Characterization and controlling thermal expansion of materials with kosnarite- and langbeinite-type structures

Accepted Manuscript Title: Characterization and controlling thermal expansion of materials with kosnarite- and langbeinite-type structures Authors: Vladimir I. Pet’kov, Alexander S. Shipilov, Anton S. Dmitrienko, Artemy A. Alekseev PII: DOI: Reference:

S1226-086X(17)30448-3 http://dx.doi.org/10.1016/j.jiec.2017.08.029 JIEC 3578

To appear in: Received date: Revised date: Accepted date:

14-2-2017 17-8-2017 18-8-2017

Please cite this article as: Vladimir I.Pet’kov, Alexander S.Shipilov, Anton S.Dmitrienko, Artemy A.Alekseev, Characterization and controlling thermal expansion of materials with kosnarite- and langbeinite-type structures, Journal of Industrial and Engineering Chemistryhttp://dx.doi.org/10.1016/j.jiec.2017.08.029 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Title Characterization and controlling thermal expansion of materials with kosnarite- and langbeinitetype structures

Authors Vladimir I. Pet’kov, Alexander S. Shipilov, Anton S. Dmitrienko, Artemy A. Alekseev

Affiliation Lobachevsky State University of Nizhni Novgorod, Gagarin Av., 23, 603950 Nizhni Novgorod, Russia

Graphical abstract

with controllable thermal expansion

Abstract MZr2(TO4)x(PO4)3 – x (M = Li, Na, K, Rb, Cs; T = As, V) solid solutions, NaFeZr(PO4)2SO4 and Pb2/3FeZr(PO4)7/3(SO4)2/3

with

mineral

kosnarite

structure

and

KPbMgTi(PO4)3,

K5/3MgE4/3(PO4)3 (E = Ti, Zr) with mineral langbeinite structure have been synthesized. According to the yielded results, which encompass thermal expansion data and crystallographic information about the structure of individual compounds and solid solutions, the meaningful selection of compounds with kosnarite and langbeinite structure for novel materials with controllable thermal expansion was carried out. The potassium-, rubidium-, and cesiumcontaining arsenates, arsenate-phosphates, vanadate-phosphates and Pb2/3FeZr(PO4)7/3(SO4)2/3 are low expansion materials (αav < 2∙10–6 K–1); sodium-zirconium arsenate and sodiumzirconium

and

lithium-zirconium

arsenate-phosphates,

vanadate-phosphates

and

K5/3MgZr4/3(PO4)3 have intermediate thermal expansion (3∙10–6 K–1 < αav < 7∙10–6 K–1); and lithium-zirconium arsenate, KPbMgTi(PO4)3, K5/3MgTi4/3(PO4)3 are the high expansion material (αav > 7∙10–6 K–1). The present results demonstrate that change of the size of alkali metal cation, anion substitution and varying of solid solution composition can produce kosnarite ceramics with controlled linear thermal expansion coefficients and extremely low thermal expansion anisotropy or langbeinite ceramics with isotropic expansion.

Keywords: Thermal expansion, X-ray difractometry, Ceramics, NZP family, Langbeinite-type materials, Near-zero expansion

Introduction Materials – structural analogues with the mineral kosnarite, КZr2(PO4)3, including a large family of NASICON, Na1 + xZr2SixP3 – xO12 solid ionic conductors, and NaZr2(PO4)3 (NZP), have attracted researchers attention owing to their high ionic conductivity, low thermal expansion, and high resistance to damage under natural corrosion conditions [1−4]. Thermophysical properties

of NZP materials can be tailored by proper ionic substitutions and materials processing techniques [2,5−7]. These properties allow them to be used in the fabrication and manufacturing of various products: thermal shock resistant ceramics for various types of engines (diesel, gasturbine), power production (catalytic combustor substrate, hot gas filters, heat exchangers, burner nozzles), metallurgical operations (molten metal handling, braze and carburizing fixtures, space research (substrates for optical devices), ceramic electrolytes, and ecologically sustainable host matrices for reliable toxic/radioactive waste immobilization [8−10]. Certain NZP materials have very low thermal expansion and expansion anisotropy and, in turn, good mechanical properties [11]. The NZP structure has anisotropic thermal expansion which depends on chemical bond strength distribution in it [11–15]. NZP framework is made up of corner-sharing LO6-octahedra (L = Nb5+, Sb5+, Ta5+, Ti4+, Zr4+, Hf4+, Ge4+, Sn4+, Mo4+, U4+, Np4+, Pu4+, Sc3+, Y3+, Ln3+ , V3+, Cr3+, Fe3+, Al3+, Ga3+, In3+, Ti3+, Mg2+, Mn2+, Cu2+, Co2+, Ni2+, Zn2+, Na+ and others) and PO4tetrahedra. The framework also contains voids of appropriate size for accommodating a variety of cations in the oxidation states from +1 to +4, predominantly with a small charge. The vast majority of compounds with NZP structure contain phosphorus as an anion-forming element. On account of high strength of the bonds in the PO4 tetrahedra and ZrO6 octahedra, thermal expansion of that structure is attributed mainly to the ions residing on extra framework sites. Models of thermal expansion of NZP phosphates relate the thermal peculiarities of such phases to the cooperative rotations of the coordination polyhedra and/or to transverse thermal movement of the oxygen atoms at their corners [16,17]. Concerning cations with smaller size than Na, such as Li, the key structural alteration of Li-NZP phases is related to the displacement of extraframework Li cations within the unit cell, bringing about variations of the thermal expansion parameters with growth of temperature [18]. Solid solutions based on the NZP compounds concerned are of considerable practical interest, because the diversity of possible combinations of cations and anions allows one to change or gradually vary their physical parameters, including their thermal expansion coefficients, with the framework remaining stable [19,20]. Thermal

expansion of the NZP materials having doubly centered hexagonal lattices can be characterized by their linear thermal expansion coefficients (LTECs) measured along (αс) and across (αа) their major threefold rotation axis and by their average LTEC: αav = (2αa + αс)/3. Heating leads to expansion/compression of their structure in different crystallographic directions. Thermal expansion anisotropy can be quantified by |αa – αс|. As a rule, αа and αс coefficients of the materials under consideration are rather large but have opposite signs, so their average LTEC is nearly zero. However, when it comes to thermal expansion anisotropy its magnitude could well reach considerable levels [2, 11]. The last factor places a certain limitation on the use of these materials because they may experience cracking in response to sharp changes in thermal load. At the same time, for certain NZP compounds their average LTEC and expansion anisotropy are both nearly zero. These are phosphates containing large cations (Cs, Sr, and Ba) in their structural voids: CsZr2(PO4)3 [13], CsHf2(PO4)3 [13], Sr0.5Zr2(PO4)3 [21], К0.5Sr0.25Zr2(PO4)3 [22], Ca0.4Sr0.1Zr2(PO4)3 [23], and Ca0.38Ba0.12Zr2(PO4)3 [14]. Their LTECs are comparable to those of the prominent low-expansion materials – zircon, cordierite, and quartz glass – which have nearly zero expansion coefficients: αav = (0.5–4.2)·10–6 K–1 in the temperature range 273–1273 K. Because of the limited number of such materials, finding compounds with a nearly zero average LTEC and expansion anisotropy continues to be a challenge. The langbeinite structural framework, as well as the NZP structure, is formed by L2(PO4)3 fragments from corner-sharing LO6-octahedra and PO4-tetrahedra [24,25]. In the structures the fragments link in a different manner. The langbeinite structure contains large cations (more than 1.5 Å ionic radii) with low oxidation degrees (+1 and +2) in the cavity sites. The main difference in thermal expansion among compounds with NZP and langbeinite structure is that they have anisotropic and isotropic thermal behavior respectively. In case of absence of anisotropy and that the object is a solid-state one, thermal expansion is determined by sole LTEC. Thermal expansion of solid-state objects is taken into consideration, whilst the materials, operating under varied temperature conditions, are projected. Differences in LTECs of materials,

which are brought into contact, lead to development of inner strains and spontaneous destruction of materials. Experimental determination of LTECs is carried out with X-ray diffractometry and dilatometry methods [26,27], though the results obtained by these methods, are varied, due to difference in physical meaning. As a rule, LTECs of crystalline compounds are defined with XRD method, whereas those of materials – with dilatometry one. In fact, X-ray diffractometry and dilatometry are complementary methods of studying of substance behavior toward changes of temperature. Thus, making comparisons among the results, obtained with such techniques, allows us to define influence of material texture on its thermal expansion and therefore to minimize it. Low demand for a polycrystalline sample, opportunity to determinate LTECs of anisotropic compounds and acute sense of microscopic defects in the sample make thermal X-ray diffractometry in many cases more preferable, rather than dilatometry. The goal of this work is to present data on the synthesis, structure and the thermal expansion behavior of MZr2(TO4)x(PO4)3 – x (M = Li, Na, K, Rb, Cs; T = As, V) solid solutions, NaFeZr(PO4)2SO4

and

Pb2/3FeZr(PO4)7/3(SO4)2/3

sulfate-phosphates,

KPbMgTi(PO4)3,

K5/3MgTi4/3(PO4)3, and K5/3MgZr4/3(PO4)3. There are wide literature data related to various methods of synthesis of the NZP-phosphates. The main approaches are based on the reactions in the solid, aqueous phases (sol–gel method, hydrothermal synthesis) and in melt. Nowadays a lot of laboratories in the world are using sol−gel method, based on the gel formation in the aqueous sols. The main advantages of this method are simple fulfillment and control of composition, size and morphology of the particles of powder, reaching the high extension of homogeneity and purity of product. Moreover, this process doesn’t demand high temperatures and takes a little time to get accomplished. In this way NZP materials with controlled thermal expansion coefficients and ultra-low thermal expansion could easily be produced with sufficient quality via sol−gel method.

Experimental

MZr2(ТO4)x(PO4)3

– x

(M = Li, Na, K, Rb, Cs; T = As, V), NaFeZr(PO4)2SO4 and

Pb2/3FeZr(PO4)7/3(SO4)2/3, KPbMgTi(PO4)3, K5/3MgTi4/3(PO4)3, K5/3MgZr4/3(PO4)3 samples have been prepared by a sol–gel process. The starting materials for their synthesis were the salts (chlorides, nitrates, carbonates) of metals M+, Mg2+ and Pb2+, FeCl3, ZrOCl2, TiOCl2, H3PO4 or NH4H2PO4, H3AsO4, NH4VO3 and H2SO4 of chemically pure grade. In accordance with the solgel method, the aqueous solutions of the salt solutions were mixed under continuous stirring at 293 K. At first after a while, solutions of acids involved were added. Then the precursor mixtures (gels) were heated at 363 K until they become completely dry. No sooner had the samples been eliminated from the trace amount of water, than they were ground in agate mortar for 30 minutes. The following step of the synthesis implied thermal treatment in unconfirmed air access at the temperatures varying from 873–1123 K. The maximal temperatures of samples’ synthesis depended on the composition and have been chosen according to the DTA data. The chemical composition and homogeneity of the samples have been checked with Xray microanalysis on a CamScan MV-2300 (Vega TS 5130MM) scanning electron microscope (accelerating voltage of 15 or 20 kV) equipped with secondary and backscattered electron YAG scintillator detectors and an energy dispersive X-ray microanalysis system (Link INCA Energy 200C Si(Li) semiconductor detector). Compositions have been calculated using the PAP correction procedure. The uncertainty in the compositions obtained was within 2 at %. The electron probe X-ray microanalysis results have showed that their grains have been uniform in composition, which coincided with the intended one to within experimental uncertainty. The chemical compositions of the samples were also confirmed by chemical analysis. The analysis routine was performed as described in [28]. Results of analyses proved that the stoichiometry of the obtained samples was close to the theoretical ones and varied within the inaccuracy of the employed methods. X-ray diffraction patterns have been obtained on Shimadzu XRD-6000 diffractometer (Ni-filtered CuKα radiation, λ = 1.54178 Å, angular range 2θ = 10°–50°). Si was used as an

external standard. Unit-cell parameters have been determined after indexing X-ray diffraction patterns and refined with least-square method. High-temperature X-ray diffraction measurements were made in the temperature range 298–1073 K at 100 K intervals on the same apparatus using a Shimadzu HA-1001 thermal accessory. The temperature has been monitored with a platinum– rhodium

thermocouple.

Low-temperature

X-ray

diffraction

measurements

of

Pb2/3FeZr(PO4)7/3(SO4)2/3 sample were made in the temperature range 153–473 K at 40–50 K intervals using Anton Paar TTK 450 thermal accessory. In this case the temperature was monitored with resistance thermometer Pt100 RTD. DTA–TG combined analysis of gels and the prepared samples has been performed in argon atmosphere via DSC-calorimeter Labsys TG-DTA/DSC in the temperature range 298−1473 K with heating and cooling speed 10 K/min. Functional composition of the samples has been confirmed by IR-spectroscopy on a Shimadzu FTIR 8400S spectrometer within the range of 400–1400 cm−1.

Results and discussion Prior to high-temperature X-ray diffraction measurements, we had assessed the phase purity of the MZr2(ТO4)x(PO4)3 – x (M = Li, Na, K, Rb, Cs; T = As, V) samples. After thermal treatment at 1123 °C, MZr2(PO4)3 (x = 0) and MZr2(AsO4)3 (x = 3) compounds and the solid solutions were homogeneous and were attributed to NZP structure. The MZr2(AsO4)x(PO4)3 – x systems are comprised of continuous series of substitutional solid solutions (0 ≤ х ≤ 3), whereas the MZr2(VO4)x(PO4)3 – x systems are comprised of limited solid solution series: 0 ≤ х ≤ 0.2 for M = Li, 0 ≤ х ≤ 0.4 for M = Na, 0 ≤ х ≤ 0.5 for M = K, 0 ≤ x ≤ 0.3 for M = Rb, and 0 ≤ x ≤ 0.2 for M = Cs. The materials are stable up to 1223–1473 K. For a particular combination of M and T cations, the X-ray diffraction patterns of the МZr2(ТO4)x(РO4)3

– x

mixed phosphates

demonstrate a gradual shift of diffraction peaks and a smooth variation in their relative intensities with increasing of х (Fig. 1). For the given combination of types of alkali metal cation

M and TO4 anion, the unit-cell parameters evaluated from indexing results for X-ray diffraction patterns of the МZr2(ТO4)x(РO4)3 – x samples gradually increase with an increase in the content of As5+ arsenic or V5+ vanadium atoms, which are larger than P5+ (Fig. 2). The following crystalline structures – MZr2(AsO4)3 (M = Li, K, Rb, Cs), MZr2(AsO4)1.5(PO4)1.5 (M = Li, Na, K, Rb, Cs) have been refined with Rietveld method and were reported in the articles [29−33]. The IR spectroscopy data for the synthesized samples show that IR spectra gradually change with the smooth alteration of phosphates compositions and amorphous impurities are absent. For example, because of the larger atomic weight and size of arsenic, the vibrational frequencies of the As–O bonds are lower than those of the P–O bonds. The IR spectra of MZr2(AsO4)1.5(PO4)1.5 (M = K, Rb, Cs) are typical for compounds belonging to the space group R 3c and exhibit bands, due to both anions – phosphate and arsenate (Fig. 3): the bands in the

1147–1010 cm−1 range are characteristic of the asymmetric stretching vibrations of the PO43− ion, and the band at ~865 cm−1 is due to the same vibrations of the AsO43− ion. The bands at ~635 and ~565 cm−1 are due to bending vibrations (ν4) of the phosphate ion, and the band at ~485cm−1 is due to ν4 vibrations of the arsenate ion. The increase in the size of the M cation in going from lithium to cesium in the МZr2(ТO4)x(РO4)3 – x solid solutions with identical T and x, as well in the MZr2(AsO4)3 (x = 3) arsenates (Fig. 2), leads to expansion of their structure along the c axis and a reduction in a cell parameter. The reason for this is that the flexible framework of the NZP structure, made up of Zr octahedra and (T,P) or As tetrahedra, which share oxygens and form infinite columns parallel to the 3 axis (с axis of the unit cell), changes in size, adjusting itself to the extra framework ions M, which differ in size and are located in the columns. An increase in the size of the M cation leads to expansion of the structure along the с axis and distortion of the tetrahedra. The internal angle O–(T,P or As)–O increases along the с axis, which reduces the separation between the parallel columns and leads to compression of the structure along the а axis.

Observed correlation among the variations in the c and a cell parameters with the size of the extra framework cation M leads us to draw some preliminary conclusions regarding the behavior of the NZP structure during heat treatment, because crystal chemical substitutions and thermal expansion influence the same atomic bonds in the structure. Clearly, heating leads to a greater increase in the length of the M–O bonds, which are weaker than the Zr–O and (T,P or As)–O bonds. Moreover, if an occupied MО6 octahedron is already elongated along the crystallographic axis с at room temperature because of the incorporation of a larger alkali metal cation, further expansion of the structure along this axis on heating will be suppressed by the (T,P or As)O4 bridge tetrahedra, which link the columns of the framework structure. The thermal expansion of the МZr2(ТO4)x(PO4)3–x (M = Li, Na, K, Rb, Cs; T = As, V) samples have been studied in the temperature range 293–1073 K. Fig. 4 shows the temperature dependences of unit-cell parameters for the MZr2(AsO4)3 (x = 3) compounds. Fig. 4 demonstrates a characteristic picture of the temperature dependence of unit-cell parameters: expansion of the structure along the с axis (because the stretching МО6 octahedra are located in the columns parallel to the с axis) and compression of the structure along the а axis (because of the correlated tilt of the AsО4 tetrahedra, which share corners with the МО6 and ZrO6 octahedra). The size of the alkali metal cation M has the overriding effect on the thermal expansion parameters of MZr2(TO4)x(PO4)3 – x (Fig. 5). An increase in the size of the M cation leads to a reduction in the magnitude of the LTECs. In case of the СsZr2(TO4)x(PO4)3

– x

samples that

contain cesium, the largest alkali metal cation, the LTECs and thermal expansion anisotropy are nearly zero (Table 1). In the MZr2(AsO4)3 arsenate series, the LTECs and thermal expansion anisotropy gradually decrease from M = Li to M = Cs (Table 1). The high αс values in the МZr2(TO4)x(PO4)3 – x (M = Li, Na) materials compared to the samples containing the larger alkali metal cations can be accounted for possible migration of the Li+ and Na+ cations from the fully occupied extra framework sites in the structural columns to vacant extra framework sites between the columns [2,18]. This process is accompanied by an increase in repulsion between

the oxygen chains around the vacant extra framework sites in the structural columns, which leads to an increase in αс. In addition to the effect of the size of the M cation on the LTECs, we have examined the effect of arsenic or vanadium substitutions for phosphorus. For KZr2(VO4)х(PO4)3–х system, an increase in vanadium content leads to a slight increase in the magnitude of the coefficient αа and a considerable increase in the coefficient αс and expansion anisotropy (Fig. 6). A nearly zero average LTEC (at х = 0.4) is accompanied by considerable thermal expansion anisotropy. Among the vanadate phosphates studied here, the minimum expansion anisotropy at a small average LTEC is only offered by СsZr2(VO4)0.2(PO4)2.8 (Table 1). For МZr2(TO4)x(PO4)3

– x

and MZr2(AsO4)3, we have observed an increase in thermal

expansion parameters in comparison with phosphate analogues (Table 1), which seems to be caused by the increase in the bond lengths between the tetrahedral atoms and oxygen and, as a consequence, by the lower strength of the bonds in the case of thermal distortion of the crystal structure. Thus, by means of combining compounds capable of forming substitutional solid solutions and varying the composition of solid solutions, materials with controlled, in particular low, LTECs and thermal expansion anisotropy could be acquired. Potassium zirconium, rubidium zirconium, and cesium zirconium arsenates, arsenate-phosphates, and vanadatephosphates can be counted as low-expansion materials. Sodium zirconium arsenate, sodium zirconium arsenate-phosphates, sodium zirconium vanadate-phosphates, and lithium zirconium vanadate-phosphates have intermediate thermal expansion, and lithium zirconium arsenate is a high-expansion material. The small LTECs of the NZP compounds and solid solutions that contain large alkali metal cations ensure extremely low anisotropy and allow one to produce crack-proof materials. The sulfate-phosphates NaFeZr(PO4)2SO4 and Pb2/3FeZr(PO4)7/3(SO4)2/3 with expected NZP-structure have been designed according to the crystallographic data. Isomorphic substitution Zr4+ + P5+ → Fe3+ + S6+ in the basic NaZr2(PO4)3 leads to NaFeZr(PO4)2SO4

formation. However, the real synthesis is quite complicated. There are some challenges connected with thermal instability of sulfate-phosphates above 873 K, caused by partial losses of sulfur as SO3. In case of Pb2/3FeZr(PO4)7/3(SO4)2/3, selected stoichiometry of the sample halts sulfur losses, binding all amount of this one in Pb(SO4)2 (melting point above 1273 K) in primal mixture. Processes followed by formation of pure NaFeZr(PO4)2SO4 phase from the gel, calcined at 363 K, have been studied with combined DTA–TG method. The loss of mass has been detected in the temperature range from 373 to 973 K (Fig. 7). There were blurred endothermic effects on the DTA-curve with minimums in 293 and 703 K. The maximal loss of mass has been registered before 473 K. The studied temperature range corresponds to decay of zirconium hydrogen phosphate (intermediate product), releasing of water, hydrogen chloride and formation of NaFeZr(PO4)2SO4. Single phase NaFeZr(PO4)2SO4 ceramic was obtained as a result of the multiple repeating of the following cycle: grinding – pressing − calcining. According to the micrograph (Fig. 8), we can conclude, that the grain size is about 1−5μm. The composition corresponded to the Na0.98(3)Fe1.02(5)Zr0.99(3)P2.06(4)S0.95(4)O12 formula and matches to the calculated one theoretically. In accordance with the X-ray diffractometric data, NaFeZr(PO4)2SO4 crystallizes in the NZP structure ( R 3c , Z = 6), lattice parameters: a = 8.750(4), c = 22.626 (9) Å, V = 1500(1) Å3. X-ray diffractometric data shows that in case of heating of NaFeZr(PO4)2SO4 above 873 K the value of lattice parameters grows. It is connected with partially loss of sulfur and smaller size of tetrahedral anion SO4 in comparison with PO4. As mentioned earlier, the stoichiometry of Pb2/3FeZrS2/3P7/3O12 can avert the losses of sulfur. In accordance with the X-ray patterns the main phase of sample was crystallizing with small amount of Fe2O3 at 923K. The extent of crystallization was going up on a par with increase of temperature up to 1053 K. At this temperature the single phase Pb2/3FeZr(PO4)7/3(SO4)2/3 with NZP structure was crystallized. The Rietveld refinement of the step scan data has been

performed by the least square method using Rietan for that compound. Table 2 summarizes the measurement conditions, unit cell parameters, and main data of the structure refinement. Fig. 9 presents fragments of the experimental, calculated, and difference X-ray diffraction patterns, as well as the line diagram of the diffraction pattern of the phosphate-sulfate. The refinement leads to a rather good agreement between the experimental and calculated diffraction pattern and yields acceptable reliability factors (Rp, Rwp) (Table 2). The phosphate-sulfate consists of the framework formed by FeO6 -, ZrO6-octahedra and SO4-, PO4-tetrahedra connected by vertices via Zr–O–P, Fe–O–P, Zr–O–S, and Fe–O–S structural bridges (Fig. 10). There is an area of interests of using of sulfate-phosphates with NZP-structure for revealing correlations composition − parameters of thermal expansion of crystal structure (controlled LTECs). The dependence between lattice parameters a, c and temperature is shown in the fig. 11. According to these ones it can be concluded that both of the lattice parameters grow with temperature increase. That is caused by correlated rotation of octahedra and tetrahedra around c axis, usual for NZP-compounds. The amounts of the linear coefficients of thermal expansion are the following: αa = 0.97∙10–6, αс = 3.24∙10–6, αav = 1.72∙10−6 K−1. This sulfatephosphate is related to the low expansion compositions group. Both of the coefficients are positive αa < αс, so it leads to little anisotropy. As we presupposed, directed combining of various anions in the NZP structure, where one of them is phosphor and the other is smaller, like sulfur, led to decrease of framework size and its cavities. Thus, the smaller and cheaper cations could well be used (like Pb in sulfatephosphates) instead of big and expensive ones (like Cs in monophosphates) for the production of ceramics with low thermal expansion and value of anisotropy. The following phosphates KPbMgTi(PO4)3, K5/3MgTi4/3(PO4)3, K5/3MgZr4/3(PO4)3 crystallize in the structural type of langbeinite mineral (K2Mg2(SO4)3, space group P213) at the temperature being 973−1073 K. In accordance with the results of X-ray diffractometry (Fig. 12), it is transparent that the following phosphates KPbMgTi(PO4)3 and K5/3MgE4/3(PO4)3 (E = Ti, Zr)

are structurally analogous to each other. The structure of KPbMgTi(PO4)3 has been refined by means of Rietveld analysis [33]. The main feature of the structure is the framework, comprised of the vertices-bound MgO6 and TiO6 octahedra and PO4 tetrahedra. K and Pb atoms are coordinated by nine oxygen atoms and completely occupy extra-framework cavities. The number of sites, being at the disposal of extra-framework cations, is equal to two per formula unit. Seemingly, in K5/3MgE4/3(PO4)3 these two sites are occupied by (5/3K+ + 1/3E4+) ions, whilst (Mg2+ + E4+) ions are distributed throughout two framework sites. Structural resemblance of langbeinite frameworks with the same cavities manifests in comparability of unit-cell parameters, with figures being 9.8540(3) Å, 9.836(3) Å and 10.196(6) Å for KPbMgTi(PO 4)3, K5/3Ti1/3MgTi(PO4)3 and K5/3Zr1/3(PO4)3 respectively. The high temperature X-ray measurement of the KPbMgTi(PO4)3, K5/3MgTi4/3(PO4)3, K5/3MgZr4/3(PO4)3 compounds showed that a lattice parameter increases with temperature growth linearly (Fig. 13). Calculated LTECs for the temperature range 298−1073 K afford us to relate these phosphates to the intermediate and high expanding materials. Due to the langbeinite structure these phosphates expand isotropically.

Conclusions We have determined the thermal expansion parameters of the MZr2(ТO4)x(PO4)3–x (T = As, V) solid solutions and MZr2(AsO4)3 (x = 3, M = Li, Na, K, Rb, Cs) compounds and Pb2/3FeZrS2/3P7/3O12 sulfate-phosphate with the NZP structure. All of the materials have been shown to have αа < αс. In the temperature range from 293 to 1073 K, the unit-cell parameter a of arsenates, and arsenate- and vanadate-phosphates is almost constant or slightly decreases with increasing temperature, whereas the c cell parameter increases. The present thermal expansion data demonstrate that the MZr2(AsO4)3 arsenates and the МZr2(AsO4)x(PO4)3–x and МZr2(VO4)x(PO4)3–x solid solutions differ in the behavior of their thermal expansion anisotropy and volume expansion. An increase in the size of the alkali metal cation in the arsenates studied

here leads to a reduction in the magnitude of their LTECs and expansion anisotropy. The decrease in anisotropy that can be reached by varying the composition (x) of the solid solutions may be accompanied by a decrease or increase in average LTEC. The minimum thermal expansion anisotropy and LTECs are offered by CsZr(AsO4)3, СsZr2(AsO4)1.5(PO4)1.5, and СsZr2(VO4)0.2(PO4)2.8. Concerning Pb2/3FeZr(PO4)7/3(SO4)2/3, in the temperature range from 153 to 473 K unit-cell parameters a and c increase slightly. This sulfate-phosphate has low volume thermal expansion with low anisotropy. Varying composition of sulfate-phosphates with preserving the stable structure fragments we can smoothly change size (and symmetry) of unit cell, precisely correct parameters of thermal expansion, using spread and accessible elements. Thus, thermophysical properties of NZP materials can be adjusted to suit materials’ need by proper ionic substitutions and opting for certain techniques of material production. Some NZP materials have very low thermal expansion and expansion anisotropy and exhibit high thermal shock resistance. Unlike materials with NZP-structure, phosphates with langbeinite structure expand isotropically and have zero expansion anisotropy. The unit-cell parameter a and the volume V of samples KPbMgTi(PO4)3, K5/3MgTi4/3(PO4)3, K5/3MgZr4/3(PO4)3 with langbeinite structure grow with temperature increasing. The nature of cations which occupy cavities in the framework of langbeinite structure has tiny influence on the deformations during the heating.

Acknowledgements This work was supported by the Russian Foundation for Basic Research (Project No. 15-0300716a).

References

[1] N. Anantharamulu, R.K. Koteswara, G. Rambabu, K.B. Vijaya, V. Radha, M. Vithal, J. Mater. Sci. 46 (2011) 2821–2837. [2] V.I. Pet’kov, Russ. Chem. Rev. 81 (2012) 606–637. [3] K. Kamali, T.R. Ravindran, N.V. Chandra Shekar, K.K. Pandey, S.M. Sharma, J. Solid State Chem. 221 (2015) 285−290. [4] V. Pet’kov, E. Asabina, V. Loshkarev, M. Sukhanov, J. Nucl. Mater. 471 (2016) 122−128. [5] V.I. Pet’kov, E.A. Asabina, Glass Ceram. 61 (2004) 233−239. [6] V.I. Pet’kov, A.S. Dmitrienko, M.V. Sukhanov, A.M. Kovalskii, J. Therm. Anal. Cal. 127 (2017) 2093−2099. [7] M. Samiee, B. Radhakrishnan, Z. Rice, Z. Deng, Y.S. Meng, S.P. Ong, J. Luo, J. Power Sources 347 (2017) 229−237. [8] NZP ceramics, a new class of materials. Amer. Ceram. Soc. Bull. 76 (1997) 71–72. [9] K. Kim, S. Yang, M.Y. Kim, M.S. Lee, J. Lim, D.R. Chang, H. Kim, J. Ind. Eng. Chem. 36 (2016) 279−283. [10] A. Bohre, K. Avasthi, V.I. Pet’kov, J. Ind. Eng. Chem. 50 (2017) 1−14. [11] V.I. Pet’kov, A.I. Orlova, Inorg. Mater. 39 (2003) 1013–1023. [12] G.E. Lenain, H.A. McKinstry, J. Alamo, D.K. Agraval, J. Mater. Sci. 22 (1987) 17–22. [13] V.I. Pet’kov, A.I. Orlova, G.N. Kasantsev, S.G. Samoilov, M.L. Spiridonova, J. Therm. Anal. Cal. 66 (2001) 623–632. [14] P. Oikonomou, C. Dedeloudis, C. J. Stournaras, C. Ftikos, J. Eur. Ceram. Soc. 27 (2007) 1253–1258. [15] E. Asabina, V. Pet’kov, P. Mayorov, D. Lavrenov, I. Schelokov, A. Kovalsky, Pure Appl. Chem. 89 (2017) 523−534. [16] J. Alamo, Solid State Ionics 63-65 (1993) 547.

[17] P. Lightfoot, D.A. Woodcock, J.D. Jorgensen, S. Short, J. Inorg. Mater. 1 (1999) 53–60. [18] M. Catti, A. Comotti, S. Di Blas, Chem. Mater. 15 (2003) 1628–1632. [19] V.I. Pet’kov, E.A. Asabina, I.A. Shchelokov, Inorg. Mater. 49 (2013) 502–506. [20] V.I. Pet’kov, A.S. Shipilov, M.V. Sukhanov, Inorg. Mater. 51 (2015) 1079–1085.

[21] S.Y. Limaye, D.K. Agraval, H.A. McKinstry, J. Am. Ceram. Soc. 70 (1987) 232–236. [22] B. Zhang, J. Guo, J. Eur. Ceram. Soc. 15 (1995) 929–933. [23] X.S. Liu, F. Li, W.B. Song, B.H. Yuan, Y.G. Cheng, E.J. Liang, M.J. Chao, Ceram. Int. 40 (2014) 6013–6020. [24] V.I. Pet’kov, E.A. Asabina, A.V. Markin, A.A. Alekseev, N.N. Smirnova, J. Therm. Anal. Cal. 124 (2016) 1535−1544. [25] K.V. Martynov, A.N. Nekrasov, A.R. Kotel’nikov, I.G. Tananaev, Glass Phys. Chem. 43 (2017) 75−82. [26] J.D. James, J.A. Spittle, S.G.R. Brown, R.W. Evans, Meas. Sci. Technol. 12 (2001) R1−R15. [27] R. Brochu, El-Yacoubi, M. Louёr, A. Serghini, M. Alami, D. Louёr, Mat. Res. Bull. 32 (1997) 15–23. [28] V.I. Pet’kov, E.A. Asabina, Indian J. Chem. 52A (2013) 350–356. [29] V.I. Pet’kov, M.V. Sukhanov, A.S. Shipilov, V.S. Kurazhkovskaya, E.Yu. Borovikova, I.Yu. Pinus, A.B. Yaroslavtsev, Inorg. Mater. 50 (2014) 263–272. [30] M.V. Sukhanov, V.I. Pet’kov, D.V. Firsov, V.S. Kurazhkovskaya, E.Yu. Borovikova, Russ. J. Inorg. Chem. 56 (2011) 1351–1357. [31] V.I. Pet’kov, A.S. Shipilov, M.V. Sukhanov, V.S. Kurazhkovskaya, E.Yu. Borovikova, Russ. J. Inorg. Chem. 59 (2014) 1201–1207. [32] V.I. Pet’kov, M.V. Sukhanov, A.S. Shipilov, V.S. Kurazhkovskaya, E.Yu. Borovikova, N.V. Sakharov, M.M. Ermilova, N.V. Orekhova, Russ. J. Inorg. Chem. 58 (2013) 1015–1021.

[33] V.I. Pet’kov, A.A. Alekseev, E.A. Asabina, E.Yu. Borovikova, A.M. Kovalskii, Russ. J. Inorg. Chem. 62 (2017) 870−878.

Figure captions

Fig. 1. X-ray diffraction patterns of KZr2(TO4)x(РO4)3 – x samples: T = (1–4) V and (5–7) As; x = (1, 5) 0, (2) 0.2, (3) 0.3, (4) 0.5, (6) 1.5, (7) 3.

Fig. 2. Composition dependences of unit-cell parameters for MZr2(AsO4)x(РO4)3 – x with M = (1) K, (2) Rb, and (3) Cs.

Fig. 3. IR spectra of MZr2(AsO4)1.5(PO4)1.5 with M = (1) K, (2) Rb and (3) Cs.

Fig. 4. Temperature dependences of the unit-cell parameters a and c for МZr2(AsO4)3 with M = (1) Li, (2) K, and (3) Cs.

Fig. 5. LTECs αа, αс, and αav and thermal expansion anisotropy |αa – αc| against the size of the alkali metal cation for MZr2(VO4)0.2(PO4)2.8.

Fig. 6. Composition dependences of the LTECs αа, αс, and αav and thermal expansion anisotropy |αa – αc| for KZr2(VO4)х(PO4)3 – х.

Fig. 7. TG (1), DTG (2) and DTA (3) curves of NaFeZr(PO4)2SO4 sample.

Fig. 8. Micrograph of NaFeZr(PO4)2SO4 sample (calcined at 873 K).

Fig. 9. Fragments of the (blue) experimental, (red dots) calculated, and (green) difference X-ray diffraction patterns, and (black) line diagram of Pb2/3FeZr(PO4)7/3(SO4)2/3 phosphate sulfate diffraction pattern.

Fig. 10. Crystal structure of Pb2/3FeZr(PO4)7/3(SO4)2/3 (amber: (Fe/Zr)O6-octahedra, sapphire: (P/S)O4-tetrahedra).

Fig. 11. Temperature dependences of the unit-cell parameters a and c for Pb2/3FeZr(PO4)7/3(SO4)2/3.

Fig. 12. X-ray diffraction patterns of KPbMgTi(PO4)3 (1) and K5/3MgTi4/3(PO4)3 (2).

Fig. 13. Temperature dependences of the unit-cell parameter a for K5/3MgTi4/3(PO4)3 (1), KPbMgTi(PO4)3 (2) and K5/3MgZr4/3(PO4)3 (3).

Table 1. Thermal expansion parameters of MZr2(TO4)x(PO4)3 – x (M = Li, Na, K, Rb, Cs; T = As, V), Pb2/3FeZr(PO4)7/3(SO4)2/3, K5/3MgTi4/3(PO4)3, KPbMgTi(PO4)3 and K5/3MgZr4/3(PO4)3. Thermal expansion coefficients, ×106 K−1 Compound αа

αс

αav

|αa − αc|

LiZr2(PO4)3

−7.4

35.5

4.4

43.0

LiZr2(AsO4)3

−6.1

47.5

9.9

53.6

LiZr2(VO4)0.2(PO4)2.8

−5.6

32.4

7.0

38.0

NaZr2(PO4)3

−5.5

22.3

3.8

27.8

NaZr2(AsO4)1.5(PO4)1.5

−5.8

19.0

2.5

24.8

NaZr2(AsO4)3

−6.7

25.0

3.9

31.7

NaZr2(VO4)0.2(PO4)2.8

−5.3

25.3

4.9

30.6

KZr2(PO4)3

−5.3

5.4

−1.7

10.7

KZr2(AsO4)1.5(PO4)1.5

−3.5

8.9

0.6

12.4

KZr2(AsO4)3

−6.2

10.6

2.1

16.8

KZr2(VO4)0.2(PO4)2.8

−4.0

6.3

−0.5

10.3

RbZr2(PO4)3

−4.2

3.6

-1.6

7.8

RbZr2(AsO4)1.5(PO4)1.5

−2.3

3.6

-0.3

5.9

RbZr2(AsO4)3

−4.4

7.9

−0.3

12.3

RbZr2(VO4)0.2(PO4)2.8

−2.3

3.3

−0.5

5.6

CsZr2(PO4)3

−0.6

0.5

−0.2

1.1

CsZr2(AsO4)1.5(PO4)1.5

0.1

3.2

1.1

3.1

Pb2/3FeZr(PO4)7/3(SO4)2/3

1.0

3.2

1.7

2.3

K5/3MgTi4/3(PO4)3

11.1

11.1

11.1

0

KPbMgTi(PO4)3

8.1

8.1

8.1

0

K5/3MgZr4/3(PO4)3

4.9

4.9

4.9

0

Table 2. Summary of crystallographic data for Pb2/3FeZr(PO4)7/3(SO4)2/3 compound. Formula Structural analogue Crystal system Space group Z

Pb2/3FeZr(PO4)7/3(SO4)2/3 PbFeZr(PO4)3 Trigonal R 3c (No. 167)

6

Unit cell parameters: a (Å)

8.6339(4)

c (Å)

23.2991(9)

V (Å3)

1504.1(1)

dcalc (g/cm3)

4.159(1)

2θ angular range (deg)

10 – 110

Total number of reflections

214

Number of refined parameters

28

Rwp (%)

3.24

Rp (%)

2.20

S

2.9103