Modelling the permeability loss of metallic iron water filtration systems 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 
27 
28 
Caré S.(a), Crane R.(b), Calabrò P.S.(c), Ghauch A.(d), Temgoua E.(e), Noubactep C.(f,g),*
(a) UniversitéParis Est, Laboratoire Navier (ENPC/IFSTTAR/CNRS), 2 allée Kepler, F- 77420 Champs sur 
Marne, France (sabine.care@ifsttar.fr). 
(b) Interface Analysis Centre, University of Bristol, 121 St. Michael’s Hill, Bristol, BS2 8BS, UK. 
(Richard.Crane@bristol.ac.uk) 
(c) Università degli Studi Mediterranea di Reggio Calabria, MECMAT, Mechanics and Materials Department, 
Faculty of Engineering, Via Graziella, loc. Feo di Vito, 89122 Reggio Calabria, Italy. (paolo.calabro@unirc.it) 
(d) American University of Beirut, Faculty of Arts and Sciences, Department of Chemistry, P.O. Box 11-0236 
Riad El Solh–1107-2020 Beirut, Lebanon. (ag23@aub.edu.lb) 
(e) University of Dschang, Faculty of Agronomy and Agricultural Science, P.O. Box 222 Dschang, Cameroon. 
(emile.temgoua@univ-dschang.com) 
(f) Angewandte Geologie, Universität Göttingen, Goldschmidtstraße 3, D - 37077 Göttingen, Germany. 
(g) Kultur und Nachhaltige Entwicklung CDD e.V., Postfach 1502, D - 37005 Göttingen, Germany. 
Correspond author; e-mail: cnoubac@gwdg.de; Tel. +49 551 39 3191, Fax. +49 551 399379 
 
Abstract 
Over the past 30 years the literature has burgeoned with in-situ approaches for groundwater 
remediation. Of the methods currently available, the use of metallic iron (Fe0) in permeable 
reactive barrier (PRB) systems is one of the most commonly applied. Despite such interest, an 
increasing amount of experimental and field observations have reported inconsistent Fe0 
barrier operation compared to contemporary theory. In the current work, a critical review of 
the physical chemistry of aqueous Fe0 corrosion in porous media is presented. Subsequent 
implications for the design of Fe0 filtration systems are modelled. The results suggest that: (i) 
for the pH range of natural waters (> 4.5), the high volumetric expansion of Fe0 during 
oxidation and precipitation dictates that Fe0 should be mixed with a non-expansive material; 
(ii) naturally-occurring solute precipitates have a negligible impact on permeability loss 
compared to Fe0 expansive corrosion; and (iii) the proliferation of H2 metabolising bacteria 
 1
may contribute to alleviate permeability loss. As a consequence, it is suggested that more 
emphasis must be placed on future work with regard to considering the Fe
29 
30 
31 
32 
33 
34 
35 
36 
37 
38 
39 
40 
41 
42 
43 
44 
45 
46 
47 
48 
49 
50 
51 
52 
53 
54 
0 PRB system as a 
physical (size-exclusion) water filter device. 
Keywords: Deep-bed filtration, Hydraulic conductivity, Modelling, Permeability loss, 
Zerovalent iron. 
Acronym List 
ITRC  Interstate Technology & Regulatory Council 
PRB  Permeable reactive barrier 
RZ  Reactive Zone 
ZVI  Zerovalent iron 
 
1 Introduction 
Permeable reactive barriers containing metallic iron as a reactive filler material (Fe0 PRBs) is 
an established technology for groundwater remediation [1-10]. At present, more than 120 Fe0 
PRBs have been installed worldwide, and effective performance has typically been reported 
[10-13]. Fe0 PRBs typically contain either pure Fe0 or a mixture of Fe0 and another material, 
such as gravel or sand. The incorporation of a secondary material is typically employed either 
to meet design requirements, cost, or to limit permeability loss. In such cases, potential 
drawbacks on the kinetics of contaminant removal must be considered [14]. However, some 
available experimental results from batch [15,16] and column [14,17] studies suggest that 
admixing pumice/sand to Fe0 is beneficial for the process of contaminant removal. Therefore, 
the recent statement of Ulsamer [13] that “there is no conclusive evidence that a sand/iron mix 
is better or worse than a pure iron barrier” can be considered the current state-of-the-art. 
In addition, the challenge of determining the fundamental mechanisms which govern 
hydraulic conductivity (permeability) loss is yet to be properly addressed [10,11,13,18,19]. At 
present it is suggested that the mechanism of permeability loss in Fe0 PRBs is due to the 
 2
accumulation of insoluble minerals within the PRB pore network [10,13]. Relevant minerals 
include siderite (FeCO
55 
56 
57 
58 
59 
60 
61 
62 
63 
64 
65 
66 
67 
68 
69 
70 
71 
72 
73 
74 
75 
76 
77 
78 
79 
80 
3), aragonite (CaCO3), and iron (hydr)oxides (e.g. Fe(OH)2, Fe(OH)3, 
FeOOH, Fe2O3, Fe3O4) [10,13,20-26]. Another mechanism reported attributes the 
permeability loss to the build-up of H2 gas, formed due to the hydrolysis of water during Fe0 
corrosion [11,27,28]. However, as H2 is a key source of energy for numerous different 
microorganism species [23,29-31], the contribution of H2 to the process of Fe0 PRB 
permeability loss has been ascribed as minor [32]. 
The theory that Fe0 PRB permeability loss is predominantly due to the accumulation of 
insoluble minerals within pore volumes was recently challenged by Henderson and Demond 
[11]. The authors cited that whilst natural groundwater constituents (e.g. carbonates) and 
contaminant species can occur in subsurface concentrations of several hundred parts per 
million (or mg per litre), the mass/volume occupied by the mineral precipitates will be minor 
compared to the large amount required to significantly impair the permeability of an average 
permeable reactive barrier system. Based on this premise they attributed the permeability loss 
to the accumulation of H2 gas, and suggested periodical venting to prevent build-up. All 
studies to date, however, have overlooked the role of the volumetric expansive iron corrosion 
products [33-36] in PRB permeability loss. 
In the current work, a multidisciplinary theoretical approach has been applied to analyse the 
relationship between the extent of Fe0 depletion and permeability loss in Fe0 beds (including 
water filters and PRBs), by linking: contemporary knowledge of the mechanisms which 
govern contaminant removal by Fe0 [37]; with mathematical modelling mass conservation 
equations. Much of the impetus for this work originates from recent work summarized in 
Noubactep et al. [38] wherein the advantages of admixing non-expansive materials to Fe0 
within Fe0 filtration systems are discussed. For the sake of clarity, the basic conservation 
equation for the oxidative dissolution of iron will be given. 
2 Conservation equation of iron corrosion at pH > 4.5 
 3
2.1 Basic conservation equation 81 
82 
83 
84 
85 
86 
87 
88 
89 
90 
91 
92 
93 
94 
95 
96 
97 
98 
99 
100 
101 
102 
103 
104 
105 
The basic constitutive equation expressing the overall conservation of mass of any chemical 
element (j) consumed in a chemical reaction relates volume (V), dry bulk density (ρ), and 
chemical composition (C) and mass fluxes (m) into or out of a system [39]: 
VpρpCj,p/100 + mj,flux = VwρwCj,w/100    (1) 
The first term of Eq. 1 expresses the mass of element j, contained in the original material 
before reaction, subscripted as p. It is given by the product of volume (V in cm3), dry bulk 
density (ρ in g/cm3), and elemental concentration (C in weight %). The mass of element j 
introduced into or out of the considered volume is indicated as mj,flux and is added to the mass 
of j in the system. Fluxes (mj,flux) are positive if they enter the system and negative if they exit 
the system. On the right-hand side of Eq. 1, the mass of element j contained in the volume of 
reaction products, subscripted w, is given by the product of the new volume, dry bulk density, 
and element concentration. 
2.2 Conservation equation of iron corrosion at pH > 4.5 
For iron corrosion, the element of concern is iron (j = Fe) which is distributed between the 
original metallic iron (Fe0 = ZVI) and various iron hydroxides and oxides (w = ox). Eq. 1 can 
therefore be written as: 
VZVIρZVICFe,ZVI + mFe,flux = VoxρoxCFe,ox   (2) 
For pH > 4.5 the solubility of iron is very low and the flux of Fe (mFe,flux) can be largely 
neglected assuming that water flow rate is slow enough that the dissolution/precipitation 
reactions are at pseudo-equilibrium. Eq. 2 can be re-written as: 
VZVIρZVICFe,ZVI = VoxρoxCFe,ox    (2a) 
Eq. 2a suggests that Vox (iron oxide) must be larger than VFe (metallic iron) because all iron 
(hydr)oxides are less dense than Fe0 (Tab. 1). 
2.3 Volumetric strain 
 4
With regard to iron corrosion driven volume changes, there are three possibilities: (i) 
volumetric compression (V
106 
107 
108 
109 
110 
111 
112 
113 
114 
115 
116 
117 
118 
119 
120 
121 
122 
123 
124 
125 
126 
127 
128 
129 
ZVI > Vox), (ii) isovolumetric transformation (VZVI = Vox), and (iii) 
volumetric expansion (VZVI < Vox). Accordingly, volumetric changes should be determined 
experimentally. This is accomplished by using the classical definition of strain, ε, the ratio of 
volume change in a process to the initial volume (Eq. 3): 
ε = (Vox – VZVI)/VZVI = (Vox/VZVI) – 1    (3) 
Eq. 3 suggests that the volumetric strain is positive because Vox/VZVI ≥ 2.1 [35]. 
In the next section, a new approach for the discussion of permeability loss will be given. This 
exercise will be based on the recent paper by Henderson and Demond [11]. 
3 Permeability loss in Fe0/H2O systems 
The purpose of this section is to discuss the relative importance of mineral precipitation, gas 
production and expansive iron corrosion for permeability loss in Fe0/H2O systems. Expansive 
iron corrosion products included rust. To this end, the species discussed by Henderson and 
Demond [11] will be considered individually (Table 1). 
A cylindrical column apparatus for Fe0 filtration has an internal diameter (D), a reactive 
length (Hrz), and a subsequent total volume Vrz (Vrz = π*D2*Hrz/4). Hrz may be a fraction of 
the length of the column apparatus (Hrz ≤ H). A column may also contain several reactive 
zones. The discussion herein is limited to a single reactive zone. The ratio of the initial 
volume of the void space (inter-particular porosity) is Φ0 and the volume of pore is Vp = 
Φ0*Vrz. The volume occupied by the solid particles Vsolid is Vsolid = (1 – Φ0)*Vrz. Solid 
particles include Fe0 and additives (e.g. gravel, pumice, sand), assumingly having the same 
size and shape (roundness or sphericity). The following cylindrical column apparatus used by 
Henderson and Demond [11] is considered: D = 5 cm, Hrz = H = 25 cm, a subsequent Vrz = 
491.1 cm3, and initial porosities (Φ0) of 0.62. Φ0 = 0.62 is also from ref. [11]. 
 5
130 
131 
132 
133 
134 
135 
136 
137 
138 
139 
140 
141 
142 
143 
144 
145 
146 
147 
148 
149 
150 
151 
152 
153 
154 
The challenge of the current work is to evaluate which quantity of each fouling species (iron 
corrosion products) is necessary to occupy the initial pore volume (Vp). 
3.1 Filling the pore volume with individual minerals and H2 gas 
In this section, Eq. (4) assumes that Fe0 is oxidized by water. The initial pore volume (Vp) is 
filled entirely by corrosion products (H2 and FeII/FeIII species): 
x Fe0  +  y H2O ⇒  FexOy  +  y H2↑  (4) 
t0 = 0  n0     0   0 
t > t0  n0*(1 – α)    n0*α/x  y*n0*α/x 
t > t0  x*n’0*(1 – α)    α*n’0   y*n’0*α 
It is considered that the number of moles (n0) of Fe0 at time t = 0 (t0) is a multiple of n’0 (n0 = 
x*n’0). Accordingly, at t0, the reactive zone contains only x*n’0 Fe0 (no oxide and no 
hydrogen). At each time t (t > t0), the residual number of moles of Fe0 is x*n’0*(1 – α), the 
number of mole of generated oxide is α*n0/x = α*n’0 and the number of mole of H2 is 
y*n0*α/x= y*n’0*α where α is the fraction of the initial amount of Fe0 which is depleted as a 
function of time (t). For iron hydroxides (Fe(OH)n) and carbonate (FeCO3), the stoichiometry 
of oxygen is taken as the value of “x” (y = x) because each mole of Fe releases one mole of 
H2 (for n = 2).  
Knowing the molar volume of individual oxides and H2 (Tab. 1), the degree of occupation of 
the initial pore volume (Vp) can be evaluated. The reactive zone is clogged when enough 
corrosion products (FexOy and H2) are produced to completely fill Vp. In other words, bed 
clogging corresponds to Eq. (5): 
VZVI + Vox + VH2 = Vrz    (5) 
The volume Vi occupied by a species i, is the product of its molar volume by the number of 
moles. The equation of the clogging can be written as (Eq. 5a): 
Vm,ZVI*x*n’0(1 – α) + Vm,ox*n’0*α + Vm,H2*n’0*y*α = Vrz  (5a) 
 6
155 
156 
157 
158 
159 
160 
161 
162 
163 
164 
165 
166 
167 
168 
169 
170 
171 
172 
173 
174 
175 
176 
177 
178 
179 
To have α values for individual oxides, it is sufficient to solve Eq. 5a. The solution is given 
by Eq. 5b: 
α = [Vrz/x*n’0 – Vm,ZVI] / [Vm,ox/x + y/x*Vm,H2 –Vm,ZVI] (5b) 
The porosity of granular sandy beds used in sand filters ranges from 0.40 to 0.50 (average of 
0.45) [41]. The porosity of the filtration bed depends on several factors including grain size, 
grain size distribution and shape (sphericity) of used particles [42]. The sphericity of the 
medium is a measure of its roundness and ranges from 0.70 (angular grains) to about 0.90 
(grains rounded by water or wind) [41,43]. 
The volume of Fe0 (VFe) in the pure Fe0 system (100 % Fe0) depends on the compactness C or 
the porosity Φ (VFe= C*VRZ = (1-Φ)* VRZ). Reported operational values for the porosity of 
Fe0 systems vary between 40 and 70 % [10,11,15]. Calculations are made for the extreme 
values of the porosity reported in peer-reviewed journal articles (36 and 62 %). Φ0 = 36 % 
corresponds to the ideal case of spherical materials [38]. The initial number of moles of Fe0 
(n0) corresponding to the extreme cases are 41.4 (Φ0 = 36 %) and 24.6 (Φ0 = 62 %). 
Calculations (Tab. 2) showed that if H2 does not escape from the reactive zone, the 
consumption of less than 0.1 % of the initial amount of Fe0 will be sufficient to clog the 
systems. If this was likely to occur, the Fe0 filtration technology would have not been 
possible. 
Calculations assuming total escape of H2 gas out of the reactive zone (Vm,H2 = 0 in Eq. 5b) 
indicate that 16 to 62 % of Fe0 can be depleted just at system clogging (Φt∞ = 0 %) when the 
initial porosity is 36 %. For Φ0 = 62 %, 46 to 100 % Fe0 could be depleted just at system 
clogging (α ≥ 0.46). In other words, the sustainability of a Fe0 filtration system depends 
strongly from its initial porosity (Φ0). 
The results herein suggest that, for Φ0 = 36 %, when the main corrosion product is Fe3O4, 
only 58 % of Fe0 is consumed just at system clogging. This value in agreement with the value 
 7
of 51 % reported in former works [38]. The difference corresponds to different values used 
for the volumetric expansion coefficient (η); η = 1.97 herein vs. η = 2.08 in ref. [37]. 
However, this approach fails to consider the in-situ generation of colloidal Fe
180 
181 
182 
183 
184 
185 
186 
187 
188 
189 
190 
191 
192 
193 
194 
195 
196 
197 
198 
199 
200 
201 
202 
203 
204 
II/FeIII species 
and their further transformation to hydroxides and oxides [44,45]. 
Eq. 5b describes a pure iron bed (100 % Fe0). In the case that Fe0 is admixed with a non 
expansive additive (e.g. gravel, pumice, sand) the initial number of moles of iron (n0) has to 
be corrected to the fraction of n0 corresponding to the volumetric proportion of Fe in the 
reactive zone, e.g. n0/2 for a system containing 50 % Fe0 (v/v) and the balance amount of a 
non porous material.  
The results from Tab. 2 suggest that, at Φ0 = 36 %, pure Fe0 beds are not sustainable as a rule 
(see section 3.2.2). For larger initial porosity (Φ0), more sustainable systems are obtained. 
This result was already theoretically achieved by admixing Fe0 and porous media (e.g. 
pumice). However, increased initial porosity as discussed here results from the geometry 
(size, shape) of used media (e.g. Fe0, sand, gravel). 
The influence of the shape of the Fe0 particles on the Fe0 bed porosity is schematically 
represented in Fig. 1 as spherical (left) and cylindrical (right) Fe0 particles (black) are 
progressively transformed to oxides (grey - rust). Fig. 1 confirms the fact that packed beds of 
spherical media are the most compact [46-49]. This delineates the importance of 
characterizing Fe0 and sand materials for their uniformity and sphericity and the resulting bed 
for its compactness (porosity).  
 Another important feature seen in Tab. 2 (α and α’ values) is that regardless from the 
abundance of Fe0 in the system, bed clogging due to H2 gas production is likely to occur prior 
to the consumption of 0.1 % Fe0. However, under the experimental conditions considered by 
Henderson and Demond [11], gas accumulation is unlikely since the solutions were pumped 
in upflow at a flow rate of 0.7 mL/min into the columns. In addition, under field conditions, 
 8
H2 consuming bacteria are ubiquitous [29]. In such cases, clogging is therefore more likely to 
result from enhanced (bio-)stimulation (biofilm growth) and not from H
205 
206 
207 
208 
209 
210 
211 
212 
213 
214 
215 
216 
217 
218 
219 
220 
221 
222 
223 
224 
225 
226 
227 
228 
229 
230 
2 accumulation. 
The estimations in this section clearly show that if H2 was primarily responsible for bed 
clogging, then it is unlikely that the Fe0 PRB technology would have been effective on 
medium to long-term timescales as observed in the field. PRB clogging would have been 
prevalent before a fraction (less than 0.1 %) of the Fe0 had corroded. However, H2 gas may 
contribute to permeability loss in association with particle ‘cementation’ (compaction) by 
nascent iron hydroxides. In this case, compaction prohibits H2 escape and increases flow 
resistance for pumped solutions.  
3.2 The process of permeability loss in Fe0/H2O systems 
In this section, a contemporary evaluation of permeability loss in the Fe0/H2O system is given. 
The methodology is explicitly presented in ref. [38]. In the current work the following 
assumptions apply: 
(i) Uniform Fe0 corrosion: the radius reduction of spherical or cylindrical Fe0 particles is the 
same for all particles;  
(ii) the volume of the reactive zone (Vrz) remains constant and the volume of granular 
materials (e.g. sand) is not modified by the corrosion process; 
(iii) Fe0 corrosion products are fluid enough to progressively fill available pore space.  
As shown in section 2.3, iron corrosion occurs with concurrent volumetric expansion (η = 
Vox/VZVI > 1). The excess volume of Fe0 imbued by corrosion product formation is given by 
Vexcess in Eq. 6. By definition, Vexcess is the difference between Vox and VZVI (Eq. 6). 
Vexcess = (η - 1) * VZVI     (6) 
The Fe0 filtration system is clogged when the volume Vexcess is equal to the initial inter 
granular voids (Vp). “VZVI” in Eq. 6 represents the volume of Fe0 in a pure Fe0 bed. However, 
as discussed in sections 1 and 3.1, Fe0 should be only a fraction of Vsolid (VZVI = τZVI*Vsolid, 
with τZVI ≤ 1). Eq. 6 can be rewritten as: 
 9
231 
232 
233 
234 
235 
236 
237 
238 
239 
240 
241 
242 
243 
244 
245 
246 
247 
248 
249 
250 
251 
252 
253 
254 
255 
256 
(η - 1) * τZVI*Vsolid = Vexcess     (6a) 
Eq. 6a suggests that, for every η value (i.e. every oxide), Vexcess is a linear function of τZVI. To 
find out at what extent τZVI contributes to complete pore filling, it is sufficient to graphically 
solve Eq. 6a for Vexcess = VP. Practically, there are two equivalent approaches: (i) solving 
Vexcess – Vp = 0 or (ii) solving Vexcess/VP = 1. The second approach is adopted in this work. 
The solution of Eq. 6a (clogging) is the interception of the line Vexcess/VP = f(τZVI) with the line 
100 % (Fig. 2). Before discussing the actual evolution of the porosity, some fundamental 
aspects for the solution of Eq. 6a will be given. 
3.2.1 Fe0 filtration systems 
To date, Fe0 particles have been widely reported as successful for water treatment [50-53]. 
However, a holistic understanding of the Fe0/H2O system is yet to be achieved.  
Fig. 2 represents the principle of Fe0 filtration beds. The origin (point O) represent a Fe0-free 
filter (e.g. activated alumina, activated carbon, gravel, pumice, sand, zeolite) and point 
I(100,100) represents an “ideal Fe0-based filter” which becomes 100 % clogged concurrent 
with 100 % Fe0 depletion (Vexcess/VP = 1). The line OI divides the graph into two halves. 
Below OI, Vexcess/VP < 100 and the system is not clogged at Fe0 depletion. Above OI, 
Vexcess/VP > 100 and the system is clogged before Fe0 depletion (a proportion of Fe0 is 
wasted). Thus, Fig. 2 can be regarded as a useful reference tool for future work within this 
field. Relevant parameters to complement Fig. 2 that will be investigated in future work 
include: (i) the intrinsic reactivity of Fe0; (ii) the shape and size of Fe0; (iii) the shape and size 
of sand; (iv) the dimensions and the geometry of the Fe0 bed; (v) the thickness of the Fe0/sand 
layer; (vi) the proportion of Fe0 in the Fe0/sand layer; and (vii) the water flow velocity.  
Point O in Fig. 2 represents all filtration designs without Fe0 (or another metallic element). 
These include conventional slow sand filters (SSF), biosand filters (BSF) and iron oxide-
coated sand filters. Considering filtration designs which entirely contain sandy materials, 
point O can be limited to BSF and SSF. SSF have been used for water treatment since 1840 in 
 10
Dijon/France by Henry Darcy [54]. BSF have been used for water treatment at household 
level for over 20 years [55,56]. However, despite intensive research on BSF, their operating 
mode is yet to be completely understood [42,43,55]. For example, there are no established 
comprehensive design criteria for BSF [41,55,57]. Accordingly, the reproducibility and 
comparison of reported results from one setting to another is problematic. To fill this gap, 
Kubare and Haarhoff [41] have provided the most recent systematic review for a rational 
design of BSF. A Fe
257 
258 
259 
260 
261 
262 
263 
264 
265 
266 
267 
268 
269 
270 
271 
272 
273 
274 
275 
276 
277 
278 
279 
280 
281 
282 
0 filtration system (Fig. 2) can be regarded as a modification of a SSF or 
BSF (point O). Therefore, it is essential to carefully develop rational and comprehensive 
engineering design criteria. In this effort, designing tools for BSF would be very helpful 
[41,55]. 
3.2.2 The role of initial porosity in Fe0 bed clogging 
The theoretical discussion of Fe0 PRB porosity until now was focused on the case of 
maximum compactness for which the initial porosity is 0.36 (Φ0 = 36 %) [32]. For such 
systems, a pure Fe0 bed is clogged when less than 60 % of the initial amount of Fe0 is 
depleted (section 3.1). According to Fig. 1, for Φ0 = 36 %, all Fe0 beds are situated above line 
OI. However, significantly larger porosity values have typically been reported in the 
literature, the highest being 62 % by Henderson and Demond [11]. Accordingly, this section 
discusses the evolution of the porosity of a conventional sand filter (0 % Fe0) as it is 
progressively transformed to a pure Fe0 filter (100 % Fe0). Particular attention is paid to the 
extreme values of the porosity (Φ0 = 36 and 62 %). The results are summarized in Fig. 3. 
The ideal line OI is not represented in Fig. 3 for clarity. Instead the point I(100,100) is 
represented. Fig. 3a (Φ0 = 0.36) shows clearly that all systems are clogged before Fe0 
depletion has been occurred. In contrast, Fig. 3b shows that, for an initial porosity (Φ0) of 
0.62, Fe0 beds are sustainable if magnetite (Fe3O4), maghemite (γ-Fe2O3) and hematite 
(Fe2O3) are the sole iron corrosion products. Additionally, it shows that ferrous hydroxide 
(Fe(OH)2) is the “ideal corrosion product” for the Fe0 PRB to clog concurrent with Fe0 
 11
depletion. With the formation of ferrous hydroxide, magnetite, maghemite and hematite being 
more prevalent in anoxic conditions, it can therefore be stated that Fe
283 
284 
285 
286 
287 
288 
289 
290 
291 
292 
293 
294 
295 
296 
297 
298 
299 
300 
301 
302 
303 
304 
305 
306 
307 
0 PRBs are most ideally 
suited for oxygen depleted or anoxic conditions. 
Magnetite (xFe = 72.4 %, Tab.1) may result from Fe(OH)2 dehydration under anoxic 
conditions and is therefore the sole mineral, that is likely to be quantitatively generated from 
anoxic Fe0 corrosion. 
3.2.3 Discussion 
The presentation until now has focused on the evolution of the permeability loss as a key 
factor for the sustainability of Fe0 PRB systems. A Fe0 filtration system is sustainable only if 
it can maintain hydraulic (permeability) performance while also remaining effective for 
pollutant removal. In other words, a permeable but non reactive Fe0 filtration system is 
useless.  
A Fe0 filtration system can be considered both a chemical and physical water filter device, 
with its efficacy dictated by progressive expansion/compression cycles during aqueous 
corrosion [52]. In a Fe0 filtration system, chemical reactions included (i) iron oxidative 
dissolution, (ii) polymerisation of iron hydroxides and, (iii) subsequent precipitation of 
hydroxides and oxides. Quantitative chemical transformations (oxidation/reduction) of 
dissolved species may also occur. However, resulted species must be removed from the 
aqueous phase by a physical process: adsorption, occlusion, size-exclusion. Accordingly, Fe0 
is not a strong reducing agent under environmental conditions as widely accepted [5-7,10]. 
More importantly, reduction is not a stand alone contaminant removal mechanism [58-61]. 
Rather, Fe0 is a generator of contaminant scavengers for reactive filtration [44,62-65]. While 
adsorptive filtration has been mostly used for metal removal [62-65], the affinity of organic 
compounds for iron hydroxide/oxides (corrosion products) is well documented [66-71]. For  
example, Saha et al. [71] investigated the adsorptive removal of seven different dyes on iron 
 12
308 
309 
310 
311 
312 
313 
314 
315 
316 
317 
318 
319 
320 
321 
322 
323 
324 
325 
326 
327 
328 
329 
330 
331 
332 
333 
oxide nanoparticles an reported on enhanced adsorption capacity of the dyes containing 
hydroxyl (-OH) (erichrome black-T, bromophenol blue, bromocresol green, and fluorescein). 
For the proper scaling of Fe0-supported sand filters as reactive filtration device, factor 
sustaining size exclusion should be understood and optimised [72]: (i) the pore size must be 
small enough for sufficient contaminant removal; or (ii) used Fe0 must be reactive enough to 
produce a sufficient amount of ‘scavengers’ as a function of time. Alternatively, the thickness 
of the Fe0 PRB can be increased to improve the devices filtration capacity. 
This highlights the importance of characterizing the intrinsic reactivity of Fe0 materials prior 
to application [73]. Ideally, the selection of a Fe0 material for a particular site should be 
governed by its intrinsic reactivity (and porosity when incorporated in the PRB system) and 
the expected impact of local geochemical (and geophysical) conditions on these factors. In 
cases where contaminant breakthrough was observed despite insignificant permeability loss, 
two explanations can be suggested: (i) the material is not reactive enough to generate 
“scavengers” in sufficient quantities, (ii) clogging of the entrance zone has disturbed the flow 
regime and preferential flow paths are created in the system. Preferential flow paths 
significantly impair the contact of flowing water with bed media (collectors, iron, sand). 
4 Conclusions 
Correlating the fundamental relationship between Fe0 PRB permeability loss and groundwater 
chemistry is extremely important for the design of sustainable Fe0 remediation systems. 
Further developments require knowledge of the intrinsic reactivity of Fe0, the rate of the 
formation of corrosion products and the role of foreign detrital minerals. Using mathematical 
modelling, the present communication challenges both the prevailing view and the 
contribution of Henderson and Demond [11]. An extensive mass balance analysis of aqueous 
iron corrosion has been used to show that volumetric expansion is the major control on 
permeability loss. It has been shown that, whilst Fe0 filtration systems (including PRBs) 
operating in anoxic (phreatic zone) conditions can exhibit limited permeability loss due to Fe0 
 13
corrosion product formation, Fe0 filtration systems operating in oxic (vadose zone) conditions 
exhibit significantly high permeability loss. It can therefore be concluded that admixing Fe
334 
335 
336 
337 
338 
339 
340 
341 
342 
343 
344 
345 
346 
347 
348 
349 
350 
351 
352 
353 
354 
355 
356 
357 
358 
359 
0 
with a non expansive materials (e.g. gravel, MnO2, pumice, sand) is a prerequisite for any 
sustainable Fe0 filtration systems operating in the near surface geosphere.  
The present work and related studies have delineated the early development of the Fe0 PRB 
technology that was marked by empirical designs [37,38,74,75-80]. Field experiences from 
more than 120 reactive barriers and an innumerable numbers of filters (including laboratory 
columns) worldwide should be used to continuously refine this innovative technology. 
Clearly the Fe0 technology should now be translated into rational engineering design criteria. 
As there are no established comprehensive design criteria for Fe0 beds, the reproducibility and 
comparison of available results is problematic. For example, despite the established 
significance of particle shape and size on the permeability, these parameters are not routinely 
given when describing operational conditions. Similarly, the initial porosity is not always 
given and the contribution of iron corrosion products to its filling was not properly addressed. 
A tentative guideline for future laboratory experiments can also be concluded: (i) assess the 
intrinsic reactivity of used Fe0, (ii) define the size and sphericity of all used materials (Fe0 and 
admixing materials), (iii) consider the surface roughness of Fe0 and sand grains, (iv) 
characterize the dimension and the composition of used columns, (v) evaluate the porosity of 
resulted columns, (vi) characterize used initial solutions (e.g. pH, Eh, O2 level, 
contamination), (vii) record the time dependant volume of the column effluent, and (viii) 
characterize the column effluent for pH, Eh, dissolved iron, target contaminants. 
Acknowledgments 
The manuscript was improved by the insightful comments of anonymous reviewers from 
CLEAN Soil, Air, Water. 
References 
[1] D. Blowes, Tracking hexavalent Cr in groundwater. Science 2002, 295, 2024–2025. 
 14
360 
361 
362 
363 
364 
365 
366 
367 
368 
369 
370 
371 
372 
373 
374 
375 
376 
377 
378 
379 
380 
381 
382 
383 
384 
[2] J.L. Jambor, M. Raudsepp, K. Mountjoy, Mineralogy of permeable reactive barriers for 
the attenuation of subsurface contaminants. Can. Miner. 2005, 43, 2117–2140. 
[3] K. Komnitsas, G. Bartzas, I. Paspaliaris, Inorganic contaminant fate assessment in zero-
valent iron treatment walls. Environ. Forensics 2006, 7, 207–217.  
[4] K. Komnitsas, G. Bartzas, I. Paspaliaris, Modeling of reaction front progress in fly ash 
permeable reactive barriers. Environ. Forensics 2006, 7, 219–231. 
[5] A.D. Henderson, A.H. Demond, Long-term performance of zero-valent iron permeable 
reactive barriers: a critical review. Environ. Eng. Sci. 2007, 24, 401–423. 
[6] A.B. Cundy, L. Hopkinson, R.L.D. Whitby, Use of iron-based technologies in 
contaminated land and groundwater remediation: A review. Sci. Tot. Environ. 2008, 
400, 42–51. 
[7] R. Thiruvenkatachari, S. Vigneswaran, R. Naidu, Permeable reactive barrier for 
groundwater remediation. J. Ind. Eng. Chem. 2008, 14, 145–156. 
[8] S. Comba, A. Di Molfetta, R. Sethi, A Comparison between field applications of nano-, 
micro-, and millimetric zero-valent iron for the remediation of contaminated aquifers. 
Water Air Soil Pollut. 2011, 215, 595–607. 
[9] M. Gheju, Hexavalent chromium reduction with zero-valent iron (ZVI) in aquatic systems. 
Water Air Soil Pollut. 2011, 222, 103–148. 
[10] ITRC, Permeable reactive barrier: Technology update. PRB-5. Washington, D.C.: 
Interstate Technology & Regulatory Council, PRB: Technology Update Team (2011). 
www.itrcweb.org (access: 09.03.2012) 
[11] A.D. Henderson, A.H. Demond, Impact of solids formation and gas production on the 
permeability of ZVI PRBs. J. Environ. Eng. 2011, 137, 689–696. 
[12] N. Muchitsch, T. Van Nooten, L. Bastiaens, P. Kjeldsen, Integrated evaluation of the 
performance of a more than seven year old permeable reactive barrier at a site 
 15
contaminated with chlorinated aliphatic hydrocarbons (CAHs). J. Contam. Hydrol. 
2011, 126, 258–270. 
385 
386 
387 
388 
389 
390 
391 
392 
393 
394 
395 
396 
397 
398 
399 
400 
401 
402 
403 
404 
405 
406 
407 
408 
[13] S. Ulsamer, A model to characterize the kinetics of dechlorination of tetrachloroethylene 
and trichloroethylene by a zero valent iron permeable reactive barrier. Master thesis, 
Worcester Polytechnic Institute, 2011, 73 pp. 
[14] E. Bi, J.F. Devlin, B. Huang, Effects of mixing granular iron with sand on the kinetics of 
trichloroethylene reduction. Ground Water Monit. Remed. 2009, 29, 56–62. 
[15] Y.H. Kim, S.-O. Ko, H.C. Yoo, Simultaneous removal of tetrachlorocarbon and 
chromium(VI) using zero valent iron. J. Korean Soc. Environ. Eng. 2002, 24, 1949–
1956. 
[16] D.-I. Song, Y.H. Kim, W.S. Shin, A simple mathematical analysis on the effect of sand 
in Cr(VI) reduction using zero valent iron. Korean J. Chem. Eng. 2005, 22, 67–69. 
[17] N. Moraci, P.S. Calabrò, Heavy metals removal and hydraulic performance in zero-
valent iron/pumice permeable reactive barriers. J. Environ. Manage. 2010, 91, 2336–
2341. 
[18] G. Bartzas, K. Komnitsas, Solid phase studies and geochemical modelling of low-cost 
permeable reactive barriers. J. Hazard. Mater. 2010, 183, 301–308. 
[19] L. Li, C.H. Benson, Evaluation of five strategies to limit the impact of fouling in 
permeable reactive barriers. J. Hazard. Mater. 2010, 181, 170–180. 
[20] M.S. Odziemkowski, T.T. Schuhmacher, R.W. Gillham, E.J. Reardon, Mechanism of 
oxide film formation on iron in simulating groundwater solutions: Raman 
spectroscopic studies. Corros. Sci. 1998, 40, 371–389. 
[21] S. Yabusaki, K. Cantrell, B. Sass, C. Steefel, Multicomponent reactive transport in an in 
situ zero-valent iron cell. Environ. Sci. Technol. 2001, 35, 1493–1503. 
 16
409 
410 
411 
412 
413 
414 
415 
416 
417 
418 
419 
420 
421 
422 
423 
424 
425 
426 
427 
428 
429 
430 
431 
432 
433 
[22] R. Venkatapathy, D.G. Bessingpas, S. Canonica, J.A. Perlinger, Kinetic models for 
trichloroethylene transformation by zero-valent iron. Appl. Catal. Environ. 2002, 37, 
139–159. 
[23] R.T. Wilkin, R.W. Puls, G.W. Sewell, Long-term performance of permeable reactive 
barriers using zero-valent iron: Geochemical and microbiological effects. Ground 
Water 2002, 41, 493–503. 
[24] R.L. Johnson, R.B. Thoms, R.O'B. Johnson, T. Krug, Field evidence for flow reduction 
through a zero-valent iron permeable reactive barrier. Ground Water Monit. Remed. 
2008, 28, 47–55. 
[25] R.L. Johnson, R.B. Thoms, R.O'B. Johnson, J.T. Nurmi, P.G. Tratnyek, Mineral 
precipitation upgradient from a zero-valent iron permeable reactive barrier. Ground 
Water Monit. Remed. 2008, 28, 56–64. 
[26] M. Odziemkowski, Spectroscopic studies and reactions of corrosion products at surfaces 
and electrodes. Spectrosc. Prop. Inorg. Organomet. Compd. 2009, 40, 385–450. 
[27] S.W. Orth, R.W. Gillham, Dechlorination of trichloroethene in aqueous solution using 
Fe0. Environ. Sci. Technol. 1996, 30, 66–71. 
[28] Y. Zhang, R.W. Gillham, Effects of gas generation and precipitates on performance of 
Fe0 PRBs. Ground Water 2005, 43, 113–121. 
[29] M.J. Wildman, P.J.J. Alvarez, RDX degradation using an integrated Fe(0)-microbial 
treatment approach. Water Sci. Technol. 2001, 43, 25–33. 
[30] S. Biswas, P. Bose, Zero-valent iron assisted autotrophic denitrification. J. Environ. Eng. 
ASCE 2005, 131, 1212–1220. 
[31] S.-M. Wang, S.-k. Tseng, Reductive dechlorination of trichloroethylene by combining 
autotrophic hydrogen-bacteria and zero-valent iron particles. Biores. Technol. 2009, 
100, 111–117. 
 17
434 
435 
436 
437 
438 
439 
440 
441 
442 
443 
444 
445 
446 
447 
448 
449 
450 
451 
452 
453 
454 
455 
456 
457 
458 
[32] T. Van Nooten, D. Springael, L. Bastiaens, Positive impact of microorganisms on the 
performance of laboratory-scale permeable reactive iron barriers. Environ. Sci. 
Technol. 2008, 42, 1680–1686. 
[33] N.B. Pilling, R.E. Bedworth, The oxidation of metals at high temperatures. J. Inst. Metals 
1923, 29, 529–591. 
[34] K. Suda, S. Misra, K. Motohashi, Corrosion products of reinforcing bars embedded in 
concrete. Corros. Sci. 1993, 35, 1543–1549. 
[35] 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. 
2008, 38, 1079–1091. 
[36] Y. Zhao, H. Ren, H. Dai, W. Jin, Composition and expansion coefficient of rust based on 
X-ray diffraction and thermal analysis. Corros. Sci. 2011, 53, 1646–1658. 
[37] C. Noubactep, Aqueous contaminant removal by metallic iron: Is the paradigm shifting? 
Water SA 2011, 37, 419–426. 
[38] C. Noubactep, S. Caré, K.B.D. Btatkeu, C.P. Nanseu-Njiki, Enhancing the sustainability 
of household Fe0/sand filters by using bimetallics and MnO2. Clean – Soil, Air, Water 
2012, 40, 100–109. 
[39] G.H. Brimhall, J.L. Christopher, C. Ford, J. Bratt, G. Taylor, O. Warin, Quantitative 
geochemical approach to pedogenesis: importance of parent material reduction, 
volumetric expansion, and eolian influx in lateritization. Geoderma 1991, 51, 51–91. 
[40] R. Balasubramaniam, K.A.V. Ramesh, P. Dillmann, Characterization of rust on ancient 
Indian iron. Curr. Sci. 2003, 85, 1546−1555. 
[41] M. Kubare, J. Haarhoff, Rational design of domestic biosand filters. J. Water Supply: 
Res. Technol. – AQUA 2010, 59, 1–15. 
[42] K.P. Panayiotopoulos, Packing of sands - a review. Soil Tillage Res. 1989, 13, 101–121. 
 18
459 
460 
461 
462 
463 
464 
465 
466 
467 
468 
469 
470 
471 
472 
473 
474 
475 
476 
477 
478 
479 
480 
481 
482 
483 
[43] J. Haarhoff, A. Vessal, A falling-head procedure for the measurement of filter media 
sphericity. Water SA 2010, 36, 97–104. 
[44] K.-M. Yao, M.T. Habibian, C.R. O’melia, Water and waste water filtration: concepts and 
applications. Environ. Sci. Technol. 1971, 5, 1105–1112. 
[45] P. Bedrikovetsky, F.D. Siqueira, C.A. Furtado, A.L.S. Souza, Modified particle 
detachment model for colloidal transport in porous media. Transp. Porous Med. 2011, 
86, 353–383. 
[46] W.O. Smith, P.D. Foote, P.F. Busang, Packing of homogeneous spheres. Phys. Rev. 
1929, 34, 1271–1274. 
[47] E. Manegold, R. Hofmann, K. Solf, Ueber Kapillarsysteme XII. I. Die mathematische 
Behandlung idealer Kugelpackungen und das Hohlraumvolumen realer 
Gerüststrukturen. Colloid Polymer Sci. 1931, 56, 142–159. 
[48] H. Iwata, T. Homma, Distribution of coordination numbers in random packing of 
homogeneous spheres. Powder Technol. 1994, 10, 79–83. 
[49] Neto, L.N., Kotousov, A. & Bedrikovetsky, P. Application of contact theory to 
evaluation of elastic properties of low consolidated porous media. Int. J. Fract. 2011, 
168, 267–276. 
[50] 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 2007, 42, 1869–1878. 
[51] M.A. Shannon, P.W. Bohn, M. Elimelech, J.G. Georgiadis, B.J. Marinas, A.M. 
Mayes,Science and technology for water purification in the coming decades. Nature 
2008, 452, 301–310. 
[52] C. Noubactep, Metallic iron for safe drinking water worldwide. Chem. Eng. J. 2010, 165, 
740–749. 
 19
484 
485 
486 
487 
488 
489 
490 
491 
492 
493 
494 
495 
496 
497 
498 
499 
500 
501 
502 
503 
504 
505 
506 
507 
508 
509 
[53] D.E. Giles, M. Mohapatra, T.B. Issa, S. Anand, P. Singh, Iron and aluminium based 
adsorption strategies for removing arsenic from water. J. Environ. Manage. 2011, 92, 
3011–3022. 
[54] H. Darcy, Les fontaines publiques de la Ville de Dijon. V. Dalmont, Paris (1856) 647 pp. 
[55] T.O. Mahlangu, B.B. Mamba, M.N.B. Momba, A comparative assessment of chemical 
contaminant removal by three household water treatment filters. Water SA 2012, 38, 
39–48. 
[56] C.E. Stauber, E.R. Printy F.A. McCarty, K.R. Liang, M.D. Sobsey, Cluster randomized 
controlled trial of the plastic biosand water filter in Cambodia. Environ. Sci. Technol. 
2012, 46, 722–728. 
[57] M. Lea, Biological sand filters: Low-cost bioremediation technique for production of 
clean drinking water. Current Protocols Microbiol. 1G.1.1-1G.1.28 2008, doi: 
10.1002/9780471729259.mc01g01s9. 
[58] C. Noubactep, Processes of contaminant removal in “Fe0–H2O” systems revisited. The 
importance of co-precipitation. Open Environ. J. 2007, 1, 9–13. 
[59] C. Noubactep, A critical review on the mechanism of contaminant removal in Fe0–H2O 
systems. Environ. Technol. 2008, 29, 909–920. 
[60] C. Noubactep, The suitability of metallic iron for environmental remediation. Environ. 
Progr. Sust. En. 2010, 29, 286–291. 
[61] C. Noubactep, The fundamental mechanism of aqueous contaminant removal by metallic 
iron. Water SA 2010, 36, 663–670. 
[62] M. Edwards, M. Benjamin, Adsorptive filtration using coated sand: A new approach for 
treatment of metal-bearing wastes. J. Water Pollut. Control Fed. 1989, 61, 1523–1533. 
[63] 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 A 2000, 35, 1021–1041. 
 20
[64] D. Dermatas, X. Meng, Removal of As, Cr and Cd by adsorptive filtration. Global Nest: 
the Int. J. 2004, 6, 73–80. 
510 
511 
512 
513 
514 
515 
516 
517 
518 
519 
520 
521 
522 
523 
524 
525 
526 
527 
528 
529 
530 
531 
532 
533 
534 
535 
[65] S. Roy, P. Bose, Modeling arsenite adsorption on rusting metallic iron. J. Environ. Eng.  
2010, 136, 405–411. 
[66] K. Hanna, Adsorption of aromatic carboxylate compounds on the surface of synthesized 
iron oxide-coated sands. Appl. Geochem. 2007, 22, 2045–2053.
[67] K. Hanna, Sorption of two aromatic acids onto iron oxides: Experimental study and 
modeling. J. Colloid Interf. Sci. 2007, 309, 419–428. 
[68] K. Eusterhues, F.E. Wagner, W. Häusler, M. Hanzlik, H. Knicker, K.U. Totsche, I. 
Kögel-Knabner, U. Schwertmann, Characterization of ferrihydrite-soil organic matter 
coprecipitates by X-ray diffraction and Mössbauer spectroscopy. Environ. Sci. Technol. 
2008, 42, 7891–7897. 
[69] K. Hanna, J.-F. Boily, Sorption of two naphthoic acids to goethite surface under flow 
through conditions.  Environ. Sci. Technol. 2010, 44, 8863–8869. 
[70] K. Eusterhues, T. Rennert, H. Knicker, I. Kgel-Knabner, K.U. Totsche, U. Schwertmann, 
Fractionation of organic matter due to reaction with ferrihydrite: Coprecipitation versus 
adsorption. Environ. Sci. Technol. 2011, 45, 527–533. 
[71] B. Saha, S. Das, J. Saikia, G. Das, Preferential and enhanced adsorption of different dyes 
on iron oxide nanoparticles: A comparative study. J. Phys. Chem. C 2011; 115, 8024–
8033.  
[72] U. Förstner, P. Jacobs, F. v. d. Kammer, Impact of natural nanophases on heavy-metal 
retention in zeolite-supported reactive filtration facilities for urban run-off treatment. 
Fresenius J. Anal. Chem. 2001, 371, 652–659.  
[73] 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. 
2009, 172, 943–951. 
 21
[74] A. Ghauch, H. Abou Assi, H. Baydoun, A.M. Tuqan, A. Bejjani, Fe0-based trimetallic 
systems for the removal of aqueous diclofenac: Mechanism and kinetics. Chem. Eng. 
J. 2011, 172, 1033–1044. 
536 
537 
538 
539 
540 
541 
542 
543 
544 
545 
546 
547 
548 
549 
550 
551 
552 
553 
[75] R.A. Crane, T.B. Scott, Nanoscale zero-valent iron: future prospects for an emerging 
water treatment technology. J. Hazard. Mater. 2012, 211–212, 112–125. 
[76] C. Noubactep, S. Caré, R.A. Crane, Nanoscale metallic iron for environmental 
remediation: prospects and limitations. Water Air Soil Pollut. 2012, 223, 1363–1382. 
[77] C. Noubactep, Metallic iron for water treatment: A knowledge system challenges 
mainstream science. Fresenius Environ. Bull. 2011, 20, 2632–2637. 
[78] C. Noubactep, E. Temgoua, M.A. Rahman, Designing iron-amended biosand filters for 
decentralized safe drinking water provision. Clean: Soil, Air, Water 2012, doi: 
10.1002/clen.201100620. 
[79] C. Noubactep, Investigating the processes of contaminant removal in Fe0/H2O systems. 
Korean J. Chem. Eng. 2012, doi: 10.1007/s11814-011-0298-8. 
[80] R. Crane, C. Noubactep, Elemental metals for environmental remediation: learning from 
hydrometallurgy. Fresenius Environ. Bull. 2012, 21, 1192–1196. 
 
 22
Table 1: Some characteristics of metallic iron (Fe0) and its main corrosion products 
commonly identified in Fe
553 
554 
555 
556 
557 
558 
0 PRBs. “x” is the weight percent of Fe in the phase. As a 
rule, oxides formed under anoxic conditions exhibit larger x values. “η” is the 
calculated coefficient of volumetric expansion. Phase parameters are compiled from 
Balasubramaniam et al. [40] and Henderson and Demond [11]. 
 
Phase Name Structure Density Vm x η 
   (g/cm3) (cm3/mol) (%) (-) 
Fe0 Iron metal bcc 7.86 7.6 100.0 - 
Fe(OH)3 FeIII hydroxide perovskite-like 3.1 34.4 52.0 4.53 
FeCO3 Siderite Trigonal 3.83 29.3 48.3 3.86 
Fe(OH)2 FeII hydroxide Trigonal 3.4 26.4 62.2 3.47 
α-FeOOH Goethite Orthorhombic 4.28 20.3 62.9 2.67 
γ-Fe2O3 Maghemite Cubic 4.69 29.1 70.0 1.91 
α-Fe2O3 Hematite Trigonal 5.3 30.1 70.0 1.98 
Fe3O4 Magnetite Cubic 5.175 45.0 72.4 1.97 
559 
560 
561 
 
 
 23
Table 2: Estimation of the extent of Fe0 depletion (α value in %) in the column of Henderson 
and Demond [11] for two values of the initial bed porosity. α and α
561 
562 
563 
564 
565 
566 
567 
568 
1 correspond to Φ0 = 36 % 
when H2 remains in the system or escapes respectively and α' and α'1 correspond to Φ0 = 62 
% when H2 remains in the system or escapes respectively. It is seen that in all cases the initial 
porosity is filled by gas when less than 0.1 % of the initial mass of Fe0 is corroded. A value of 
100 % is related to a system which is not clogged when Fe0 is depleted.  
 
 
Name Formula α values (%) 
  α α1 α’ α’1
Maghemite Fe2O3 0.01 62 0.04 100 
Magnetite Fe3O4 0.01 58 0.04 100 
Hematite Fe2O3 0.01 57 0.04 100 
Goethite FeOOH 0.02 34 0.06 98 
Ferrous hydroxide Fe(OH)2 0.02 24 0.06 70 
Siderite FeCO3 0.02 20 0.06 57 
Ferric hydroxide Fe(OH)3 0.02 16 0.06 46 
569 
570 
 
 24
Figure captions 570 
571 
572 
573 
574 
575 
576 
577 
578 
579 
580 
581 
582 
 
Figure 1: Comparison of the evolution of porosity loss in a Fe0 bed filled with spherical (left) 
and cylindrical (right) particles. The compactness is maximal for spherical 
particles. The roundness or sphericity of used materials (Fe0 and additives) should 
be routinely characterized as this is crucial for the initial porosity.  
Figure 2: Types of Fe0-based filters for water treatment. The point O(0.0) represents a Fe0 
free filter (e.g. biosand filter, iron oxide-amended sand). The point I(100,100) 
correspond to a filter which is clogged just at Fe0 depletion. 
Figure 3: Evolution of the residual porosity as function of the volumetric proportion of Fe0 is 
the filter for the two extreme values of the initial porosity (Φ0 = 0.36 and 0.62). It is 
seen that for Φ0 = 0.36 no filter is sustainable. For Φ0 = 0.62, filter operating under 
strictly anoxic conditions are sustainable. 
 25