-% Fe0 in the reactive wall in Borden (Ontario 
Canada). This proportion corresponds to less than 10 vol-% Fe0 showing that the 
demonstration wall at Borden is highly permeable. Accordingly, system clogging due to 
expansive iron corrosion is not expected because the available pore space is by far larger than 
the volume of iron corrosion products at Fe0 depletion. As discussed in section 4, most Fe0 
PRBs content a zone with 100 % Fe0. In some cases “equalization zones” and a “sacrificial 
pre-treatment zone” exist in which Fe0 is mixed with gravel or sand. In recent barriers mixing 
Fe0 and sand is considered as a tool to save expense for Fe0 media [51]. However, the proper 
consideration of the expansive nature of Fe0 corrosion shows that mixing Fe0 and inert 
material is a prerequisite for long service life. 
Westerhoff and James [50] could evidence the difficulty of performing long-term experiment 
with 100 % Fe0 beds. A weight-base 50:50 Fe0:sand bed could perform accurately for several 
months (12 months). Similarly, household 100 % Fe0 filters were abandon because of rapid 
clogging [10,32]. The calculations above rationalize the current renaissance of Fe0 filter 
technology for household filters [52] and its use for small scale water facilities [42,53]. Fe0 
filter technology is regarded as a flexible and affordable technology, which could enable the 
achievement of the Millenium Development Goals (MDGs) for water. This simple technology 
could even enable to achieve universal access to safe drinking water within some few years 
[52]. 
6  General discussion  
6.1 Fe0 bed as adsorptive size-exclusion system 
 12
The consideration of Fe0 beds as adsorptive size-exclusion systems arises from the strong 
adsorptive properties of in-situ generated iron corrosion products [54]. Iron is progressively 
corroded (uniform corrosion) in the whole bed. Contaminants are removed by adsorption, co-
precipitation and size exclusion within the whole bed. Removed contaminants could be 
further chemically transformed (oxidized or reduced). A contaminant that is not removed in 
the entrance zone could be removed deeper in the Fe
306 
307 
308 
309 
310 
311 
312 
313 
314 
315 
316 
317 
318 
319 
320 
321 
322 
323 
324 
325 
326 
327 
328 
329 
330 
331 
0 bed (deep-bed filtration). This 
behaviour is illustrated the best by simple experiments by Leupin and Hug [34]. These 
authors performed an As removal experiment in a series of four filters. Each filter contained 
1.5 g iron and 60 g sand. The system with a total of 6 g iron could efficiently filtered 36 L of 
an aqueous solution containing 500 mg As/L. A close consideration of the filtration efficacy 
pro filtration event showed that less than 20 % (100 mg As/L) of the initial arsenic was 
removed during the first filtration; a much larger fraction (≥ 200 mg As/L) was removed 
during the second filtration, arsenic removal continued during the third and fourth filtration. It 
is important to note that Fe0 was not depleted in the experiments and the filters were not 
clogged. Accordingly, further As removal occurred even though As breakthrough ([As] > 50 
μg/L) was observed. The Fe0 weight percent of 2.4 was necessarily too low for efficient 
filtration in a single event, but has the advantage to avoid the clogging of the filter. However, 
this experiment demonstrated the deep-bed filtration nature of individual Fe0 beds which 
could equally be demonstrated with four sample ports on a single bed. 
For the further illustration of deep-bed filtration nature of Fe0 beds, Fig. 3 compares the 
breakthrough of contaminants in a granular activated carbon (GAC) bed and a Fe0 bed. To be 
treated, water is applied directly to the upper end and allow to flow through the packing bed 
by gravity. 
GAC is inert in water and the adsorption capacity is consumed only by contaminants which 
displace H2O from adsorption sites. Accordingly, the region where contaminant adsorption 
takes place is called the mass-transfer zone (or adsorption front). The region above the 
 13
332 
333 
334 
335 
336 
337 
338 
339 
340 
341 
342 
343 
344 
345 
346 
347 
348 
349 
350 
351 
352 
353 
354 
355 
356 
adsorption front is the saturated zone and the region below is the virgin zone. As a function of 
time, the saturated zone moves through the bed and approaches the end [54]. The adsorption 
bed is exhausted when no more satisfactorily decontamination is achieved. 
On the contrary, in a Fe0 bed, the whole bed is available as sorption, co-precipitation and size 
exclusion system. A sort of “adsorption front” may exist because of increased oxidizing 
agent’s levels in the inflowing solution. However in the whole bed H2O corrodes Fe0 
producing corrosion products for efficiency contaminant removal. Contaminant removal may 
thus occur deeper in the Fe0 bed from the initial stage of bed service on. 
6.2 Significance for system design 
The scientific community has long been searching for common underlying mechanisms for 
the process of contaminant removal in Fe0/H2O systems that provide a confidence for design 
that is non-site-specific [57,58]. This was the idea behind introducing specific reaction rate 
constant (kSA). kSA values are regarded as a more general reactivity descriptor of contaminants 
with Fe0. They are also believed to allow intersystem comparisons [56]. However, there are 
two major problems with the kSA concept: (i) it is contaminant specific, and (ii) it is based on 
the concept of reductive transformation which is definitively not determinant for the process 
of the removal of several contaminants. 
While previous efforts were directed at achieving a significant body of removal rate for 
individual contaminants in order to enable non-site specific bed design, the present study 
suggests that site-specificity will govern material selection. For example, if contaminated 
water is carbonate-rich, it could be advantageous to use a relative low reactive material which 
corrodibility will be sustained by carbonates. Accordingly, if available Fe0 is classified for 
specific conditions, treatability studies may only be required to fine-tune design criteria for 
the optimal Fe0 bed performance.  
7 Concluding remarks 
 14
This study clearly delineates the important role of volumetric expansion of corroding iron for 
the process of contaminant removal in Fe
357 
358 
359 
360 
361 
362 
363 
364 
365 
366 
367 
368 
369 
370 
371 
372 
373 
374 
375 
376 
377 
378 
379 
380 
381 
382 
0 beds and the sustainability of Fe0 beds. 
Sustainability is primarily warrant by admixing Fe0 with non reactive additives to avoid or 
delay porosity loss. The characterization of Fe0 beds by the volumetric Fe0:additive ratios and 
the bed sizes provide a clear starting point for the design of future laboratory, pilot, and field-
scale studies aiming at characterizing remediation Fe0 beds. This certainly has economic 
implications for Fe0 bed design as the use of too high Fe0 amount (e.g. > 60 vol-%) has to be 
avoided. Most importantly results will be more comparable, accelerating progress in 
technology development. 
The most important result from the calculations of this study is that, for a given Fe0 amount, 
necessary for efficient decontamination at a specific contaminated site, building a thicker 
barrier in which iron represents a volumetric proportion of 30 to 45 % is more advantageous 
than a thin barrier containing more than 60 vol-% iron. A further useful tool to extend Fe0 bed 
service live is to use porous additives which allow avoiding/delaying bed clogging. 
The installation of thicker reactive walls in the underground is certainly coupled with elevated 
investment costs. However, thickening the barrier is essential for barrier sustainability (deep-
bed filtration). For household filters and Fe0 beds in water treatment plants [42,53] the 
achievement of multi-filtration is an easier task as for instance, several beds could let to 
operate in series. 
Finally, it should be highlighted that the very first reactive wall constructed at Borden 
(Ontario, Canada) for the demonstration of the feasibility of the new technology contained 
less that 10 vol-% (Fe0) and could never been clogged because the porosity of the system 
could not be filled by expansive iron corrosion products. In other words because of 
insufficient system analysis, the Fe0 reactive wall technology was demonstrated on a very 
permeable system but operating walls are necessarily less permeable. Moreover, mixing Fe0 
and sand was considered as a tool to reduced Fe0 costs [51,52]. It is now demonstrated, that 
 15
mixing Fe0 with inert additives is even the prerequisite for sustainability. It is hoped that the 
huge literature on deep-bed filtration [59,60] will now be used for the further development of 
iron wall technology. 
383 
384 
385 
386 
387 
388 
389 
390 
391 
392 
393 
394 
395 
396 
397 
398 
399 
400 
401 
402 
403 
404 
405 
406 
407 
Acknowledgments 
Sven Hellbach (student research assistant) is acknowledged for technical assistance. The 
manuscript was improved by the insightful comments of anonymous reviewers from 
Chemical Engineering Journal. 
References 
[1] J.W. Lee, D.K. Cha, Y.K. Oh, K.B. Ko, S.H. Jin, Wastewater screening method for 
evaluating applicability of zero-valent iron to industrial wastewater, J. Hazard. Mater. 180 
(2010), 354–360. 
[2] L. Li, C.H. Benson, Evaluation of five strategies to limit the impact of fouling in 
permeable reactive barriers, J. Hazard. Mater. 181 (2010) 170–180. 
[3] M.I. Litter, M.E. Morgada, J. Bundschuh, Possible treatments for arsenic removal in Latin 
American waters for human consumption, Environ. Pollut. 158 (2010) 1105–1118. 
[4] J. Bundschuh, M. Litter, V.S.T. Ciminelli, M.E. Morgada, L. Corneo, S.G. Hoyos, J. 
Hoinkis, M.T. Alarcon-Herrera, M.A. Armienta, P. Bhattacharya, Emerging mitigation needs 
and sustainable options for solving the arsenic problems of rural and isolated urban areas in 
Latin America e A critical analysis, Water Res. (2010) doi:10.1016/j.watres.2010.04.001. 
[5] S.F. O´Hannesin, R.W. Gillham, Long-term performance of an in situ "iron wall" for 
remediation of VOCs, Ground Water 36 (1998) 164–170. 
[6] K.C.K. Lai, I.M.C. Lo, V. Birkelund, P. Kjeldsen, Field monitoring of a permeable 
reactive barrier for removal of chlorinated organics, J. Environ. Eng. 132 (2006) 199–210. 
[6] A.D. Henderson, A.H. Demond, Long-term performance of zero-valent iron permeable 
reactive barriers: a critical review. Environ. Eng. Sci. 24 (2007) 401–423. 
 16
408 
409 
410 
411 
412 
413 
414 
415 
416 
417 
418 
419 
420 
421 
422 
423 
424 
425 
426 
427 
428 
429 
430 
431 
432 
[6] D.H. Phillips, T. Van Nooten, L. Bastiaens, M.I. Russell, K. Dickson, S. Plant, J.M.E. 
Ahad, T. Newton, T. Elliot, R.M. Kalin, Ten year performance evaluation of a field-scale 
zero-valent iron permeable reactive barrier installed to remediate trichloroethene 
contaminated groundwater, Environ. Sci. Technol. 44 (2010), 3861–3869. 
[9] T.K.K. Ngai, R.R. Shrestha, B. Dangol, M. Maharjan, S.E. Murcott, Design for 
sustainable development - Household drinking water filter for arsenic and pathogen treatment 
in Nepal, J. Environ. Sci. Health A42 (2007) 1879–1888. 
[10] D. Pokhrel, B.S. Bhandari, T. Viraraghavan, Arsenic contamination of groundwater in 
the Terai region of Nepal: An overview of health concerns and treatment options, Environ. 
Int. 35 (2009) 157–161. 
[11] C. Noubactep, Processes of contaminant removal in “Fe0–H2O” systems revisited: The 
importance of co-precipitation, Open Environ. J. 1 (2007) 9–13. 
[12] C. Noubactep A critical review on the mechanism of contaminant removal in Fe0–H2O 
systems, Environ. Technol. 29 (2008) 909–920. 
[13] C. Noubactep, The suitability of metallic iron for environmental remediation, Environ. 
Progr. (2010)DOI 10.1002/ep.10406. 
[14] B.D.M. Painter, Optimization of permeable reactive barriers for the remediation of 
contaminated groundwater. Dissertation, Lincoln University, New Zealand (2005). 
[15] S. Caré, Q.T. Nguyen, V. L'Hostis, Y. Berthaud, Mechanical properties of the rust layer 
induced by impressed current method in reinforced mortar, Cement Concrete Res. 38 (2008) 
1079–1091. 
[16] L. Li, C.H. Benson, E.M. Lawson, Impact of mineral fouling on hydraulic behavior of 
permeable reactive barriers, Ground Water 43 (2005) 582–596. 
[17] L. Li, C.H. Benson, E.M. Lawson, Modeling porosity reductions caused by mineral 
fouling in continuous-wall permeable reactive barriers, J. Contam. Hydrol. 83 (2006) 89–121. 
 17
[18] D.C McMurty, R.O. Elton, New approach to in-situ treatment of contaminated 
groundwaters, Environ. Progr. 4/3 (1985) 168
433 
434 
435 
436 
437 
438 
439 
440 
441 
442 
443 
444 
445 
446 
447 
448 
449 
450 
451 
452 
453 
454 
455 
456 
–170. 
[19] G.W. Reynolds, J.T. Hoff, R.W. Gillham, Sampling bias caused by materials used to 
monitor halocarbons in groundwater, Environ. Sci. Technol. 24 (1990) 135–142. 
[20] J.P. Muegge, P.W. Hadley, An evaluation of permeable reactive barrier projects in 
California, Remediation 20 (2009) 41–57. 
[21] S.J. Morrison, D.R. Metzler, B.P. Dwyer, Removal of As, Mn, Mo, Se, U, V and Zn 
from groundwater by zero-valent iron in a passive treatment cell: reaction progress modelling, 
J. Contam. Hydrol. 56 (2002) 99–116. 
[22] S. Goldberg, S.M. Lesch, D.L. Suarez, Predicting molybdenum adsorption by soils using 
soil chemical parameters in the constant capacitance model. Soil Sci. Soc. Am. J. 66 (2002) 
1836–1842. 
[23] L. Brinza, L.G. Benning, P.J. Statham, Adsorption studies of Mo and V onto ferrihydrite, 
Miner. Mag. 72 (2008) 385–388. 
[24] S.-H. Kang, W. Choi, Response to comment on “Oxidative degradation of organic 
compounds using zero-valent iron in the presence of natural organic matter serving as an 
electron shuttle”, Environ. Sci. Technol. 43 (2009) 3966–3967. 
[25] P.G. Tratnyek, A.J. Salter, Response to comment on “Degradation of 1,2,3-
trichloropropane (TCP): Hydrolysis, elimination, and reduction by iron and zinc”, Environ. 
Sci. Technol. 44 (2010) 3198–3199. 
[26] C. Noubactep, Elemental metals for environmental remediation: Learning from 
cementation process, J. Hazard. Mater. 181 (2010) 1170–1174. 
[27] C. Noubactep, Characterizing the discoloration of methylene blue in Fe0/H2O systems, J. 
Hazard. Mater. 166 (2009) 79–87. 
 18
457 
458 
459 
460 
461 
462 
463 
464 
465 
466 
467 
468 
469 
470 
471 
472 
473 
474 
475 
476 
477 
478 
479 
480 
481 
[28] C. Noubactep, A.-M.F. Kurth, M. Sauter, Evaluation of the effects of shaking intensity 
on the process of methylene blue discoloration by metallic iron, J. Hazard. Mater. 169 (2009) 
1005–1011. 
[29] A. Ghauch, H. Abou Assi, A. Tuqan, Investigating the mechanism of clofibric acid 
removal in Fe0/H2O systems, J. Hazard. Mater. 176 (2010) 48–55. 
[30] A. Ghauch, H. Abou Assi, S. Bdeir, Aqueous removal of diclofenac by plated elemental 
iron: Bimetallic systems, J. Hazard. Mater. (2010), doi:10.1016/j.jhazmat.2010.05.139. 
[31] A.H. Khan, S.B. Rasul, A.K.M. Munir, M. Habibuddowla, M. Alauddin, S.S. Newaz, A. 
Hussam, Appraisal of a simple arsenic removal method for groundwater of Bangladesh, J. 
Environ. Sci. Health A35 (2000) 1021–1041. 
[32] A. Hussam, A.K.M. Munir, A simple and effective arsenic filter based on composite iron 
matrix: Development and deployment studies for groundwater of Bangladesh, J. Environ. Sci. 
Health A 42 (2007) 1869–1878. 
[33] A. Hussam, Contending with a development disaster: SONO filters remove arsenic from 
well water in Bangladesh, Innovations 4 (2009) 89–102. 
[34] O.X. Leupin, S.J. Hug, Oxidation and removal of arsenic (III) from aerated groundwater 
by filtration through sand and zero-valent iron, Wat. Res. 39 (2005) 1729–740. 
[35] O.X. Leupin, S.J. Hug, A.B.M. Badruzzaman, Arsenic removal from Bangladesh tube 
well water with filter columns containing zerovalent iron filings and sand, Environ. Sci. 
Technol. 39 (2005) 8032–8037. 
[36] H. Chiew, M.L. Sampson, S. Huch, S. Ken, B.C. Bostick, Effect of groundwater iron and 
phosphate on the efficacy of arsenic removal by iron-amended biosand filters, Environ. Sci. 
Technol. 43  (2009) 6295–6300. 
[37] V. Tellen, G. Nkeng, S. Dentel, Improved filtration technology for pathogen reduction in 
rural water supplies. Water 2 (2010) 285–306. 
 19
482 
483 
484 
485 
486 
487 
488 
489 
490 
491 
492 
493 
494 
495 
496 
497 
498 
499 
500 
501 
502 
503 
504 
505 
506 
[38] C. Noubactep, S. Caré, On nanoscale metallic iron for groundwater remediation. J. 
Hazard. Mater. (2010) DOI: 10.1016/j.jhazmat.2010.06.009. 
[39] G. Bartzas, K. Komnitsas, Solid phase studies and geochemical modelling of low-cost 
permeable reactive barriers. J. Hazard. Mater. (2010) doi:10.1016/j.jhazmat.2010.07.024. 
[40] C. Noubactep, S. Caré, Enhancing sustainability of household water filters by mixing 
metallic iron with porous materials. Chem. Eng. J. (2010) DOI: 10.1016/j.cej.2010.06.012. 
[41] C. Noubactep, S. Caré, F. Togue-Kamga, A. Schöner, P. Woafo, Extending service life 
of household water filters by mixing metallic iron with sand, Clean (2010) (Accepted 
manuscript # clen.201000177). 
[42] A.M. Gottinger, Chemical-free arsenic removal from potable water with a ZVI-amended 
biofilter. Master thesis, University of Regina (Saskatchewan, Canada) (2010) 90 pp. 
[43] S. Tuladhar, L.S. Smith, SONO filter: An excellent technology for save water in Nepal. 
SOPHEN 7 (2009) 18–24. 
[44] S. Tuladhar, Er.B. Man Shakya, A study on the performance of SONO filter in reducing 
different drinking water quality parameters of ground water: A case study in Ramgram 
municipality of Nawalparasi District, Nepal. Paper presented at the Regional Conference on 
Appropriate Water Supply, Sanitation and Hygiene (WASH) Solution for Informal 
Settlements and Marginalized Communities, Kathmandu, Nepal, May 19-21, 2010, (2010) 
297-310. 
[45] J.M. Triszcz, A. Porta, F.S. García Einschlag, Effect of operating conditions on iron 
corrosion rates in zero-valent iron systems for arsenic removal. Chem. Eng. J. 150 (2009) 
431–439. 
[46] C. Noubactep, T. Licha, T.B. Scott, M. Fall, M. Sauter, Exploring the influence of 
operational parameters on the reactivity of elemental iron materials, J. Hazard. Mater. 172 
(2009) 943–951. 
 20
[47] C. Noubactep, G. Meinrath, P. Dietrich, B. Merkel, Mitigating uranium in groundwater: 
Prospects and limitations,
507 
508 
509 
510 
511 
512 
513 
514 
515 
516 
517 
518 
519 
520 
521 
522 
523 
524 
525 
526 
527 
528 
529 
530 
531 
532 
 Environ. Sci. Technol. 37 (2003) 4304–4308. 
[48] D. Burghardt, A. Kassahun, Development of a reactive zone technology for simultaneous 
in situ immobilisation of radium and uranium, Environ. Geol. 49 (2005) 314–320. 
[49] A. Nur, G. Mavko, J. Dvorkin, D. Galmudi, Critical porosity; a key to relating physical 
properties to porosity in rocks, The Leading Edge 17 (1998) 357–362. 
[50] P. Westerhoff, J. James, Nitrate removal in zero-valent iron packed columns. Wat. Res. 
37 (2003) 1818–1830. 
[51] E. Bi, J.F. Devlin, B. Huang, Effects of mixing granular iron with sand on the kinetics of 
trichloroethylene reduction, Ground Water Monit. Remed. 29 (2009) 56–62. 
[52] C. Noubactep, A. Schöner, P. Woafo, Metallic iron filters for universal access to safe 
drinking water. Clean 37 (2009) 930–937. 
[53] A.M. Gottinger, D.J. Wild, D. McMartin, B. Moldovan, D. Wang, Development of an 
iron-amended biofilter for removal of arsenic from rural Canadian prairie potable water, 
(2010) Retrieved from: http://www.mainstreamwater.com/Gottinger%20et%20al.pdf. (Acces 
06.05.2010). 
[54] E. Sikora, D.D. Macdonald, The passivity of iron in the presence of 
ethylenediaminetetraacetic acid I. General electrochemical behavior, J. Electrochem. Soc. 147 
(2000) 4087–4092. 
[55] P. Le Cloirec, C. Faur, Adsorption of organic compounds onto activated carbon - 
applications in water and air treatments, In Environmental Remediation T.J. Bandosz (editor). 
Interf. Sci. Technol. 7 (2006) 375–419. 
[56] T.L. Johnson, M.M. Scherer, P.G. Tratnyek, Kinetics of halogenated organic compound 
degradation by iron metal, Environ. Sci. Technol. 30 (1996) 2634–2640. 
[57] K.L. McGeough, R.M. Kalin, P. Myles, Carbon disulfide removal by zero valent iron, 
Environ. Sci. Technol. 41 (2007) 4607–4612.
 21
533 
534 
535 
536 
537 
538 
539 
540 
541 
[58] R.W. Gillham, Discussion of Papers/Discussion of nano-scale iron for dehalogenation. 
by Evan K. Nyer and David B. Vance (2001), Ground Water Monit. Remed. 21, 41–54. 
Ground Water Monit. Remed 23 (2003) 6–8. 
[59] R.P. Batycky, H. Brenner, Thermal macrotransport processes in porous media. A review, 
Adv. Water Resour. 20 (1997) 95–110. 
[60] A.A. Shapiro, P.G. Bedrikovetsky, A stochastic theory for deep bed filtration accounting 
for dispersion and size distributions, Physica A (2010) doi:10.1016/j.physa.2010.02.049. 
 
 22
Table 1: Mass of material necessary to completely fill the hypothetical treatment units with 
100 % metallic iron. The fundamental porosity of Φ
541 
542 
543 
0 = 36 % is assumed and the 
value of 7,800 kg/m3 is taken for the specific weight of Fe0. 
Unit Vunit VFe Vpores mFe
 (m3) (m3) (m3) (kg) 
Filter 0.004 0.0026 0.0014 20.08 
Wall 19.4 12.4 6.97 96,645
544 
545 
546 
 
 
 23
Table 2: Mass (mwall or mfilter in kg) of iron and weight proportion of consumed iron (P in %, 
same value for the wall or the filter) leading to porosity loss in the hypothetical 
field reactive wall and household filter as function of the nature of corrosion 
products. Φ
546 
547 
548 
549 
550 
551 
r is the residual porosity (in this case Φr = 0 and iron is not completely 
consumed, P < 100%). Voxid/VFe values are expansive coefficients from Ref. [15]. 
 
Oxid Voxid/VFe mwall mfilter P Φr 
  (kg) (kg) (%) (%) 
1/2 Fe2O3 2.08 50,336 10.45 52.1 0 
1/3 Fe3O4 2.12 48,538 10.08 50.2 0 
γ-FeOOH 3.03 26,779 5.56 27.7 0 
β-FeOOH 3.48 21,920 4.55 22.7 0 
Fe(OH)2 3.75 19,768 4.11 20.5 0 
α-FeOOH 3.91 18,681 3.88 19.3 0 
Fe(OH)3 4.2 16,988 3.53 17.6 0 
Fe(OH)3.3H2O 6.4 10,067 2.09 10.4 0 
552 
553 
 
 24
Table 3: Weight proportion P of consumed iron leading to porosity loss (Eq. 3) or residual 
porosity Φ
553 
554 
555 
556 
557 
558 
559 
r (Eq. 4) as function of the nature of corrosion products for 
Fe0:sandstone with a volumetric ratio 50:50. Voxid/VFe values are expansive 
coefficient from Ref. [15]. The critical porosity of sandstone is 40% and its 
specific weight is 2.0 kg/m3. The results are the same for the reactive wall and the 
household filter.  
 
Oxid Voxid/VFe P Φr
  (%) (%) 
1/2 Fe2O3 2.08 100 14.2 
1/3 Fe3O4 2.12 100 12.9 
γ-FeOOH 3.03 75.1 0 
β-FeOOH 3.48 61.5 0 
Fe(OH)2 3.75 55.5 0 
α-FeOOH 3.91 52.4 0 
Fe(OH)3 4.2 47.7 0 
Fe(OH)3.3H2O 6.4 28.2 0 
560 
561 
562 
 
 
 25
Figure 1  562 
563 
564 
 
 
(a) volumetric ratio Fe0:quartz 100:0 (b) volumetric ratio Fe0:quartz 50:50 
565 
566 
567 
 
 
 26
Figure 2 567 
568  
0 20 40 60 80 100
0
20
40
60
80
100
 Additive (vol-%)
 quartz
 sandstone
 activated carbon
 pumice
ad
di
tiv
e 
/ [
w
t-%
]
Fe0 / [vol-%]
 569 
570 
571 
572 
573 
 
 
 
 27
Figure 3 573 
574  
 575 
576 
 28
Figure captions576 
577 
578 
579 
580 
581 
582 
583 
584 
585 
586 
587 
588 
589 
590 
591 
592 
593 
 
Figure 1: Schematic illustration of the impact of mixing Fe0 and quartz for the long-term 
reactivity of Fe0 beds (clogging). When Fe0 is mixed with quartz more iron corrodes and the 
initial porosity if progressively filled with porous iron oxides for water multi-filtration. 
 
Figure 2: Variation of the weight percent of additive materials as function of the Fe0 
volumetric ratio. Due to the differences in density, there is no linear dependence. The depicted 
variation of the wt-ratio depends on the material density. Used density values are: Fe0: 7.80 
g/cm3, quartz: 2.65 g/cm3, sandstone: 2.00 g/cm3, activated carbon: 1.47 g/cm3, and pumice 
0.64 g/cm3. 
 
Figure 3: Comparison of the evolution of contaminant loading in granular activated carbon 
(GAC - up) and Fe0 (down) filters. The evolution of the GAC filters is virgin - preloaded 
(reaction front) and saturated carbon. For the Fe0 filters a reaction front may exist due to 
increased O2 in the influent but iron corrosion by H2O (or H+) occurs uniformly in the whole 
column. The light grey shadow indicates progressive Fe0 corrosion by water. 
 
 29