1
Manuscript for Journal of Radioanalytical and Nuclear Chemistry
Investigating the Release of Co-precipitated Uranium from Iron Oxides
Noubactep C.*(a), Schöner A. (b), Dienemann H. (c), Sauter M.(a)
(a) Centre of Geosciences - Applied Geology; Goldschmidtstrasse 3, D - 37077 Göttingen;
(b) Department for Geology, FSU Jena; Burgweg 11, D - 07749 Jena;
(c ) Institute of General Ecology and Environmental Protection of Tharandt, Pienner Straße 8,
D - 01737 Tharandt.
(*) corresponding author: cnoubac@gwdg.de;
Tel. +49 551 39 3191, Fax: +49 551 399379
Abstract:
The removal of uranium (VI) in zerovalent iron permeable reactive barriers and wetlands can
be explained by its association with iron oxides. The long term stability of this immobilized U
is yet to be addressed. The presented study investigates the remobilization of U(VI) from iron
oxides via diverse reaction pathways (acidification, reduction, complex formation) in the
laboratory. Prior uranium co-precipitation experiments were conducted under various
conditions. The addition of various amounts of a pH-shifting agent (pyrite), an iron
complexing agent (EDTA) or an iron (III) reduction agent (TiCl3) yielded in uranium
remobilisation to concentrations above the US EPA maximum allowed contaminant level
(MCL = 30 µg/L). This study demonstrates that U(VI) release in nature will strongly depend
on the conditions and the mechanism of its fixation by geological materials.
Key Words: co-precipitation, iron oxides, pyrite, remobilization, uranium, reactive barriers.
2
Introduction
Iron oxides, ubiquitous in soils and sediments, are known to play an important role in the
mobility (retardation and transport) of many organic and inorganic contaminants in natural
environments. The retention property is primarily due to their large surface area, their strong
adsorptive properties and their high adsorptive capacity for both cationic and anionic species.
1,2
Iron (hydr)oxides readily eliminated inorganic contaminants from the aqueous phase via
different competing mechanisms: adsorption, co-precipitation, and reduction on green rust.3
The extent of co-precipitatio n depends on the bio-geochemical conditions, in particular on the
reactivity of iron oxides and the contact time of the contaminant with them. Generally it can
be assumed that co-precipitation is favoured when the iron (III) concentration is high and the
contaminant concentration is low. 2,4-8 In this case the contaminant is first adsorbed onto
amorphous ferrihydrite (e.g. Fe(OH)3) and co-precipitated with it as it is aged.9-11 Depending
on the geometrical characteristics of the contaminant relative to iron, stoichometrical co-
precipitates can be formed (e.g. FeAsO4.xH2O, Fe1-xCrx(OH)3). Whether the associated
contaminant is primarily structurally incorporated or surface-adsorbed is important for the
long term stability of the fixed contaminant.
Several active and passive remediation strategies aim at immobilizing long-term
contaminants. Among the passive strategies in which contaminant retention by interactions
with iron oxides are important, wetlands, permeable reactive barriers and natural attenuation
can be listed.2,12 Under relevant natural conditions the interactions between iron oxides,
organic components, and microorganis ms will increase the contaminant removal. For
example, Zuyi et al.13 showed an elevated U(VI) sorption by iron oxides in the presence of
fulvic acids. The important impact of microorganisms for the contaminant retention has been
3
demonstrated for example by Behrends & Cappellen14 and Nico et al.15. These authors have
shown that Shewanella putrefaciens can induce the reduction of U(VI) to U(IV). However,
the long term optimisation of microbiological processes under natural conditions is yet to be
properly investigated. Biological processes are not experimentally addressed in this study.
A zerovalent iron (Fe°, ZVI) permeable reactive barrier is a typical case where the ratio
Fe:contaminant is high to favour co-precipitation. Table 1 summarises some relevant
processes governing the fate of iron.
Iron corrosion (Eq.1) produces ferrous iron (Fe2+) that can be further oxidized to ferrihydrite
(Eq. 2, Fe(OH)3) which is an excellent trace metal (e.g. As, Cr, U) adsorbent. Ferrous and
ferric ions from Eq.1 and Eq.2 can react to build green rust (Eq. 3) which is known for their
reductive capacity.16 Ferrihydrite from Eq.2 then aged and transformed into several iron
oxides (Eq. 4) entrapping the contaminant in their mass (co-precipitation). These processes
are probable in wetlands with elevated iron (III) concentration and relative low pH values.5
Several processes are capable of releasing co-precipitated metal ions from iron oxides back
into solution. These processes are of special interest concerning the mobility of contaminants
in natural systems. Some relevant remobilisation processes are listed in table 1;17,18 they
include: (1) acidic oxide dissolution (Eq. 5 & 6), (2) oxide dissolution by complex formation
(Eq. 7), and (3) abiotic (or biotic) reductive dissolution of oxides (Eq. 8). The understanding
of these individual processes is crucial in the comprehension of the mobility of metal ions in
the geosphere and therefore, the prediction of long term stability of (inorganic) contaminants
that are associated with iron oxides.
The remobilisation of metals by synthetic anthropogenic chelating agents such as EDTA
(ethylenediaminetetraacetic acid) and NTA (nitrilotriacetic acid) has been addressed because
of their potential for increasing the solubilisation and remobilisation of heavy metals from
aquatic sediments or from aquifer material during the infiltration from river water to
4
groundwater.18 The metal (re)mobilisation by acidification has been mostly discussed in the
context of acid mine drainage, AMD.19 The investigation of reductive transformation of iron
oxides and their impact on trace metal mobility and remediation is currently under
investigation.20,21
The geochemical conditions of the system “U(VI)-iron oxides” can be summarized as
follows: low pH destabilizes iron oxides; chelates (e.g. EDTA) decrease U(VI) sorption and
dissolve iron oxides; iron hydroxides are dissolved at low redox potential (EH) but U(VI)
reduction is favoured, and high carbonate levels provide favourable conditions for uranium
mobilization.5 To date, the association of U(VI) with iron oxides has been mostly investigated
with the objective to understand its environmental retardation process,1,10 or to develop
efficient decontamination processes for contaminated steels.9 Since it was demonstrated that
the U(VI) retention in reactive barriers is not necessarily the result of a “reductive
precipitation”12,20,22 there is an high need for the understanding of the fate of U(VI) as iron
oxides undergo transformations, in particular dissolution.
The present study aims at investigating the influence of fundamental mechanisms of iron
oxides transformation on the U(VI) removal from corrosion products. For this purpose, iron
complexation by EDTA, acidification by pyrite and reductive dissolution by Ti(III) were
tested and their effects on the remobilisation of co-precipitated U(VI) from iron corrosions
products were recorded. Experimental results were discussed for their implications on the
long-term efficiency of two passive remediation techniques: permeable reactive barrier and
wetlands.
Experiments
Materials
The used ZVI is a scrap iron from MAZ (Metallaufbereitung Zwickau, Co.). Its
elemental conditions are determined as 3.52% C, 2.12% Si, 0.93% Mn, 0.66% Cr, and
5
92.77% Fe. The materials were fractionated by sieving. The fraction 1.0-2.0 mm was used
without any further pretreatment. ZVI is used as U(VI) reducing agent.
Pyrite was crushed and sieved. Five different particle sizes (di, mm) of pyrite were
used: 0.063 < d1 < 0.125; 0.125 < d2 < 0.250; 0.250 < d3 < 0.315; 0.315 < d4 < 0.63; 0.63 <
d5 < 1.0. The elemental composition is 40% Fe, 31.4% S, 6.7% Si, 0.5% Cl, 0.15% C, and
<0.01% Ca. The material served as a pH-shifting agent, diminishing the sorptive reactivity of
corrosion products, enhancing the solubility of U(VI), promoting the co-precipitation of
U(VI) with iron corrosion products in long term experiments.11
Fixation Experiments, Desorption with Na2CO3
The experimental procedure for the fixation experiments, the desorption by 0.1M
Na2CO3 and the analytical method is described in detail elsewhere11,12 and will not be repeated
here. Fixation studies consisted in different not shaken experiments for given duration with 5
g/L ZVI, and 15 g/L FeS2: ZVI and additives were allowed to react in sealed sample tubes
containing 20.0 mL of an uranium solution (20 mg/L or 0.084 mM) at laboratory temperature
(about 20° C). The tubes (16 ml graded) were filled to the total volume to reduce the head
space in the reaction vessels. The contact vessels were allowed to equilibrate in darkness to
avoid photochemical side reactions, the initial pH was ~7.2. Desorption experiments were
conducted in a 0.1 M Na2CO3 solution for about 14 h. The experiments were performed in
triplicates. The mean values are presented.
The experiments were conducted in closed essay tubes under non controlled O2 (and CO2)
pressure. It is certain that PO2 was less than the atmospheric pressure. It can be assumed that
U(VI) removal mainly occurred under very low O2 partial pressure, because iron corrosion
(and pyrite dissolution) is O2 consuming. Note that strictly anoxic conditions are not expected
in the majority of ZVI reactive walls, since the technology is yet applicable for rather shallow
plumes: 15 - 22 m (50 - 70 feet deep; US EPA 199823). Therefore, working at low oxygen
6
level (PO2 < PO2,atm and PO2 ≠ 0) is a good simulation for groundwater situations at several
sites.
Remobilisation Experiments with Pyrite, EDTA, HCl and TiCl3
Previous fFixation experiments were conducted for one, two or three months with ZVI (15
g/L) alone or the system “ZVI + FeS2 (d2)”, containing 25 g/L FeS2. The remobilisation then
occurred for a given duration or as a function of time through the addition of defined amounts
of additives: pyrite (1g or 50 g/L), EDTA (10 mM) and TiCl3 (1.25 %). The selection of these
reactants was motivated by previous results11, 12 and literature data from Heron et al.24 and
Ford 25. The aim was to achieve different dissolution grads of corrosion products.
Analytical Method
Analysis for U was performed by inductively coupled plasma mass spectrometry (ICP-MS) at
the Institute of Geosciences of the University of Jena. All chemicals used for experiments and
analysis were of analytical grade. The pH value was measured by combination glass
electrodes (WTW Co., Germany). The electrode were calibrated with five standards following
a multi-point calibration protocol26 and in agreement with the new IUPAC recommendation.27
All experiments were performed in triplicate. Error bars given in figures represent the
standard deviation from the triplicate runs.
Results
The experiments were compared on basis of the final U(VI) concentration (C in µg/L), the
total fixation Ptot (in %) defined by Eq. 9:
Ptot = 100% * (1 - (C/C0)) (9)
where C0 is the initial concentration of uranium in solution, while C gives the uranium
concentration after the experiment. The percent recovery, Prev, of uranium after the end of the
7
experiment (recovery with 0.1 M Na2CO3, 0.01 M EDTA, TiCl3, or pyrite (d i)) is calculated
by Eq. (10)
)(
)(%100
00
100
CCV
VVCPrev
−
−
= (10)
where V0 gives the initial volume, and V1 the volume after removing solution for uranium
analysis at the end of the fixation experiment.
Uranium fixation and remobilization with 0.1 M Na2CO3
< Figure 1 >
Figure 1 shows a typical kinetic curve for the uranium total fixation and reversible fixation
with 0.1 M Na2CO3 under the experimental conditions of this study. It can be seen that after
three weeks (24 days) the total fixation of aqueous uranium was almost completely achieved
(Ptot > 99 %), whereas the reversibility of the uptake as achieved with 0.1 M Na2CO3 was less
than 10 % (Prev) after one month. Based on this observation, a minimal fixation duration of
one month was selected for the further experiments aiming at investigating the uranium
release from iron oxides by processes likely to occur in nature. In one experiment co-
precipitation was promoted by the addition of pyrite (FeS2, d2) and an experimental duration
of three months. As discussed elsewhere,12 Na2CO3 is not able to dissolve nor to transform
iron oxide. On the other hand, the use of 0.1 M Na2CO3 as remobilizing agent has been shown
to be inadequate since its employment yields to elevated Na+ concentration at high pH values
and therefore to a likely formation of sodium uranates such as Na2UO4. Uranates formation
induces an underestimation of the reversibility of U removal since U in uranates is supposed
to be irreversibly fixed onto iron oxides.27 This study investigates some plausible scenarios
likely to occur in nature to gain a realistic idea on the reversibility of co-precipitated uranium
both in wetlands and reactive barriers.
8
Effects of Various Factors on the Mobilization of Co-precipitated Uranium
There are two major possible pathways that can induce the release of co-precipitated uranium:
• local change of the solution chemistry; changes in pH or EH for example by infiltration
of waters from acidic or reduced zones. Alternatively this change can be the result of
the weathering of available minerals (e.g. FeS2).
• dissolutive transformation of iron oxides (e.g. complex formation by infiltrating
chelating agents, biotic and abiotic iron oxide reduction).
To gain an impression on the fate of co-precipitated uranium as corrosion products are
transformed in the environment, calculated amounts of target additives were added to reaction
vessels after two months of uranium fixation to achieve the given final concentration of the
transformation agents. Mobilization agents were pyrite (d3, d4), 0.1 M Na2CO3 as reference
desorption agent for laboratory investigations, 10 mM EDTA as an environmentally relevant
complexing agent, and 1.25 % TiCl3 as iron oxide reducer. Note that EDTA can increase the
dissolved concentration of U(VI) by two processes: by remobilization of adsorbed or
precipitated U(VI) and by dissolution of iron oxides.
< Figure 2 >
Figure 2 summarizes the results of U(VI) recovering (Prec) by the enumerated agents for an
experimental duration of 14 hours. The U(VI) removal efficiency varied from 0.3% to 40%
depending on the treatment agent. As shown in Fig. 2, the two different particle sizes of the
used pyrite gave the same recovery efficiency of about 0.3 % (the lowest). The recovery
efficiency by EDTA was about one half of that of Na2CO3 (1.3 %) for the experimental
duration. However, it can be expected that the recovery efficiency for EDTA will increase
with the time since the complexation kinetics by EDTA depends on the crystallization grade
of iron oxides.18 Finally the recovery efficiency for the system including TiCl3 was about 40
% for the 14 hours. It should be emphasized that, although the reduction of all Fe(III)
9
contained in the available amount of corrosion products in each vessel is completed, a total
recovery of U(VI) can not be expected since U(VI) also adsorbs onto Fe(II) colloids.12
The above experiments show that partial or total dissolution of iron oxides in wetlands or
reactive barriers will be associated with a release of sequestrated U(VI) into the environment.
Note that the experiment with TiCl3 solely shows the fate of co-precipitated U(VI)
qualitatively, as iron oxides are reduced. A quantitative characterization is almost impossible
under the experimental conditions of this work because of the complicated interactions
between U(VI) and Fe(II) and Fe(III) with increasing pH.12 However new concepts have to be
developed to characterize the fate of co-precipitated U(VI) in the environment as physical,
chemical or biological transformations of corrosion products occur.
Effect of local acidification
The effect of local acidification was studied by two sets of experiments. The first used various
particle sizes of pyrite (d1 to d5) and both fixation and recovery experiments were conducted
for 30 days (Fig. 3). The second set used a pyrite particle size d5 (0.63 < d (mm) < 1.0), the
fixation experiment was conducted for 30 days and the recovery experiment was conducted
for 0 to 50 days (Fig. 4).
< Figure 3 >, < Figure 4 >
Figure 3 shows that the addition of pyrite either induces the uranium release or delays its
uptake. Towards the end of the experiment, the U(VI) concentration in the reference system
(ZVI alone, 60 days fixation) was 6.1 µg/L and, together with that in the systems with FeS2
(d1) and FeS2 (d5), was below the US EPA maximum contaminant level (MCL = 30 µg/L). In
all other systems (FeS2:2d3 , d3, d4, d2) uranium concentration was above the MCL. This result
suggests that a local acidification can release co-precipitated U(VI) in wetlands and
permeable barriers. From the variation of the pH value with the particle size (Fig. 3) it is
apparent that, the smaller the particles (d1 being the smallest), the lower the pH value. This
10
implies that reactivity increases with decreasing particle size in accordance with other
observation in literature.29 This tendency is not confirmed by the order of efficiency of U(VI)
release: d1 < d5 < 2 d3 < d3 < d4 < d2. These results are not surprising since the impact of pyrite
is twofold and conflicting: acidifying the solution (therefore promoting U(VI) release) as
discussed and adsorbing uranium (lowering U(VI) concentration). The results for both
systems with FeS2 (d3) illustrates this clearly. In fact, the system “FeS2 (d3)” with 25 g/L
pyrite induced a smaller pH decrease than “FeS2 (2 d3)” with 50 g/L pyrite. The U(VI) release
was lesser in the system with 50 g/L pyrite due to the adsorption onto pyrite material. The
behaviour of the system for 25 g/L pyrite (d5) was further investigated for 50 days (Fig. 4).
Figure 4 shows that the variation of pH with time was not uniform. The pH first decreased as
a result the addition of pyrite (FeS2 dissolution) from an initial value of 8.54 to a minimum of
6.52 after 3 days and again increased to 7.34 at the end of the experiment (day 50). The
evolution of the uranium concentration was not synchronous. The uranium concentration
decreased in all cases as a result of pyrite addition. The decrease was initially uniform from
the beginning to day 19, then a progressive increase occurred until day 40 and the
concentration decreased again to the end value (37 µg/L > MCL). It should be noted that the
variation of the pH value was not noticeable, a smaller particle size of pyrite (e.g. d2 or d3)
would have permitted a better discussion of the processes. Nevertheless, the competing
processes (fixation through adsorption and mobilization through acidification) governing the
U(VI) release could be addressed. The large variation within the triplicates (error bars on Fig.
4) provide an impression of the likely complexity of the involved processes.
Effects of reductive transformation of iron oxides
Investigating reductive transformation of iron oxides and their impact on trace metal (e.g.
As5+, Cr6+, U6+) mobility in the environment receives an increasing interest since the major
uptake mechanism of several contaminants by Fe° materials is not their chemical
11
reduction.2,20,30 It is apparent from Figure 2 that co-precipitated uranium (30 days with ZVI),
approximately 40 % could be resolubilized within 14 hours in the presence of 1.25 % TiCl3.
To further investigate this finding another fixation experiment was conducted for three
months in the presence of 25 g/L FeS2 (d2) (system “ZVI + FeS2 (d2)”). The uranium release
was recorded as a function of time (for 57 hours). Figure 5 shows that the uranium release
initially rapidly increased with time (first 12 hours) and then reached a plateau (hour 12 to
36). After that the release rate tended to decrease. The error bars show that the standard
deviation (s) of the triplicates was very large within the first six hours and at the end of the
experiment (s > 4 %). Maximum remobilization efficiency in this experiment (27 %) was less
than the 40 % in the absence of pyrite (14 hour experiment with 30 day fixation). As
discussed elsewhere,11,20 the U(VI) uptake was delayed in the system with pyrite (ZVI +
FeS2), yielding to a progressive U(VI) co-precipitation with aging iron oxides around pH 4.
This result clearly indicates that the datum at which U(VI) is associated with native iron oxide
(reactivity, crystallization degree) is essential for the stability of the co-precipitated U(VI).
Future investigations will have to address this aspect.
< Figure 5 >
Discussion and conclusions
The widespread evidence for various degrees of irreversible uptake of contaminants by soils
and soil minerals have been reported.4,7,8,9,17 Recent results in the context of groundwater
remediation with so called passive techniques have demonstrated the importance of
irreversible contaminant removal by iron oxides.2,11,12,30 However existing transport codes do
not typically account for irreversible contaminant uptake by the rock/soil matrix.5
This work demonstrates that the remobilization of U(VI) from iron oxides is determined by
three factors: (1) the age and the crystallinity of iron oxide, (2) the contact time of the
contaminant with iron oxides, and (3) kinetics of the co-precipitation reaction. Typically, if
12
U(VI) is adsorbed onto aged corrosion products, the fixation mechanism is ion exchange and
the reaction is almost completely reversible. Sorption onto amorphous iron hydroxides
(Fe(OH)3) is often observed to be irreversible over time spans exceeding years.5 This
irreversibility is promoted if amorphous Fe(OH)3 is generated and allowed to age in the
presence of U(VI); in this manner, U(VI) is entrapped it in the matrix of aging Fe(OH)3: that
is the process of co-precipitation.6,11 Therefore U(VI) can only be released if the oxide is
destroyed (dissolutive transformations).
The rate and extent of U(VI) dissolution in the individual systems depend on its association
with the oxide. Eng et al.9 reported that U(VI) present as oxyhydroxide or polyuranate species
undergoes rapid dissolution followed by a slow dissolution of iron, while inner-sphere
complexation of U(VI) with iron resulted in concomitant dissolution of U(VI) and Fe. A
thorough understanding of the association of uranium with iron oxides at the molecular level
is useful for the prediction of the long term stability of co-precipitated U(VI) in passive
decontamination processes.
Some scenarios of oxide dissolution are presented in this study. The acidic dissolution
through pyrite weathering and the dissolution through complex formation are surely possible
in an aquifer. The reductive dissolution, as investigated in this study (TiCl3 at low pH, Fig. 5),
is solely of qualitative importance. Better results can be obtained with selected reducing
agents, efficient at neutral pH values (e.g. 0.008 M Ti3+ in 0.05 M EDTA at pH 6,23). In nature
it can be expected that enzymatic oxide reduction will play a more important role than abiotic
reduction.
To access the long term stability of co-precipitated U(VI) in any specific case, a fundamental
understanding of the likely range of groundwater compositions over time and their effect on
iron oxides in the future is needed. Among the factors to be considered the following are very
important: (1) weathering of soil minerals, (2) atmospheric inputs, and (3) biological activity.
13
Finally, since drastic changes in the compositions of natural waters are more an exception
than the rule,5 it can be considered that the factors favouring U(VI) co-precipitation (in
wetlands or reactive barriers) will be maintained far into the future. However, continuous
surveillance and monitoring of the groundwater is needed in order to detect and evaluate
eventual U(VI) release from the barrier zone. Moreover, alternatives for a satisfactorily U(VI)
mitigation downstream from the barrier have to be envisaged.
Acknowledgments
The scrap iron („S69“) was kindly purchased by the branch of the MAZ (Metallaufbereitung
Zwickau, Co) in Freiberg. The work was supported by the Deutsche Forschungsgemeinschaft
(DFG), contract Sa 501/15-1.
References
1. Ch.- K. Hsi, D. Langmuir, Geochim. Cosmochim. Acta 49 (1985) 1931.
2. W. R. Richmond, M. Loan, J. Morton, G. M. Parkinson, Environ. Sci. Technol. 38 (2004)
2368.
3. H. C. B. Hansen, S. Guldenberg, M. Erbs, C. Koch, Appl. Clay Sci. 18 (2001) 81.
4. C. C. Ainsworth, J. L. Pilou, P. L. Gassman, W. G. Van Der Sluys, Soil Sci. Soc. Am. J.
58 (1994) 1615.
5. P. V. Brady, B. P. Spalding, K. M. Krupka, R. D. Waters, P. Zhang, D. . Borns, W. D.
Brady, Sandia report. SAND99-0464, (1999) 146 pp.
6. M. C. Duff, J. U. Coughlin, B. D. Hunter, Geochim. Cosmochim. Acta 66 (2002) 3533.
7. M. F. Schultz, M. M. Benjamin, J. F. Ferguson, Environ. Sci. Technol. 21 (1987) 863.
8. R. K. Schulz, H. H. Riedel, Soil Sci. 91 (1961) 262.
9. C. W. Eng, G. P. Halada, J. A. Francis, J. C. Dodge, J. Gillow Surf. Interf. Anal. 35
(2003) 525.
14
10. S. J. Morrison, R. R. Spangler, V. S. Tripathi, J. Cont. Hydrol. 17 (1995) 333.
11. C. Noubactep, G. Meinrath, P. Volke, H.-J. Peter, P. Dietrich, B. Merkel, In: B. J. Merkel,
B. Planer-Friedrich, C. Wolkersdorfer, (Eds.), Uranium in the Aquatic Environment.
Springer, Berlin, (2002) 577.
12. C. Noubactep, G. Meinrath, P. Dietrich, B. Merkel, Environ. Sci. Technol. 37 (2003)
4304.
13. T. Zuyi, C. Taiwei, D. Jinzhou, D. Xiong, G. Yingjie, Appl. Geochem. 15 (2000) 133.
14. T. Behrends, v. P. Cappellen, Abstract. International Workshop on biochemical Processes
involving iron minerals in natural waters 16.-21. Nov. 2003 Centro Stefano Franscini
Monte Vertà Schwitzerland. (2003) 9.
15. P. S. Nico, S. G. Benner, S. E. Fendorf, Abstract. International Workshop on biochemical
Processes involving iron minerals in natural waters 16.-21. Nov. 2003 Centro Stefano
Franscini Monte Vertà Schwitzerland. (2003)
16. P. Refait, M. Abdelmoula, F. Trolard, J.-M. R. Génin, J.J., Ehrhardt , G. Bourrié, Amer.
Miner. 86 (2001) 731.
17. B. R. Coughlin, A.T. Stone, Environ. Sci. Technol. 29 (1995) 2445.
18. B. Nowack, F. G. Kari, H. G. Krüger, Water, Air, Soil Poll. 125 (2001) 243.
19. R. S. Hedin, G. R. Watzlaf, R.W. Nairn, J. Environ. Qual. 23 (1994) 958.
20. C. Noubactep, Dissertation, TU Bergakademie Freiberg, Wiss. Mitt. Institut für Geologie
der TU Bergakademie Freiberg, Band 21, 140 pp, ISSN1433-1284.
21. E. E. Roden, J. M. Zachara, Environ. Sci. and Technol. 30 (1996) 1618.
22. L. J. Matheson, W. C. Goldberg, W. D. Bostick, L. Harris, in Handbook of groundwater
remediation using permeable reactive barriers: Applications to radionuclides, trace metals,
and nutrients. D. Naftz, S. J. Morrison, C. C. Fuller, J. A. Davis (eds.), Academic Press,
San Diego, Calif. (2002) 343-367.
15
23. USEPA. “Permeable Reactive Barrier Technologies for Contaminant Remediation.” EPA
600-R-98-125. September 1998.
24. G. Heron, H.T. Christensen, J. Tjeli, Environ. Sci. Technol. 28 (1994) 153.
25. G. R. Ford, Environ. Sci. Technol. 2002 , 36, 2459–2463.
26. G. Meinrath, P. Spitzer, Mikrochim. Acta 135 (2000) 155.
27. R. P. Buck, S. Rondinini, A. K. Covington, F. G. K. Baucke, C. M. A. Brett, M. F.
Camoes, M. J. T. Milton, T. Mussini, R. Naumann, K. W. Pratt, P. Spitzer, G. S. Wilson,
Pure Appl. Chem. 74 (2002) 2169.
28. C. Noubactep, J. Sonnefeld, M. Sauter, Grundwasser 10 (2005) accepted
29. P. Reiche, A survey of weathering processes and products. University of New Mexico
Press; Albuquerque, (1950) 95 pp.
30. J. Farrell, J. Wang, P. O’Day , M. Conklin, Environ. Sci. Technol. 35 (2001) 2026.
16
Table 1: Some relevant reactions for the elucidation of the mechanism of co-precipitated
U(VI) release from iron oxides. GR = green rust.
Process Reaction equation Eq.
Iron corrosion Fe° ⇔ Fe2+ + 2 e- (1)
Fe(OH)3 formation 2 Fe2+ + ½ O2 + 5 H2O ⇔ 2 Fe(OH)3 + 4 H+ (2)
GR formation (1-x)Fe2+ + xFe3+ + (2+x)OH- ⇔ [Fe2+1-xFe3+x(OH)2]x[x OH-] (3)
Fe(OH)3 aging Fe(OH)3 ⇔ FeOOH, (Fe3O4, Fe2O3) (4)
Acidic dissolution Fe(OH)3 + 3 H+ ⇔ Fe3+ + 3 H2O (5)
FeOOH + 3 H+ ⇔ Fe3+ + 2 H2O (6)
Compl. dissolution FeOOH + EDTA + 3 H+ ⇔ Fe(EDTA)3+ + 2 H2O (7)
Red. dissolution FeOOH + Ti3+ + H+ ⇔ Fe2+ + 2 OH- + Ti4+ (8)
17
Noubactep et al. Figure 1
0 12 24 36 48 60
0
15
30
45
60
75
90
105
[U(VI)]0 = 20 mg/L
[ZVI] = 15 g/L
total fixation
reversible fixation
fix
at
io
n
/ [
%
]
elapsed time / [days]
18
Noubactep et al. Figure 2
FeS2 (d1) FeS2 (d2) EDTA Na2CO3 TiCl3
0
10
20
30
40
[U(VI)]0 = 20 mg/L
60 days of fixation contact
14 hours of recovering contact
FeS2 (d3): 0.250 < d (mm) < 0.315)
FeS2 (d4): 0.315 < d (mm) < 0.63)
[EDTA] = 0.01 M
[Na2CO3] = 0.1 M )
[TiCl3] = 1.25 %
0.3 % 0.3 % 0.6 % 1.3 %
36 %
P r
ev
/
[%
]
remobilizing agent
19
Noubactep et al. Figure 3
Ref. d1 d5 MCL 2 d3 d3 d4 d2
0
14
28
42
56
70
4.68
5.31
4.91
4.196.55
4.55
8.41
[U(VI)]0 = 20 mg/L
ZVI : 15 g/L
30 days of fixation contact
30 days of recovering contact
recovering material: pyrite (di)
di : d1 < d2 < d3 < d4 < d5
U
ra
ni
um
/
[m
g/
L]
Particle size of pyrite
20
Noubactep et al. Figure 4
0 12 24 36 48
20
40
60
80
100
120
7.34
7.10
6.81
6.99
6.97
6.96
6.52
8.54
[U(VI)]
0
= 20 mg/L
ZVI: 15 g/L
FeS2 (d4): 25 g/L
30 days of fixation contact
FeS2 (d5): 0.63 - 1.0 mm
U
ra
ni
um
/
[m
g/
L]
elapsed time / [days]
21
Noubactep et al. Figure 5
0 12 24 36 48 60
0
6
12
18
24
30
[U(VI)]0 = 20 mg/L
ZVI : 15 g/L
FeS2 (d2): 25 g/L
90 days of fixation contact
recovering with TiCl3: 1.25 %0.69
0.84
1.09
1.56
1.49
1.56
1.79
P r
ev
/
[%
]
elapsed time / [hours]
22
Figure captions
Figure 1: Evolution of total and reversible uranium (VI) fixation from aqueous solution by
scrap iron (ZVI) as a function of time. The recovery experiments were conducted
in 0.1 M Na2CO3. Error bars give standard deviations (triplicate experiments). The
lines are not fitting functions, they simply join the data points to facilitate
visualization.
Figure 2: Percent recovery Prev of uranium from ZVI and corrosion products by different
remobilizing agents for 14 hours. Prev = 0.3 % corresponds to a concentration of 60
µg/L (> 30 µg/L, MCL of the US EPA). Error bars provide standard deviations
(triplicate experiments).
Figure 3: Uranium remobilization (ppb or µg/L) from ZVI and corrosion products by
different particle sizes of 25 g/L pyrite (di). One experiment (2 d3) was conducted
with a double amount of pyrite d3 (50 g/L). The reference consisted in an
accompanying experiment without pyrite addition (Ptot = 99.97 %). MCL = 30
µg/L is the maximum contaminant level of the US EPA. The values on the bars
indicated the final pH. Error bars provide standard deviations (triplicate
experiments).
Figure 4: Uranium remobilization (ppb or µg/L) from ZVI and corrosion products by 25 g/L
pyrite (d5) as function of time. The experimental point at t = 0 (pH = 8.54)
represents the solution at the end of the fixation experiment (no pyrite addition; Ptot
= 99.4 %). The values on the curve indicated the pH. Error bars provide standard
deviations (triplicate experiments). The lines are not fitting functions, they simply
join the data points to facilitate visualization.
Figure 5: Uranium remobilization from ZVI and corrosion products by 1.25 % TiCl3 as
function of time. The fixation experiment was conducted for three months with 15
23
g/L ZVI and in the presence of 25 g/L pyrite (d2) in order to favor U(VI) co-
precipitation. The values on the curve indicated the pH. Error bars provide
standard deviations (triplicate experiments). The lines are not fitting functions,
they simply join the data points to facilitate visualization.