1
1 
2 
3 
4 
5 
6 
7 
8 
9 
10 
11 
12 
13 
14 
15 
16 
17 
18 
Effect of pumice and sand on the sustainability of granular iron beds for the 
aqueous removal of CuII, NiII, and ZnII  
Stefania Bilardia, Paolo S. Calabròa, Sabine Caréb,*, Nicola Moracia, Chicgoua Noubactepc,d
aUniversità degli Studi Mediterranea di Reggio Calabria, MECMAT, Mechanics and Materials Department, Faculty of 
Engineering, Via Graziella, loc. Feo di Vito, 89122 Reggio Calabria, Italy. 
bUniversité Paris-Est, Laboratoire Navier (UMR 8205), CNRS, ENPC, IFSTTAR, F-77455 Marne-la-Vallée, France. 
cAngewandte Geologie, Universität Göttingen, Goldschmidtstraße 3, D-37077, Göttingen, Germany. 
dKultur und Nachhaltige Entwicklung CDD e.V., Postfach 1502, D-37005 Göttingen, Germany 
* corresponding author: e-mail: sabine.care@ifsttar.fr; Tel.: + 33 140435466; Fax: + 33 140435450. 
 
Running Title: Admixing pumice and sand to reactive Fe0 sustains long-term metal removal. 
 
Acronym List 
PRB  Permeable Reactive Barrier 
 
Keywords: Hydraulic conductivity, Reactive barriers, Pumice, Sand, Zerovalent iron. 
 
  
2
18 
19 
20 
21 
22 
23 
24 
25 
26 
27 
28 
29 
30 
31 
32 
33 
34 
35 
36 
37 
38 
39 
40 
41 
42 
Abstract  
Current knowledge of the basic principles underlying the design of Fe0 beds is weak. The 
volumetric expansive nature of iron corrosion was identified as the major factor determining the 
sustainability of Fe0 beds. This work attempts to systematically verify developed concepts. Pumice 
and sand were admixed to 200 g of Fe0 in column studies (50:50 volumetric proportion). Reference 
systems containing 100 % of each material have been also investigated. The mean grain size of the 
used materials (in mm) were 0.28 (sand), 0.30 (pumice) and 0.50 (Fe0). The five studied systems 
were characterized (i) by the time dependent evolution of their hydraulic conductivity 
(permeability) and (ii) for their efficiency for aqueous removal of CuII, NiII, and ZnII (about 0.30 M 
of each). Results showed unequivocally that (i) quantitative contaminant removal was coupled to 
the presence of Fe0, (ii) additive admixture lengthened the service life of Fe0 beds, and (iii) pumice 
was the best admixing agent for sustaining permeability while the Fe0/sand column was the most 
efficient for contaminant removal. The evolution of the permeability was well-fitted by the 
approach that the inflowing solution contained dissolved O2. The achieved results are regarded as 
starting point for a systematic research to optimise/support Fe0 filter design. 
1 Introduction 
Permeable reactive barriers (PRBs) containing metallic iron (Fe0) as reactive medium have been 
developed during the past two decades to an established technology for groundwater remediation 
[1-11]. The original PRB technology containing granular Fe0 has been expanded to the injection of 
nano-scale Fe0 for source remediation [8-11]. To date, more than 180 Fe0 PRBs have been installed 
worldwide [8,11]. Successful accomplishment of remedial goals has been typically reported. At 
some few sites, system failures were recorded [8,12]. Reported failures were attributed to design 
shortcomings due to poor site characterization (reason 1), poor design selection (reason 2) or 
installation at sites where the technology is not an appropriate choice (reason 3) [8,12,13]. 
However, there is clear evidence that the physico-chemistry of the Fe0/H2O system was not properly 
  
3
43 
44 
45 
46 
47 
48 
49 
50 
51 
52 
53 
54 
55 
56 
57 
58 
59 
60 
61 
62 
63 
64 
65 
66 
67 
considered [14]. Accordingly, design shortcomings may have reasons different from or additional to 
reasons 1, 2 and 3. 
A major concern of Fe0 PRBs is related to the reduction of the hydraulic conductivity (permeability 
loss) with time [15-17]. Laboratory and field data have also demonstrated diminished Fe0 reactivity 
with time [9, 18]. Consequently, the sustainability of Fe0 PRBs in terms of both Fe0 reactivity and 
system permeability has been extensively discussed during the past 15 years [1,18-24]. Reported 
results are confusing and even conflicting as demonstrated below for trichloroethene (TCE). 
O’Hannesin and Gillham [1] reported on successful TCE (268 mg/L) reductive degradation by a 
Fe0/sand mixture containing 22 % Fe0 by weight (laboratory and field test). Bi et al. [21] tested 
several weight Fe0/sand mixtures (25/75, 50/50, 75/25, 85/15 and 100/0) for TCE (up to 60 mg/L) 
treatment and reported that the system with less than 50 % Fe0 was not efficient (laboratory test). 
Ruhl et al. [18] evaluated four dual mixtures (Fe0/anthracite, Fe0/gravel, Fe0/pumice and Fe0/sand) 
for TCE (about 10 mg/L) treatment (laboratory test). The used masses of additives varied from 24.4 
g for pumice to 104.3 g for gravel. The used mass of Fe0 was 100 g resulting in Fe0 weight ratios 
varying from 49 % for gravel to 80 % for pumice. Ruhl et al. [18] concluded that tested dual 
systems are not applicable for TCE treatment but “might be applicable for the removal of heavy 
metals”. 
The three examples reveal that researchers use varying experimental procedures to characterize 
processes in Fe0/H2O systems (see Tab. 1) [1,21,25-33]. These procedures differ for instance in Fe0 
intrinsic reactivity, Fe0 pre-treatment, Fe0 mass, Fe0 particle size and shape, used admixing 
additives and their proportions, duration of the experiments, nature and concentration of the 
contaminant, buffer application, solution flow velocity and water chemistry. As a result, many 
different reports for the same compound are available in the literature (even for the same Fe0). 
Water and dissolved inorganic constituents (Ca2+, HCO3-, Mg2+, O2, PO43-, SO42-) react with iron 
species (Fe0, FeII and FeIII) to form precipitates that progressively fill the inter-particular porosity 
  
4
68 
69 
70 
71 
72 
73 
74 
75 
76 
77 
78 
79 
80 
81 
82 
83 
84 
85 
86 
87 
88 
89 
90 
91 
92 
within a Fe0 filter. The potential of these in situ generated precipitates to limit the permeability and 
the efficiency of Fe0 PRBs filters has been clearly documented [1,4,6-8]. However, the role of Fe0 
oxidation by water has not always been properly considered and the role of gas (H2) formation in 
porosity/permeability loss has been sometimes overestimated [15,16]. Recent theoretical works 
demonstrated that a Fe0-based filter should be considered as a system in which iron is corroded 
mostly by water and the micro-pollutants are sequestrated in the matrix of precipitation corrosion 
products [14,34-36]. This view corroborates concordant reports regarding Fe0 filters as a long-term 
sink for C, S, Ca, Si, Mg, and N [12,37-39]. 
The present work is an attempt to improve the design of Fe0 filtration systems based on recent 
theoretical studies [14]. In the present work, the efficiency of five different systems (A to E) for 
aqueous contaminant removal is tested in column studies. The volumetric composition of individual 
systems is given as: (A) 100 % sand, (B) 100 % pumice, (C) 100 % Fe0, (D) 50:50 Fe0:pumice, and 
(E) 50:50 Fe0:sand. The model solution contained about 0.30 M of CuII, NiII, and ZnII. The 
evolution of the systems is characterized by determining the (i) extend of contaminant removal, and 
(ii) evolution of hydraulic conductivity. 
2 Materials and methods  
2.1  Chemicals and solutions 
Copper(II) nitrate hydrate (purity 99.999), nickel(II) nitrate hexahydrate (purity 99.999) and zinc(II) 
nitrate hexahydrate (purity 99.000) were obtained from Sigma-Aldrich. All chemicals used for 
experiments and analysis were of analytical grade. The used solutions were obtained by dissolving 
copper nitrate, nickel nitrate and zinc nitrate in distilled water. The molar concentration of the 
resulting solution was as follows: 0.27 M Cu, 0.29 M Ni and 0.37 M Zn. The corresponding mass 
concentrations are 17 mg/L Cu, 17 mg/L Ni, and 23 mg/L Zn. 
2.2 Solid materials 
2.2.1 Porosity of binary granular media 
  
5
93 
94 
The total porosity Φ of a binary granular medium composed of two kinds of particles P1 and P2 
(here P1 corresponds to Fe0 particles) is given by :  
V
M
V
MM1
V
V
V
VV1f
rz
2sa2
2
rz
2sa21sa1
rz
2a
2
rz
2a1a
pppper
ρϕ+ρ+ρ−=Φ
ϕ++−=ϕ+Φ=Φ
/.]//[
.][.int
   (1) 95 
96 Where: (i) Φinter is inter-particular porosity (ii) Vai, Mai, ρai are respectively the apparent volume of 
the particles i, the mass and the apparent specific weight, and ϕ2  is the intra-particular porosity of 
the particles 2 with 
97 
02 =ϕ  for non porous particles (sand) and 02 ≠ϕ  for porous particles (pumice) 
and (iii) V
98 
99 
100 
101 
102 
103 
104 
105 
106 
107 
108 
109 
110 
111 
112 
113 
114 
rz is the total packing volume of the granular medium.  
2.2.2 Metallic iron (Fe0) 
The used Fe0 is of the type FERBLAST RI 850/3.5, distributed by Pometon S.p.A., Mestre - Italy. 
The powdered material contents mainly iron (> 99.74 %). Identified impurities included mainly Mn 
(0.26 %), O, S and C. The material is characterized by uniform grain size distribution. The 
coefficient of uniformity U (ratio between the diameters corresponding to 60 and 10 % finer in the 
grain size distribution) is 2. The mean grain size (d50) is about 0.5 mm and the initial porosity of 
used Fe0 medium has been estimated to be Φ0 = 49.6 % (see Tab. 2, Eq. 1). 
2.2.3 Pumice 
The used pumice originates from Lipari (Aeolian Islands, Sicily – Italy). Its mineralogical 
composition was determined as follows: SiO2: 71.75 %; Al2O3: 12.33 %; K2O: 4.47 %; Na2O: 3.59 
%; Fe2O3: 1.98 %; moreover it contains about 4 % of bound water entrapped in the pumice structure 
during the sudden cooling of magma and traces of other compounds (e.g. CaO, SO3, MgO, TiO2, 
FeO, MnO, P2O5). The material is characterized by uniform grain size distribution. The coefficient 
of uniformity U is 1.4. The mean grain size (d50) is about 0.3 mm. This type of pumice has been 
chosen since it was the available fraction closest to Fe0 in dimension. The initial porosity of the 
  
6
115 pumice granular medium has been estimated to be Φ0 = 73.3 % and the inner porosity of the pumice 
(intra particular porosity ϕpp ) to be 41 % (Tab. 2, Eq. 1) through Mercury Intrusion Porosity (MIP) 
measurements under the hypothesis that the relative density (packing) of granular mixtures in the 
columns and during MIP experiments were the same. 
116 
117 
118 
119 
120 
121 
122 
123 
124 
125 
126 
127 
128 
129 
130 
131 
132 
133 
134 
135 
136 
137 
138 
2.2.4 Sand 
The used quartz sand was obtained from a river quarry. The sand was carefully washed with 
distilled water and sieved before use. The material is characterized by uniform grain size 
distribution. The coefficient of uniformity U is 2.2. The mean grain size (d50) is about 0.28 mm. The 
material was used without any further characterization. The initial porosity of the sand medium has 
been estimated to be Φ0 = 45.0 % (Tab. 2, Eq. 1). 
2.3 Column operation  
Laboratory scale polymethyl methacrylate (Plexiglas) columns were operated in up-flow mode. The 
influent solution was pumped upwards from a single PE bottle using a precision peristaltic pump 
(Ismatec, ISM930). The flow rate was maintained constant at a value of 0.5 mL/min. Tygon tubes 
were used to connect inlet reservoir, pump, columns and outlet. Five plexiglas columns (50 cm 
long, 5.0 cm inner diameter) were used in the experiments (Fig. 1). 
Five different systems were investigated (Systems A through E) (Tab. 3). System A was the 
operational reference system containing only sand (0 % Fe0), System B was the second operational 
reference containing only pumice (0 % Fe0) and system C was a pure iron bed (100 % Fe0). The 
volumetric proportion of Fe0 in the 2 other systems (D, E) was 50 %. In systems C to E, the mass of 
iron was fixed to 200 g. This mass represented either 100 % of the reactive zone (rz) or the relevant 
volumetric proportion of rz (Fig. 1, Tab. 3). In system B the pumice volume was set to be the same 
occupied by Fe0 in system C while the pumice mass was obviously the same as in system E. The 
total porosity of the all systems varies between 45 % and 73 % (Tab. 2, Eq. 1). 
  
7
139 
140 
141 
142 
143 
144 
145 
146 
147 
148 
149 
150 
151 
152 
153 
154 
155 
156 
157 
158 
159 
160 
161 
162 
163 
The hydraulic conductivity [40] was determined during the column tests, by either constant-head (k 
> 10-6 m/s) or variable-head (k < 10-6 m/s) permeability methods, at selected dates to assess the 
permeability of the systems. The experiments were performed at room temperature (21 ± 4 °C). 
Samples for analysis were collected at periodic intervals and the experiments where prolonged until 
contaminant breakthrough (systems A and B) or a significant loss of the hydraulic conductivity 
(systems C to E) was observed. 
2.4  Analytical methods 
Samples from the columns were centrifuged at 3000 rpm (ALC, PK121 Multispeed Centrifuge). 
The supernatant was vacuum filtered through a 0.45 μm glass filters. The aqueous concentrations of 
Cu, Fe, Ni and Zn were then measured by Atomic Absorption Spectrophotometry (AAS - Shimadzu 
AA – 6701F; wavelengths: Cu 324.75 nm, Ni 232.00 nm, Zn 213.86 nm, Fe 248.33 nm) using air-
acetylene flame and according to conventional Standard Methods [41]. The used AAS device was 
calibrated  using three operational standard solutions covering the expected concentration range of 
the samples (after dilution if applicable). Each operational standard solution was prepared by an 
appropriate dilution of a 1000 ppm (Cu(NO3)2, Fe(NO3)3, Ni(NO3)2, Zn(NO3)2) certified atomic 
absorption stock solution from Merck (Germany). The minimum correlation coefficient of 
calibration curves was of 0.997. The pH value was measured by combination glass electrodes 
(WTW GmbH, inolab pH/Cond 720). 
2.4.1 MIP measurements 
MIP measurements have been carried out using a Micromeritics instrument apparatus type 
(AutoPore IV 9500). The instrument is capable of a minimum intruding pressure of 3.4 kPa and a 
maximum pressure of 227 MPa, so that the pore radius ranges from 2.7 nm to 180 μm.  
For pumice particles the measured pore data allow determining the inter-particular and intra-
particular porosities of the pumice particles, the apparent specific weight ρas (defined as the ratio of 
the mass and the apparent volume of the pumice particles) and the specific weight ρs (defined as the 
  
8
164 
165 
166 
167 
168 
169 
170 
171 
172 
173 
174 
175 
176 
177 
178 
179 
180 
181 
182 
183 
184 
185 
ratio of the mass and the volume of the solid phase of the pumice particles). 
2.5 Expression of the experimental results 
In order to characterize the magnitude of tested systems for contaminant removal, the removal 
efficiency (E) and the specific removal (Es) were calculated using Eq. 2 and Eq. 3 [33].  
E = mrem/min*100    (2) 
Es = mrem/mFe*100    (3) 
where min is the mass of contaminant flowed into the column, mrem is the mass of removed 
contaminant, and mFe the mass of Fe0 present in the column. 
2.6 Evaluation of the residual porosity 
When iron corrodes, porous oxide layers are formed at the Fe0/H2O interface. The volume of the 
corrosion product (Voxide) is higher than that of the original metal (VFe). The ratio (η) between the 
volume of expansive corrosion product and the volume of iron consumed in the corrosion process is 
called ‘‘coefficient of volumetric expansion’’ [42,43]. Generally, Voxide is 2.08 to 6.40 times larger 
than V0 (2.08 ≤ η ≤ 6.40 for free expansion).  
At any time (t > 0), Voxide can be calculated using Eq. 4: 
Voxide = η*(V0 –Vt)      (4) 
Where η is the coefficient of volumetric expansion, (V0 –Vt) is the consumed Fe0 volume with V0 
the initial volume of Fe0 and Vt its residual Fe0 at time t.  
The effective volumetric expansion ΔV (Eq. 5) corresponding to the volume of pores that is 
occupied by iron corrosion products is a measure of the extent of porosity loss.  
ΔV = (η - 1)* (V0 –Vt)      (5) 
The residual porosity of the system at time t (Φ(t)) may be estimated by (Eq. 6):  
V
VV1t
rz
t0
0
)).(()( −−η−Φ=Φ     (6) 186 
187 Where  is the initial porosity of the reactive zone given in Tab. 3, and VΦ0 rz is the volume of the 
  
9
188 
189 
190 
191 
192 
193 
194 
195 
196 
197 
198 
199 
200 
201 
202 
203 
204 
205 
206 
207 
208 
209 
210 
211 
212 
reactive zone.  
3 Results and discussion 
3.1 pH variation and Fe breakthrough 
Metal ions are known to be removed from the aqueous phase in packed Fe0 beds by adsorption, co-
precipitation and adsorptive size-exclusion when the pH > 4.5 [44-48]. Figure 2a clearly 
demonstrates that the pH value of the initial solution (t = 0) and that of the effluent from all 
columns was larger that 5.5. This suggests that contaminant removal could be quantitative (see Tab. 
4) if the residence time is sufficient to enable the formation of enough iron corrosion products for 
contaminant retention in the column. A hint that quantitative contaminant removal was likely is 
given by the evolution of the iron concentration (Fig. 2b). Fig. 2b clearly shows that the effluent 
iron concentration was less than 0.2 mg/L and reached values close to up to 1 mg/L only shortly 
before clogging for the system with 50 % pumice (system E). This observation could be attributed 
to accelerated transport through preferential flow paths [26]. 
3.2 Metal breakthrough 
Fig. 3 and Tab. 4 summarize the results of contaminant removal in systems containing Fe0. It is 
clear from Fig. 3a that no Cu breakthrough occurs. Ni breakthrough occurs first (Fig. 3b). In fact Ni 
breakthrough occurs before day 8 in the system with 100 % Fe0. Zn is the next less retained metal 
with a breakthrough occurring at day 10 in the system with 100 % Fe0 (Fig. 3c). The observed order 
of removal efficiency corresponds to the selectivity sequence for iron oxides and soils: Ni < Zn < 
Cu [49-51]. For example, Fontes and Gomes [50] found that in competitive adsorption CuII 
maintains its strong affinity with the surface, while NiII and ZnII are displaced from the surface. This 
observation corroborates the view that species with higher affinity to iron oxides are better treated 
by Fe0 filters [23,24,52,53]. 
Another important result from Fig. 3 is that no contaminant breakthrough was observed in the 
system with 50 % sand (system D). This system is less porous than the system with 50 % pumice 
  
10
213 
214 
215 
216 
217 
218 
219 
220 
221 
222 
223 
224 
225 
226 
227 
228 
229 
230 
231 
232 
233 
234 
235 
236 
(Tab. 3). The differential behaviour of the systems with pumice and sand (D and E) illustrates the 
dilemma of sustaining efficiency (maximum Fe0 ratio) while maintaining permeability, for instance 
by using porous pumice in place of sand [33]. As discussed in details elsewhere [14,36] this 
dilemma could be solved by using an appropriate thickness of the Fe0-based layer for each relevant 
additive (e.g. activated carbon, anthracite, gravel, pumice, sand) to achieve water treatment under 
site specific conditions. Relevant site specific parameters include the nature of contaminant, the 
water chemistry and the water flow velocity. In other words, a proper design (reason 2, § 1) should 
be extended to the width of the Fe0 PRB, the nature of the admixing agent (e.g. type and grain size 
distribution) and the proportion of Fe0 therein. 
The last important issue on contaminant breakthrough concerns the suitability of specific removal 
(Eq. 3) for a dynamic system in which reactive species are progressively generated. Es values from 
Tab. 4 show that the lowest specific removal (1.74 mg Ni/g Fe0) was obtained in system C (100 % 
Fe0). While this result seems contradictory, it corroborates the view that iron corrosion is self-
inhibitory and that decreasing the proportion of Fe0 is a powerful tool to increase sustainability 
([54] and ref. cited therein). Tab. 5 shows that correcting Es by considering the extent of Fe0 
depletion at tlimit (Es,eff) restores the intuitive trend that “the greater the adsorbent amount, the larger 
the Es value“. Accordingly, the highest Es,eff values were obtained in system C (absolute value) 
which clogged at first. This result corroborates previous findings that filtration systems containing a 
100 % Fe0 layer are efficient but not sustainable [55,56]. Moreover, the fact that the effective 
specific removals for the three systems are similar (4.0 ≤ Es,eff ≤ 6.8) is a hint that the calculation of 
the consumed iron is right. Note that, Es,eff values are derived from Es values on the basis of the 
extent of Fe0 depletion at tlimit (Es, tlimit), not at the depletion at the end of the experiment. This result 
means that no significant breakthrough was observed before tlimit. 
3.3 Hydraulic conductivity 
  
11
237 
238 
239 
240 
241 
242 
243 
244 
245 
246 
247 
248 
249 
250 
251 
252 
The results presented in Fig. 4 clearly demonstrate that the hydraulic conductivity decreases with 
time for the systems containing Fe0 particles (systems C, D, E) then remains constant at time tlimit 
(Tab. 5). 
The initial hydraulic conductivity K0 for all systems is about 5.10-4 m/s. The hydraulic conductivity 
tends to about the limit value Klimit = 5.10-9 m/s at time tlimit. The results show that the decrease of 
the permeability is down to about five orders of magnitude K0/Klimit = 1.10-5 (-). The time tlimit 
seems to depend on the investigated system (Tab. 5). The data in Tab. 5 clearly indicate that Fe0 
admixture with sand and pumice resulted in extended service life. The longest service life was 
observed for the system with pumice particles and is consistent with the fact that intra-particle 
porosity has contributed to increased permeability [57,58].  
Among the proposed models in the literature, the Kozeny-Carman equation is often considered to 
evaluate the evolution of the hydraulic conductivity [59]. This equation was developed after 
considering a porous material as an assembly of capillary tubes and yielded the hydraulic 
conductivity K as function of the porosity Φ, the specific surface S (m2/kg of solids) and a factor C 
to take into account the shape and tortuosity of channels. The first approximation is to accept the 
Kozeny-Carman equation [60-64]: 
)(.)(.)( Φ−
Φ−
Φ
Φ=
1
01
0
KtK
23
0   (6) 253 
254 
255 
256 
257 
258 
where K0 is the initial hydraulic conductivity and Φ0 the initial porosity.  
For the evaluation of the residual porosity Φ as a function of the time t, uniform corrosion for 
spherical particles with initial radius R0 (here R0 = 500 μm) is assumed. Individual particles corrode 
independently with the same kinetics until material depletion. Under these assumptions, it is 
considered that the actual radius R(t) of Fe0 particles varies linearly with time t according to:  
t
tRRtR
depletion
0
0
,
.)(
∞
−=   (7) 259 
where t depletion,∞  is the time at Fe0 depletion.  260 
261 
262 
From Eq. 5, 6 and 7, it is possible to simulate the decrease of the hydraulic conductivity as a 
function of time (Figure 4b-d). The modelling has been applied for the coefficient of volumetric 
expansion η = 6.40 in coherence with high O2 levels and for two times at Fe0 depletion ( t depletion,∞ = 
50 days and 62.5 days) [43]. It can be noticed that the maximum volume of Fe
263 
264 
265 
0 which may corrode 
is the one which leads to clogging (VFe,clogging) and is expressed by:  
1
V=V rz0gingcFe −η
Φ .
log, .  (8) 266 
267 At time tlimit it is assumed that the volume of consumed Fe0 tends to the one leading to the clogging 
of the column ( ).(. log,VV0011Vt gingcFe−=  in calculations) and remains constant for t > tlimit. Under 
these assumptions, the permeability K
268 
269 
270 
271 
272 
limit corresponding to the time tlimit is reached at the clogging 
of the columns (Φ ≈ 0, see Tab. 5). The obtained results show that there is a good agreement 
between experiment and modelling concerning the kinetic of the decrease of the permeability with 
time and the Klimit value. It can be noticed that the linear law for corrosion process (Eq. 7) with the 
two considered times ( t depletion,∞ ) as a first approximation of corrosion kinetic, allows to well 
reproduce the decrease of the permeability at the beginning of the filtration process. However, the 
evolution of the permeability around t
273 
274 
275 
276 
277 
278 
279 
280 
281 
282 
283 
limit can not be accurately predicted.  
The proposed modelling is a first attempt to predict the time-dependent decrease of the hydraulic 
conductivity (permeability loss) on the basis of the volumetric expansion of corroding iron. This 
work shows that the evolution of the hydraulic conductivity may be predicted without considering 
the evolution of the tortuosity or the specific surface in the Kozeny-Carman equation and is the 
consequence of the filling of the porosity by expansive iron corrosion products.  
3.4 Discussion 
The achieved experimental results and the proposed modelling show that there is a significant effect 
of the inner porosity of the pumice (system E). This effect is a clogging delay compared to the 
 
 
12
  
13
284 
285 
286 
287 
288 
289 
290 
291 
292 
293 
294 
295 
296 
297 
298 
299 
300 
301 
302 
303 
304 
305 
306 
307 
308 
Fe0/sand system (Fig. 4). This result is explained by the internal porosity of pumice particles which 
may store iron corrosion products, delaying the filling of the inter-granular porosity. Although 
theoretically sound [58,65,66], this hypothesis has to be confirmed in future works, for instance, 
using X-ray micro-tomography to probe inner porosity of the pumice specimen and considering 
various pumice material with differential pore connectivity.  
For a better understanding of the evolution of the initial porosity as iron corrosion proceeds, it is 
imperative to couple imaging (visualization) and mathematical modelling. The first attempt to 
visualize the deposition of iron particles (nano-scale) in the context of remediation with Fe0 was 
recently published [67]. It is expected that the use of X-ray microtomography visualization (and 
other appropriate techniques) will enable a better understanding of the effects of corrosion products 
on the bed clogging and to interpret the evolution of the residual porosity. 
An increase of the sustainability of the Fe0 bed is noticed (Tab. 5). More iron could be consumed at 
the time tlimit. The extent of Fe0 depletion is increased by using admixtures. This result corroborates 
the view that admixing Fe0 with non-expansive material is a tool to induce sustainability [14,24]. 
Accordingly, the repeatedly reported cost reduction (Fe0 costs) ([21] and ref. cited therein) should 
be regarded as a positive side-effect. In other words, while using admixtures, material wastage [12] 
is avoided and service life is increased. It seems that the Fe0 proportion in efficient real systems 
should be lower than 50 % (1:1, v/v) used here [23,24]. In fact, the efficiently permeable reactive 
barrier at Borden (Ontario, Canada) contained only 22 % Fe0 (w/w) [1]. On the other hand, while 
testing Fe0 for viruses and bacteriophages removal from drinking, Shi et al. [68] found out that a 
sand filter containing only 15 % Fe0 (w/w) was very efficient for microbe removal. The design of 
Shi et al. [68] consisted in a column packed with sand (sand filter) containing a reactive Fe0/sand 
layer (50:50, v/v). While the Fe0/sand ratio was the same as the one discussed here, this work and 
related studies propose that parameters such as the characteristics of the column, the mass of Fe0, 
the chemical reactivity of Fe0, the thickness of the Fe0/sand layer, the proportion of Fe0, the relative 
  
14
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 
size of used particles (δ values), the porosity of the additives are routinely given to enable the 
comparison of results.  
4 Conclusions 
The concept that dual Fe0/inert additive systems are more sustainable than pure Fe0 systems for 
water treatment is validated using pumice and sand as additives and CuII, NiII and ZnII as model 
contaminants. As expected the sand system was more efficient for contaminant removal and the 
pumice system more permeable. The order of contaminant removal efficiency corresponds to the 
selectivity sequence for iron oxides. This observation corroborates the view that chemical reduction 
(if applicable) is of secondary important for the process of contaminant removal. 
The presented work highlights the volumetric expansive nature of iron corrosion as the most 
important clogging factor in water treatment. Filter clogging is demonstrated to be related to pore 
blocking in the inlet zone. Therefore, reliable strategies have to be developed to design sustainable 
Fe0 filters under environmental conditions (water works). Systematic research at laboratory scale is 
needed to understand the impact of various factors on the clogging process. These factors include: 
the temperature, the nature of Fe0 (chemical reactivity) and used additives (reactivity, porosity), the 
shape and dimension of Fe0 and additives, the relative dimensions of Fe0 and additives (δ values) 
and the quality of the inflowing aqueous solution (pH, O2 level, HCO3-, humic substances, 
contaminants). The possibility to use various Fe0 materials of different reactivity in the treatment 
chains should be carefully considered. For example, a readily reactive Fe0 material (e.g. powdered) 
can be used in the first column(s) as O2 scavenger and substituted when clogging occurs. In such a 
configuration, less reactive materials (e.g. granular) are used in subsequent columns for effective 
water treatment. 
The net output of such a systematic research will be the minimization of uncertainties on the long 
term efficiency (sustainability) of Fe0-based filtration systems, including nano-scale Fe0. The lack 
of systematic approach has already led to difficulties in finding research funding [69] after more 
  
15
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 
than 15 years intensive research. Moreover, there is currently a hindrance in the spreading of this 
efficient technology in Europe [70]. At the term, a small number of avoidable failures could result 
in dismissal of a proven efficient technology. 
Acknowledgements 
The authors are grateful to Dott Giuseppe Panzera for its essential assistance during this research 
activity. 
This paper is the results of the joint work of the authors. 
References  
[1] S.F. O´Hannesin, R.W. Gillham, Long-term performance of an in situ "iron wall" for 
remediation of VOCs. Ground Water 1998, 36, 164–170. 
[2] K. Komnitsas, G. Bartzas, I. Paspaliaris, Inorganic contaminant fate assessment in zero-valent 
iron treatment walls. Environ. Forensics 2006, 7, 207–217.  
[3] K. Komnitsas, G. Bartzas, I. Paspaliaris, Modeling of reaction front progress in fly ash 
permeable reactive barriers. Environ. Forensics 2006, 7, 219–231. 
[4] 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. 
[5] R. Thiruvenkatachari, S. Vigneswaran, R. Naidu, Permeable reactive barrier for groundwater 
remediation. J. Ind. Eng. Chem. 2008, 14, 145–156. 
[6] G. Bartzas, K. Komnitsas, Solid phase studies and geochemical modelling of low-cost 
permeable reactive barriers. J. Hazard. Mater. 2010, 183, 301–308. 
[7] 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. 
[8] R.W. Gillham, Development of the granular iron permeable reactive barrier technology (good 
science or good fortune). In "Advances in environmental geotechnics : proceedings of the 
International Symposium on Geoenvironmental Engineering in Hangzhou, China, 
  
16
359 
360 
361 
362 
363 
364 
365 
366 
367 
368 
369 
370 
371 
372 
373 
374 
375 
376 
377 
378 
379 
380 
381 
September 8-10, 2009"; Y. Chen, X. Tang, L. Zhan (Eds); Springer Berlin/London (2010) 
5–15. 
[9] 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. 
[10] M. Gheju, Hexavalent chromium reduction with zero-valent iron (ZVI) in aquatic systems. 
Water Air Soil Pollut. 2011, 222, 103–148. 
[11] 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). 
[12] R.W. Puls, Long-term performance of permeable reactive barriers: lessons learned on design, 
contaminant treatment, longevity, performance monitoring and cost – an overview. In: I. 
Twardowska et al. (eds.), Soil and Water Pollution Monitoring, Protection and Remediation, 
3–23 (2006) 221–229. 
[13] S.J. Morrison, P.S. Mushovic, P.L. Niesen, Early breakthrough of molybdenum and uranium in 
a permeable reactive barrier. Environ. Sci. Technol. 2006, 40, 2018–2024.
[14] S. Caré, R. Crane, P.S. Calabro, A. Ghauch, E. Temgoua, C. Noubactep, Modelling the 
permeability loss of metallic iron water filtration systems. Clean – Soil, Air, Water 2012, doi: 
10.1002/clen.201200167. 
[15] Y. Zhang, R.W. Gillham, Effects of gas generation and precipitates on performance of Fe° 
PRBs. Ground Water 2005, 43, 113–121. 
[16] 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. 
  
17
382 
383 
384 
385 
386 
387 
388 
389 
390 
391 
392 
393 
394 
395 
396 
397 
398 
399 
400 
401 
402 
403 
404 
405 
406 
[17] L. Carniato, G. Schoups, P. Seuntjens, T. Van Nooten, Q. Simons, L. Bastiaens, Predicting 
longevity of iron permeable reactive barriers using multiple iron deactivation models. J. 
Contam. Hydrol. 2012, doi:10.1016/j.jconhyd.2012.08.012. 
[18] A.S. Ruhl, N. Ünal, M. Jekel, Evaluation of two-component Fe(0) fixed bed filters with porous 
materials for reductive dechlorination. Chem. Eng. J. 2012, 209, 401–406. 
[19] P.D. Mackenzie, D.P. Horney, T.M. Sivavec, Mineral precipitation and porosity losses in 
granular iron columns. J. Hazard. Mater. 1999, 68, 1–19.  
[20] D. Sarr, Zero-valent-iron permeable reactive barriers - how long will they last? Remediation 
2001, 11, 1–18. 
[21] 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. 
[22] 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. 
[23] K. Miyajima, Optimizing the design of metallic iron filters for water treatment. Master thesis, 
University of Göttingen, 2012, 70 pp. 
[24] K. Miyajima, C. Noubactep, Effects of mixing granular iron with sand on the efficiency of 
methylene blue discoloration. Chem. Eng. J. 2012, 200–202, 433–438. 
[25] C. Su, R.W. Puls, In situ remediation of arsenic in simulated groundwater using zerovalent 
iron: laboratory column tests on combined effects of phosphate and silicate. Environ. Sci. 
Technol. 2003, 37, 2582–2587. 
[26] P. Westerhoff, J. James, Nitrate removal in zero-valent iron packed columns. Water Res. 2003, 
37, 1818–1830. 
[27] D.I. Kaplan, T.J. Gilmore, Zero-valent iron removal rates of aqueous Cr(VI) measured under 
flow conditions. Water Air Soil Pollut. 2004, 155, 21–33. 
  
18
407 
408 
409 
410 
411 
412 
413 
414 
415 
416 
417 
418 
419 
420 
421 
422 
423 
424 
425 
426 
427 
428 
429 
[28] S. Bang, G.P. Korfiatis, X. Meng, Removal of arsenic from water by zero-valent iron. J. 
Hazard. Mater. 2005, 121, 61–67. 
[29] 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. 
2005, 39, 8032–8037. 
[30] K. Komnitsas, G. Bartzas, K. Fytas, I. Paspaliaris, Long-term efficiency and kinetic evaluation 
of ZVI barriers during clean-up of copper containing solutions. Miner. Eng. 2007, 20, 1200–
1209. 
[31] 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. 
[32] J. Li, Y. Li, Q. Meng, Removal of nitrate by zero-valent iron and pillared bentonite. J. Hazard. 
Mater. 2010, 174, 188–193. 
[33] 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. 
[34] C. Noubactep, An analysis of the evolution of reactive species in Fe0/H2O systems. J. Hazard. 
Mater. 2009, 168, 1626–1631. 
[35] C. Noubactep, S. Caré, Designing laboratory metallic iron columns for better result 
comparability. J. Hazard. Mater. 2011, 189, 809–813. 
[36] C. Noubactep, E. Temgoua, M.A. Rahman, Designing iron-amended biosand filters for 
decentralized safe drinking water provision. Clean: Soil, Air, Water 2012, 40, 798–807. 
[37] 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. 
  
19
430 
431 
432 
433 
434 
435 
436 
437 
438 
439 
440 
441 
442 
443 
444 
445 
446 
447 
448 
449 
450 
451 
452 
453 
[38] 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. 
[39] S. Tuladhar, L.S. Smith, SONO filter: An excellent technology for save water in Nepal. 
SOPHEN 2009, 7, 18–24. 
[40] K.H. Head, G.P. Keeton, Permeability, shear strength & compressibility tests, In: Manual of 
Soil Laboratory Testing, vol. 2. Whittles Publishing: United Kingdom, 2008. 
[41] APHA, AWWA, WEF, Standard Methods for the examination of water and wastewater, 21st 
ed. American Public Health Association, Washington D.C. (USA), 2005. 
[42] 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. 
[43] 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. 
[44] C. Noubactep, Processes of contaminant removal in “Fe0–H2O” systems revisited. The 
importance of co-precipitation. Open Environ. J. 2007, 1, 9–13. 
[45] C. Noubactep, A critical review on the mechanism of contaminant removal in Fe0–H2O 
systems. Environ. Technol. 2008, 29, 909–920. 
[46] C. Noubactep, The fundamental mechanism of aqueous contaminant removal by metallic iron. 
Water SA 2010, 36, 663–670. 
[47] C. Noubactep, The suitability of metallic iron for environmental remediation. Environ Prog. 
Sustain Energy 2010, 29, 286–291. 
[48] C. Noubactep, Aqueous contaminant removal by metallic iron: Is the paradigm shifting? Water 
SA 2011, 37, 419–426. 
  
20
454 
455 
456 
457 
458 
459 
460 
461 
462 
463 
464 
465 
466 
467 
468 
469 
470 
471 
472 
473 
474 
475 
476 
477 
478 
[49] P.C. Gomes, M.P.F. Fontes, A.G. Silva, E.S. Mendonca, A.R. Netto, Selectivity sequence and 
competitive adsorption of heavy metals by Brazilian soils. Soil Sci. Soc. Am. J. 2001, 65, 
1115–1121. 
[50] M.P.F. Fontes, P.C. Gomes, Simultaneous competitive adsorption of heavy metals by the 
mineral matrix of tropical soils. Appl. Geochem. 2003, 18, 795–804. 
[51] Y.N. Vodyanitskii, The role of iron in the fixation of heavy metals and metalloids in soils: a 
review of publications. Eurasian Soil Sci. 2010, 43, 519–532. 
[52] 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.  
[53] T.B. Scott, I.C. Popescu, R.A. Crane, C. Noubactep, Nano-scale metallic iron for the treatment 
of solutions containing multiple inorganic contaminants. J. Hazard. Mater. 2011, 186, 280–
287. 
[54] 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. 
[55] A. Hussam, Contending with a Development Disaster: SONO Filters Remove Arsenic from 
Well Water in Bangladesh. Innovations, 2009, 4, 89–102. 
[56] 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. 
2009, 35, 157–161. 
[57] P.S. Calabrò, N. Moraci, P. Suraci, Estimate of the optimum weight ratio in zero-valent 
iron/pumice granular mixtures used in permeable reactive barriers for the remediation of 
nickel contaminated groundwater. J. Hazard. Mater. 2012, 207–208, 111–116. 
[58] C. Noubactep, S. Caré, Enhancing sustainability of household water filters by mixing metallic 
iron with porous materials. Chem. Eng. J. 2010, 162, 635–642. 
  
21
479 
480 
481 
482 
483 
484 
485 
486 
487 
488 
489 
490 
491 
492 
493 
494 
495 
496 
497 
498 
499 
500 
501 
[59] B. Courcelles, A. Modaressi-Farahmand-Razavi, D. Gouvenot, A. Esnault-Filet, Influence of 
precipitates on hydraulic performance of permeable reactive barrier filters. Int. J. Geomech. 
2011, 11, 142–151. 
[60] P.C. Carman, Fluid flow through a granular bed. Trans. Inst.Chem. Eng. 1937, 15, 150–166. 
[61] W.D. Carrier III, Goodbye, Hazen; Hello, Kozeny-Carman. J. Geotechn. Geoenviron. Eng. 
2003, 129, 1054–1056. 
[62] R.P. Chapuis, Predicting the saturated hydraulic conductivity of sand and gravel using effective 
diameter and void ratio. Canadian Geotech. J. 2004, 41, 787–795. 
[63] P.W.J. Glover, E. Walker, Grain-size to effective pore-size transformation derived from 
electrokinetic theory. Geophysics 2009, 74, E17–E29. 
[64] Y. Chen, R. Hu, W. Lu, D. Li, Ch. Zhou, Modelling coupled processes of non steady seepage 
flow and non linear deformation for a concrete-faced rockfill dam. Comput. Struct. 2011, 89, 
1333–1351. 
[65] C. Noubactep, S. Caré, Dimensioning metallic iron beds for efficient contaminant removal. 
Chem. Eng. J. 2010, 163, 454–460. 
[66] 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 – Soil, Air, Water 2010, 38 
951–959. 
[67] Q. Wang, J.-H. Lee, S.-W. Jeong, A. Jang, S. Lee, H. Choi, Mobilization and deposition of iron 
nano and sub-micrometer particles in porous media: A glass micromodel study. J. Hazard. 
Mater. 2011, 192, 1466–1475. 
[68] C. Shi, J. Wei, Y. Jin, K.E. Kniel, P.C. Chiu, Removal of viruses and bacteriophages from 
drinking water using zero-valent iron. Separ. Purif. Technol. 2012, 84, 72–78. 
  
22
502 
503 
504 
505 
506 
507 
508 
509 
[69] N.C. Mueller, J. Braun, J. Bruns, M. Cerník, P. Rissing, D. Rickerby, B. Nowack, Application 
of nanoscale zero valent iron (NZVI) for groundwater remediation in Europe. Environ. Sci. 
Pollut. Res. 2011, 19, 550–558. 
[70] A.S. Ruhl, A. Weber, M. Jekel, Influence of dissolved inorganic carbon and calcium on gas 
formation and accumulation in iron permeable reactive barriers. J. Contam. Hydrol. 2012, 
142–143, 22–32. 
 
  
23
509 
510 
511 
512 
Table 1: Variability of the operational conditions for column experiments as illustrated by the 
dimension of the column (H, D), the nature of contaminants (X), the nature of additives, the Fe0 
mass, and the relative dimension of particles (δ). 
 
X H D Additive Fe
0 dFe(∗) dadditive(∗) δ (∗∗) Ref. 
 (cm) (cm)  (g) (mm) (mm) (-)  
TCE. PCE 50 3.8 Sand n.a. 0.25 1.315 0.19 [1] 
As 31 2.5 Sand 75 0.42 0.275 0.65 [25] 
NO3- 30 5 Sand 1636 0.3 n.a. (-) [26] 
Cr 10 2.5 Sand n.a. 1.45 0.638 0.44 [27] 
As 17.8 5.1 Sand 400 0.15 0.4 0.38 [28] 
As 4 0.1 Sand 1.5 n.a. 0.5 (-) [29] 
Cu 45 5 Sand 525 0.7 0.8 0.88 [30] 
TCE 40 1.59 Sand 80 1.355 0.118 0.09 [21] 
As n.a. n.a. Sand n.a. 0.5 0.5 1.0 [31] 
NO3- 20 2 Sand 90 0.1 0.55 0.19 [32] 
Cu. Ni 50 5 Pumice 240 0.5 0.30 0.6 [33] 
Cu. Ni, Zn 50 5 Sand 200 0.5 0.28 0.6 present 
Cu. Ni, Zn 50 5 Pumice 200 0.5 0.30 0.6 present 
 513 
514 
515 
516 
517 
(∗) dFe and dadditive are the the mean grain size of Fe0 and the additive respectively. 
(∗∗) δ is the diameter ratio of the smaller particle size to the larger one (Fe0 or additive). 
 
  
24
517 
518 
Table 2: Characteristics of used materials tested by Mercury Intrusion Porosity (MIP). 
 
 Fe0 Pumice Sand 
Specific weight ρs (g/cm3) 7.78 1.92 2.60 
Apparent specific weight ρas (g/cm3) 7.78 1.14 2.60 
Compactness C(-) 0.51 0.45 0.55 
Inter particular porosity Φinter (%) 49.6 54.8 45.0 
Intra particular porosity ϕpp  (%) - 41.0 - 
Porosity Φ0 (%) 49.6 73.3 45.0 
519 
520 
 
  
25
520 
521 
Table 3: Main characteristics of the studied columns. “rzeff” is the measured reactive zone. The 
estimated porosity is also given.  
System Media Fe0 Fe0 Additive Duration rzeff. Porosity*
  (vol %) (g) (g) (days) (cm) (%) 
A sand 0.0 0.0 1060 28 40.0 45.05 
B pumice 0.0 0.0 27 28 2.6 72.6 
C Fe0 100.0 200 0.0 17** 2.6 49.6 
D Fe0 + sand 50.0 200 76.4 15** 5.2 46.05 
E Fe0 + pumice 50.0 200 27.0 28** 5.0 59.5 
* in this values the internal porosity of the pumice is also included. 522 
523 
524 
525 
526 
527 
**-marked systems were stopped because of excessive permeability loss. 
 
 
 
  
26
527 Table 4: Magnitude of contaminant removal in investigated systems. 
System min (mg) E (%) Es (mg/g) 
 Ni Cu Zn Ni Cu Zn Ni Cu Zn 
A 685 685 927 57.3 99.8 65.7  n.a.  
B 685 685 927 57.2 51.5 52.1  n.a.  
C 367 367 497 94.7 99.9 99.8 1.74 1.83 2.48 
D 367 367 497 99.3 99.8 99.9 1.82 1.83 2.48 
E 612 612 828 93.3 99.9 99.6 2.86 3.06 4.12 
 528 
529 
530 
 
  
27
530 Table 5: Modelled characteristics of Fe0-containing columns. The extent of Fe0 depletion (%) is 
given by: 
V
VV100consumedFe
0
t0 −= .%  where (V0 –Vt) is the consumed Fe0 volume with V0 the 
initial volume of Fe
531 
532 
533 
0 and Vt its residual Fe0 at time t. 
 
System tlimit (*) Fe0 depletion Fe
0
residual Φ/Φ0 (**) Es.eff (mg/g) (***)
 (days) (%) (g) (%) Ni Cu Zn 
Fe0 7.5 17.1 168.8 4.6 5.0 5.0 6.8 
Fe0:sand 10.0 33.2 133.6 2.0 4.0 4.1 5.5 
Fe0:pumice 16.0 41.4 117.2 1.4 4.3 4.4 6.0 
534 
535 
536 
537 
538 
539 
540 
 
(*) “tlimit” is the time which corresponds to a constant hydraulic conductivity. 
(**) Φ/Φ0 is the residual porosity. Φ0 is the initial porosity of the column. 
(***) Es,eff is the specific removal. 
 
 
  
28
540 
541 
542 
543 
544 
545 
546 
547 
548 
549 
550 
551 
552 
Figure captions 
 
Figure 1: Schematic diagram of the experimental design. Used materials were (i) Fe0 (0 or 200 g), 
(ii) pumice (0 or 27 g), and (iii) quartz gravel (0 or 1060 g). 
 
Figure 2: Time-dependant evolution of the pH value (a) and the iron concentration (b) of column 
effluent. The lines are not fitting functions, they simply connect points to facilitate visualization. 
 
Figure 3: Magnitude of Cu (a), Ni (b), and Zn (c) breakthrough from the columns containing Fe0. 
The lines are not fitting functions, they simply connect points to facilitate visualization. 
 
Figure 4: Time-dependant evolution of the hydraulic conductivity in the three systems containing 
Fe0 and the reference system pumice: (a) experimental K/K0 values; (b, c and d) relative 
permeability K/K0, experimental and modelled values. η = 6.4, model 1: t depletion,∞  = 50 days, and 
model 2: 
553 
t depletion,∞  = 62.5 days. For the three Fe0-systems, the value at time tlimit has been adjusted so 
that the residual content of Fe
554 
555 
556 
557 
0 is slightly superior to the Fe0 volume which is not consumed at 
clogging.