Dissolution of ZrO2 based pyrochlores in the acid pH range: A macroscopic and electron microscopy study

Applied Geochemistry 49 (2014) 31–41

Contents lists available at ScienceDirect

Applied Geochemistry journal homepage: www.elsevier.com/locate/apgeochem

Dissolution of ZrO2 based pyrochlores in the acid pH range: A macroscopic and electron microscopy study S. Finkeldei ⇑, F. Brandt, K. Rozov, A.A. Bukaemskiy, S. Neumeier, D. Bosbach Forschungszentrum Jülich, Institute of Energy and Climate Research, Nuclear Waste Management and Reactor Safety (IEK – 6), 52425 Juelich, Germany

a r t i c l e

i n f o

Article history: Available online 1 July 2014

a b s t r a c t The dissolution kinetics of ZrO2–Nd2O3 polycrystalline pyrochlore and defect fluorite ceramic powders under acidic conditions were observed following a combined macroscopic and electron microscopic (SEM) approach. Dynamic dissolution experiments were carried out with a variation of temperature and pH as well as the chemical composition within the ZrO2–Nd2O3 system. SEM observations indicate a preferential leaching at the grain boundaries for all experiments. A preferential release of Nd during the initial stages of dissolution, which is several orders of magnitude higher than the Zr release, was measured by ICP-MS. At steady state, the normalised Nd-rate approaches the Zr based dissolution rate within one order of magnitude, becoming congruent for most experiments. Zr-based BET surface area normalised steady state rates at c(H+) = 0.1 N and 90 °C are in the range between 4  107 and 3  106 g m2 d1, indicating no significant influence of the transition pyrochlore to defect fluorite and the chemical composition on the macroscopic dissolution rate. Based on the different numbers of grain boundaries per surface area in the pyrochlore compared to the defect fluorite, a slightly higher ‘‘true’’ dissolution rate could be assumed for the defect fluorite. The influence of pH and temperature variations to the dissolution rates of defect fluorite and pyrochlore are similar and in the range observed for other multioxide materials. Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction Plutonium and the minor actinides (MA = Am, Cm, Np) are the main contributors to the long-term radiotoxicity of spent fuel. Due to their inherent stability, oxidic crystalline materials e.g., pyrochlores seem to be very promising candidates among crystalline materials for the conditioning of the MA. Zirconia based pyrochlores combine a high radiation resistance of the zirconates with a high chemical durability of the pyrochlore group minerals in aqueous environments (Lumpkin, 2006). Pyrochlores belong to the group of fluorite-derived structures. Within the fluorite structure AX2 with A = Ca, Ba, Sr, Cd, U and X = F, Cl, O (Riedel, 2004), the A-cation forms a face centred cubic structure and the anions are located within the tetrahedral vacancies. The defect fluorite structure (A, B)4O7 is closely related to the fluorite structure, but has two different cationic positions. Compared to the fluorite structure one eighth of the oxygen is removed, causing vacancies within the anionic sub-lattice. These vacancies are randomly distributed in the defect fluorite structure. Finally, ordering of these vacancies within the anionic sub-lattice leads ⇑ Corresponding author. Tel.: +49 2461 61 9281; fax: +49 2461 61 2450. E-mail address: s.fi[email protected] (S. Finkeldei). http://dx.doi.org/10.1016/j.apgeochem.2014.06.014 0883-2927/Ó 2014 Elsevier Ltd. All rights reserved.

to the pyrochlore structure, i.e., the pyrochlore structure is a superstructure of the defect fluorite structure. Besides a large number of ‘‘stoichiometric’’ pyrochlores A2B2X6Y (Y = O2, OH, F, (Ewing et al., 2004)) e.g., titanates (A2Ti2O7) or zirconates (A2Zr2O7), several pyrochlore solid solutions exist. These ‘‘non-stoichiometric’’ pyrochlores can be formed by a coupled cationic exchange on the A and B positions in combination with the anions to counterbalance the cationic charge. As a result the number of oxygen anions and vacancies of these solid solutions varies, leading to the A2xB1xO2+x formula. Several kinetic studies on the dissolution of pyrochlore have been reported, especially on titanate based compositions (Icenhower, 2003; Pöml et al., 2007; Roberts et al., 2000; Strachan et al., 2006). However, it is known that these pyrochlores are not stable against radiation damage i.e., they become amorphous. In contrast, the ZrO2 based pyrochlores undergo a phase transition to the less ordered defect fluorite structure due to radiation damage. In the literature mainly static dissolution studies exist for ZrO2 based pyrochlores (Hayakawa and Kamizono, 1993a,b; Kamizono et al., 1991). Previously, Finkeldei et al. (2013) compared the dissolution rates of one Nd2O3–ZrO2 based pyrochlore and one defect fluorite sample at pH = 1 and 90 °C, with Nd as a surrogate for the MA,

32

S. Finkeldei et al. / Applied Geochemistry 49 (2014) 31–41

following a dynamic experimental approach. In this study, the dissolution kinetics of the defect fluorite and pyrochlore solid solution series has been extended. Moreover, a variation of pH and temperature was carried out to compare the response of these two crystal structures towards these parameters in more detail. A special focus was laid on microscopic observations to determine where the dissolution actually takes place. 2. Materials and methods 2.1. Experiments Seven powder samples – four pyrochlore (22.0, 25.0, 33.3 and 35.1 mol% Nd2O3) and three defect fluorite (12.9, 15.6 and 17.6 mol% Nd2O3) powders were synthesised, characterised and introduced into dissolution experiments. 2.1.1. Synthesis and characterisation of pyrochlore and defect fluorite ceramics The ceramic powders for the dissolution experiments were synthesised via a wet chemical synthesis route to yield highly homogeneous materials. The synthesis of the powder was carried out according to Chen (Chen et al., 2009) and similar to the samples presented in Finkeldei et al. (2013). 0.1 mol L1 aqueous solutions of ZrOCl28H2O and Nd(NO3)36H2O (99.9% purity, Alfa Aesar) were prepared. The concentration was checked by inductively coupled plasma optical emission spectrometry (ICP-OES) because the chemicals are of hygroscopic character. The desired amounts of the metal salt solutions were mixed in a beaker and stirred for 5 min and transferred into a dropping funnel. The hydroxides were precipitated simultaneously by dropping the solution into liquid concentrated ammonia (25% NH4OH, Merck, p.a.). The precipitate was separated by centrifugation for 15 min at 4000 rpm. The supernatant was checked for complete precipitation by addition of a few drops of NH4OH (Finkeldei et al., 2013). Additionally, an alternative precipitation route using gaseous ammonia was applied to obtain a stoichiometric pyrochlore (PY-2). The precipitate of both routes was then washed with deionised water until the washing water was free of chloride and nitrate (QUANTOFIX dipsticks, Macherey–Nagel). The precipitate was dried overnight at 120 °C and ground thoroughly using a mortar and pistel. A calcination step was applied in air for 2 h at 600 °C for the fine ground powder in an alsintÒ sintering crucible. The heating and cooling rate was 4.8 °C/min. A second grinding step was applied before batches of 2.5 g of the powder were pressed in a pellet with 10 mm in diameter by a cold press with 50 kN. Afterwards, each pellet was transferred into a hot uniaxial press (HP W 5, FCT Systeme GmbH) and hot pressed at 1450 °C with 3.9 kN in argon atmosphere. The surface of the pellet was polished to remove contaminations from the die. To adjust the oxygen stoichiometry a resintering step was applied at 1600 °C for 5 h. The heating and cooling rate was 5.3 °C/min. The obtained pellets were crushed and separated into two fractions 180–100 lm and <100 lm by wet sieving. The 180–100 lm fraction was washed to remove residual fines and dried at 120 °C over night. The fine fraction was used for X-ray measurements. 2.1.2. Characterisation Each sample was characterised by powder X-ray diffraction (XRD) in order to determine if the sample consisted of pyrochlore or defect fluorite. XRD measurements were carried out in a range from 10 to 90° 2h using a D4 (h–2h geometry) or D8 (h–h geometry) instrument from Bruker AXS GmbH. The morphology and chemical composition of individual grains of the 100–180 lm fraction was characterised by scanning electron

microscopy (SEM) before and after the dissolution experiments using a FEI Quanta 200F equipped with energy dispersive X-ray spectroscopy (EDX). The specific surface area was measured previous to the dissolution experiments using a single-point N2 BET-method (Brunauer et al., 1938). 2.1.3. Dissolution experiments All dissolution experiments presented here were dynamic single pass flow-through experiments. Dynamic experiments have been used in many kinetic studies (Dacheux et al., 2010; Oelkers and Poitrasson, 2002; Rimstidt and Dove, 1986) because the reaction rates can be measured at steady state conditions without any fitting procedure to a presumed rate law. Two different experimental setups were used, depending on the temperature of the experiment: (1) experiments at temperatures below 100 °C were set up in a perfluoroalkoxy polymer (PFA) reactor and (2) experiments at temperatures higher than 100 °C were set up in a pressurised titanium reactor. The setup for the lower temperatures was adapted from Neeway et al. (2011) and Dacheux et al. (2010), and has been recently used in similar studies by Brandt et al. (2013) and Finkeldei et al. (2013). This experiment consisted of a peristaltic pump, which passed the inlet solution through a PFA reactor (Savillex) with a total volume of 50 mL. About 900 mg of the prepared powder were placed in the PFA flow-through reactor and 45 mL of the acid were added. A peristaltic pump continuously exchanged the solution within the reactor which entered the reactor at the bottom near the ceramic powder and left at the top of the reactor. Because of the high density of pyrochlore (6.363 g cm3, (Vandijk et al., 1984)) and defect fluorite and the coarse grain size used within the experiments no filters were used in the setup to keep the ceramic powder inside the flow-through reactor. A flow rate of 0.15 mL min1 was set and monitored regularly before sampling. Icenhower et al. (2000) described the dependence of the dissolution rate on the flow rate for very low flow rates. Here the aim was to determine the forward dissolution rate. Therefore a flow rate of 0.15 mL min1 was chosen. This flow rate is orders of magnitude higher than those of Icenhower and it can be assumed that the forward dissolution rate was measured and the rate was independent of the chosen flow rate. Due to the expected low dissolution rates (Icenhower et al., 2000), stirring was not applied during the flow-through experiments. This type of experiments was typically run for 80 d using solutions comprised of deionised water and HCl (Merck). Blank tests were run before the dissolution experiments to clean reactors and tubes and derive blank concentrations of Zr and Nd. During the experiments, samples of the outflow solution were taken regularly for chemical analyses of Nd and Zr, filtered through ultrafilters (10,000 Da, Advantec) and prepared for ICP-MS analyses. Prior to sampling, each filter was flushed with 6 mL of the effluent which were discarded before the final filtration step took place. ICP-MS analyses of the Nd and Zr concentrations were carried out using an ICP-MS ELAN 6100 DRC (PerkinElmer SCIEX) instrument. For the measurements, the samples were acidified by adding 1 vol% HNO3 suprapurÒ (Merck Millipore). A separate hydrothermal mixed flow experiment was constructed for the experiments carried out with 0.1 N HNO3 at temperatures of 110 °C. The reactor system is similar to that described by Berger et al. (1994), Dove and Crerar (1990) and Oelkers and Poitrasson (2002). The experimental setup consisted of a reactor vessel (Parr; V = 50 mL), a high pressure liquid chromatography pump and a computer controlled pressure valve. Inside the reactor vessel was equipped with a pressure gauge, an externally driven stirring system and a thermocouple. The pressure gauge was connected to a controller and computer system which

33

S. Finkeldei et al. / Applied Geochemistry 49 (2014) 31–41

regulated the outflow of the reactor via the pressure valve. The outflow of the reactor was filtered through a Ti cylinder frit with a pore size of 2 lm and an in-line Ti-filter with 0.5 lm pore size to avoid the loss of fines. All parts that were in contact with solution were made of polyether ether ketone (PEEK), polytetrafluoroethylene (PTFE) or Ti to avoid corrosion. To start the dissolution experiment, 1 g of the respective powder sample and 35 mL of 0.1 N HNO3 (Merck) were placed inside the reaction vessel. The reactor was pressurised with He and heated to 110 °C. The reactor was continuously stirred at a low stirring rate of 100 rpm to keep the solution homogeneous. During the experiment the pressure was kept constant by pumping fresh solution into the reactor at pump rates of 0.5–0.6 mL min1 and continuously releasing solution through the pressure valve. This type of experiments was run for 30 d. Samples were taken, filtered and analysed in the same way as for the experiments at 70 and 90 °C. All experimental parameters are summarised in Table 1.

2.1.4. Determination of normalised dissolution rates The theoretical background of the determination of normalised dissolution rates from dynamic experiments is described in detail in the literature e.g., Dove and Crerar (1990). The normalised dissolution rate based on the release of an element i from a ceramic can be described by the following equation (Brandt et al., 2003; Chou and Wollast, 1984; Dove and Crerar, 1990; Nagy and Lasaga, 1992; Oelkers and Poitrasson, 2002):

Rate ¼

F  Dc i A  m  vi

ð1Þ

where F is the fluid flow rate (L d1), A is the specific surface area of the ceramic powder (m2 g1), M is the mass of ceramic powder in the reactor (g), mi is the stoichiometric number of moles of the element i (Nd or Zr) in one mole of pyrochlore or defect fluorite (–), Dci is the concentration difference between the outflow and inflow concentrations of element i in solution (g L1). As a result, this rate represents the total dissolution rate of the complete pyrochlore based on the release rates of the cations Nd and Zr. In the ideal case, which is called congruent dissolution, the rates calculated from both cations are identical. Non-ideal dissolution is observed when the rates calculated from Nd and Zr deviate beyond the total uncertainties of A, Dci, m and F which is about 40% for all experiments. This error was based on error propagation. Steady state rates were calculated from average concentrations of each experiment after a steady state was reached. Typically a steady state was reached after 20–30 days, depending on the temperature and pH conditions.

2.2. Thermodynamic calculations In order to evaluate the deviation from thermodynamic equilibrium at the conditions of the dissolution experiments, the thermodynamic modelling with GEM-Selektor package (Kulik; Kulik et al., 2013; Wagner et al., 2012) was carried out. Particularly, the aims of this modelling were: (1) to estimate the solubility of Nd2Zr2O7 at elevated temperatures in aqueous acid environments and (2) to explore the possibility of precipitation of other secondary phases, like Nd2O3, Nd(OH)3, Zr(OH)4, ZrO2, during the dissolution experiments carried out in the present study. The activity coefficients of aqueous species were calculated with the aid of the Davies model (Davies, 1962). Thermodynamic data (standard molar Gibbs free energies, enthalpies, entropies and heat capacity values) for Nd2O3, Nd(OH)3 and Nd2Zr2O7 solids were taken from Jacob et al. (1998), Lutique et al. (2003a,b), Diakonov et al. (1998) and Lide (2007). Thermodynamic data for other solids and aqueous complexes were taken from implemented NAGRA-PSI (Hummel et al., 2002) and SUPCRT92/Slop98 (Shock et al., 1997) databases. 3. Results and discussion Dissolution rates of minerals are often described by empirical equations derived from the transition-state theory similar to the one shown below (Lasaga, 1995): n

n

Rate ¼ k0  Areactiv e  eEa =RT  aHHþ  Pai i  f ðDGr Þ þ

ð2Þ

where k0 is the intrinsic rate constant, Areactive is the reactive surface n þ area, Ea is the apparent activation energy, aHHþ is the hydronium ion n activity in solution, Pai i are the ions in solution, f(DGr) is the deviation from thermodynamic equilibrium. One feature of this approach was the possible separation of variables. This would imply the possibility of deriving the rate dependence for each variable independently in order to deduce dissolution rates by extra- or interpolation to the desired conditions. From a fundamental point of view, this type of model is only valid if just one mechanism occurs at one defined surface site and does not change within the observed parameter interval. In complex systems such as the ceramics of this dissolution study the empirical approach can only be used to study the total dissolution rate regardless of the heterogeneity of surface sites which may depend on the different crystallographic planes and the occurrence of kinks, steps and surface roughness. A recent publication by Luettge et al. (2013) shows that for an individual mineral, depending on these effects, different microscopically observed dissolution steps may lead to a significant variation of the macroscopic dissolution rate. A deconvolution of the total dissolution rate to

Table 1 Overview of the experimental conditions for all experiments. Nd2O3 (mol%)

Crystal structure (–)

Temperature (°C)

c(HCl) (mol L1)

Experimental setup (–)

33.3 33.3 33.3 33.3 33.3 15.6 15.6 15.6 15.6 15.6 17.6 22.0 25.0 35.1

Pyrochlore Pyrochlore Pyrochlore Pyrochlore Pyrochlore Defect fluorite Defect fluorite Defect fluorite Defect fluorite Defect fluorite Defect fluorite Pyrochlore Pyrochlore Pyrochlore

70 90 90 90 110 70 90 90 90 110 90 90 90 90

0.10 0.01 0.10 1.00 0.10 0.10 0.01 0.10 1.00 0.10 0.10 0.10 0.10 0.10

PFA reactor PFA reactor PFA reactor PFA reactor Ti reactor, HNO3 PFA reactor PFA reactor PFA reactor PFA reactor Ti reactor, HNO3 PFA reactor PFA reactor PFA reactor PFA reactor

34

S. Finkeldei et al. / Applied Geochemistry 49 (2014) 31–41

the individual microscopic mechanisms and rates is not possible based on macroscopic observations. However, in this study the aim was to compare the dissolution behaviour of Nd2O3–ZrO2 based ceramics under various conditions and the influence of the pyrochlore and defect fluorite crystal structure. 3.1. Deviation from equilibrium For the thermodynamic modelling two input recipes presented in Tables 2 and 3, reflecting the conditions of experiments at 90 °C and pH  1 were used. The bulk composition shown in Table 2 corresponds to the assumption of the complete dissolution of Nd2Zr2O7. Here, the aim was to calculate if there are any boundary conditions at which the dissolution of pyrochlore would stop due to saturation effects. On the other hand, in order to check in details the possibility of the precipitation of other secondary phases and to evaluate the degree of over- or undersaturation in liquids at typical steady state conditions, the alternative recipe (Table 3) was designed. In this case the system definition included only concentrations of Nd and Zr which were determined in the outlet solution of the dynamic dissolution experiments. Results of calculations show that the theoretical values of total dissolved molalities of Nd and Zr after complete dissolution of 0.1 g of Nd2Zr2O7 in acidic solution at 90 °C (Table 2) correspond to 3.43  104 and 7.04  1011 mol kg1, respectively (Table 4). This observation may indicate that Nd2Zr2O7 solid has to be dissolved incongruently, i.e., a complete dissolution would take place with the simultaneous precipitation of a secondary phase. The thermodynamically stable phase at conditions of such experiments is ZrO2 (baddeleyite) and the value 7.04  1011 mol kg1 corresponds to the aqueous solubility of ZrO2 at pH = 1.128 and T = 90 °C. At the

same time, the value 3.43  104 mol kg1 corresponds to the total amount of Nd in 0.1 g of Nd2Zr2O7. From this thermodynamic modelling follows that pyrochlore which is metastable at room temperature would eventually completely transform to baddeleyite if none of the reactants is removed from the system. In order to predict compositions of the aqueous solution at the conditions of the dynamic dissolution experiments, the outflow concentrations of Nd and Zr were used (Nd = 5.80  108 and Zr = 3.80  108 mol kg1; Table 3). The results of these calculations show that at conditions of thermodynamic equilibrium the total dissolved concentrations of these metals have to be equal to 5.18  108 and 7.17  1011 mol kg1, respectively (Table 5) and the pH value in aqueous media will be equal to 1.12. These calculations indicate that the ZrO2 is supersaturated and should eventually precipitate. In order to numerically express the deviation of the solid–liquid system from the thermodynamic equilibrium (i.e., oversaturation or undersaturation state) it is necessary to introduce the so-called ‘‘saturation index’’ (SI) which relates the calculated ion activity product in solution (IAP) to the solubility product (KSP; (Parkhurst and Appelo, 1999)):

SI ¼ logðIAP=K SP Þ

ð3Þ

The SI has been estimated for the ZrO2 phase because the thermodynamic calculations presented above demonstrated that this phase has to precipitate. For this purpose the reaction of ZrO2(cr) dissolution may be formulated as:

ZrO2 ðcrÞ þ 2H2 OðlÞ ¼ Zr4þ þ 4OH

ð4Þ

The solubility product of zirconium oxide (KSP,ZrO2) is written by using aqueous activities of Zr4+ and OH species at equilibrium:

K SP;ZrO2 ¼ fZr4þ g  fOH g4 Table 2 System definition for GEM-Selektor calculation of the dissolution of Nd2Zr2O7. Parameter

Value

H2O Nd2Zr2O7 HCl Temperature Pressure

1000 g 0.1 g 0.1 mol 90 °C 1 bar

Table 3 System definition used for GEM-Selektor interpretation of results of dynamic dissolution experiments with Nd2Zr2O7 pyrochlore. Parameter

Value

H2O Nd Zr HCl Temperature Pressure

1000 g 5.80  108 mol 3.80  108 mol 0.1 mol 90 °C 1 bar

Table 4 Thermodynamic modelling results of Nd2Zr2O7 dissolution in a batch from which no reactantsa are removed.

a

mNdaq (mol kg1)

mZraq (mol kg1)

pH

Solids (mol ZrO2(cr))

3.43  104

7.04  1011

1.128

3.43  104

Reactants are: Nd2O3(cr), Nd(OH)3(cr), Zr(OH)4(am), ZrO2(cr).

ð5Þ

The ionic activity products of the current solution can be calculated by neglecting the precipitation of ZrO2(cr) and by calculating of aqueous activities of Zr4+ and OH aqueous species and activities from the actually measured aqueous composition. In Table 6 the comparison of activity coefficients obtained from estimations of KSP,ZrO2 and IAPZrO2 and SI of ZrO2(cr) in the dynamic dissolution experiment are presented. As seen from Table 6 there is a significant supersaturation for ZrO2 (SI = 2.65). The solubility of monoclinic and cubic modifications of ZrO2(cr) in aqueous solutions at room temperature has been investigated experimentally in long-term experiments (250 days) at pH = 9 (Pouchon, 2000; Pouchon et al., 2001). Dissolutions tests with very short ageing time (24 h) have been performed at acidic conditions by Kovalenko and Bagdasarov (1961). Sheka and Pevzner (1960) have carried out dissolution experiments (45 days of ageing) at strongly alkaline conditions (pH = 13–14). Bilinski et al. (1966) used the opposite approach and performed experiments referred to precipitates separated from strongly oversaturated solutions with pH = 1.5–6.5. Curti and Degueldre (2002) demonstrated that the dataset of dissociation constants selected by Baes and Mesmer (1986) allows constructing the solubility curve for the system ZrO2(cr)–H2O as a function of pH. Moreover, this solubility curve reproduces reasonably well all experimental results obtained from dissolution (i.e., Table 5 Thermodynamic modelling results based on averaged typical Nd and Zr aqueous concentrations in outlet solutions of experiments at pH  1 and 90 °C. mNdaq (mol kg1)

mZraq (mol kg1)

pH

Solids (mol ZrO2(cr))

5.80  108

7.17  1011

1.123

2.19  108

35

S. Finkeldei et al. / Applied Geochemistry 49 (2014) 31–41

Table 6 Estimated values of activities for Zr4+, OH aqueous species, solubility product (KSP,ZrO2), ionic activity product (IAPZrO2) of ZrO2(cr), saturation index (SI) of ZrO2(cr) at conditions of the dynamic dissolution experiment (T = 90 °C, P = 1 bar). ZrO2(cr) is not precipitating

ZrO2(cr) in equilibrium with aqueous solution

SI

(Zr4+)

(OH)

IAPZrO2

(Zr4+)

(OH)

KSP,ZrO2

4.61  1012

4.97  1012

2.81  1057

1.04  1014

4.97  1012

6.35  1060

undersaturation) experiments. Nevertheless, such curve deviates significantly from total dissolved concentrations of Zr measured in experiments of Bilinski et al. (1966). Curti and Degueldre (2002) explained this phenomenon by the process when in oversaturated solutions the nucleation of ZrO2 precipitates takes place through condensation of dissolved polymers. This transformation of polymers (depolymerisation) into a crystalline anhydrous oxide is a very slow process and the equilibrium between ZrO2 and aqueous media is hardly achievable. Therefore, in the case of dissolution studies with Nd2Zr2O7 the similar effect of slow depolymerisation cannot be neglected, i.e., ZrO2 may not precipitate within the time scale of the experiments due to kinetic reasons. Experimental observations indicate that the aqueous solutions taken during the experiments are stable for several months. No relevant concentration differences could be detected for solutions measured immediately after sampling and several weeks later. 3.2. Temporal evolution and stoichiometry of metal release The temporal evolution of the Nd- and Zr-concentrations of all presented dissolution experiments followed the typical trend of dynamic dissolution experiments with silicates and oxides as described in the literature (Brandt et al., 2003; Finkeldei et al., 2013; Horlait et al., 2012; Oelkers and Poitrasson, 2002). Starting from high initial rates, the dissolution rates decreased until a steady state was reached, i.e., the rate became constant with time (Fig. 1). In general, the higher initial dissolution rates can be attributed to (1) high-energy surface sites and (2) the enhanced dissolution of fine particles (Casey and Bunker, 1990; Chou and Wollast, 1984; Horlait et al., 2012; Knauss and Thomas, 1989; Oelkers and Poitrasson, 2002; Wehrli, 1989). Another explanation for the enhanced initial rates could be a change of the dominating dissolution mechanism e.g., from the dissolution at grain boundaries to the dissolution on steps, kinks and etch pits on the grain surfaces (Luettge et al., 2013).

2.65

The approach of steady state depends on the experimental conditions and was fastest at 110 °C and c(H+) = 0.1 M, where a steady state of the Zr release was reached after about 10–20 days. At 70 °C a steady state of the Zr release is typically reached after 20–30 days. In some experiments, the Nd-release is oscillating even after the plateau of the steady state is reached. For these experiments alternating mechanisms may cause two distinct Nd release rates. Therefore, not an average rate but the two boundary values are given for these experiments in Table 8 accordingly. Typical for all experiments is a strongly incongruent initial release of Zr and Nd (Fig. 1). During this initial stage Nd was released at a systematically higher rate which deviated several orders of magnitude from the Zr released rate. With time, the Nd-based dissolution rate became more and more congruent with the Zr-based rate (Fig. 1). The initial incongruent dissolution is typical for mixed oxides (Horlait et al., 2012; Szenknect et al., 2012; Tocino et al., 2013). For some minerals the incongruent dissolution can be attributed to distinctively different structural units e.g., the brucite-like octahedral layer of clay minerals versus the TOT (tetrahedral–octahedral–tetrahedral) layer (Brandt et al., 2003). In the case of pyrochlore and defect fluorite, possible differences in the Zr–O and Nd–O bonding strengths may be responsible for the higher initial release rate of Nd. Since the powders used for the experiments are polycrystalline ceramics, surface inhomogeneities due to sintering cannot be excluded. Such inhomogeneities have been observed for other multioxides leading to an enrichment of one cation in a surface layer of a few nanometers which may then lead to incongruent dissolution (Szenknect et al., 2012). In addition, Szenknect et al. (2012) present a model for the dissolution of Cerium (IV)–Neodymium(III) oxides which is also incongruent at the early stages of the dissolution and congruent at a later steady state. This model is related to the evolving microstructure during the early stages of dissolution and discussed below. 3.3. Effect of microstructure

-3

10

Nd Zr -1

Normalised concentration (mol L )

-4

10

-5

10

-6

10

-7

10

-8

10

-9

10

0

5

10

15

20

25

30

35

time (d) Fig. 1. Temporal evolution of normalised Nd and Zr release concentrations from a dissolution experiment with a pyrochlore (25 mol% Nd2O3, 90 °C and 0.1 N HCl). The concentrations are normalised to the pyrochlore composition.

For ceramics, which are considered as potential nuclear waste forms, the presence of pores is indispensable, as the a-decay of the embedded radionuclides will lead to a release of gaseous He. This gas needs to be stored inside the pores or released through them. Otherwise gas formation would lead to the evolution of cracks within the matrix (Wiss et al., 2006). Therefore, the ceramics produced for this study were all porous materials with a typical porosity of about 10%. It is known from the literature, that the microstructure plays a key role concerning the dissolution kinetics (Hingant et al., 2009; Luettge et al., 2013; Szenknect et al., 2012). Luettge et al. (2013) have discussed the difference between the specific surface area and the reactive surface sites due to the different activation energies of e.g., kink sites, grain boundaries and edges. Within this article the dissolution rate was normalised to the specific surface area from N2-BET measurements. Even though the BET surface area does not necessarily reflect the reactive surface area, this normalisation step is commonly applied. Deriving the true reactive surface area would necessitate in situ microscopic observation of the dissolution which is beyond the scope of this study. However, SEM studies previous and after the dissolution experiment were

36

S. Finkeldei et al. / Applied Geochemistry 49 (2014) 31–41

carried out for all samples to gain some insight where the dissolution actually takes place at the powdered samples. In the following we discuss two selected compositions in detail – one defect fluorite and one ‘‘stoichiometric’’ pyrochlore – with respect to the microstructural changes after dissolution experiments under acidic conditions. Additionally the effect of porosity was studied. Typical for the defect fluorite samples was a well-developed microstructure with straight, well-defined grain boundaries and low intragranular porosity, as can been seen for the selected 12.9 mol% Nd2O3 sample (Fig. 2a and b). The typical primary grain size for this defect fluorite sample was between 5 and 10 lm. The starting material exhibited a preferential formation of intergranular pores. The structure of the 100–180 lm particles remained the same after the dissolution experiment (Fig. 2c). The image at higher magnification in Fig. 2d indicates a preferential dissolution at the triple junctions and grain boundaries. The observed dissolution at the grain boundaries may be due to a higher defect density at the grain boundaries compared to the bulk material. The microstructure of all pyrochlore samples differed significantly from the defect fluorite powders. Typically the primary grains of all pyrochlores obtained by coprecipitation with liquid ammonia were smaller and the pyrochlores were less dense compared to the defect fluorite. A typical powder of this group (PY-1) is shown in Fig. 3a and b. This sample consisted of primary grains with a size of 1–2 lm which were sintered together. The 100–180 lm particles consist of dense agglomerates of these 1–2 lm grains and regions with a higher porosity. After dissolution at 90 °C in 0.1 M HCl for 82 days a significant change of the microstructure was visible (Fig. 3c). Besides the 100–180 lm fraction a second fraction with a grain size of 1–2 lm was observed. During the dissolution experiment small amounts of the ceramic disintegrated, which led to the presence of this fine fraction. The disintegration was most probably caused by the weakening of the grain boundaries. Besides the activation of grain boundaries due to a defect accumulation, other mechanisms are described in the recent literature:

(1) chemical inhomogeneities at the grain boundaries and (2) the formation of leached layers. Chemical inhomogeneities produced during sintering may result in a Nd enrichment at pores and grain boundaries (Bellière et al., 2006; Lei et al., 2002; Luo et al., 2006). A favoured dissolution at these sites can explain the preferential Nd release which was observed especially during the initial phase of the dissolution experiments. On the other hand, the formation of leached layers could be caused by a weaker binding energy between Nd–O compared to Zr–O. A model for this kind of dissolution mechanism has been proposed by Szenknect et al. (2012). Thermodynamic calculations and experimental observations indicate that ZrO2 is supersaturated during the experiments but does not precipitate as a separate phase. Therefore the incongruent release of Zr and Nd may also be due to a partial transformation of pyrochlore to ZrO2 or Zr(OH)4 at the surface. In all above cases the grain boundaries will be weakened at some point leading to the disintegration. This disintegration is accompanied by a supply of fresh surfaces which may dominate the dissolution kinetics. At the fresh surfaces a preferential release at the pores and grain boundaries as well as a leached layer formation or transformation will again lead to a preferential Nd release. Therefore the observed incongruent dissolution of the Nd2O3–ZrO2 ceramics can be the result of any of the mechanisms described above, acting continuously on the surface. This would explain the oscillating concentrations of the Nd release observed in some experiments when the Zr release has already reached a steady state. However, the disintegration is not expected to have a significant impact on the specific surface area, because the starting powder already had a high level of porosity and was therefore already dominated by the surface areas of the primary grains. Moreover, the dissolution rate of the individual grains should be lower than at the grain boundaries as indicated by the preferential leaching of the grain boundaries. Therefore the contribution to the total dissolution rate of this fine but not very reactive fraction should be insignificant.

Fig. 2. SEM images of the 100–180 lm fraction of a defect fluorite powder with 12.9 mol% Nd2O3. The pictures (a) and (b) are taken previous to the dissolution experiment at a lower (a) and higher magnification (b). The powder after the alteration experiment is depicted in (c) while the picture (d) is taken again at a higher magnification.

S. Finkeldei et al. / Applied Geochemistry 49 (2014) 31–41

37

Fig. 3. SEM images of the 100–180 lm fraction of the Nd2Zr2O7 powder (PY-1) previous to the alteration experiment at a lower (a) and higher magnification (b) and after the dissolution experiment at a lower (c) and higher magnification (d).

The effect of porosity was studied by comparing two stoichiometric pyrochlores, PY-1 and PY-2. PY-2 was obtained by an alternative precipitation route (Section 2.1.1). This approach allowed an adjustment of the microstructure. In Fig. 4a and b a second sample of the stoichiometric pyrochlore (PY-2) is shown before the dissolution experiment. The 100–180 lm particles also consist of grains

of a size of 1–2 lm and the particles are also porous. In contrast to PY-1, the texture of PY-2 consists of a homogeneous network of 1–2 lm grains. After the dissolution experiment, the SEM images of PY-2 do not show a disintegration of the sample (Fig. 4c and d). The normalised Nd and Zr dissolution rates of PY-2 and PY-1 are compared in Table 7. Here, the effect of the microstructure causes a

Fig. 4. SEM images of the 100–180 lm fraction of the Nd2Zr2O7 powder (PY-2) previous to the alteration experiment at a lower (a) and higher magnification (b) and after the dissolution experiment at a lower (c) and higher magnification (d). This sample has been prepared by a different synthesis route compared to the samples in Fig. 3.

38

S. Finkeldei et al. / Applied Geochemistry 49 (2014) 31–41

Table 7 Comparison of Nd- and Zr-based normalised dissolution rates of a stoichiometric pyrochlore with liquid ammonia (PY-1) and gaseous ammonia (PY-2) as precipitating agent. Pyrochlore sample

Rate (Nd) (g m2 d1)

Rate (Zr) (g m2 d1)

PY-1 PY-2

4  106 2  106

3  106 1  106

significantly higher dissolution rate of the less stable microstructure of PY-1 by a factor of 2–3. If the microstructure is less stable, there are more grain boundaries accessible for dissolution, leading to a higher proportion of reactive surface sites compared to the total surface area. 3.4. Influence of pyrochlore/defect fluorite chemical composition and crystal structure The influence of the chemical composition and crystal structure has been studied for seven different samples. Three samples with the defect fluorite crystal structure and four with the pyrochlore crystal structure. For all compositions a highly incongruent initial dissolution was observed. The initial release rates for the Zr-based rates of the defect fluorite samples are very similar and within the limit of error (Fig. 5a). For the Nd-rich pyrochlores an increase in the Zr-based rates is observed. In Fig. 5 it can be seen, that the Zr-based steady state rates are already close to the Zr-based initial rates from Fig. 5a. In all experiments, the Nd-based rate approaches the Zr-based dissolution rate at steady state. Fig. 5a shows a strongly preferential release of Nd no matter which chemical composition or crystal structure is studied. At steady state conditions a decrease of several orders of magnitude was observed for the Nd-based rates compared to the initial rates, leading to a congruent or close to congruent dissolution. From the presented experiments it appears there is no systematic effect on the dissolution rate due to the phase transformation from the defect fluorite to the pyrochlore induced by changes of the Nd/Zr ratio. However, as it was shown in the previous section, the primary grain size of the defect fluorite sample was larger than those of the pyrochlores. Therefore, using the same amount of a defect fluorite sample and a pyrochlore might also lead to a higher dissolution rate of the pyrochlore samples, because there is a higher concentration of grain boundaries (Fig. 3d). Zr-based pyrochlores are known to undergo a phase transformation due to radiation damage to the defect fluorite structure. From the presented results it can be learned that a radiation damage induced phase transition to the defect fluorite will most probably lead to dissolution rates which are within the same order of magnitude as the ones of the original pyrochlore. 3.5. Influence of c(H+) In order to compare the effect of pH on the dissolution rate of pyrochlore and defect fluorite, one sample of each was chosen. Fig. 6 shows the c(H+) dependence in the acidic regime at 90 °C which is similar for the defect fluorite and the pyrochlore ceramic. Here, the relationship between c(H+) and the dissolution rate is presented. The concentration of H+ is used as it was experimentally verified. The corresponding pH as calculated with GEMS-PSI for the given temperatures deviates from pH = 0, 1 and 2 by less than 0.1 logarithmic units. The pH dependence of the dissolution rate was found to be linear for the presented pH regime. An increase in the pH leads to a decrease of the dissolution rates for both samples. The slope for the pH dependent release as a function of log(Rate) is in the range

Fig. 5. Nd (black) and Zr (red) based dissolution rates of pyrochlore and defect fluorite powders for seven different chemical compositions. Dissolution rates for the initial phase are given in (a) whereas rates for the steady state are shown in (b). All data are collected under the same experimental conditions: 90 °C and c(HCl) = 0.1 N. Open symbols assign data which were taken from Finkeldei et al. (2013) which were obtained at the same conditions. Steady state rates are average values. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

of n = 0.43–0.67 for the Nd- and Zr-based dissolution rates of both samples. The pH-dependency of the pyrochlore and defect fluorite dissolution rates is similar to many natural minerals such as silicates where cation exchange and breaking of metal–oxygen bonding was identified as the rate controlling step in the pH-regions below the point of zero charge (Brantley, 2003; Oelkers, 2001). Depending on the pH at which the point of zero charge occurs, the n may decrease towards the neutral pH-range. For many minerals, a minimum rate was observed close to the point of zero charge (Brantley, 2003).

3.6. Effect of temperature In order to compare the effect of temperature on the dissolution rate of pyrochlore and defect fluorite, the same samples were chosen as in Section 3.5. Experiments were conducted with c(H+) = 0.1 mol L1 at 70 °C, 90 °C and 110 °C. Since the release of Zr appears to control the steady state dissolution rates, only the Zr-based dissolution rates are depicted in Fig. 7. The general trend of the dissolution rates between 110 °C and 90 °C is similar for

39

S. Finkeldei et al. / Applied Geochemistry 49 (2014) 31–41 -8

Pyrochlore

-9

Defect Fluorite -10

ln (Rate)

-11 -12 -13 -14 -15 -16 -17 0.0026

0.0027

0.0028

0.0029

0.0030

-1

1/T (K ) Fig. 7. Temperature dependence of Zr-based dissolution rates for a defect fluorite (15.6 mol% Nd2O3, open symbols) and pyrochlore (33.3 mol% Nd2O3, closed symbols).

range between 90 °C and 150 °C. From their data, a similar temperature dependency as for the pyrochlore in this study can be derived. 4. Conclusions

Fig. 6. c(H+) dependence of the normalised dissolution rate in the acidic regime for a defect fluorite (15.6 mol% Nd2O3); with a slope for the Nd-based rate of 0.56; and 0.51 for the Zr-based rate (a) and a stoichiometric pyrochlore with a slope for the Nd-based rate of 0.43; and 0.67 for the Zr-based rate (b).

pyrochlore and defect fluorite, with a higher dissolution rate obtained for the pyrochlore. At 70 °C the dissolution rate becomes similar for both ceramic types. Commonly, activation energies for the dissolution rates are calculated according to Arrhenius’ law:

Rate  eEA =RT

ð6Þ

This approach assumes a single dissolution mechanism which stays the same all through the temperature range of interest. Within the Arrhenius plot of Fig. 7, a linear dependence of the ln(Rate) with the temperature could be constructed for the pyrochlore, whereas this would no longer be possible for the defect fluorite sample. Instead, the temperature dependence for the defect fluorite dissolution rate could be described by a non-linear function. From the microscopic observations discussed in Section 3.3 a complex system of several potentially interacting dissolution regimes can be assumed which may not have the same activation energy. Therefore, the Arrhenius approach may not be appropriate for the prediction of the dissolution rate as a function of temperature in this type of system. Kamizono et al. (1991) reported dissolution rates of La–Zrbased pyrochlores obtained in deionised water in the temperature

We have presented a study of the dissolution kinetics of Nd2O3– ZrO2 pyrochlore and defect fluorite ceramics under acidic and far-from equilibrium conditions with respect to the ceramic. In all dissolution experiments, initial dissolution rates were found to be strongly incongruent suggesting either the formation of a Zr-rich secondary phase, leached layers or Nd-enriched inhomogeneities at the grain boundaries. The microscopic analysis of the ceramic powders after the dissolution experiments shows a preferential dissolution at the grain boundaries which may be therefore considered to dominate the surface reactivity under acidic conditions. Slightly lower dissolution rates were determined for the defect fluorite compared to the pyrochlore ceramics, which may be due to the higher concentration of grain boundaries in the pyrochlore samples. Thus, related to the reactive surface area, the stability of defect fluorite may be very similar to the stability of pyrochlore. Therefore, the expected transformation of ZrO2 based pyrochlore to defect fluorite due to radiation damage should not cause any significant change in the dissolution kinetics. The pH-dependency of the dissolution rate was observed to be similar for one selected pyrochlore and one defect fluorite sample at 90 °C with n = 0.5 and 0.7. This range has also been observed in the literature for oxides and silicates for which a cation-exchange mechanism was reported to be the rate controlling step. A clear linear dependence of the ln(Rate) with the temperature was not found within this study. Due to the heterogeneity of the surface reactivity observed microscopically it becomes questionable whether the Arrhenius approach should be used to extrapolate the temperature dependence to lower temperatures. However, from the new data in this study it can be assumed that the dissolution rates at near neutral and low temperature conditions relevant to high level nuclear waste disposal would be extremely low and could therefore not be quantified by the macroscopic method applied here. Advanced microscopic methods may be needed to examine the dissolution under such rather mild conditions and to gain a deeper understanding about the different mechanisms which control the total dissolution rate.

40

S. Finkeldei et al. / Applied Geochemistry 49 (2014) 31–41

Table 8 Dissolution rates of pyrochlore and defect fluorite samples obtained at 90 °C and c(H+) = 0.1 N. Nd2O3

Crystal structure

Initial rate (Nd) (g m2 d1)

Steady state Ratemax (Nd) (g m2 d1)

Steady state Ratemin (Nd) (g m2 d1)

Initial rate (Zr) (g m2 d1)

Steady state Ratemax (Zr) (g m2 d1)

Steady state Ratemin (Zr) (g m2 d1)

Defect fluorite Defect fluorite Defect fluorite Pyrochlore Pyrochlore Pyrochlore Pyrochlore Pyrochlore

1.0  103 7.9  104 3.9  104 2.4  103 2.2  104 4.4  106 1.2  103 2.0  103

4.2  106 6.5  106 2.4  106 1.2  105 1.7  106 2.0  106 4.4  106 1.2  105

1.1  106 5.0  107 1.3  106 2.4  106 8.3  107

2.0  106 1.9  106 3.3  106 2.1  106 2.2  106 9.2  106 1.6  105 2.0  105

5.0  107 1.8  106 6.9  107 1.0  106 1.0  106 1.3  106 2.9  106 2.8  106

4.7  107 6.7  107 5.4  107

(mol%) 12.9a 15.6 17.6 22.0 25.0 33.3 (PY-2) 33.3 (PY-1) 35.1 a

3.6  107

Taken from Finkeldei et al. (2013).

Acknowledgements We would like to thank Sigrid Schwartz-Lückge, Volker Bader, Fabian Sadowski, Martina Klinkenberg and Jakob Dellen for their analytical support. The authors would also like to thank Andreas Lüttge und Cornelius Fischer for inspiring discussions. This work was partly supported by the Federal Ministry of Education and Research (BMBF); Support Code: 02NUK021A.

Appendix A See Table 8.

References Baes, C.F., Mesmer, R.E., 1986. The Hydrolysis of Cations. Robert E. Krieger Publishing Company, Malabar, FL. Bellière, V., Joorst, G., Stephan, O., de Groot, F.M.F., Weckhuysen, B.M., 2006. Phase segregation in ceriumlanthanum solid solutions. J. Phys. Chem. B 110, 9984– 9990. Berger, G., Claparols, C., Guy, C., Daux, V., 1994. Dissolution rate of a basalt glass in silica-rich solutions – implications for long-term alteration. Geochim. Cosmochim. Acta 58, 4875–4886. Bilinski, H., Branica, M., Sillen, L.G., 1966. Precipitation and hydrolysis of metallic ions. II. Studies on the solubility of zirconium hydroxide in dilute solutions and in 1 M NaClO4. Acta Chem. Scand. 20, 853–861. Brandt, F., Bosbach, D., Krawczyk-Bärsch, E., Arnold, T., Bernhard, G., 2003. Chlorite dissolution in the acid pH-range: a combined microscopic and macroscopic approach. Geochim. Cosmochim. Acta 67, 1451–1461. Brandt, F., Neumeier, S., Schuppik, T., Arinicheva, Y., Bukaemskiy, A., Modolo, G., Bosbach, D., 2013. Conditioning of minor actinides in lanthanum monazite ceramics: a surrogate study with Europium. Prog. Nucl. Energy, http:// dx.doi.org/10.1016/j.pnucene.2013.07.019. Brantley, S.L., 2003. Reaction kinetics of primary rock-forming minerals under ambient conditions. In: Drever, J.I. (Ed.), Fresh Water Geochemistry, Weathering, and Soils, vol. 5 of Treatise on Geochemistry, Turekian, K.K., Holland, H.D. (Eds.), Pergamon Press, Oxford, pp. 73–117. Brunauer, S., Emmett, P.H., Teller, E., 1938. Adsorption of gases in multimolecular layers. J. Am. Chem. Soc. 60, 309–319. Casey, W.H., Bunker, B., 1990. Leaching of mineral and glass surfaces during dissolution. In: Hochella, M.F., Bancroft, G.M. (Eds.), Mineral Water Interface Geochemistry. Mineralogical Society of America, Washington, D.C, pp. 397–426. Chen, H.F., Gao, Y.F., Liu, Y., Luo, H.J., 2009. Coprecipitation synthesis and thermal conductivity of La2Zr2O7. J. Alloys. Compd. 480, 843–848. Chou, L., Wollast, R., 1984. Study of weathering of albite at room-temperature and pressure with a fluidized-bed reactor. Geochim. Cosmochim. Acta 48, 2205– 2217. Curti, E., Degueldre, C., 2002. Solubility and hydrolysis of Zr oxides: a review and supplemental data. Radiochim. Acta 90, 801–804. Dacheux, N., de Kerdaniel, E.D.F., Clavier, N., Podor, R., Aupiais, J., Szenknect, S., 2010. Kinetics of dissolution of thorium and uranium doped britholite ceramics. J. Nucl. Mater. 404, 33–43. Davies, C.W., 1962. Ion Association. Butterworths, Washington, D.C.. Diakonov, I.I., Ragnarsdottir, K.V., Tagirov, B.R., 1998. Standard thermodynamic properties and heat capacity equations of rare earth hydroxides: II. Ce(III)-, Pr-, Sm-, Eu(III)-, Gd-, Tb-, Dy-, Ho-, Er-, Tm-, Yb-, and Y-hydroxides. Comparison of thermochemical and solubility data. Chem. Geol. 151, 327–347. Dove, P.M., Crerar, D.A., 1990. Kinetics of quartz dissolution in electrolyte-solutions using a hydrothermal mixed flow reactor. Geochim. Cosmochim. Acta 54, 955– 969.

Ewing, R.C., Weber, W.J., Lian, J., 2004. Nuclear waste disposal – pyrochlore (A2B2O7): nuclear waste form for the immobilization of plutonium and ‘‘minor’’ actinides. J. Appl. Phys. 95, 5949–5971. Finkeldei, S., Brandt, F., Bukaemskiy, A., Neumeier, S., Modolo, G., Bosbach, D., 2013. Synthesis and dissolution kinetics of zirconia based ceramics. Prog. Nucl. Energy, http://dx.doi.org/10.1016/j.pnucene.2013.07.020. Hayakawa, I., Kamizono, H., 1993a. Durability of an La2Zr2O7 waste form containing various amounts of simulated HLW elements. J. Nucl. Mater. 202, 163–168. Hayakawa, I., Kamizono, H., 1993b. Durability of an La2Zr2O7 waste form in water. J. Mater. Sci. 28, 513–517. Hingant, N., Clavier, N., Dacheux, N., Barre, N., Hubert, S., Obbade, S., Taborda, F., Abraham, F., 2009. Preparation, sintering and leaching of optimized uranium thorium dioxides. J. Nucl. Mater. 385, 400–406. Horlait, D., Tocino, F., Clavier, N., Dacheux, N., Szenknect, S., 2012. Multiparametric study of Th1xLnxO2x/2 mixed oxides dissolution in nitric acid media. J. Nucl. Mater. 429, 237–244. Hummel, W., Berner, U., Curti, E., Pearson, F.J., Thoenen, T., 2002. Nagra/PSI chemical thermodynamic data base 01/01. Radiochim. Acta 90, 805–813. Icenhower, J.P., McGrail, B.P., Schaef, H.T., Rodriguez, E.A., 2000. Dissolution kinetics of titanium pyrochlore ceramics at 90 °C by single pass flow-through experiments. Mater. Res. Soc., 373–378. Icenhower, J.P., Strachan, D.M., Lindberg, M.M., Rodriguez, E.A., Steele, J.L., 2003. Dissolution kinetics of titanate-based ceramic waste forms: results from singlepass flow tests on radiation damaged specimens. In: Prepared for the US Energy Department of Energy under Contract DE-AC06-76RL01830. Pacific Northwest National Laboratory. Jacob, K.T., Dasgupta, N., Waseda, Y., 1998. Composition-graded solid electrolyte for determination of the Gibbs energy of formation of lanthanum zirconate. J. Am. Ceram. Soc. 81, 1926–1930. Kamizono, H., Hayakawa, I., Muraoka, S., 1991. Durability of zirconium-containing ceramic waste forms in water. J. Am. Ceram. Soc. 74, 863–864. Knauss, K.G., Thomas, J.W., 1989. Muscovite dissolution kinetics as a function of pH and time at 70 °C. Geochim. Cosmochim. Acta 53, 1493–1501. Kovalenko, P.N., Bagdasarov, K.N., 1961. Opredelenie Rastvorimosti Gidrookisi Tsirkoniya. Zh. Neorg. Khim. 6, 534–538. Kulik, D.A., GEM-Selektor. Research package for thermodynamic modelling of aquatic (geo)chemical systems by Gibbs Energy Minimization (GEM). . Kulik, D.A., Wagner, T., Dmytrieva, S.V., Kosakowski, G., Hingerl, F.F., Chudnenko, K.V., Berner, U.R., 2013. GEM-Selektor geochemical modeling package: revised algorithm and GEMS3K numerical kernel for coupled simulation codes. Comput. Geosci. 17, 1–24. Lasaga, A.C., 1995. Fundamental approaches in describing mineral dissolution and precipitation rates. In: White, A.F. (Ed.), Chemical Weathering Rates of Silicate Minerals. Mineralogical Society of America, Washington, D.C., pp. 23–86. Lei, Y., Ito, Y., Browning, N.D., Mazanec, T.J., 2002. Segregation effects at grain boundaries in fluorite-structured ceramics. J. Am. Ceram. Soc. 85, 2359–2363. Lide, D., 2007. CRC Handbook of Chemistry and Physics, 88th ed. CRC Press, Boca Raton, FL. Luettge, A., Arvidson, R.S., Fischer, C., 2013. A stochastic treatment of crystal dissolution kinetics. Elements 9, 183–188. Lumpkin, G.R., 2006. Ceramic waste forms for actinides. Elements 2, 365–372. Luo, M.-F., Yan, Z.-L., Jin, L.-Y., He, M., 2006. Raman spectroscopic study on the structure in the surface and the bulk shell of CexPr1xO2d mixed oxides. J. Phys. Chem. B 110, 13068–13071. Lutique, S., Javorsky, P., Konings, R.J.M., van Genderen, A.C.G., van Miltenburg, J.C., Wastin, F., 2003a. Low temperature heat capacity of Nd2Zr2O7 pyrochlore. J. Chem. Thermodyn. 35, 955–965. Lutique, S., Konings, R.J.M., Rondinella, V., Somers, J., Wiss, T., 2003b. The thermal conductivity of Nd2Zr2O7 pyrochlore and the thermal behaviour of pyrochlorebased inert matrix fuel. J. Alloys. Compd. 352, 1–5. Nagy, K.L., Lasaga, A.C., 1992. Dissolution and precipitation kinetics of gibbsite at 80 °C and pH 3: the dependence on solution saturation state. Geochim. Cosmochim. Acta 56, 3093–3111. Neeway, J., Abdelouas, A., Grambow, B., Schumacher, S., 2011. Dissolution mechanism of the SON68 reference nuclear waste glass: new data in dynamic system in silica saturation conditions. J. Nucl. Mater. 415, 31–37.

S. Finkeldei et al. / Applied Geochemistry 49 (2014) 31–41 Oelkers, E.H., 2001. General kinetic description of multioxide silicate mineral and glass dissolution. Geochim. Cosmochim. Acta 65, 3703–3719. Oelkers, E.H., Poitrasson, F., 2002. An experimental study of the dissolution stoichiometry and rates of a natural monazite as a function of temperature from 50 to 230 °C and pH from 1.5 to 10. Chem. Geol. 191, 73–87. Parkhurst, D.L., Appelo, C.A.J., 1999. PHREEQC 2.15 A Computer Program for Speciation, Batch-reaction, One-dimensional Transport and Inverse Geochemical Calculation. Water-Resources Investigation Report 99-4259. Pöml, P., Menneken, M., Stephan, T., Niedermeier, D.R.D., Geisler, T., Putnis, A., 2007. Mechanism of hydrothermal alteration of natural self-irradiated and synthetic crystalline titanate-based pyrochlore. Geochim. Cosmochim. Acta 71, 3311– 3322. Pouchon, M.A., 2000. Contribution to the Study of a Zirconia based Nuclear Fuel for Plutonium use in Light Water Reactor. Ph.D. Thesis. University of Geneva, Switzerland. Pouchon, M.A., Curti, E., Degueldre, C., Tobler, L., 2001. The influence of carbonate complexes on the solubility of zirconia: new experimental data. Prog. Nucl. Energy 38, 443–446. Riedel, E., 2004. Anorganische Chemie, vol. 6. Auflage, Walter de Gruyter, Berlin. Rimstidt, J.D., Dove, P.M., 1986. Mineral solution reaction-rates in a mixed flow reactor – wollastonite hydrolysis. Geochim. Cosmochim. Acta 50, 2509–2516. Roberts, S.K., Bourcier, W.L., Shaw, H.F., 2000. Aqueous dissolution kinetics of pyrochlore, zirconolite and brannerite at 25, 50, and 75 °C. Radiochim. Acta 88, 539–543. Sheka, I.A., Pevzner, T.V., 1960. Rastvorimost Gidrookisei Tsirkoniya I Gafniya V Rastvorakh Edkogo Natra. Zh. Neorg. Khim. 5, 2311–2314.

41

Shock, E.L., Sassani, D.C., Willis, M., Sverjensky, D.A., 1997. Inorganic species in geologic fluids: correlations among standard molal thermodynamic properties of aqueous ions and hydroxide complexes. Geochim. Cosmochim. Acta 61, 907– 950. Strachan, D.M., McGrail, B.P., Scheele, R.A., Rodriguez, E.A., Icenhower, J.P., 2006. Dissolution kinetics of pyrochlore ceramics for the disposition of plutonium. Am. Mineral. 91, 39–53. Szenknect, S., Mesbah, A., Horlait, D., Clavier, N., Dourdain, S., Ravaux, J., Dacheux, N., 2012. Kinetics of structural and microstructural changes at the solid/solution interface during dissolution of cerium(IV)–neodymium(III) oxides. J. Phys. Chem. C 116, 12027–12037. Tocino, F., Szenknect, S., Mesbah, A., Clavier, N., Dacheux, N., 2013. Dissolution of uranium mixed oxides: the role of oxygen vacancies vs the redox reactions. Prog. Nucl. Energy, http://dx.doi.org/10.1016/j.pnucene.2013.09.014. Vandijk, M.P., van Dijk, J.H.H., ter Maat, G., Roelofs, H., Bosch, G.M.H., van de Velde, P.J., Gellings, A.J., Burggraaf, 1984. Electrical and catalytic properties of some oxides with the fluorite or pyrochlore structure. Mater. Res. Bull. 19, 1149– 1156. Wagner, T., Kulik, D.A., Hingerl, F.F., Dmytrieva, S.V., 2012. GEM-Selektor geochemical modeling package: TSolMod library and data interface for multicomponent phase models. Can. Mineral. 50, 1173–1195. Wehrli, B., 1989. Monte Carlo simulations of surface morphologies during mineral dissolution. J. Colloid Interface Sci. 132, 230–242. Wiss, T.A.G., Hiernaut, J.P., Damen, P.M.G., Lutique, S., Fromknecht, R., Weber, W.J., 2006. Helium behaviour in waste conditioning matrices during thermal annealing. J. Nucl. Mater. 352, 202–208.