Characterization and densification of defect pyrochlore oxide powders ABi₂Ta₅O₁₆ (A=Na, Tl)

Abstract

Defect pyrochlore oxide powders ABi₂Ta₅O₁₆, in the pseudo-binary systems BiTaO₄–ATa₃O₈ (A = Na, Tl), were prepared by solid state method. The structural study showed that all oxides crystallize in a cubic system with the space group Fd3m, the lattice parameter “a” were determined using Rietveld refinement method. For systematic study on densification from powders, the samples were pressed with a uniaxial pressure into pellets; sintering temperature, holding time and heating rate were optimized. Techniques, including X-ray diffraction, IR-Raman, MEB, dilatometry, were employed to investigate the structure and the morphology of the synthesized powders and sintered materials. The dielectric characteristics, relative permittivity and dielectric losses (tgδ), determined at room temperature are comparable to those of other pyrochlores.

1. Introduction

Among metallic oxides, compounds with general formula, A₂B₂O₆O′, where A and B are different cation species with different oxidation states (e.g. A³⁺ B⁴⁺ or A²⁺ B⁵⁺) and O and O′ are anions, represent a family of phases isostructural to the mineral pyrochlore (NaCa) (NbTa)O6F/(OH). The pyrochlore structure possesses cubic symmetry with space group (Fd3m, Oh7), no. 227 and number of formula units (Z = 8). Four crystallographically nonequivalent kinds of atoms: A, B, O and O′ atoms occupy the 16d, 16c, 48f and 8b sites, respectively. The smaller B cations (16c) are six coordinated and located in a distorted octahedron formed by six oxygen ions O (48f). The larger A cations (16d) are eight-coordinated, located within a distorted cubic polyhedron formed by six O (48f) and 2 O’ (8b). The structural formula is often written as B₂O₆•A₂O', which emphasizes that the arrangement consists of two interpenetrating networks of vertex-linked octahedra (B₂O₆) sharing corners, and a cuprite-like A₂O' tetrahedron [1]. This latter network is not essential for the stability of the structure and hence pyrochlore structure tolerates vacancies at A and O′ sites giving defect pyrochlore with formula A_2-x□_xB₂O₆O′_1-y□'_y giving rise to different stoichiometries from AB₂O₆ to A₂B₂O₆O'. It is known that AB₂O₆ type defect pyrochlore are formed in the case where the A cation is a large monovalent metal such as Cs⁺, Rb⁺, Tl⁺ or K⁺ [2]. The site occupancy of the A cations in defect pyrochlore is insertion because the structural considerations reveal that in addition to the 16d and 8b sites there in another possible site (32e) depending on the size, charge and polarizability of the A cation either full or fractional occupation of any of the three sites is possible. Studies showed that the A ions in RbTa₂O₅F occupy the 8b positions whereas in TlNb₂O₅F the Tl cations are delocalized from the ideal 16d position to occupy the 32e positions statistically [2]. Pyrochlore mixed oxides have many technological applications thanks to theirs wide spectrum of properties (electrical, dielectric, magnetic, magneto-resistive, optical, electro-catalytic etc.) and can also be used as electrolytes or anode materials for solid electrolyte fuel cells (SOFC) and fixatives for radioactive waste due to theirs oxygen nonstoichiometry and to the distribution of cations in the structure [1]. Recently, Bi-based pyrochlores in the Bi₂O₃–ZnO–M₂O₅ (X = Sb, Ta and Nb) ternary systems have triggered great research interests owing to their relatively low sintering temperatures and excellent dielectric properties [3, 4, 5]. Some pyrochlore compounds occurring in the A₂O–Bi₂O₃- M₂O₅ system (A = Ag, Na, K, Tl and M = Nb, Ta) are attractive candidates for temperature-stable, low-loss, high-permittivity dielectric applications [6].

2. Experimental

The compounds ABi₂Ta₅O₁₆ (A = Na, Tl) were prepared by usual solid state reaction from starting reagents Bi₂O₃ (Sigma-Aldrich, 99.9%), Ta₂O₅ (Sigma-Aldrich, 99%) Na₂CO₃ and Tl₂CO₃ (Sigma-Aldrich, 99.9%) Weighted quantities of the reactants were intimately mixed and ground in an agate mortar. In order to avoid a possible oxidization of Tl⁺ to Tl³⁺, all compounds were treated under nitrogen atmosphere. Three thermal processings with intermittent regrinding at 300 °C (6 hours) (to provoke the departure of CO₂ and in the same time to avoid volatilization of Tl₂CO₃ [7], 800 °C (12 hours) and 950–1000 °C (12 hours) were necessary to obtain the final compounds. X-ray powder diffractograms were obtained using Cu Kα line of a D5000 Siemens diffractometer equipped with a back monochromator. The data were recorded between 10 and 120 (2θ) in steps of 0.04° with a count time of 72 s. Temperature programmed X-ray diffraction was performed on a Siemens D5000 diffractometer fitted out with an ANTON PARR furnace (CHTK10) and a linear detector (Elphyse 14°). The powder of each prepared compound was ground and mixed with 5% of polyvinyl alcohol and polyethylene glycol and pressed using a uniaxial pressure 200 MPa into pellets with a diameter of 10 mm and thickness of about 2 mm. The sintering behavior of the powders was investigated by dilatometric method. The powders were uniaxially pressed at 200 MPa and the compacts were sintered in the vertical dilatometer (SETARAM TMA92) up to 1400 °C under air sweeping with heating rates of 3 °C/min. The relative densities were also calculated from comparisons of densities which are determined geometrically and theoretically. The surface morphologies of the samples were examined by scanning electron microscopy (SEM). The IR spectra were recorded in the range of 400–4000 cm⁻¹ with 4 cm⁻¹ resolution using a Fourier-Transformed BOMEN MX spectrometer from dicks containing 2% of the sample and 98% of KBr or CsCl. Infrared active phonon modes are attributed to the vibrations of ions in the crystal lattice. Raman measurements were done in a DILOL-XY type Raman spectrometer equipped with a triple monochromator in order to examine lattice vibration and hence local structure of materials. All spectra were obtained at room temperature. The excitation source is an argan-ion laser (5W) and the emission line used is 514.5 nm (Spectra Physics, 20 mW). Before sintering, the pellets were heated at 500 °C to evaporate the binder. After being electrode with post-fired silver paste on two major surfaces of the disk pellets, dielectric properties were measured at room temperature as a function of the frequency (1000 Hz - 10MHz) by using a HP 4194A bridge of impedance analyzer.

3. Results and discussion

3.1. Structural analyses

3.1.1. XRD diffraction

The phase structures of the ABi₂Ta₅O₁₆ (A = Na, Tl) samples were determined by XRD patterns, as shown in Fig. 1. All characteristic peaks belong to a cubic pyrochlore and are fully indexed based on space group, Fd3m [8]. No impurity phases were detected in the XRD pattern of the three compounds NaBi₂Ta₅O₁₆ and TlBi₂Ta₅O₁₆ indicating that the sample was phase-pure pyrochlore. These results are shown in Table 1. The lattice parameter of the ABi₂Ta₅O₁₆ (A = Na, Tl) compounds determined from the Rietveld refined XRD pattern are a = 10.5013 (1) Å and 10.5144 (1) Å, respectively, for NaBi₂Ta₅O₁₆ and TlBi₂Ta₅O₁₆.

Fig. 1.

XRD patterns of NaBi₂Ta₅O₁₆ (a) and TlBi₂Ta₅O₁₆ (b).

Table 1.

X-ray powder data observed (obs.) and calculated (calc.) for NaBi₂Ta₅O₁₆ and TlBi₂Ta₅O₁₆.

	NaBi2Ta5O16			TlBi2Ta5O16
hkl	dobs. (Å)	dcalc. (Å)	I (%)	dobs. (Å)	dcalc. (Å)	I (%)
111	6.06	6.06	16	6.07	6.07	9
220	3.713	3.712	<1	3.718	3.717	1
311	3.166	3.166	11	3.170	3.170	21
222	3.0315	3.0314	100	3.0353	3.0352	100
400	2.6253	2.6253	35	2.6286	2.6286	29
331	2.4092	2.4091	1	2.4122	2.4121	1
422	2.1435	2.1436	<1	2.1463	2.1462	<1
511	2.0209	2.0210	2	2.0235	2.0235	3
440	1.8563	1.8564	41	1.8587	1.8587	38
531	1.7750	1.7750	4	1.7773	1.7772	6
442	-	1.7502	-	1.7524	1.7524	<1
620	1.6604	1.6604	<1	1.6625	1.6624	<1
533	1.6014	1.6014	1	1.6034	1.6034	4
622	1.5831	1.5831	36	1.5851	1.5851	33
444	1.5157	1.5157	10	1.5176	1.5176	9
551	1.4704	1.4705	2	1.4723	1.4723	3
731	1.3671	1.3671	3	1.3689	1.3688	4
800	1.3127	1.3127	5	1.3143	1.3143	4
733	1.2829	1.2829	<1	1.2845	1.2845	1
644	-	1.2735	-	1.2751	1.2751	<1

3.1.2. Infrared spectroscopy

The group theoretical analysis of the normal vibration of the ideal pyrochlore structure predicts 26 normal modes [9]: ^A_1g^+E_g^+2F_1g^+4F_2g^+3A_2u^+3E_u^+8F_1u+4F_2u. Only A_1g, E_g, 4F_2g and 8F_1u (one of eight F_1u modes refers to acoustic vibrations) are Raman active and infrared active respectively. The IR spectra for NaBi₂Ta₅O₁₆ and TlBi₂Ta₅O₁₆, are represented in Fig. 2 We note at first that these spectra are similar to those of others pyrochlores prepared in [10].

Fig. 2.

Infrared transmission spectra of NaBi₂Ta₅O₁₆ (a) and TlBi₂Ta₅O₁₆ (b).

The Infrared vibrational frequencies observed at around 840 and 490 cm⁻¹ attributed to the Bi–O′ stretching can be assigned on the one hand to the shorter bond distance (2.20 Å) and on the other hand to longer (2.40 Å) bond distance Bi–O’ [11]. The static disorder of Bi and O’ is associated with tendency of lone pair electron of Bi (6s²) to be stereo-chemically active [11]. The band in the region 560–700 cm⁻¹ can be ascribed to Ta–O stretching vibration. These later are usually broadened with shoulder probably due to distortions in the TaO₆ octahedra as for Cd₂Ta₂O₇ [12]. The vibrational modes at about 390–400 cm⁻¹ have been assigned to fundamental Bi–O stretching. The absorption bond at 3500 cm⁻¹ is related to O–H stretching vibrations, whereas the bond near to 1630 cm⁻¹ represents H⋯O⋯H bonding vibration of water.

3.1.3. Raman diffusion

Based on the group theoretical analysis the ideal pyrochlore structure has 26 normal vibrating modes, among these 26 modes only A_1g, E_g, 4F_2g modes are Raman-active (R). From the Raman spectra which presented Fig. 3(a) and (b), the Raman peaks are observed at 149, 200, 237, 290, 343, 448, 612 and 655 cm⁻¹ for NaBi₂Ta₅O₁₆ and at 150, 196, 231, 288, 237, 435, 477, 601 and 658 cm⁻¹ for TlBi₂Ta₅O₁₆. These observed Raman peaks can be assigned to symmetry species by referencing the previous literatures on diverse pyrochlore oxides [13, 14, 15, 16]. It is noticed that more than six Raman bands (predicted for the ideal pyrochlore structures) are observed, which implies the existence of some additional displacements of the A and O′ sites in the bismuth pyrochlore structure [14]. The sharp peak with lower frequency (150 cm⁻¹) is assigned to a O′-Bi-O′ bending vibration and corresponds to F_1u mode. Similar mode is observed in the Raman spectra of other cubic bismuth pyrochlores such as Bi_1.5Zn_0.92Nb_1.5O_6.92, Bi_1.5ZnTa_1.5O₇, and Bi_1.5MgTa_1.5O₇ [14].

Fig. 3.

Raman spectra of NaBi₂Ta₅O₁₆ (a) and TlBi₂Ta₅O₁₆ (b).

The ideal A₂B₂O₇ (or A₂B₂O₆O′) cubic pyrochlore can be described as two interpenetrating networks A₂O' (network of A₄O′ tetrahedra as seen in anticristobalite Cu₂O type structure) and B₂O₆ (3D-network of corner shared BO₆ octahedra). The variations occurred in A atoms environment and in A-O′ bond length can be the consequence of A and O′ atoms displacement from their ideal positions in the O'A₂ network. The observed peak at about 237 cm⁻¹ is attributed to F_2g mode, corresponding to the stretching vibration of O–B–O bond. The frequency at around of 343 cm⁻¹ is assigned to the stretching vibration of A-O bond, which is assigned to the (E_g ‏and F_2g) mode. The A_1g mode observed at 477 cm⁻¹ is attributed to the stretching vibration of the B–O bond in BO₆ octahedra. For many pyrochlores, the F_2g mode is observed at high frequency band around 600 cm⁻¹, for examples, at 598 cm⁻¹ for Er₂Mn₂O₇ [17], at 590 cm⁻¹ for Yb₂Ti₂O₇ [13], and at 590 cm⁻¹ for La₂Zr₂O₇ [18]. In this work, the frequency band assigned to F_2g mode is observed in the range 601–612 cm⁻¹, while the A_1g mode at 655 cm⁻¹ corresponds to the breathing mode of the O octahedra [19, 20]. All the observed Raman modes can be attributed to the relaxation of the selection rules and they show a reduction in symmetry of these samples due to the displacement of the bismuth ions in Bi-based pyrochlore.

3.2. Densification

3.2.1. Raw powder

The morphology of the raw powder of NaBi₂Ta₅O₁₆ and TlBi₂Ta₅O₁₆ obtained by solid state reaction is illustrated in Fig. 4 (a) and (b). It consists of grain sizes of around 1μm. we note also in the raw powder the presence of agglomerates, large grains surrounded by smaller ones. This phenomenon seems to be related to a pre-sintering generated by the method of preparation.

Fig. 4.

SEM images powders of NaBi₂Ta₅O₁₆ (a) and TlBi₂Ta₅O₁₆ (b).

3.2.2. Binder influence

Different samples from the powder have been cold-pressed without adding organic binder did not make good pellets of good mechanical strength. In order to improve the raw quality of the pellets, an addition of binder consisting of a mixture of ultrapure water, polyvinyl alcohol (PVA 4/125) and polyethylene glycol (PEG 1500) has been added to the powder.

3.2.3. Pressure influence

Various powder samples of NaBi₂NbTa₅O₁₆ and TlBi₂Ta₅O₁₆ compounds were pressed into pellets between 50 and 400 MPa giving density before sintering between 60% and 70%. By using the same thermal cycle, pellets were sintered and give density between 70% and 92%. Fig. 5 (a) and (b) shows the typical evolution of density versus pressure before and after sintering (a) for NaBi₂NbTa₅O₁₆ and (b) for TlBi₂NbTa₅O₁₆. We can notice that on the one hand the density in raw pellets increases regularly with pressure and on the other hand the density of the sintered materials showed the highest density located around 200 MPa. Density of sintered materials did not increase any further at higher pressing, above 200 MPa. This latter was retained as optimal pressure, so all powders were uniaxially pressed at 200 MPa.

Fig. 5.

Pressure versus density before and after sintering for NaBi₂NbTa₅O₁₆ (a) and TlBi₂Ta₅O₁₆ (b).

3.2.4. Elimination of binder

Before sintering, it is necessary to begin with the first step of elimination of binder. According to commercial data on the thermal characteristics of PVA and PEG, we started the sintering process at 500 °C with low heating rate to burn organic matter without cracking pellets [21, 22]. For this reason, the choice of the heating rate is set at 5 °C/min with holding at 500 °C for 1 hour to ensure complete elimination of binder.

The density percentage D (%) after sintering was carried out using the following formula:

$$D=\frac{{\rho }_{exp}}{{\rho }_{th}}$$

The theoretical density (ρ_th) was calculated from the diffraction patterns applying the formula:

$${\rho }_{th}=\frac{Z\times M}{N\times V}$$

Where:

• •M: Molar mass of sample
• •Z: Number of formula units
• •N: Avogadro number
• •V: Volume of the cell

The experimental density (ρ_Exp) was determined from diameter and thickness of pellets applying the formula:

$${\rho }_{exp}=\frac{m_p}{\pi \times {\left(\frac{\varnothing }{2}\right)}^2\times e}$$

Where:

• •m_p: Mass of pellet
• •ϕ: Diameter of pellet
• •e: Thickness of pellet

3.2.5. Sintering temperature influence

After using heating rate at (5 °C/min), different attempt to determine sintering temperatures were applied to the pellets corresponding to NaBi₂Ta₅O₁₆ and TlBi₂Ta₅O₁₆. We note that between 1000 and 1400 °C, the density rises with temperatures and passes for both samples from 70% to 90% for NaBi₂Ta₅O₁₆ and from 73% to 92% for TlBi₂Ta₅O₁₆ but in parallel with this increase in density; we noted a slight loss of mass observed and became significant at 1400 °C. So, we have to find exactly optimal sintering temperatures. We have used dilatometry curves which consist to measure the evolution of ΔL/Lo as a function of temperature Fig. 6 (a) and (b).

Fig. 6.

Dilatometry curves of NaBi₂NbTa₅O₁₆ (a) and TlBi₂Ta₅O₁₆ (b).

Analysis of these curves leads to distinguish two regions:

• -The first region, between room temperature and 1000 °C, characterized by a weak variation of ΔL/Lo (no shrinkage).
• -The second region, between 1000 and 1400 °C for the two compounds indicate notable variations of ΔL/Lo characterized by the beginning and ending of shrinkage process.

During the sintering process we have found that at optimal sintering temperature, sintering reaches its maximum. Treatments carried above 1400 °C do not affect density but only contribute to increase mass losses of pellets. So, the optimal sintering temperature chosen for both pressed samples to increase densification and to decrease losses of mass, was 1330 °C and 1380 °C respectively for NaBi₂Ta₅O₁₆ and TlBi₂Ta₅O₁₆.

3.2.6. Influence of the holding time

The effects of holding time (HT) between 1330 and 1380 °C for both compounds was investigated and for one hour at optimal temperatures leads to a reasonable densification (90–92%). Longer heating times do not improve the density, but promote the sublimation of the material elements.

3.2.7. Thermal cycle

Given the influence of the parameters studied above sintering heating, heating rate and heating time, we performed the thermal cycle shown in Fig. 7.

Fig. 7.

Sintering thermal cycle.

A first step for one hour at 500 °C allowing the total elimination of the binder used and a second step also for one hour at sintering temperature to reach good density of pellets. The characteristics of the thermal sintering cycles used are summarized in Table 2.

Table 2.

Characteristics of the thermal sintering adopted for ABi₂MTa₅O₁₆ (A = Na, Tl) compounds.

Compounds	Temperature (°C)	Holding time (h)	Densification (%)
NaBi2Ta5O16	1330	1	90
TlBi2Ta5O16	1380	1	92

The morphological examination under the scanning electron microscope (SEM) of the sintered pellets, reveals the presence of relatively large grains (∼5 μm) showing angles of 60–120°, indicating a good densification of the materials (Fig. 8 (a) and (b)).

Fig. 8.

SEM images of NaBi₂NbTa₅O₁₆ (a) and TlBi₂Ta₅O₁₆ (b) sintered at about 1350 °C.

3.3. Dielectric properties

Fig. 9 and Fig. 10 show frequency and temperature dependence of dielectric properties (dielectric constant and dielectric loss tanδ) for NaBi₂Ta₅O₁₆ (a) and for TlBi₂Ta₅O₁₆ (b) compounds. The examination of these curves shows that the dielectric constants at low frequencies from 1 kHz to 10 MHz exhibit a frequency independent characteristic and dielectric constants values are in the range of 107–110. Table 3 shows the dielectric constant temperature coefficients.

Fig. 9.

Frequency dependence of dielectric constant vs. temperature of NaBi₂Ta₅O₁₆ (a) and TlBi₂Ta₅O₁₆ (b).

Fig. 10.

Frequency dependence of dielectric losses vs. temperature of NaBi₂Ta₅O₁₆ (a) and TlBi₂Ta₅O₁₆ (b).

Table 3.

Dielectrics characteristics of ABi₂MTa₅O₁₆ (A = Na, Tl) compared with others in the literature.

Compound	Relative density (%)	ε	tgδ	τε = Δε/εΔT.106 (K−1) (ppm/°C)	Reference
Bi1.5Nb1.5Zn0.92O6.92	94	178	-	−647	[23]
Bi3.5Mg1.8Ta2.7O13.8	-	84	0.001	−328	[9]
RbBi2Ta5O16	-	70	0.266	-	[24]
NaBi2Ta5O16	90	110	0.284	−1327	(This work)
TlBi2Ta5O16	92	107	0.045	−953	(This work)

3.3.1. Raw powder

We have mentioned in an earlier work [8] that the plots of dielectric loss for NaBi₂Ta₅O₁₆ exhibit maximum of tanδ which shifted towards higher temperature when frequency increases, in this paper we continue the explication of this phenomenon.

3.3.2. Temperature dependence of dipolar polarization in NaBi₂Ta₅O₁₆

The analysis of the dielectric losses (tgδ) of the compound NaBi₂Ta₅O₁₆ presented in Fig. 10 shows the presence of a maximum which varies with temperature and with frequency. The maximum tgδ value shifts, due to increasing temperature from, to lower values. This phenomenon is related to dielectric relaxations and is considered as intrinsic property for some materials [26, 27].

In Table 4 we have reported values of temperatures T_max and of frequencies F_max corresponding to the maximum tgδ for the compound NaBi₂Ta₅O₁₆. This phenomenon varies with reciprocal of measuring temperature obeying Arrhenius relation, given by: F_max° = °F_oExp (-Ea/KT_max). The slope of the log (F_max) vs. f (1/T_max) (K⁻¹) plot in Fig. 11 will be equal to (-Ea/k) as given in Arrhenius relation. Relaxation activation energies were determined due to experimental measurement of the dielectric loss peaks as a function of temperature.

Table 4.

Frequency and temperature corresponding to the maximum of tgδ for NaBi₂Ta₅O₁₆.

Fmax.(Hz).10−4	1	5	10	50	100	200	500	1000
Tmax. (K)	294	338	358	414	442	470	506	546

Fig. 11.

Frequency Logarithmic temperature dependence of the frequency location at the dielectric losses peak for compound NaBi₂Ta₅O₁₆ heated from 25 °C to 500 °C.

From this plot we can calculate the activation energy corresponding to the relaxation phenomenon: Ea = (0.16 ± 0.01) eV.

There are several other important parameters which can be also obtained from dielectric data. τ_o, relaxation time can be calculated from the data such as shown in figure 13. Values of 10⁻¹² to 10⁻¹⁴s are indicative that the relaxation process is ionic in origin. However, very low values of τ_o are suggestive that the relaxation process is electronic space charge polarization [25, 26, 27]. In our case, τ_o = 0.6 10⁻¹³ s (τ_o = 10⁻¹² s for the pyrochlore phase Pb_1,83Nb₁,₇₂Mg_0,29O_6,39).

4. Conclusions

Defect pyrochlore oxide powders NaBi₂Ta₅O₁₆ (A = Na, Tl) have been prepared by the solid state method and characterized by powder XRD and infrared and Raman spectroscopies. Both materials crystallize in the cubic crystal lattice with Fd3m space group. The infrared and Raman bands observed for these compositions are consistent with the Raman spectra expected for defect pyrochlores, ABB’O_7-x. The morphology of sintering ceramics indicates a good densification of the materials. The dielectric constants of the samples are in the range of 107–110 together with a negative temperature coefficient of permittivity. Dielectric losses (tgδ) of the compound NaBi₂Ta₅O₁₆ presented maximum versus temperature and frequency, this phenomenon is related to dielectric relaxations. The activation energy Ea and relaxation time τ_o corresponding to this phenomenon were found to be 0.16 eV and 0.6 10⁻¹³ s, respectively. Such dielectric and electrical properties make the compound NaBi₂Ta₅O₁₆ attractive in the applications of ceramic capacitor and resonators.

Declarations

Author contribution statement

Omar Ait Sidi Ahmed: Conceived and designed the experiments; Analyzed and interpreted the data; Wrote the paper.

Abdeslam Chagraoui, Abdennajib Moussaoui, Sylvie Villain, Abdelmjid Tairi, Lamia Bourja, Hajar Ait Oulahyane: Analyzed and interpreted the data.

Bouchaib Manoun: Analyzed and interpreted the data; Contributed reagents, materials, analysis tools or data.

Funding statement

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Competing interest statement

The authors declare no conflict of interest.

Additional information

No additional information is available for this paper.