(Na, Zr) and (Ca, Zr) Phosphate-Molybdates and Phosphate-Tungstates: I–Synthesis, Sintering and Characterization

Submicron-grade powders of Na1-xZr2(PO4)3-x(XO4)x compounds (hereafter referred to as NZP) and Ca1-xZr2(PO4)3-x(XO4)x compounds (hereafter, CZP), X = Mo, W (0 ≤ x ≤ 0.5) were obtained by sol-gel synthesis. The compounds obtained were studied by X-ray diffraction phase analysis and electron microscopy. An increase in the W or Mo contents was shown to result in an increase in the unit cell volume of the NZP and CZP crystal lattices and in a decrease in the coherent scattering region sizes. Thermal expansion behavior at high temperatures of synthesized NZP and CZP compounds has been investigated. The dependencies of the parameters a and c on the heating temperature, as well as the temperature dependence of the crystal lattice unit cell volume V in the range from the room temperature up to 800 °C, were obtained. The dependencies of the average thermal expansion coefficient (αav) and of the volume coefficient (β) on the W and Mo contents in the compositions of NZP and CZP compounds were studied. Ceramics Na1-xZr2(PO4)3-x(XO4)x with relatively high density (more than 97.5%) were produced by spark plasma sintering (SPS). The increase in the W or Mo contents in the ceramics leads to an increase in the relative density of NZP and to a decrease of the optimum sintering temperature. The mean grain size in the NZP ceramics decreases with increasing W or Mo contents. The study of strength characteristics has revealed that the hardness of the NZP ceramics is greater than 5 GPa, and that the minimum fracture toughness factor was 1 MPa·m1/2.


Introduction
In this paper, the inorganic compounds with the NaZr 2 (PO 4 ) 3 (NZP-type) structure are studied extensively in connection with the prospects of application for solving the problems of immobilization of the highly active components of radioactive waste (HLW) [1][2][3][4]. In particular, the Mo-containing fractions of the radioactive waste solidify as inclusions in glass, together with other nuclides, for example, the wedge-shaped glass ceramics in the borosilicate glasses [5][6][7]. When more than 10 wt.% of Mo is incorporated into a multicomponent highly radioactive waste, Mo can form powellite-a mineral of Ca molybdate, CaMoO 4 . This affects the chemical resistance of the glasses, adversely affecting the solubility of molybdates [7]. A similar situation takes place with W. tages of SPS allow a significant expansion of applications in the development of ceramic matrices for the immobilization of high-level waste (HLW) and in the transmutation of minor actinides [4,27,[51][52][53][54][55][56].
This work is devoted to the synthesis and investigation of properties of the NZPframework orthophosphates with partial substitution of the P ions by Mo and W and various cations in the structure voids. These ceramics may potentially be applied to the immobilization of the Mo-and W-containing fractions of radioactive waste.

Materials and Methods
Solid solutions of the Na 1-x Zr 2 (PO 4 ) 3-x (XO 4 ) x and Ca 1-x Zr 2 (PO 4 ) 3-x (XO 4 ) x types, where X = Mo, W, and x = 0.1, 0.2, 0.3, 0.4, 0.5, were chosen for investigation. The compounds were synthesized by the sol-gel method. A sample of Zr oxychloride was dissolved in a mixture of solutions of ammonium molybdate or tungstate and Na or Ca nitrate. A 1 M solution of ammonium dihydrogen phosphate was added dropwise to the resulting solution in the course of continuous stirring. As a result, gel-like precipitates were formed. The reagents were mixed in the stoichiometric proportions. For more complete precipitation, a saltingout agent (ethyl alcohol) was added. After brief stirring, the gel was dried at 90 • C for 1 day. The gel-like sediment was dried in a drying cabinet and then was dispersed in an agate mortar, heated up to 600 • C, dispersed again, and heated up to 800 • C. The resulting powders were investigated by X-ray diffraction (XRD) between the stages of annealing.
The ceramics were sintered from the obtained powders by SPS using Dr. Sinter™ model SPS-625 (SPS SYNTEX ® , Kanagawa, Japan). Prior to sintering, the powders were ground in an agate mortar for 5 min, and pressed into a graphite mold at a pressure of 70 MPa. No binders were used for the preparation of the initial powder green body. The binderless powders were placed in a graphite mold with an inner diameter of 12.8 mm and an outer diameter of 30 mm, and heated up by passing millisecond pulses of high-power direct electric current (up to 5 kA). Sintering was performed in a vacuum. The effective temperature (T 1 ) was measured with a Chino IR-AHS2 pyrometer (Chino Corporation, Tokyo, Japan) focused on the graphite mold surface. Based on previous studies, and on the comparison of the infrared pyrometer readings (T 1 in Figure 1) to those of a thermocouple attached to the ceramic specimen surface, the values of T1 were used to calculate the actual specimen temperature (T) using the following empirical relation: T = A·T 1 − B, where A and B are empirical constants which depend on the heating rate V h (Figure 1 The method of comparing temperatures T and T 1 is described in [60]. To improve the contact of the powder sample with the graphite mold and to compensate for the difference in the thermal expansion coefficients of graphite and the ceramics, a graphite foil was placed inside the graphite mold. A two-stage heating regime was used-stage A: heating up to 570-580 • C with the heating rate V h1 = 100 • C/min, and stage B: heating up to the sintering temperature T s with the heating rate V h2 = 50 • C/min. Holding at the sintering temperature T s was absent (t s = 0 s). The average uniaxial stress applied was~65 MPa. In order to avoid cracking of the ceramic specimen, multistage pressure regulation was applied during sintering: (stage I) a gradual increasing of the uniaxial stress P from 35-38 MPa up to 64-65 MPa during the rapid heating stage up to 570-580 • C, (stage II) holding at 64-66 MPa at 580-840 • C, (stage III) reducing the pressure to 60 MPa at 900 • C, and (stage IV) gradually increasing the pressure again to~65 MPa to the moment of the end of shrinkage. The typical scheme of the stress and temperature regulation during SPS is presented in Figure 1. The samples were cooled down together with the setup (stage C, Figure 1). As a result of sintering, dense ceramic specimens without macrodefects were obtained. holding at 64-66 MPa at 580-840 °C, (stage III) reducing the pressure to 60 MPa at 90 and (stage IV) gradually increasing the pressure again to ~65 MPa to the moment o end of shrinkage. The typical scheme of the stress and temperature regulation during is presented in Figure 1. The samples were cooled down together with the setup (sta Figure 1). As a result of sintering, dense ceramic specimens without macrodefects obtained. The temperature curves of the effective shrinkage (Leff) were measured during si ing using the built-in dilatometer of the SPS-625. To account for thermal expansion o mold (L0), a separate series of experiments on heating up an empty mold was carried True shrinkage was calculated as L = Leff -L0 ( Figure 2). Using the dependencies L(T temperature dependencies of the shrinkage rate were calculated: S = ∆L/∆t. The optimal sintering temperature Ts for each ceramic was selected on the ba analysis of the dependencies Leff(T). The samples were heated up to the temperature responding to the end of the stage of intensive powder shrinkage. Due to the contribu of the thermal expansion L0(T) on the Leff(T) curve, stage III arose in the temperatur pendence of true shrinkage L(T), within which true shrinkage of the ceramics al didn't change ( Figure 2). The change in temperature during stage III was ~50 °C, c sponding to the time interval of ~1 min at the heating rate Vh2 = 50 °C/min. In the co of heating, we tried to minimize the duration of stage III, since intensive grain gr takes place in the ceramics at high temperatures (T > Ts). The coarse-grain ceramics very low resistance against thermal shock, and the samples are often destroyed du cooling down. Additionally, the sintering at high temperatures can lead to p The temperature curves of the effective shrinkage (L eff ) were measured during sintering using the built-in dilatometer of the SPS-625. To account for thermal expansion of the mold (L 0 ), a separate series of experiments on heating up an empty mold was carried out. True shrinkage was calculated as L = L eff -L 0 ( Figure 2). Using the dependencies L(T), the temperature dependencies of the shrinkage rate were calculated: S = ∆L/∆t.
The optimal sintering temperature T s for each ceramic was selected on the basis of analysis of the dependencies L eff (T). The samples were heated up to the temperature corresponding to the end of the stage of intensive powder shrinkage. Due to the contribution of the thermal expansion L 0 (T) on the L eff (T) curve, stage III arose in the temperature dependence of true shrinkage L(T), within which true shrinkage of the ceramics almost didn't change ( Figure 2). The change in temperature during stage III was~50 • C, corresponding to the time interval of~1 min at the heating rate V h2 = 50 • C/min. In the course of heating, we tried to minimize the duration of stage III, since intensive grain growth takes place in the ceramics at high temperatures (T > T s ). The coarse-grain ceramics have very low resistance against thermal shock, and the samples are often destroyed during cooling down. Additionally, the sintering at high temperatures can lead to partial dissociation of the elements (W, Mo) from the ceramic surfaces and to more intensive saturation of the ceramic sample surfaces with the mold (see [50,[61][62][63][64][65][66][67]).
The XRD phase analysis of the ceramics was performed using Bruker ® D8 Discover™ X-ray diffractometer (Bruker Corp., Billerica, MA, USA) in symmetric Bragg-Brentano geometry. An X-ray tube with a Cu cathode (CuK α radiation) was used. The goniometer radius was 30 cm. All experiments were performed in the same conditions. The specimens, comprising cylindrical tablets of 12 mm diameter and 3 mm height, were placed on the goniometer table. At the beginning of each experiment, adjustment of the specimen height by the beam splitting method was performed with 0.1 mm slits both on the primary beam and in front of the detector. When acquiring the XRD curves, the primary beam size was limited in the equatorial plane by a 0.6 mm slit, and in the axial plane by a 12 mm slit. The XRD curves were acquired in the θ/2θ-scan mode in the angle range from 10 • to 70 • in 2θ.
The step in angle was 0.1 • . LynxEYE linear position-sensitive detector with 192 independent acquisition channels and a 2 • angle aperture in 2θ was used. The dwell time was 2 s, which was equivalent to the effective dwell time about 40 s due to the use of the position-sensitive detector. The thermal expansion in the temperature range 25-800 • C was investigated using Panalytical ® X'Pert Pro™ high-temperature X-ray diffractometer (Malven PANAlytical B.V., Almelo, The Netherlands) with Anton Paar ® HTK-1200N high-temperature chamber (Anton Paar GmnH, Graz, Austria) with a step of 100 • C. The unit cell parameters were calculated with the Rietveld method using Topas 4.2 software. aterials 2023, 16, x FOR PEER REVIEW dissociation of the elements (W, Mo) from the ceramic surfaces and to mor uration of the ceramic sample surfaces with the mold (see [50,[61][62][63][64][65][66][67]). The XRD phase analysis of the ceramics was performed using Bruker  X-ray diffractometer (Bruker Corp., Billerica, USA) in symmetric Bragg-Br try. An X-ray tube with a Cu cathode (CuKα radiation) was used. The gon was 30 cm. All experiments were performed in the same conditions. The sp prising cylindrical tablets of 12 mm diameter and 3 mm height, were place ometer table. At the beginning of each experiment, adjustment of the spec the beam splitting method was performed with 0.1 mm slits both on the and in front of the detector. When acquiring the XRD curves, the primary limited in the equatorial plane by a 0.6 mm slit, and in the axial plane by a XRD curves were acquired in the θ/2θ-scan mode in the angle range from 1 The step in angle was 0.1°. LynxEYE linear position-sensitive detector with ent acquisition channels and a 2° angle aperture in 2θ was used. The dwe which was equivalent to the effective dwell time about 40 s due to the use sensitive detector. The thermal expansion in the temperature range 25-800 gated using Panalytical  X'Pert Pro high-temperature X-ray diffractometer Alytical B.V., Almelo, The Netherlands) with Anton Paar  HTK-1200N hig chamber (Anton Paar GmnH, Graz, Austria) with a step of 100 °C. The unit c The microstructure of the powders and ceramics was investigated using Jeol ® JSM-6490 scanning electron microscope (SEM, Jeol Ltd., Tokyo, Japan) and Jeol JEM-2100 transmission electron microscope (TEM, Jeol Ltd., Tokyo, Japan). The mean particle size (D P ), the size of agglomerates (D A ), and grain size (d) were calculated by the section method using GoodGrains 2.0 software (UNN, Nizhny Novgorod, Russia). To determine the mean particle sizes, investigations of the sizes of all particles of the powders presented in the images obtained by TEM were conducted. When determining the mean grain sizes, we measured 100 or more grains presented in the images obtained by SEM. When determining D P and d, three photographs (images) or more were analyzed. The mean sizes were calculated with GoodGrains 2.0 software in the automated mode, on the base of analysis of the histograms of the size distributions of the particles/grains. The uncertainties of determining D P and d were calculated as the standard deviations.
The density of the obtained ceramics was measured by hydrostatic weighing in distilled water using Sartorius ® CPA 225D balance (Sartorius AG, Göttingen, Germany). The uncertainty of the density measurement was ±0.001 g/cm 3 . The theoretical density of the ceramics (ρ th ) was calculated on the basis of the XRD investigations. The microhardness (H v ) of the ceramics was measured using Duramin ® Struers-5 hardness tester (Struers LLC, Cleveland, OH, USA). The load was 200 g. The minimum fracture toughness Materials 2023, 16, 990 6 of 24 factor (K IC ) was calculated according to the Palmquist method from the length of the largest radial crack forming by the Vickers pyramid during the indentation of the ceramics: where P is the indenter load [g], c is the mean distance from the imprint center to the end of the crack, and E is the elastic modulus of the material [GPa]. When calculating the magnitude of K IC , we used the lengths of 10 or more of the longest cracks from 10 different indents formed on the ceramic sample surfaces when loading by Vickers pyramid.

Synthesis and Characterization of the Powders
The powder samples were light-colored. According to the SEM results, the submicrongrained powders were agglomerated; there were some large agglomerations up to 10-20 µm in size in the Ca 1-x Zr 2 (PO 4 ) 3-x (XO 4 ) x powders (denoted as "CZP" in work) ( Figure 3a tilled water using Sartorius  CPA 225D balance (Sartorius AG, Göttingen, Germany). The uncertainty of the density measurement was ± 0.001 g/cm 3 . The theoretical density of the ceramics (ρth) was calculated on the basis of the XRD investigations. The microhardness (Hv) of the ceramics was measured using Duramin  Struers-5 hardness tester (Struers LLC, Cleveland, USA). The load was 200 g. The minimum fracture toughness factor (KIC) was calculated according to the Palmquist method from the length of the largest radial crack forming by the Vickers pyramid during the indentation of the ceramics: where P is the indenter load [g], c is the mean distance from the imprint center to the end of the crack, and E is the elastic modulus of the material [GPa]. When calculating the magnitude of KIC, we used the lengths of 10 or more of the longest cracks from 10 different indents formed on the ceramic sample surfaces when loading by Vickers pyramid.

Synthesis and Characterization of the Powders
The powder samples were light-colored. According to the SEM results, the submicron-grained powders were agglomerated; there were some large agglomerations up to 10-20 μm in size in the Ca1-xZr2(PO4)3-x(XO4)x powders (denoted as "CZP" in work) ( Figure  3a) and ~50 μm in size in the Na1-xZr2(PO4)3-x(XO4)x powders (denoted as "NZP") ( Figure  3c,e) after grinding in the agate mortar. No considerable effect of the W and Mo contents on the granulometric composition and morphology of the powders was observed.  Figure  5). The Ca1-xZr2(PO4)3-x(MoO4)x powders were large-grained and agglomerated strongly. The individual particle sizes ranged from ~50-100 nm to ~300 nm and didn't depend on the Mo content ( Figure 4). The Na0.5Zr2(PO4)2.5(XO4)0.5 powders were more finely dispersed and agglomerated. The particle sizes ranged from ~20-30 nm to ~50-100 nm (  (e) (f)    According to the XRD phase analysis data ( Figures 6 and 7), the compounds under study crystallized in the NaZr 2 (PO 4 ) 3 type structure, hexagonal syngony, space group R3c (analog of NaZr 2 (PO 4 ) 3 [37]) for the Na-containing samples and space group R3 (analog of Ca 0.5 NaZr 2 (PO 4 ) 3 [68]) for the Ca-containing ones. The XRD patterns of the samples with Ca at x > 0.2 for phosphate-molybdates and x > 0.3 for phosphate-tungstates contained a significant number of auxiliary (secondary) phase reflections. For this reason, these samples were not studied further.  According to the XRD phase analysis data (Figures 6 and 7), the compounds under study crystallized in the NaZr2(PO4)3 type structure, hexagonal syngony, space group R3 � c (analog of NaZr2(PO4)3 [37]) for the Na-containing samples and space group R3 � (analog of Ca0.5NaZr2(PO4)3 [68]) for the Ca-containing ones. The XRD patterns of the samples with Ca at x > 0.2 for phosphate-molybdates and x > 0.3 for phosphate-tungstates contained a significant number of auxiliary (secondary) phase reflections. For this reason, these samples were not studied further.
The particle (coherent scattering region, CSR) sizes DXRD were estimated using the Scherer equation: where K = 0.9 is the Scherer constant, λ is the wavelength of the X-ray radiation, β is the full width at half maximum of the XRD peak (in radians), and θmax is the diffraction angle study crystallized in the NaZr2(PO4)3 type structure, hexagonal syngony, space group R3 � c (analog of NaZr2(PO4)3 [37]) for the Na-containing samples and space group R3 � (analog of Ca0.5NaZr2(PO4)3 [68]) for the Ca-containing ones. The XRD patterns of the samples with Ca at x > 0.2 for phosphate-molybdates and x > 0.3 for phosphate-tungstates contained a significant number of auxiliary (secondary) phase reflections. For this reason, these samples were not studied further.  The particle (coherent scattering region, CSR) sizes DXRD were estimated using the Scherer equation: where K = 0.9 is the Scherer constant, λ is the wavelength of the X-ray radiation, β is the full width at half maximum of the XRD peak (in radians), and θmax is the diffraction angle The particle (coherent scattering region, CSR) sizes D XRD were estimated using the Scherer equation: where K = 0.9 is the Scherer constant, λ is the wavelength of the X-ray radiation, β is the full width at half maximum of the XRD peak (in radians), and θ max is the diffraction angle corresponding to the (116) XRD peak maximum. The analysis of the XRD results demonstrated that the particle sizes D XRD of the Na 1- strated that the particle sizes DXRD of the Na1-xZr2(PO4)3-x(MoO4)x powders decreas notonously from 61 nm down to 19 nm with increasing Mo content from 0.1 up to from 44 nm down to 23 nm with increasing W content from 0.1 up to 0.5 (Figure considerable effects of W and Mo content on the DXRD of the Ca1-xZr2(PO4)3-x(XO4)x p were observed; the mean particle sizes were ~60 nm for the W-containing powd ~47 nm for the Ca1-xZr2(PO4)3-x(MoO4)x ones. From the analysis of the XRD, TEM, and SEM results comes the conclusion th tial sintering of several nanoparticles into submicron ones takes place in the co synthesis, at the last stage of high-temperature annealing (800 °C, 20 hrs). This con was confirmed by comparing the results of investigations of the mean particle s XRD, TEM, and SEM. The analysis of the XRD results shows that the mean CSR the NZP and CZP compounds were several tens of nanometers (DXRD < 60 nm). The of TEM and SEM show the particles to be agglomerated strongly; the agglomer submicron and micron sizes consist of separate nanoparticles (Figures 2-4).
The unit cell parameters for the solid solutions were determined from the XR (the results are presented in Table 1). Table 1. Crystallographic characteristics of the phosphates Na1-xZr2(PO4)3-x(XO4)x (den "NZP") and Ca0.5(1-x)Zr2(PO4)3-x(XO4)x (denoted as "CZP").  From the analysis of the XRD, TEM, and SEM results comes the conclusion that partial sintering of several nanoparticles into submicron ones takes place in the course of synthesis, at the last stage of high-temperature annealing (800 • C, 20 h). This conclusion was confirmed by comparing the results of investigations of the mean particle sizes by XRD, TEM, and SEM. The analysis of the XRD results shows that the mean CSR sizes in the NZP and CZP compounds were several tens of nanometers (D XRD < 60 nm). The results of TEM and SEM show the particles to be agglomerated strongly; the agglomerates of submicron and micron sizes consist of separate nanoparticles (Figures 2-4).
The unit cell parameters for the solid solutions were determined from the XRD data (the results are presented in Table 1). Table 1. Crystallographic characteristics of the phosphates Na 1-x Zr 2 (PO 4 ) 3-x (XO 4 ) x (denoted as "NZP") and Ca 0.5(1-x) Zr 2 (PO 4 ) 3-x (XO 4 ) x (denoted as "CZP"). The composition dependences of the unit cell parameters for the phosphate-molybdates and phosphate-tungstates are presented in Figure 9. Substitution of phosphate anions (PO 4 ) 3− (R (P 5+ ) = 0.17 Å) by larger molybdate anions (MoO 4 ) 2− (R (Mo 6+ ) = 0.41 Å) and tungstate ions (WO 4 ) 2− (R(W 6+ ) = 0.44 Å) leads to an increase in the unit cell parameters. One can see from Figure 9 that increasing the W content in the Na 1-x Zr 2 (PO 4 ) 3-x (XO 4 ) x compounds results in a greater increase in the lattice parameter c and the unit cell volume V than increasing the Mo content. No essential differences in the values of the lattice parameter a for the phosphate-molybdates and phosphate-tungstates Na 1-x Zr 2 (PO 4 ) 3-x (XO 4 ) x were observed. For the Ca-phosphate-tungstates Ca 1-x Zr 2 (PO 4 ) 3-x (WO 4 ) x , no dependencies of the unit cell parameters a and c, nor the unit cell volume V, on the W content were found. In the case of Ca 1-x Zr 2 (PO 4 ) 3-x (MoO 4 ) x , the increase in the Mo content up to x = 0.2 resulted in a linear increase in all unit cell parameters and, hence, in a linear increase in the unit cell volume V.
The composition dependences of the unit cell parameters for the phosphate-molybdates and phosphate-tungstates are presented in Figure 9. Substitution of phosphate anions (PO4) 3-(R (P 5+ ) = 0.17 Å) by larger molybdate anions (MoO4) 2-(R (Mo 6+ ) = 0.41 Å) and tungstate ions (WO4) 2-(R(W 6+ ) = 0.44 Å) leads to an increase in the unit cell parameters. One can see from Figure 9 that increasing the W content in the Na1-xZr2(PO4)3-x(XO4)x compounds results in a greater increase in the lattice parameter c and the unit cell volume V than increasing the Mo content. No essential differences in the values of the lattice parameter a for the phosphate-molybdates and phosphate-tungstates Na1-xZr2(PO4)3-x(XO4)x were observed. For the Ca-phosphate-tungstates Ca1-xZr2(PO4)3-x(WO4)x, no dependencies of the unit cell parameters a and c, nor the unit cell volume V, on the W content were found. In the case of Ca1-xZr2(PO4)3-x(MoO4)x, the increase in the Mo content up to x = 0.2 resulted in a linear increase in all unit cell parameters and, hence, in a linear increase in the unit cell volume V. To study the behavior of the obtained compounds upon heating, the XRD patterns of the samples were recorded at elevated temperatures (RT-800 • C). The XRD data were used to calculate the values of the unit cell parameters at different temperatures. The temperature dependencies of the cell parameters are shown in Figures 10 and 11. One can see from Figure 10 a minor decrease (~0.02 Å) in the unit cell parameter a and a considerable increase in the parameter c (~0.2-0.4 Å) with increasing temperature (from 20 up to 800 • C) for Na 1-x Zr 2 (PO 4 ) 3-x (XO 4 ) x . This leads to an increase in the unit cell volume V when heating. For Ca 1-x Zr 2 (PO 4 ) 3-x (XO 4 ) x , a greater increase in the parameter c was observed with a similar increase in temperature (Figure 11), which resulted in a smaller increase in the unit cell volume of Ca 1-x Zr 2 (PO 4 ) 3-x (XO 4 ) x as compared to Na 1-x Zr 2 (PO 4 ) 3-x (XO 4 ) x .
To study the behavior of the obtained compounds upon heating, the XRD patterns of the samples were recorded at elevated temperatures (RT-800 °C). The XRD data were used to calculate the values of the unit cell parameters at different temperatures. The temperature dependencies of the cell parameters are shown in Figures 10 and 11. One can see from Figure 10 a minor decrease (~0.02 Å) in the unit cell parameter a and a considerable increase in the parameter c (~0.2-0.4 Å) with increasing temperature (from 20 up to 800 °C) for Na1-xZr2(PO4)3-x(XO4)x. This leads to an increase in the unit cell volume V when heating. For Ca1-xZr2(PO4)3-x(XO4)x, a greater increase in the parameter c was observed with a similar increase in temperature (Figure 11), which resulted in a smaller increase in the unit cell volume of Ca1-xZr2(PO4)3-x(XO4)x as compared to Na1-xZr2(PO4)3-x(XO4)x.   The above dependencies were used to calculate the values of the axial (αa and αc), average (αav), and volume (β) thermal expansion coefficients (CTEs), as well as the thermal expansion anisotropy (Δα) for the phosphate-molybdates and phosphate-tungstates under study (Table 2, Figures 12 and 13). The analysis of these data shows the decreasing of αa and αc with increasing W and Mo contents in Na1-xZr2(PO4)3-x(XO4)x. At the same time, the mean value of αav changes slowly. The volume CTE (β) decreases slowly as the thermal expansion anisotropy (Δα) increases. In general, a similar character of the CTE dependencies on the composition was observed for the Ca1-xZr2(PO4)3-x(MoO4)x compounds with the increase I n the Mo content ( Figure 12). For the W-containing phosphates Ca1-xZr2(PO4)3x(WO4)x, no considerable effect of the W content on all thermal expansion parameters (αa, αc, αav, β, and Δα) was observed. Table 2. Thermal expansion coefficients [in °C −1 ] of phosphates Na1-xZr2(PO4)3-x(XO4)x (denoted as "NZP") and Ca0.5(1-x)Zr2(PO4)3-x(XO4)x (denoted as "CZP").

X
x α a ·10 6 α c ·10 6 α av ·10   One can note a trend for the thermal expansion parameters to approach zero with decreasing population of the extraframe positions in the structure of the phosphate-molybdates and phosphate-tungstates under study. Therefore, novel ceramic materials with zero CTE and high thermal shock resistance can potentially be developed on the basis of the phosphates with predefined W and/or Mo contents studied in the present work. One can note a trend for the thermal expansion parameters to approach zero with decreasing population of the extraframe positions in the structure of the phosphate-molybdates and phosphate-tungstates under study. Therefore, novel ceramic materials with zero CTE and high thermal shock resistance can potentially be developed on the basis of the phosphates with predefined W and/or Mo contents studied in the present work. One can note a trend for the thermal expansion parameters to approach zero with decreasing population of the extraframe positions in the structure of the phosphatemolybdates and phosphate-tungstates under study. Therefore, novel ceramic materials with zero CTE and high thermal shock resistance can potentially be developed on the basis of the phosphates with predefined W and/or Mo contents studied in the present work.

Sintering and Characterization of the Ceramics
The ceramic samples with high relative densities were obtained from the Na-containing compounds by SPS. The ceramic samples have no visible macro-nor microcracks. Examples of temperature curves L(T) and S(T) are shown in Figure 14. The shapes of the curves L(T) are typical for SPS of the nano-and submicron NZP-type powders [27,51,53]. The analysis of the temperature curves of the shrinkage L(T) for the phosphatemolybdates (Figure 14a,c) shows an increase in the maximum shrinkage L max and of the shrinkage rate S max with increasing Mo content from x = 0.1 up to 0.3. At the same time, an increase in the optimal sintering temperature was observed, which is manifested as the shift of the curves L(T) and S(T) towards higher heating temperatures. With a further increase in the Mo content up to x = 0.5, a shift of the curves L(T) and S(T) towards lower sintering temperatures was observed. In our opinion, such a behavior is due, first of all, to the appearance of some auxiliary phases during sintering, and to the increasing concentration of the auxiliary phases with increasing Mo concentration (see above).
In the Na 1-x Zr 2 (PO 4 ) 3-x (WO 4 ) x phosphates, a minor increase in the shrinkage and a decrease in the optimal sintering temperature corresponding to the completion of the intensive shrinkage stage with increasing W content were observed (Figure 14b). At the same time, an increase in the maximum shrinkage rate of nanopowders S max of more than three times, and a minor shift of the temperatures, during which the maximum shrinkage rates S max were manifested towards higher heating temperatures, were observed (Figure 14d).
The average full SPS process times (t sps ), including the heating times, were 13 min for the phosphate-molybdates Na 1-x Zr 2 (PO 4 ) 3-x (MoO 4 ) x and 16 min for the phosphatetungstates Na 1-x Zr 2 (PO 4 ) 3-x (WO 4 ) x . The characteristics of the sintering process (the heating rates V h , the sintering temperatures T s , and the full SPS process time t sps ), the relative densities achieved, the microhardness values, and the minimal fracture toughness factors for the ceramics obtained are presented in Table 3. One can see from Table 3 that the increasing of the W and Mo contents resulted in an increasing of the relative density of the Na 1-x Zr 2 (PO 4 ) 3-x (XO 4 ) x ceramics. Densities close to the theoretical one were ensured for almost all ceramics. In the ceramics with increased W and Mo contents, the relative densities exceeded 100% slightly. In our opinion, this result originates from the auxiliary (secondary) phases contained in the powders synthesized (see the XRD phase analysis results above). The results of XRD phase analysis of the ceramics also show the secondary phases to be present in the ceramics (Figure 15). In the W-containing ceramics, the presence of Zr 2 (WO 4 )(PO 4 ) 2 secondary phase was observed (Figure 15a), in the Mo-containing ones CaCO 3 and Al 3 O 0.34 Zr 5 , secondary phases were found (Figure 15b). One can see from Table 3 that in the NZP ceramics with high Mo and W contents, one can provide a high density at reduced sintering temperatures. We suggest the that an increase in Mo and W contents will result in a decrease in the sintering activation energy for the Na 1-x Zr 2 (PO 4 ) 3-x (XO 4 ) x ceramics, X = Mo, W (0 ≤ x ≤ 0.5). This allows sintering the ceramics with high values of x up to high densities at lower temperatures. The decrease in the sintering activation energy was manifested indirectly also in a decrease in the temperature at the beginning of the intensive shrinkage stage (Stage II) in Figure 14 at low heating temperatures. all ceramics. In the ceramics with increased W and Mo contents, the relative densities exceeded 100% slightly. In our opinion, this result originates from the auxiliary (secondary) phases contained in the powders synthesized (see the XRD phase analysis results above). The results of XRD phase analysis of the ceramics also show the secondary phases to be present in the ceramics (Figure 15). In the W-containing ceramics, the presence of Zr2(WO4)(PO4)2 secondary phase was observed (Figure 15a), in the Mo-containing ones CaCO3 and Al3O0.34Zr5, secondary phases were found (Figure 15b).  Figure 16) and phosphate-tungstates Na 1-x Zr 2 (PO 4 ) 3-x (WO 4 ) x ( Figure 17) with various Mo and W contents. One can see in Figure 16 a decrease in the mean grain size (Figure 16b,d) and an increase in the microstructure uniformity of the ceramic phosphate-molybdates (Figure 16a,c) with increasing Mo content. Note that there were some areas with abnormally large grains in the sintered ceramics with low Mo contents (x = 0.1). The sizes of these grains were~5-10 µm (Figure 16a), greater than the mean grain size of the ceramic matrix (Figure 16b) by an order of magnitude. In the ceramics with increased Mo contents (x = 0.5), the areas with abnormally large grains were almost absent (Figure 16c), whereas the mean grain sizes in the sintered ceramic phosphate-molybdates were close to 100-200 nm (Figure 16d). A similar effect has been observed when analyzing the results of SEM investigations of the ceramic phosphate-tungstates with different W contents. As one can see in Figure 17, an increase in the W content resulted in a decrease in the mean grain sizes in the ceramics down to~100-150 nm. There are few areas with abnormally large grains, the sizes of which is not greater than 2-5 µm in the ceramics with x = 0.5 (Figure 17d).
The XRD phase analysis evidenced the phase composition of the ceramics not to change after sintering ( Figure 15). A minor increase in the intensities and decrease in the widths of the XRD peaks have been observed that is typical for sintering the nano-and submicron powders, which is associated with the grain growth.
As one can see from Table 3, the increase in the Mo and W contents resulted in a minor increase in the microhardness of the ceramics up to 5.3-5.6 GPa. At the same time, the minimum fracture toughness factor almost did not change and exceeded 1 MPa·m 1/2 as a rule, which is a typical value for the mineral-like ceramics with the NZP structure [4,56].  3. The dependencies of the crystal lattice constants a and c as well as of the unit cell volume V on the heating temperature were studied by high-temperature X-ray diffraction. The dependencies of the mean thermal expansion coefficient (TEC) (α av ) and of the volume coefficient (β) on the W and Mo contents in NZP and CZP compounds were investigated. The mean values of TEC (α av ) for the Mo- The bulk of the NZP ceramic specimens obtained by SPS were characterized by high relative densities (ρ rel > 97.5%). The full time of the SPS process, including the heating and cooling-down times, is~11.5-16.5 min for the Mo-containing ceramics and~15.1-18.1 min for the W-containing ones. The values of microhardness (H v ) of the ceramics ranged from 3.6 to 5.6 GPa, and the minimal fracture toughness factor (K IC ) ranged from 0.9 to 1.6 MPa m 1/2 . The ceramics with the NZP structure have fine-grained microstructure; abnormal grain growth was observed in the ceramics with increased W contents. In the NZP ceramics with high Mo and W contents, one can provide high density at reduced sintering temperature. Institutional Review Board Statement: Not applicable.