- © 2016 Mineralogical Society of America
The mineral loparite is the principal ore of the light rare earth elements (LREE) with solid solution compositions in the quaternary system perovskite-lueshite-tausonite-loparite (CaTiO3-NaNbO3-SrTiO3-Na0.5Ln0.5TiO3) (Ln = La, Ce, Pr, Nd). In this study, perovskite solid solutions Ca1−2xNaxLaxTiO3 (0 ≤ x ≤ 0.50) along the perovskite-loparite join were synthesized using solid-state sintering methods. XRD analysis indicates that as the Na+La content increases, the structure changes from orthorhombic to tetragonal. The enthalpies of formation at 298 K from the constituent oxides and from the elements have been determined using high-temperature oxide melt solution calorimetry in molten 3Na2O·4MoO3 at 973 K. Enthalpy of formation from oxides becomes more exothermic with increasing Na+La content, suggesting a stabilization effect of the substitution Ca2+ → 0.5Na+ + 0.5La3+. The observed trend of increasing thermodynamic stability with decreasing structural distortion is in agreement with that seen in many other ABO3 perovskites. The thermodynamic stability of perovskite solid solutions Ca1−2xNaxLaxTiO3 (0 ≤ x ≤ 0.50) along the CaTiO3-Na0.5La0.5TiO3 join provides insights into the natural occurrence of loparite minerals (La,Na,Sr,Ca)(Ti,Nb,Ta,Fe3+)O3.
Loparite, a complex perovskite with general formula (Ce,Na,Sr,Ca) (Ti,Nb,Ta,Fe3+)O3, is of significance ranging from mineralogy to advanced materials. From the geoscience point of view, the mineral loparite is the principal ore of the light rare-earth elements (LREE) found, for example, in Russia (Hedrick et al. 1997), Canada (Chakhmouradian et al. 2007; Platt 1994), South America (Haggerty and Mariano 1983), as well as the U.S.A. (Mitchell and Chakhmouradian 1999). It is a ubiquitous accessory mineral in agpaitic alkaline lithologies, with particular interest in petrogenetic investigations (Mitchell et al. 2011). Most recently, loparite-(Ce) has also been found with promising ferroelectric properties (Popova et al. 2014).
Loparite was first described by Wilhelm Ramsay in 1890 from the agpaitic syenites of the Lovozero alkaline complex in Kola Peninsula of Russia (Ramsay and Hackman 1894) and the chemical composition was analyzed later (Bykova 1941). Loparite belongs to the perovskite family where the A-site may accommodate alkali, alkaline earth, and rare earth cations, ranging from monovalent to trivalent. The structure of loparite-(Ce) minerals has been determined (Mitchell et al. 2000a; Mitchell and Chakhmouradian 1996) using single-crystal X-ray diffraction and is often described to have pseudocubic symmetry. Niobian calcian loparite-(Ce) from the Khibina complex in Russia has been found with orthorhombic structure, while calcian niobian loparite-(Ce) from the Lovozero complex in Russia and strontian calcian loparite-(Ce) from the Bearpaw Mountains in Montana are tetragonal.
Naturally occurring perovskites are of importance as indicators of geochemical evolution and as ores for rare earths. Loparite minerals, with the ideal end-member chemical formula Na0.5Ln0.5TiO3 (Ln = rare earth), belong to the perovskite group of rare earth minerals that also includes CaTiO3 (perovskite), NaNbO3 (lueshite), and SrTiO3 (tausonite). Among naturally occurring perovskite titanates, the quaternary system CaTiO3-NaNbO3-SrTiO3-Na0.5Ln0.5TiO3 (Ln = La, Ce, Pr, Nd) represents much of the compositional variations. In many occurrences, a solid solution is found between Na0.5Ln0.5TiO3 and other perovskite end-members. Figure 1 depicts the quaternary system perovskite (CaTiO3)-lueshite (NaNbO3)-tausonite (SrTiO3)-loparite (Na0.5Ln0.5TiO3). There has been considerable research carried out within this quaternary system. Thermodynamic and crystal-chemical properties of several constituent binary systems have been investigated, including tausonite (SrTiO3)-loparite (Na0.5La0.5TiO3) (Mitchell et al. 2000b), lueshite (NaNbO3)-tausonite (SrTiO3) (Xu et al. 2005), as well as perovskite (CaTiO3)-tausonite (SrTiO3) (Navi et al. 2012). Nevertheless, research on other rare-earth-containing perovskite mineral systems containing the loparite end-member (Na0.5La0.5TiO3) is still relatively scarce, particularly in terms of thermodynamic properties.
The solid solution Ca1−2xNaxLaxTiO3 (0 ≤ x ≤ 0.50) is an A-site substituted perovskite series between perovskite (CaTiO3) and loparite (Na0.5La0.5TiO3) end-members. Structure of both the end-members is shown in Figure 2. Synthetic Na0.5La0.5TiO3 has been described structurally and is claimed to have the undistorted cubic perovskite structure with the space group , and La3+ and Na+ ions randomly distributed at the 12-fold cuboctahedral site (Brous et al. 1953). The crystal structure of the natural mineral loparite has also been determined as the cubic modification for the first time by X-ray diffraction analysis (Zubkova et al. 2000). In contrast, other studies showed that Na0.5La0.5TiO3 may crystallize with either orthorhombic (Pnma) (Mitchell and Chakhmouradian 1998) or rhombohedral symmetry (Mitchell et al. 2000b). These observed differences might be due to the different synthesis routes and/or slight variations in the Na/La ratio (Mitchell et al. 2000b). The other end-member in the CaTiO3-Na0.5La0.5TiO3 series, CaTiO3, has been studied extensively (e.g., Buttner and Maslen 1992; Sasaki et al. 1987). At room temperature, CaTiO3 exhibits the orthorhombic space group Pbnm. This structure is derived from the aristotype by octahedral tilting about the three fourfold axes of the cubic subcell. Phase transitions have been reported from the room-temperature orthorhombic (Pbnm) structure to a tetragonal (I4/mcm) polymorph at temperatures in the range 1373–1423 K, followed by transformation to the cubic aristotype at 1523 ± 10 K (Guyot et al. 1993; Redfern et al. 1989).
In this study, we synthesized the CaTiO3-Na0.5La0.5TiO3 rare earth perovskites solid solution by solid-state reaction. We analyzed structures of the intermediate members in the series by X-ray powder diffraction with particular emphasis on the phase transition and changes in symmetry. One purpose of this study is, therefore, to obtain more insights into the structures and the influence of the A-site substitution on the phase transition. Another goal is to determine thermodynamic properties in the system Ca1−2xNaxLaxTiO3 (0 ≤ x ≤ 0.5) by measuring heats of formation using high-temperature oxide melt solution calorimetry. The energetics of these substitutions could help understand the A-site substitutions and the occurrences of these rare earth perovskite minerals in nature as well as their processing as ores of rare earths and niobium.
Powder samples in the system, Ca1−2xNaxLaxTiO3 (0 ≤ x ≤ 0.5), were made from stoichiometric amounts of CaCO3, Na2CO3, TiO2, and La2O3 by standard solid-state synthesis. The reagents were dried for 5 h at 1273 K for TiO2 (Aldrich 99.99%), 773 K for CaCO3 (Alfa Aesar 99.9%), 1273 K for La2O3 (Alfa Aesar 99.9%), and 423 K for Na2CO3 (Alfa Aesar 99.9%). Stoichiometric mixtures with the addition of 2 wt% excess Na2CO3 (to compensate for sodium vaporized at high temperature) were mixed and ground in an agate mortar, then heated in air initially from 773 to 1073 K at 30 K/h, held at this temperature for 12 h, and then cooled to ambient. After grinding in an agate mortar, the samples were fired again in air at 1473 K for at least 72 h with periodic regrinding, finally rapidly cooled in air to room temperature. The calcined powder was cold pressed and sintered at 1473 K for 12 h. The sintered pellets were crushed and annealed at 973 K for diffraction and calorimetric experiments.
Room-temperature powder X-ray diffraction (XRD) of the synthesized products was carried out on a Bruker D8 diffractometer (Bruker-AXS Inc., Madison, Wisconsin) operated at an accelerating voltage of 40 kV and an emission current of 40 mA with a rotating sample stage using CuKα radiation (λ = 0.15406 nm). Data were recorded between 20 and 130° 2θ with steps of 0.02° and counting time of 10 s per step. XRD data were then analyzed using the Rietveld method with GSAS software. Refined parameters were background coefficients, histogram scale factors, lattice parameters, and profile parameters.
The chemical compositions were analyzed using wavelength-dispersive electron probe microanalysis using a Cameca SX100 instrument (Cameca Inc., Gennevilliers, France) on polished samples. The instrument was operated at an accelerating voltage of 15 kV, with a 20 nA beam current and a spot size of 1 μm. The compositions were calculated from an average of 10 data points per sample. The standards used were NaAlSi3O8 (albite), LaPO4, CaO, and TiO2.
High-temperature drop solution calorimetry in molten sodium molybdate (3Na2O·4MoO3) at 973 K was performed in a custom built Tian-Calvet twin calorimeter. The details of the calorimeter and methodology have been described previously (Navrotsky 1977, 1997, 2014). Prior to calorimetry, compounds were stored in a desiccator. For each measurement, a ~5 mg pellet was loosely pressed, weighed on a microbalance, and dropped from room temperature into 3Na2O·4MoO3 melt at 973 K. The calorimeter was calibrated using the heat content of corundum. Oxygen was bubbled through the solvent at 5 mL/min and flushed over it at 50 mL/min to stir the melt and retain oxidizing conditions. The methodology was the same as that used in our laboratory for other perovskites (Xu et al. 2005; Feng et al. 2016).
Structure and chemical analysis
The formation of a phase with perovskite structure was confirmed by XRD for all samples and no secondary phases were detected. Sodium deficit has been observed in high-temperature synthesis of sodium compounds, resulting from the fact that Na2CO3 could decompose into Na2O and CO2 at 851 °C. The vapor pressure of Na2O is relatively high (Brewer and Margrave 1955), which could explain the sodium deficit. A systematic Na deficit was found in initial syntheses using nominally stoichiometric mixtures of reagents. The addition of 2 wt% excess Na2CO3 to compensate for sodium vaporized at high temperatures gave good final chemical compositions as seen from Table 1. From the microprobe analysis, the chemical compositions were found to be close to nominal with a systematic Na deficit on the order of 1% (which is within experimental error of the nominal composition) and homogeneity at the 2% level. Figure 3 shows X-ray diffraction data and calculated patterns for Ca1−2xNaxLaxTiO3 (0 ≤ x ≤ 0.50). All the phases were perovskites with rhombohedral or tetragonal symmetry. In ABO3 perovskites, structural distortions from the ideal aristotype can often be described in terms of tilting of rigid BO6 octahedra. Structure models for the Rietveld refinements for the end-members are taken from the literature (Li et al. 2007). For the solid solutions, the LeBail fit was done using both Pbnm and I4/mcm space groups and the ones with the lowest goodness of fit parameters were chosen for the Rietveld refinements. Since there is no long-range cation ordering in these materials, all three cations (La, Na, Ca) were assumed to be distributed randomly in the A-site. With increasing sodium and lanthanum content, a transition from orthorhombic Pbnm to tetragonal I4/mcm structure was observed. The cell parameters obtained from the refinements of the entire solid solution are listed in Table 1 and plotted in Figure 4. The cell-edge lengths a and b of the Pbnm phase converge near x = 0.375 with increasing x, indicating a phase transformation to the tetragonal I4/mcm phase at this composition.
Enthalpies of formation
The measured enthalpies of drop solution, ΔHds, of Ca1−2xNaxLaxTiO3 (0 ≤ x ≤ 0.50) were used for determination of the enthalpy of formation from binary oxides, ΔHf,ox, and from elements, ΔHf,el. Table 2 shows the thermochemical cycles used for the calculation of formation enthalpies of Ca1−2xNaxLaxTiO3 (0 ≤ x ≤ 0.50) from binary oxides and from elements. The results on enthalpies of rare earth perovskites Ca1−2xNaxLaxTiO3 (0 ≤ x ≤ 0.50) at 298 K are summarized in Table 3. The drop solution enthalpies as a function of composition are shown in Figure 5.
The powder XRD analysis at room temperature confirms a phase transition in Ca1−2xNaxLaxTiO3 (0 ≤ x ≤ 0.50) solid solutions. With increasing sodium and lanthanum content, a phase transition from orthorhombic Pbnm to tetragonal I4/mcm structure has been observed. Other A-site substituted perovskite solid solutions show analogous transformations. For example, in the tausonite (SrTiO3)-lueshite (NaNbO3) series, the structures of the perovskite solid solution Na1−xSrxNb1−xTixO3 (0 ≤ x ≤ 1) change from orthorhombic to tetragonal and finally to cubic (Xu et al. 2005). In the tausonite (SrTiO3)-loparite (Na1/2La1/2TiO3) series, the structures of the solid solution Sr1−2xNaxLaxTiO3 (0 ≤ x ≤ 0.50) change from cubic to tetragonal and finally to rhombohedral (Mitchell et al. 2000b). In the perovskite (CaTiO3)-tausonite (SrTiO3) series, the structures of solid solutions CaxSr1−xTiO3 (0 ≤ x ≤ 1) change from orthorhombic to tetragonal (Navi et al. 2012). As mentioned above, the commonly accepted space group for the end-member Na1/2La1/2TiO3 seems to be tetragonal I4/mcm (Li et al. 2007), not the rhombohedral reported by Mitchell et al. (2000a). Our structure refinements confirm this as well. Both Pbnm to I4/mcm and Pbnm to R3c transitions are of first order (Howard and Stokes 1998). Although the transition from Pbnm directly to has been observed in some cases, such as the B-site substituted perovskites in the LaNiO3–LaFeO3 system (Falcon et al. 1997), and the A- and B-site substituted perovskites in the LaCrO3–SrTiO3 series (Kennedy et al. 2000), it has not been reported in A-site substituted perovskites.
The enthalpy of formation of Ca1−2xNaxLaxTiO3 (0 ≤ x ≤ 0.50) from the oxides is plotted in Figure 6a. With increasing Na+La substitution at the A-site, the enthalpy of formation becomes more exothermic, implying greater stability relative to the binary oxides. The distortion from cubic symmetry can be rationalized using the Goldschmidt tolerance factor, t = (rA + rO)/√2(rB + rO) (Goldschmidt 1926). Figure 6b shows a linear trend between the enthalpy of formation and the Goldschmidt tolerance factor, consistent with previous trends in perovskite stability (Navrotsky 2000; Xu et al. 2005; Zlotnik et al. 2015). Interestingly, there is no change in slope near the phase transformation (space group change), suggesting that the enthalpy of the phase transition is small, probably 1 kJ/mol or less, as seen in other perovskites (Navrotsky 2000; Xu et al. 2005; Zlotnik et al. 2015).
XRD analysis indicates that as the Na+La content increases, the structure changes from orthorhombic to tetragonal near x = 0.375. The formation enthalpies from oxides become more exothermic with increasing Na+La content, suggesting a stabilization effect of the substitution Ca2+ → 0.5Na+ + 0.5La3+. The trend of increasing thermodynamic stability with decreasing structural distortion is consistent with that seen in many other ABO3 perovskites, reflecting increasing stability as the tolerance factor approaches unity.
Natural loparite minerals occur as perovskite solid solutions, mainly in the quaternary system perovskite-lueshite-tausonite-loparite (CaTiO3-NaNbO3-SrTiO3-Na1/2Ln1/2TiO3) (Ln = La, Ce, Pr, Nd) (see Fig. 1). It is interesting to compare the energetics of the various substitutions along these joins. The substitution along the CaTiO3-Na0.5La0.5TiO3 join can be written as the following, (1)which can also be written in terms of oxide components (2)
Since there is negligible enthalpy of mixing along this solid solution, one can take the difference in enthalpy of formation of the end-members as the enthalpy of the substitution, namely ΔH = ΔHf,ox (Na0.5La0.5TiO3) − ΔHf,ox (CaTiO3) = −16.0 kJ/mol. Xu et al. (2005) investigated the energetics of perovskite solid solutions along the NaNbO3-SrTiO3 join and found that the formation enthalpy becomes more exothermic with increasing Na+Nb content, suggesting stabilization with respect to component oxides for the substitution, (3)or, relative to oxide components, (4)
Again, with an observed negligible heat of mixing along the solid solution ΔH = ΔHf,ox (NaNbO3) − ΔHf,ox (SrTiO3) = −40.3 kJ/mol. The energetics of other substitutions, calculated analogously from the differences in enthalpies of formation of end-member perovskites, are shown in Table 4. The stabilizing effect, relative to binary oxides, of the sodium-containing substitutions reflects the greater basicity of Na2O than of CaO or SrO. The substitution of Na+La is energetically favorable for Ca but not for Sr. Since the alkali and alkaline earth ions can only occupy the A-sites, the maximum extent of substitution depends on whether the charge balancing cation of higher valence occupies A-sites (like the rare earths) or the B-sites (like niobium), with greater substitution possible in the latter case (compare Na0.5La0.5TiO3 and NaNbO3 as end-members). The stabilizing effect of the substitution of Sr for Ca reflects the greater basicity of SrO than CaO. The Na+Nb substitution is more stabilizing than the Na+La substitution.
These energetic trends can be compared to substitution patterns seen in natural loparite minerals. The compositions of loparite-(Ce) from the layered complex at Lovozero, Russia (Pakhomovsky et al. 2014), generally follow the schemes of isomorphic replacements. The loparites, although variable in composition, fall in a relatively narrow composition range, with factor analysis suggesting the following two major substitutions: 2Ca2+ + Ti4+ = Sr2+ + Na+ + Nb5+ at constant rare earth content and Ln3+ + Ti4+ = Ca2+ + Nb5+ at constant Na and Sr contents. These are indeed energetically stabilizing substitutions as can be calculated from the data in Table 4. Suk et al. (2013) studied loparite from a complex of differentiated rocks and from the Lovozero alkaline massif. They also carried out experimental studies of loparite formation in complex silicate-molten salt systems including sodium carbonate, chloride, fluoride, and sulfate at 673–1473 K and 1–2 kbar. The results show that the composition of loparites depends on the physicochemical conditions of their formation (fluid/melt composition) and that loparite can crystallize in a wide range of temperatures. The synthetic loparite crystals are zoned with variations in Ti, Nb, La, Ce, Y, Ca, and Sr contents and are similar to those in the natural systems. Higher alkalinity of the fluid appears to initially favor loparite formation. This is in accord with the thermochemical data in Table 4, which show that the Na+Ca and Na+Nb substitutions are energetically favorable. Kogarko et al. (2002) observed that the composition of the cumulus loparite changes systematically upward through the Lovozero intrusion with an increase in Na, Sr, Nb, and Th and decrease in REE and Ti. At later stages of evolution, when concentrations of alkalis and volatiles in the fluid reached very high levels, loparite reacted with the residual melt to form various other minerals incorporating rare earths. From our research, we found that the synthesis of sodium-rich compounds (x > 0.5) in the Ca1−2xNaxLaxTiO3 was not successful. We could conclude that the Ca2+ ↔ 0.5Na+ + 0.5Ln3+ substitution is complete at x = 0.5 when the A-site is completely substituted. This substitution is in competition with the Ca2+ + Ti4+ = Na+ + Nb5+ substitution. Again, since sodium can only enter the A-site, the ability of perovskite to accommodate Na+Nb may become limited at very high alkalinity. Therefore additional sodium, niobium, and lanthanum concentrations in the reacting medium would have to react to form other compounds, as we observed.
This work was supported by the National Science Foundation, Division of Earth Sciences with contributions from the NSF-wide SUSchem initiative ( Grant EAR-1321410).
- Manuscript Received February 26, 2016.
- Manuscript Accepted May 2, 2016.