Ionic Conductivity of Strontium Fluoroapatites Co-doped with Lanthanides

Abstract

Britholites derivatives of apatite’s that contain lanthanium and neodymium in the serial compounds Sr8La2−xNdx(PO4)4(SiO4)2F2 with 0 ≤ x ≤ 2 were subject of the present investigation. The solid state reaction was the route of preparing these materials. Several techniques were employed for the analysis and characterization of the synthesized powders. The chemical analysis results indicated that molar ratio Sr+La+NdP+Si was of about 1.67 value of a stoichiometric powder. The X-ray diffraction data showed single-phase apatites crystallizing in hexagonal structure with P63/m space group were successively obtained. Moreover, the substitution of lanthanium by neodymium in strontium phosphosilicated fluorapatite was total. This was confirmed by the a and c lattice parameters contraction when (x) varies coherently to the sizes of the two cations. The infrared spectroscopy and the 31P NMR (MAS) exhibited the characteristic bands of phosphosilicated fluorapatite. The pressureless sintering of the material achieved a maximum of 89% relative density. The sintered specimens indicated that the Nd content as well as the heating temperature affected the ionic conduction of the materials and the maximum was 1.73 × 10−6 S cm−1 obtained at 1052 K for x = 2.

1. Introduction

The phosphosilicate apatites containing a coupled substitution of the divalent cation by a trivalent lanthanide or a tetravalent actinide ion and the trivalent groupment PO₄ by a tetravalent SiO₄ groupment in the general formula Me(XO₄)6Y₂ (Me: divalent cation; XO₄: anionic groupment and Y: monovalent anion) allow to obtain materials called britholite [1, 2, 3, 4]. Such materials were found in the natural nuclear reactors Alko of Gabon which demonstrated that they are storing some radionuclides such as uranium U, thorium Th, plutonium Pu and minor actinides like neptinium Np, americium Am and curium Cm [5, 6, 7, 8]. Moreover, silicate based apatite samples were found to contain up to 50 wt% of lanthanides (La, Ce, Nd) and actinides (U, Th) in Ouzzal site of Algeria [9]. Hence, britholites were considered as natural nuclear waste disposal and allowing the confinement of radionuclides and some fission byproducts produced by the nuclear industry [10, 11, 12]. In fact, many studies indicated that britholites are able to confine radionuclides with continuous irradiation for millions of years with conserved structure and thermal and chemical stability [13, 14]. Indeed, due to the stability and flexibility of their structure, apatites offer many possibilities for substitutions. Moreover, britholite materials favored many cationic and anionic substitutions in their crystallographic structure. These later might be in a total or limited range [15, 16, 17]. Therefore, these substitutions are governed by the ionic sizes, the valence, the electronegativity and the polarizability [18]. In this context, several processes have been developed for the preparation of these materials containing various elements such as actinides and lanthanides via solid state reaction or mechanical synthesis [19, 20, 21, 22, 23, 24, 25, 26, 27].

On the other hand, many investigations have revealed that britholites might be a good ionic conductor for their use in fuel cells. The conductivity was proved as a thermal process at intermediate temperature range 400–900°C [28, 29, 30, 31, 32]. Therefore, the electrical properties allow using the materials as a solid electrolyte in solid oxide fuel cells (SOFCs) [33, 34].

Like-apatite, phosphosilicate apatites have a hexagonal structure and a space group P63/m [15, 35]. Their framework is built on the sixth XO₄ groups and the Me is divided between two crystallographic sites: four are located in the site Me(1), coordination 9, and six other are located in the site Me(2), coordination 7. Hence, in order to highlight the capacity of these materials to store non radioactive elements similar to radionuclides as well as their potentialities as ionic conductors, the sintered materials series Sr₈La_2−xNd_x(PO₄)₄(SiO₄)₂F₂ with 0 ≤ x ≤ 2 were investigated.

2. Materials and methods

2.1 Powder synthesis

A solid state method was adopted to prepare strontium fluorobritholites compounds Sr₈La_2−xNd_x(PO₄)₄(SiO₄)₂F₂ with 0 ≤ x ≤ 2 [36]. The starting reagents: strontium fluoride SrF₂ (99.99%. Merck)), strontium carbonate SrCO₃ (≥99.00% Fluka), silica SiO₂ (Prolabo), lanthanum and neodymium oxide (La₂O₃·Nd₂O₃₎ (99.99% Merck) and strontium diphosphate (Sr₂P₂O₇) were used. The reaction equation (1) is the following:

$$\begin{array}{l}{\text{3SrCO}}_3+{\text{2Sr}}_2{\mathrm{P}}_2{\mathrm{O}}_7+{\mathrm{SrF}}_2+{\text{2SiO}}_2+\frac{2-\mathrm{x}}{2}{\mathrm{La}}_2{\mathrm{O}}_3+\frac{\mathrm{x}}{2}{\mathrm{Nd}}_2{\mathrm{O}}_3\\ \to {\mathbf{Sr}}_{\mathbf{8}}{\mathbf{La}}_{\mathbf{2}-\mathbf{x}}{\mathbf{Nd}}_{\mathbf{x}}{\left({\mathbf{PO}}_{\mathbf{4}}\right)}_{\mathbf{4}}{\left({\mathbf{SiO}}_{\mathbf{4}}\right)}_{\mathbf{2}}{\mathbf{F}}_{\mathbf{2}}+{\text{3CO}}_2\uparrow \text{with}0\le \mathrm{x}\le 2.\end{array}$$E1

Sr₂P₂O₇ was synthesized by the following reaction at 900°C:

$${\text{2SrCO}}_3+2{\left({\mathrm{NH}}_4\right)}_2{\mathrm{HPO}}_4\to {\mathrm{Sr}}_2{\mathrm{P}}_2{\mathrm{O}}_7+{\text{2CO}}_2+{\text{3H}}_2\mathrm{O}+{\text{4NH}}_3$$E2

SrCO₃ (>96% Riedel de Haen), Gd₂O₃ (>99.5% Prolabo), Nd₂O₃ (>99.5% Prolabo) SiO₂ (>99.5% Alfa), SrF₂ (>99.5% Prolabo) and (NH₄)₂HPO₄ (>99% Acros Organics) were used as raw materials. For each composition the molar ratio (Sr + La + Nd)/(P + Si) and the obtained quantity of each composition should be respectively 1.67 and 1.5 × 10⁻³ moles. Before synthesis, each quantity of lanthanum and neodymium oxides given in Table 1 was furnaced at 1000°C for 12 h to avoid the formation of Ln-hydroxide. Then, the solid mixture was milled and homogenized in an agate mortar for about 30 min, and then cold pressed under 100 MPa into pellets (30 and 3 mm). During sintering, the pellets were sintered in the temperature range 1250–1450°C in a carbolyte type furnace with controlled argon atmosphere. The temperature varied with 50°C for each value of x. The sintering cycle is shown in Figure 1. The heating and cooling rate was of 10°C min⁻¹. In the following sections, the samples will be named SrLa_2−xNd_xF where x is the substituted Nd rate.

Figure 1.

Thermal cycle used for strontium fluorobritholite sintering.

Reactants	SrLa2F	SrLa1.5Nd0.5F	SrLa1Nd1F	SrLa0.5Nd1.5F	SrNd2F
La2O3	0.4887	0.3665	0.2443	0.1221	—
Nd2O3	—	0.1261	0.2523	0.3785	0.5047

Table 1.

Masses in grams of lanthanum and neodymium oxides used in the synthesis of ${\mathit{Sr}}_{\mathit{8}}{\mathit{La}}_{\mathit{2}-\mathit{x}}$Nd_x(PO₄)₄(SiO₄)₂F₂ with 0 ≤ x ≤ 2.

2.2 Analysis and characterization

A PANalytical X’pert Pro diffractometer with a KαCu anode (λ = 1.54 Å) operating 40 kV and 40 mA was the apparatus used for the XRD patterns recording. The scans range was between 10 and 70° (2θ) with a step size of 0.02°. The crystallite size of the powder D_hkl was calculated using the (300) and (002) reflections following Debye Sheerer equation [37]:

$$\mathrm{D}=\frac{\mathrm{K\lambda }}{{\mathrm{\beta }}^{1/2}\text{cosθ}}$$E3

Needs to remember that λ is the X-ray wavelength of the monochromatic X-ray beam. For the apatitic crystallites K is a constant equal to 0.9. β^1/2 is the full width at half maximum of the selected reflection and θ is the Bragg’s diffraction angle.

The Fourier transformed infrared (FTIR)-attenuated total reflection (ATR) spectra were performed at room temperature on a Perkin Elmer spectrometer in the spectral range 4000–400 cm⁻¹.

The chemical analysis of Sr., P, Si, La and Nd ions in the synthesized samples was determined via an inductively coupled plasma atomic emission spectroscopy (ICP-AES) (JY-Horiba Ultima-C spectrometer). The samples were thus previously mixed with 99.9% lithium metaborate, fused at 1000°C for 25 min and dissolved in HCl (0.6 M). The fluoride content in the synthesized samples was measured by a specific ion-selective electrode.

The complex impedance measurements were performed on pellets sintered at between 1250 and 1450°C for 24 h. Their densities varied from 72 to 83% of the theoretical density as a function the sintering temperature. The two faces of the pellets were coated with a silver paint and then two platinum wires electrodes linked them to a Hewlett-Packard 4192-A impedance analyzer. The measurements were recorded with the temperatures variation from 450 to 780°C and frequencies from 10 Hz to 13 MHz.

3. Results and discussion

3.1 Chemical content

The samples’ quantitative chemical analyses are shown in Table 2. As observed there is a satisfactory agreement between the elements amount determined from the analyses and those introduced in the starting. As a consequence, the experimental formula was close to the theoretical ones. The $\frac{\mathrm{Sr}+\mathrm{La}+\mathrm{Nd}}{\mathrm{P}+\mathrm{Si}}$ molar ratios are very close to the theoretical value of 1.667 for stoichiometric apatite.

Samples	Sr	La	Nd	P	Si	F	Molar ratio $\frac{\mathrm{Sr}+\mathrm{La}+\mathrm{Nd}}{\mathrm{P}+\mathrm{Si}}$
SrLa2F	7.96	1.97	—	3.98	1.97	1.99	1.668
SrLa1.5Nd0.5F	7.96	1.47	0.47	3.98	1.96	1.98	1.666
SrLa1Nd1F	7.97	0.98	0.97	3.99	1.96	1.96	1.667
SrLa0.5Nd1.5F	7.99	0.47	1.48	3.99	1.98	1.97	1.664
SrNd2F	7.98	—	1.97	3.98	1.98	1.97	1.669

Table 2.

Number of atoms per unit cell of Sr₈La_2−xNd_x(PO₄)₄(SiO₄)₂F₂ (0 ≤ x ≤ 2).

3.2 X-Ray diffraction analysis

Figure 2 showed the XRD patterns of all compositions. It is evident that the samples were single apatite phase. By comparaison to the JCPDS 17-0609 file data for the strontium fluorapatite, the samples are characteristic of the hexagonal symmetry and the P63/m space group. No additional diffraction lines relative to supplementary phases were detected in any of the patterns. However, the presence of very small quantities of impurities was not excluded. The XRD patterns of the Figure 3 indicated that when the substitution level increased, the peaks slightly shift towards the high 2θ angles indicating a contraction of the unit cell. This contraction, which agrees with the Nd³⁺ radius (^VIr_Nd³⁺ = 0.983 Å) [38], that is smaller than that of La³⁺ (^VIr_La³⁺ = 1.032 Å) [39] confirms the incorporation of Nd³⁺ into the apatite structure.

Figure 2.

DRX spectra of strontium fluorbritholites Sr₈La₂Nd_x(PO₄)₄(SiO₄)₂F₂ with (0 ≤ x ≤ 2).

Figure 3.

Radiation (300) of fluorbritholites Sr₈La_2−xNd_x(PO₄)₄(SiO₄)₂F₂ (0 ≤ x ≤ 2).

As shown in Figure 4 and Table 3, the cristallographic parameters calculated using the Fullprof program without any structural refinement of the all compositions depended on the substitution level. In fact, if Nd content rose, a and c decreased. The calculated cristallographic parameters were similar to that existing in the literature [40, 41]. Moreover the evolution of the lattice parameters was linear in accordance with the Vegard’s law:

Figure 4.

Lattice parameters as a function neodymium level in the Sr₈La_2−xNd_x(PO₄)₄(SiO₄)₂F₂ with (0 ≤ x ≤ 2).

Sample	a (Å)	c (Å)	V (Å^3)	D300 (Å)	D300 (Å)
SrLa2F	9.735(2)	7.281(2)	597.55(2)	304(3)	387(4)
SrLa1.5Nd0.5F	9.730(3)	7.278(3)	596.69(2)	278(2)	362(3)
SrLa1Nd1F	9.725(2)	7.271(2)	595.43(2)	254(3)	347(2)
SrLa0.5Nd1.5F	9.720(3)	7.267(2)	594.57(3)	237(3)	326(3)
SrNd2F	9.717(3)	7.263(2)	593.87(3)	223(4)	307(4)

Table 3.

Crystallographic parameters of strontium fluorbritholites Sr₈La_2−xNd_x(PO₄)₄(SiO₄)₂F₂ (0 ≤ x ≤ 2).

$$\mathit{a}=-0.0092\mathrm{x}+9.7346\mathrm{Ǻ};\mathrm{\sigma }\left(\mathrm{a}\right)=2.5\times {10}^{-3}\mathrm{Ǻ}\left(\mathrm{\sigma }:\text{standard deviation}\right)$$

$$\mathit{c}=-0.00\text{9x}+7.2806\mathrm{Ǻ};\mathrm{\sigma }\left(\mathrm{c}\right)=2\times {10}^{-3}\AA$$E4

indicates the existence of a continuous solid solution in the explored substitution domain.

The rational parameters that govern the site occupation are the nature, the electronegativities, the valences and the polarizabilities of the ions. The bibliography studies’ results indicated that like those observed in natural phosphosilicate apatites, the substituted cations in the apatite structure had preferential occupation for Me(2) sites [42, 43, 44, 45]. Thus, it could be concluded that La³⁺ and Nd³⁺ ions substituting Sr²⁺ with 0 ≤ x ≤ 2 in our studied samples were subsequently preferentially localized in Me(2) sites.

3.3 FTIR spectroscopy

The FTIR spectra of the samples were given in Figure 5. The identification of all the bands was done by comparison with un- and substituted strontium fluorapatite the previously reported in the literature [40, 41]. The characteristic absorption bands of SiO₄ and PO₄ were observed [41].

Figure 5.

Infrared spectra of Sr₈La_2−xNd_x(PO₄)₄(SiO₄)₂F₂ (0 ≤ x ≤ 2).

The PO₄characteristic bands observed at 1072–1024 cm ⁻¹ coincide with to the asymmetric stretching mode (υ₃). The band at 943 cm⁻¹ is attributed to the symmetrical stretching mode (υ₁). The asymmetric bending mode (υ₄) bands appeared in the 538–562 cm⁻¹ range and the symmetric one (υ₂) are shown at 455 cm⁻¹. The bands observed at 909–943 (υ₃), 832–870 (υ₁), about 538 (υ₄) and 460–498 cm⁻¹ (υ₂) were assigned to SiO₄. Moreover as the neodymium amount in the samples increased a shift of the PO₄ and SiO₄ absorption bands towards the low wave numbers was detected. This was attributed to a reduction in size of the lattice inducing an increase in anion-anion repulsion (PO₄ vs. SiO₄) [45]. This observation was in good agreement with those obtained by diffraction of X-rays and confirms that the neodymium substitution reduced the lattice size.

3.4 ³¹P NMR spectroscopy

In the Figure 6 are represented the ³¹P NMR-MAS spectra. A single isotropic signal was observed for all the spectra. It indicated also that a unique crystallographic site for the PO₄ tetrahedron in the apatite structure was present. However a slight chemical shift towards the lower values was observed as well as a broadening of the peaks was attributed to the Nd substitution. This fact was related to a disorder induced in the apatite network caused by the substitution of La by Nd. This was previously seen with doped with rare earth apatite’s [46, 47, 48].

Figure 6.

³¹P NMR-MAS spectra of fluorbritholites Sr₈La_2−xNd_x(PO₄)₄(SiO₄)₂F₂ with (0 ≤ x ≤ 2).

3.5 Materials sintering

Materials densification optimization has been performed by sintering the synthesized samples in the temperatures range 1250–1500°C with a fixed holding time of 6 h. Relative density $d_{\mathit{re}}$ was calculated using the equation:

$$d_{\mathit{re}}=\frac{{\mathrm{\rho }}_{\text{the}}}{{\mathrm{\rho }}_{\exp }}\times 100$$E5

where the theoretical density ρ_the was calculated using the equation:

$${\mathrm{\rho }}_{\text{the}}=\frac{\mathrm{ZM}}{\mathrm{NaV}}$$E6

(Z: number of molecules/unit cell, M: molecular weight Na: Avogadro number and V: volume of the unit cell) and experimental density determined from the mass and the dimension of sintered pellets by means of the equation

$${\mathrm{\rho }}_{\text{the}}=\frac{m}{\pi {\mathit{hr}}^2}$$E7

Figure 7 shows that relative density of the sintered samples strictly depends on sintering temperatures as well as on Nd content. An irregular trend was noted and the highest relative density 89% was obtained with x = 2 Nd content when sintered only at 1250°C. The remaining samples presents lower than densifications ratios obtained at higher temperatures. From these data, it can be deduced that the grains morphology and size modification strongly depends on Nd content and sintering temperature. The Nd doping should improve the materials densification by reducing the porosity. This was confirmed by the percentage porosity of the higher densified samples calculated by the following equation:

Figure 7.

Relative density versus sintering temperature of Sr₈La_2−xNd_x(PO₄)₄(SiO₄)₂F₂ with (0 ≤ x ≤ 2).

$$p=\left(1-\mathit{dr}\right)\times 100$$E8

As plotted on Figure 8, the porosity of the samples decreased as Nd content increased. This result muched the evolution of the relative density suggested to increase when crystallite size is reduced (i.e. grain size). This should promotes the materials densifications by eliminating the intergranular porosity.

Figure 8.

Porosity versus Nd content of maximum densified samples.

The microstructure of the samples given on Figure 9 is closely coherent with the densification rates as well as porosity. Indeed, the micrographs show a progressive removal of the porosity when the Nd rate rises. Thus with x = 0 the microstructure is of intergranular aspect revealing the presence of abundant porosity. With x = 0.5, although some pores persist on the surface the porosity was reduced,. When x = 1 the open porosity has almost disappeared and only the closed porosity remains, reflecting the 89% densification.

Figure 9.

Micrographs of sintered samples Sr₈La_2−xNd_x(PO₄)₄(SiO₄)₂F₂ (a) x = 0.0; (b) x = 0.5; (c) x = 1.0.

$${\text{Sr}}_8{\text{La}}_{\mathrm{2-x}}{\text{Nd}}_{\text{x}}{\left({\text{PO}}_4\right)}_4{\left({\text{SiO}}_4\right)}_2{\text{F}}_2\text{with}(0\le \text{x}\le 2).$$E9

3.6 Impedance spectroscopy

The ionic conductivity of the samples was determined between 400 and 800°C with a step of 20°C by complex impedance plots. Thus, for each sample, 20 complex impedance plots (plane, Z″ vs Z′) were plotted. The intercept of the semicircular arcs with the real axis allow obtaining the bulk resistance R. The ionic conductivity of the sintered samples was calculated from the equation:

$$\mathrm{\sigma }=\frac{\mathrm{e}}{\mathrm{SR}}$$E10

The thickness and the area of the sample were e and S, respectively. Figure 10 reprinted the ionic conductivity σ versus the neodymium substitution. The first deduction is that σ depends on this substitution and particularly at higher temperatures. The curves obtained at 604 (877 K) and 482°C (755 K) indcated that the measured conductivity was about 4.4 × 10⁻⁷ S cm⁻¹_. By contrary with the increase of Nd content, σ rose up to 1.73 × 10⁻⁶ S cm⁻¹ at 779°C (1052 K). Hence, the electric conductivity of the samples depend onthe Nd substituted level.

Figure 10.

Ionic conductivity versus neodymium content.

The total activation energy of the samples was obtained from the Arrhenius equation:

$$\mathrm{\sigma T}=\mathrm{A}{\mathrm{e}}^{-\frac{\mathrm{Ea}}{\mathrm{KT}}}$$E11

The parameters to define are the pre-exponential factor A, activation energy Ea, Boltzmann constant k and absolute temperature T, respectively. Figure 11 shows an Arrhenius-type plot indicating that the electrical conduction of the materials is activated by heating. The σ values were slightly different from those found in the literature [49, 50]. The difference might have resulted from the preparation and sintering methods reflected by the difference in densification ratios (range 72–83%). The slope in the Arrhenius plots versus temperatures gives the activation energy. This later parameter increased when Nd level rose reaching a maximum of 1.1 eV when x = 1 then decreased to 0.91 eV (Table 4). Moreover a slight break in slope for x ≥ 1 was detected in the Arrhenius plots. This was related to the Sr/Nd▬F bond likely to the work of Njema and al [49]. In fact, in Sr₈La_2−xNd_x(PO₄)₄(SiO₄)₂F₂ with (0 ≤ x ≤ 2) samples, the mobility of F⁻ along the c axis ensure the charge motion. Thus as the Nd-doping increased the F⁻ mobility was enhanced and the conductivity was improved. The fluoride ions motion along the structure should be related to the neodymium polarizability slightly higher than its of lanthanium. Herein, the polarizabilities of lanthanum and neodymium were 4.82, 5.01 Å³, respectively [51, 52]. Laghzizil and al emphasized the improved fluoride mobility in the presence of polarizable cations localized in Me(2) site [53, 54].

Figure 11.

Plots of LnσT versus 1000/T of fluorbritholites Sr₈La_2−xNd_x(PO₄)₄(SiO₄)₂F₂ with (0 ≤ x ≤ 2).

x	0	0.5	1	1.5	2
Ea (eV)	0.87	0.95	1.1	1.03	0.91

Table 4.

Activation energy of Sr₈La_2−xNd_x(PO₄)₄(SiO₄)₂F₂ with (0 ≤ x ≤ 2).

4. Conclusion

Strontium fluorbritholites Sr₈La_2−xNd_x(PO₄)₄(SiO₄)₂F₂ with (0 ≤ x ≤ 2) were successfully prepared by reaction in the solid state. Characterization by several techniques revealed that all the powders were composed of a single apatite phase. The lattice parameters a and c were inversely proportional to Nd content. The La-Nd substitution was totally coherent with the ionic size of lanthanum and neodymium ions. The chemical analysis showed that for all the compositions the ratio $\frac{\mathrm{Sr}+\mathrm{La}+\mathrm{Nd}}{\mathrm{P}+\mathrm{Si}}$ was close to the stoichiometric value (1.667) indicating the stoichiometry of the powders. The Infrared absorption apectroscopy study confirmed the presence of the absorption bands relative to PO₄ and SiO₄ groups. The ³¹P NMR (MAS) showed the presence of a unique isotropic signal confirming the existence of a single crystallographic site for the phosphorus nucleus. The materials’ ionic conductivity measured via impedance spectroscopy was found to be heating-dependent. The maximum of 1.73 × 10⁻⁶ S cm⁻¹ is obtained with higher Nd level and at a temperature of around 1052 K.

Conflict of interest

The authors declare no conflict of interest.