Defect perovskites (He2–x□x)(CaZr)F6 can be prepared by inserting helium into CaZrF6 at high pressure. They can be recovered to ambient pressure at low temperature. There are no prior examples of perovskites with noble gases on the A-sites. The insertion of helium gas into CaZrF6 both elastically stiffens the material and reduces the magnitude of its negative thermal expansion. It also suppresses the onset of structural disorder, which is seen on compression in other media. Measurements of the gas released on warming to room temperature and Rietveld analyses of neutron diffraction data at low temperature indicate that exposure to helium gas at 500 MPa leads to a stoichiometry close to (He1□1)(CaZr)F6. Helium has a much higher solubility in CaZrF6 than silica glass or crystobalite. An analogue with composition (H2)2(CaZr)F6 would have a volumetric hydrogen storage capacity greater than current US DOE targets. We anticipate that other hybrid perovskites with small neutral molecules on the A-site can also be prepared and that they will display a rich structural chemistry.

CaZrF6 has recently been shown to combine strong negative thermal expansion (NTE) over a very wide temperature range (at least 10–1000 K) with optical transparency from mid-IR into the UV range. Variable-temperature and high-pressure diffraction has been used to determine how the replacement of calcium by magnesium and zirconium by niobium(IV) modifies the phase behavior and physical properties of the compound. Similar to CaZrF6, CaNbF6 retains a cubic ReO3-type structure down to 10 K and displays NTE up until at least 900 K. It undergoes a reconstructive phase transition upon compression to ∼400 MPa at room temperature and pressure-induced amorphization above ∼4 GPa. Prior to the first transition, it displays very strong pressure-induced softening. MgZrF6 adopts a cubic (Fm3̅m) structure at 300 K and undergoes a symmetry-lowering phase transition involving octahedral tilts at ∼100 K. Immediately above this transition, it shows modest NTE. Its’ thermal expansion increases upon heating, crossing through zero at ∼500 K. Unlike CaZrF6 and CaNbF6, it undergoes an octahedral tilting transition upon compression (∼370 MPa) prior to a reconstructive transition at ∼1 GPa. Cubic MgZrF6 displays both pressure-induced softening and stiffening upon heating. MgNbF6 is cubic (Fm3̅m) at room temperature, but it undergoes a symmetry-lowering octahedral tilting transition at ∼280 K. It does not display NTE within the investigated temperature range (100–950 K). Although the replacement of Zr(IV) by Nb(IV) leads to minor changes in phase behavior and properties, the replacement of the calcium by the smaller and more polarizing magnesium leads to large changes in both phase behavior and thermal expansion.

CaZrF6 and CaHfF6 display much stronger negative thermal expansion (NTE) (αL100 K ∼ −18 and −22 ppm·K–1, respectively) than ZrW2O8 and other corner-shared framework structures. Their NTE is comparable to that reported for framework solids containing multiatom bridges, such as metal cyanides and metal–organic frameworks. However, they are formable as ceramics, transparent over a wide wavelength range and can be handled in air; these characteristics can be beneficial for applications. The NTE of CaZrF6 is strongly temperature-dependent, and first-principles calculations show that it is largely driven by vibrational modes below ∼150 cm–1. CaZrF6 is elastically soft with a bulk modulus (K300K) of 37 GPa and, upon compression, starts to disorder at ∼400 MPa. The strong NTE of CaZrF6, which remains cubic to <10 K, contrasts with cubic CoZrF6, which only displays modest NTE above its rhombohedral to cubic phase transition at ∼270 K. CaZrF6 and CaHfF6 belong to a large and compositionally diverse family of materials, AIIBIVF6, providing for a detailed exploration of the chemical and structural factors controlling NTE and many opportunities for the design of controlled thermal expansion materials.

The cubic ReO3-type material ScF3 exhibits strong isotropic negative thermal expansion (NTE) over a wide temperature range while remaining cubic. Control of its thermal expansion was investigated by forming Sc1-xTixF3 solid solutions, which were characterized by synchrotron powder diffraction at ambient pressure from 100 to 500 K. TiF3 is fully soluble in ScF3 at a synthesis temperature of 1338 K. The temperature for the cubic-to-rhombohedral phase transition in Sc1-xTixF3 varies linearly with composition (above 100 K), and, at large x, the transition is clearly first-order. The rhombohedral phase for each composition examined exhibits strongly positive thermal expansion, while the expansion of the cubic phase (between 420 and 500 K) is negative for all Sc1-xTixF3.

Materials with the cubic ReO3-type structure are, in principle, excellent candidates for negative thermal expansion (NTE). However, many such materials, including TaO2F, do not display NTE. It is proposed that local distortions away from the ideal structure, associated with the need to accommodate the different bonding requirements of the disordered O/F, contribute to the occurrence of near zero thermal expansion rather than NTE. The local structure of TaO2F is poorly described by an ideal cubic ReO3-type model with O and F randomly distributed over the available anion sites. A supercell model featuring −Ta–O–Ta–O–Ta–F– chains along ⟨1 0 0⟩, with different Ta–O and Ta–F distances and O/F off-axis displacements, gives much better agreement with pair distribution functions (PDFs) derived from total X-ray scattering data for small separations (<8 Å). Analyses of PDFs derived from variable temperature measurements (80 to 487 K), over different length scales, indicate an average linear expansion coefficient of close to zero with similar contributions from the geometrically distinct Ta–O—Ta and Ta–F—Ta links in TaO2F.Keywords: anion site disorder; near zero thermal expansion; pair distribution functions; tantalum oxyfluoride;

The role of MO4 (M = W, Mo) orientational disorder in the thermal expansion and compressibility of ZrW2O8 and ZrMo2O8 was investigated via in situ powder X-ray diffraction at elevated temperature and pressure. A dramatic reduction in the bulk modulus of α-ZrW2O8, which has ordered WO4 tetrahedra at room temperature, from 65 GPa at room temperature to 47 GPa at 386 K was observed to be concomitant with the onset of a reversible WO4 orientational disordering upon compression. Additionally, the coefficient of thermal expansion (CTE) of the α phase became more negative upon compression within the temperature range in which pressure-dependent disorder was observed; αl, over the range 298 to 386 K, was ∼−11 ppm K−1 at 35 MPa but ∼−16 ppm K−1 at 276 MPa. No softening upon heating or change in CTE upon compression was observed for ZrW2O8 above the order → disorder phase transition temperature. Cubic ZrMo2O8 has a disordered arrangement of MoO4 tetrahedra at all temperatures and pressures accessed in this study. Its bulk modulus was independent of temperature, and its CTE was insensitive to pressure, much like β-ZrW2O8. The stability/metastability of the cubic and orthorhombic phases upon heating above room temperature and compression is discussed, with a focus on changes in the thermodynamics and kinetics of the cubic ↔ orthorhombic transition.

The hydration of tricalcium silicate (C3S) is accelerated by pressure. However, the extent to which temperature and/or cement additives modify this effect is largely unknown. Time-resolved synchrotron powder diffraction has been used to study cement hydration as a function of pressure at different temperatures in the absence of additives, and at selected temperatures in the presence of retarding agents. The magnitudes of the apparent activation volumes for C3S hydration increased with the addition of the retarders sucrose, maltodextrin, aminotri(methylenephosphonic acid) and an AMPS copolymer. Pressure was found to retard the formation of Jaffeite relative to the degree of C3S hydration in high temperature experiments. For one cement slurry studied without additives, the apparent activation volume for C3S hydration remained close to ~ − 28 cm3 mol− 1 over the range 25 to 60 °C. For another slurry, there were possible signs of a decrease in magnitude at the lowest temperature examined.

A sample cell for the simultaneous measurement of synchrotron X-ray powder diffraction and ultrasound shear-wave reflection data from cement slurries is described. White cement slurries at 25 and 50 °C with 0–3% bwoc CaCl2 were studied to illustrate the potential of the apparatus. The decrease in reflected S-wave amplitude, in dB, showed a linear correlation with C3S hydration. CaCl2 retarded the development of G′ and G″ relative to the extent of C3S hydration. At short times, there was a correlation between the time evolution of both G′ and G″, and the amount of precipitated CH seen by diffraction, which was almost independent of CaCl2 concentration and temperature. CaCl2 addition resulted in a decrease in the amount of CH visible to X-rays, relative the degree of C3S hydration. This may indicate a change in C–S–H gel C:S ratio or the presence of nanoscale CH that could not be seen by diffraction.

Scandium trifluoride maintains a cubic ReO3 type structure down to at least 10 K, although the pressure at which its cubic to rhombohedral phase transition occurs drops from >0.5 GPa at ∼300 K to 0.1−0.2 GPa at 50 K. At low temperatures it shows strong negative thermal expansion (NTE) (60−110 K, αl ≈ −14 ppm K−1). On heating, its coefficient of thermal expansion (CTE) smoothly increases, leading to a room temperature CTE that is similar to that of ZrW2O8 and positive thermal expansion above ∼1100 K. While the cubic ReO3 structure type is often used as a simple illustration of how negative thermal expansion can arise from the thermally induced rocking of rigid structural units, ScF3 is the first material with this structure to provide a clear experimental illustration of this mechanism for NTE.

Zr2(MoO4)(PO4)2 is orthorhombic (Sc2W3O12 structure) from 9 to at least 400 K, and shows anisotropic volume negative thermal expansion (αa=−8.35(4)×10−6 K−1; αb=3.25(3)×10−6 K−1; αc=−8.27(5)×10−6 K−1 in the range 122–400 K) similar in magnitude to A2M3O12 (M—Mo or W) with large A3+. The contraction on heating is associated with a pattern of Zr–O–Mo/P bond angle changes that is somewhat similar, but not the same as that for Sc2W3O12. On heating, the most pronounced reductions in the separation between the crystallographic positions of neighboring Zr and P are not associated with significant reductions in the corresponding Zr–O–P crystallographic bond angles, in contrast to what was seen for Sc2W3O12.Negative thermal expansion in Zr2(MoO4)(PO4)2 is associated with a complex pattern of structural changes that is not the same as that previously reported for Sc2(WO4)3.

Negative thermal expansion materials can experience significant stresses when they are used in composites. Under ambient conditions Zr2(WO4)(PO4)2 displays anisotropic negative thermal expansion (NTE) (αv=−14.0(10)×10−6K−1, αa=−7.9(5)×10−6K−1, αb=2.5(5)×10−6K−1, αc=−8.7(2)×10−6K−1 at 0 GPa). The effect of hydrostatic pressure on its thermal expansion characteristics was investigated by neutron diffraction between 300 and 60 K at pressures up to 0.3 GPa. No phase transitions were observed in the pressure and temperature range examined. The material was found to have a bulk modulus, B0, of 61.3(8) GPa at ambient temperature, and unlike some other NTE materials, pressure had no detectable effect on thermal expansion (αv=−14.2(8)×10−6K−1, αa=−7.9(3)×10−6K−1, αb=2.9(5)×10−6K−1, αc=−9.2(2)×10−6K−1 at 0.3 GPa).

CeP2O7, a close structural relative of ZrP2O7, and many other MIVX2O7 (X – P, V, As) phases, forms on heating Ce(HPO4)2·xH2O between ∼300 and 600 °C and decomposes by oxygen loss at higher temperatures. In-situ X-ray diffraction measurements showed, for some precursor batches, the formation of CeP2O7 in two distinct stages. At room temperature, CeP2O7 is pseudocubic, but probably triclinic, displaying positive thermal expansion below ∼115 °C. Above this temperature, there is a transition to cubic symmetry (Pa3̅) with the linear coefficient of thermal expansion going to zero at ∼450 °C and becoming negative at higher temperatures. CeP2O7 probably undergoes two phase transitions on compression below ∼10.5 GPa. The phase existing between 0.65 and ∼5 GPa is soft with an average bulk modulus of ∼18 GPa.

TaO2F, with a ReO3-type structure, has been studied at up to 12.8 GPa using monochromatic synchrotron powder diffraction and diamond anvil cells. Two-phase transitions at ∼0.7 and 4 GPa were observed on compression. Below ∼0.7 GPa the cubic material was found to have a bulk modulus (K0) of 36(3) GPa (Kp fixed at 4.0), similar to that reported for NbO2F but much smaller than that of ReO3. Immediately above 0.7 GPa on compression, the diffraction data were not fully consistent with a VF3-type structure as previously proposed for NbO2F. On decompression, the data between 8 and 4 GPa could be satisfactorily attributed to a single R-3c phase with a VF3-type structure and an average bulk modulus of 60(2) GPa.

Time-of-flight powder neutron diffraction combined with Rietveld analysis was used to study Sc2W3O12 at pressures of up to 0.6 GPa in a large volume helium gas cell. Orthorhombic (Pnca) Sc2W3O12, bulk modulus 32(3) GPa, transforms to a softer monoclinic (P21/a) structure, 11.8(8) GPa, on compression at between 0.25 and 0.30 GPa. This transition shows significant hysteresis and involves a volume reduction of ∼1.5%. The volume reduction during compression of the orthorhombic phase is much less anisotropic than that induced by heating this negative thermal expansion phase at ambient pressure. While the volume reduction on heating orthorhombic Sc2W3O12 is associated with pronounced changes in ScOW angles, there are no major changes in these angles on compression to similar densities. The structural changes seen at the pressure induced orthorhombic to monoclinic transition in Sc2W3O12 are similar, although not identical, to those seen for the equivalent thermally induced transition in Sc2Mo3O12.

A series of compounds (M0.5IIIM′0.5V)P2O7, MIIIM′VMIIIM′V=AlTa, FeTa, GaTa, InNb, YNb, NdTa, and BiTa that are close structural relatives of cubic ZrP2O7 were prepared. Annealing samples with MIIIM′VMIIIM′V=InNb or YNb at temperatures above 600 °C did not lead to any long-range cation ordering. The thermal expansion characteristics of samples quenched from 1000 °C with MIIIM′VMIIIM′V=AlTa, InNb and YNb were investigated by high-temperature powder diffraction over the temperature range 25–600 °C. There are no lattice constant discontinuities in this range, unlike ZrP2O7. (Al0.5Ta0.5)P2O7 and (In0.5Nb0.5)P2O7 show linear coefficients of thermal expansion (CTEs) of 11.5(2)×10−6 and 11.8(2)×10−6 K−1, respectively. These values are similar to that for the low-temperature ZrP2O7 structure. However, the linear CTE for (Y0.5Nb0.5)P2O7 (4.8(2)×10−6 K−1) is similar to that of the high-temperature form of ZrP2O7.The average linear coefficient of thermal expansion (∼25–600 °C) for M0.5III(Nb/Ta)0.5P2O7 decreases dramatically on going from In to Y. The behavior of the Y compound is similar to that of ZrP2O7 at high temperatures (>290 °C), whereas the In compound is similar to ZrP2O7 at low temperature.

A combination of in situ high-pressure X-ray diffraction and Mo K-edge XANES was used to examine the changes in local structure that occur as the negative thermal expansion material cubic zirconium molybdate becomes amorphous on compression and, hence, provide insight into the mechanism of amorphization. Amorphization started at ∼1.7 and by 4.1 GPa the sample was glass like by diffraction. The XANES data shows that the pressure induced amorphization at 4.1 GPa does not involve a complete change of molybdenum coordination from tetrahedral, in the starting phase, to approximately octahedral as would be expected if the amorphous material were a metastable intermediate well advanced along the pathway towards decomposition into a mixture of MoO3 and ZrO2. Recrystallization of the amorphous material at 600 °C and 4.1 GPa produced a sample that was not a simple mixture of known MoO3 and ZrO2 polymorphs nor a known polymorph of ZrMo2O8.

The negative thermal expansion material cubic ZrMo2O8 can be prepared by the carefully controlled dehydration of ZrMo2O7(OH)2·2H2O. The quality of the end product is dependent upon the route used to prepare the ZrMo2O7(OH)2·2H2O. The influence of both the Zr∶Mo ratio in the solution used to prepare the hydrate and the type of acid used in its preparation (HCl, H2SO4, CH3CO2H, HNO3 and HClO4) are examined. Not all acids provide media suitable for the preparation of ZrMo2O7(OH)2·2H2O. Cubic ZrMo2O8 could only be obtained from precipitates containing the hydrate. Starting solutions with a Zr∶Mo ratio in excess of 1∶2 were necessary to avoid coprecipitation of amorphous MoO3. The nature of the acid used in the hydrate synthesis affects the morphology of both the hydrate precursor and the cubic ZrMo2O8 particles. Many of the syntheses examined led to cubic ZrMo2O8 contaminated with amorphous impurities. Highly crystalline pure material could be obtained by using a perchloric acid medium for the synthesis of the hydrate precursor.

Highlights•The order-disorder phase transition temperatures in ZrV2O7 and HfV2O7 are strongly pressure dependent (∼700 K.GPa).•The high temperature (disordered) phase of ZrV2O7 is much stiffer than the ambient temperature (ordered) phase.•Compression reduces the magnitude of the negative thermal expansion in the high temperature phase of ZrV2O7.Low or negative thermal expansion (NTE) has been previously observed in members of the ZrP2O7 family at temperatures higher than their order-disorder phase transitions. The thermoelastic properties and phase behavior of the low temperature superstructure and high temperature negative thermal expansion phases of ZrV2O7 and HfV2O7 were explored via in situ variable temperature/pressure powder x-ray diffraction measurements. The phase transition temperatures of ZrV2O7 and HfV2O7 exhibited a very strong dependence on pressure (∼700 K GPa), with moderate compression suppressing the formation of their NTE phases below 513 K. Compression also reduced the magnitude of the coefficients of thermal expansion in both the positive and negative thermal expansion phases. Additionally, the high temperature NTE phase of ZrV2O7 was found to be twice as stiff as the low temperature positive thermal expansion superstructure (24 and 12 GPa respectively).Graphical abstractThe temperature at which ZrV2O7 transforms to a phase displaying negative thermal expansion is strongly pressure dependent. The high temperature form of ZrV2O7 is elastically stiffer than the low temperature form.

The role of MO4 (M = W, Mo) orientational disorder in the thermal expansion and compressibility of ZrW2O8 and ZrMo2O8 was investigated via in situ powder X-ray diffraction at elevated temperature and pressure. A dramatic reduction in the bulk modulus of α-ZrW2O8, which has ordered WO4 tetrahedra at room temperature, from 65 GPa at room temperature to 47 GPa at 386 K was observed to be concomitant with the onset of a reversible WO4 orientational disordering upon compression. Additionally, the coefficient of thermal expansion (CTE) of the α phase became more negative upon compression within the temperature range in which pressure-dependent disorder was observed; αl, over the range 298 to 386 K, was ∼−11 ppm K−1 at 35 MPa but ∼−16 ppm K−1 at 276 MPa. No softening upon heating or change in CTE upon compression was observed for ZrW2O8 above the order → disorder phase transition temperature. Cubic ZrMo2O8 has a disordered arrangement of MoO4 tetrahedra at all temperatures and pressures accessed in this study. Its bulk modulus was independent of temperature, and its CTE was insensitive to pressure, much like β-ZrW2O8. The stability/metastability of the cubic and orthorhombic phases upon heating above room temperature and compression is discussed, with a focus on changes in the thermodynamics and kinetics of the cubic ↔ orthorhombic transition.