High-Pressure Low-Temperature Study of LnCa4O(BO3)3 (Ln = Pr, Nd, Gd, Er, Tm) Authors: Fatiha Azrour, Romain Viennois, Jérôme Long, Dominique Granier, Fapeng Yu, Shujun Zhang, Frederico Alabarse, Mickael Beaudhuin, Jérôme Rouquette Published: 2024 ABSTRACT AbstractWe investigated the site-disorder and thermal expansion of LnCa4O(BO3)3 (Ln = Pr, Nd, Gd, Er, Tm) through a single crystal structural study conducted between 100 and 300 K. Additionally, a high-pressure synchrotron X-ray diffraction study at low temperatures was conducted to determine the compressibility of both ordered and disordered prototype oxyborate lanthanides, specifically NdCa4O(BO3)3 and ErCa4O(BO3)3. The study revealed distinct behaviors consistent with their respective ionic radii. Furthermore, satellite reflections at 50 K, along with the onset of a triple a-parameter in the Nd crystal, indicate that the ground state of this extensively studied family may be different from what was previously reported. Introduction Lanthanide calcium oxyborates (LnCa4O(BO3)3) stand out as a fascinating class of materials with potential for high-temperature sensing applications in addition to their optical properties.1,2 Unlike many other piezoelectric materials, LnCa4O(BO3)3 exhibits no reported phase transitions, suggesting remarkable thermal stability and the possibility of retaining their piezoelectric properties up to their high melting points around 1700 K. This unique characteristic makes them particularly attractive for harsh environments where conventional piezoelectrics struggle.1,3−10 Crystallographically,11,12LnCa4O(BO3)3 belong to the monoclinic polar space group Cm with the general formula Ln(M1)2(M2)2O(BO3)3, Figure 1. Their structure features three distinct cation sites (Ln, M1, and M2).13 Each site is coordinated by six oxygen atoms in a distorted octahedral arrangement. Interestingly, these octahedra connect through shared edges, forming prominent cation chains that run parallel to the c-axis. These chains are further bridged by planar BO3 groups oriented roughly perpendicular to the [001] direction. Typically, trivalent lanthanide cations (Ln3+) occupy the Ln site, while calcium (Ca2+) consistently fills both the M1 and the M2 positions. Structure of LnCa4O(BO3)3 of general formula Ln(M1)2(M2)2O(BO3)3. Ln and Ca cations are shown in orange and blue, while boron and oxygen are depicted in green and red, respectively. Many studies have revealed the presence of cation disorder on the Ln and M1 sites.14,15 This phenomenon, which influences physical properties like electrical conductivity, thermal expansion, and heat capacity, is dependent on the size of the lanthanide ion.16−19 Smaller lanthanide ions appear to promote a greater degree of cation disorder, particularly within the 800–1100 K temperature range across all LnCa4O(BO3)3. This suggests that cation disorder serves as a mechanism to minimize internal stress within the crystal structure, particularly along the c-axis where the Ln, M1, and M2 sites reside.18 Despite their promise for high-temperature sensors, a critical knowledge gap exists regarding the behavior of LnCa4O(BO3)3 under extreme conditions relevant to space exploration and planetary science, for example.20 This study aims to address this gap by comprehensively characterizing a series of LnCa4O(BO3)3 crystals. In this study, we will determine the level of cation disorder within each LnCa4O(BO3)3 crystal as it influences the material’s properties. We will further study the thermal expansion behavior of the LnCa4O(BO3)3 between 100 and 300 K, and we will analyze the compressibility behavior of typical ordered and site disordered LnCa4O(BO3)3, i.e., NdCa4O(BO3)3 and ErCa4O(BO3)3, down to 20 K to reveal how their structure responds to high pressure and low-temperature, a crucial factor for potential planetary science applications. Experimental Section Materials Piezoelectric LnCa4O(BO3)3 crystals were produced by using the Czochralski technique. High-purity (99.99%) CaCO4, Ln3O3 (or Pr6O11), and H3BO3 powders were used as starting materials for the growth of LnCa4O(BO3)3 crystals. The powders were weighed according to their nominal compositions, thoroughly mixed for 24 h, and then calcined at 1000 °C to decompose H3BO3 and CaCO3. The resulting calcined powders were remixed, pressed into tablets, and subjected to a second calcination step at 1100–1200 °C for 10 h to synthesize polycrystalline LnCa4O(BO3)3 compounds. An iridium crucible was employed as the growth vessel during the crystal growth process. More details regarding the crystal growth can be found in ref. (16) Structural Characterization LnCa4O(BO3)3 crystals were mounted on MiTeGen microloops and were subjected to X-ray crystallographic analysis using molybdenum Kα radiation (λ = 0.71073 Å) at ambient temperature and a Bruker D8 venture diffractometer. Data collection lasted for a total of 7–8 h using ω-scans. The Bruker SAINT software with a narrow-frame algorithm was used to integrate the collected frames. The analysis confirmed a monoclinic unit cell with Cm space group. Reflections were collected up to a resolution of about 0.40 Å resolution for a data completeness greater than 97.5%, with a high ratio of intensities greater than 2σ(F2), indicating good data quality. Refinement was performed by using the final XYZ-centroids of the high-intensity reflections. Data were corrected for absorption effects using the Multi-Scan method (SADABS).21 Site Disorder The final stage of the analysis involved a refinement process to achieve the best fit between the experimental and theoretical model of the crystal structure using 92 variables using the Bruker SHELXTL Software Package.22 Details can be obtained on the uploaded different CIF files available at the CCDC.23 The occupancy of cation sites was refined by assuming an ideal molar ratio of Ln3+ to Ca2+ and complete occupancy of the same crystallographic sites by both cations. Based on these assumptions, the SUMP, EXYZ, and EADP constraints were applied to the Ln, M1, and M2 atomic positions within the SHELXL software.22 All refinement processes successfully converged with R1 lower than 0.043, indicating good agreement between the calculated and observed data. Thermal Expansion The same LnCa4O(BO3)3 crystals and diffractometer were utilized in this experiment. Temperature control was achieved using an Oxford Cryostream 700 system, ranging from 100 to 300 K. To determine the unit cell parameters, 360 ω-scan frames were taken at χ = 0 with an exposure time of 5 s per frame. This process involved analyzing the centroids of approximately 3895 reflections above 20 σ(I) to a resolution of 0.67 Å, resulting in a total exposure time of 0.55 h per temperature point. In order to compare our measurements with those obtained using dilatometry, we will take into account the standardized axes for piezoelectric materials for monoclinic symmetry with β-angle different from 90°. The IEEE standard on piezoelectricity defines a specific convention for relating physical axes to crystallographic axes in piezoelectric materials.24 Following this convention, the Y and Z physical axes are aligned with the crystallographic b and c axes, respectively. The X physical axis then completes a right-handed orthogonal system perpendicular to the YZ plane. In the present study, we will therefore show the temperature variation of a.sin β(X), b(Y), c(Z), and V which are compatible with α11, α22, α33, and αV in reported dilatometry measurements. High-Pressure Synchrotron X-ray Diffraction Study at Low Temperatures Precharacterized, high-quality single crystals were used: one Nd-based compound exhibiting cation ordering and another Er-based compound exhibiting site disorder. To investigate the structural behavior of Nd- and Er-based oxyborate crystals under pressure and low temperature, single-crystal X-ray diffraction experiments were conducted at the ELETTRA XPRESS beamline. The setup employed 0.4956 Å wavelength photons, a PILATUS 3S 6 M detector, and a 110° large-aperture membrane-type diamond anvil cell (DAC). Helium served as the pressure-transmitting medium within the DAC at the XPRESS beamline. The crystals were loaded into 300 μm diameter, 60 μm thick metal gaskets along with pressure and temperature calibrants. These calibrants included a ruby sphere for pressure calibration25 inside the sample chamber along with gold powder for internal temperature calibration,26Figure 2a, and an additional ruby sphere on the diamond’s outer surface. a) Diamond anvil cell (DAC) before He-loading. b) Scattering limitation (4θ = 52°) of the high pressure-low temperature diffraction experiment. Isothermal measurements were performed at three temperatures (300, 100, and 50 K) using ω-scans at χ = 0° ± 30° for each pressure–temperature point. Lower temperature measurements were attempted without achieving isothermal conditions. Data processing, including indexing, reduction, and absorption corrections, was carried out using CrysalisPro software.27 The use of the cryostat (Figure 2b) significantly limited the scattering geometry and the accessible reciprocal space in the present experiment. The diffracted beam was restricted to a maximum 4θ value of 52°, resulting in a data completeness of only 26%. This limitation, combined with the low symmetry (monoclinic Cm space group) of the LnCa4O(BO3)3 crystals and the requirement to measure Friedel pairs for the polar space group, hindered a complete analysis of the structural behavior. Consequently, only unit cell parameters and volume could be extracted from the current study. These parameters were used to fit a second-order Birch–Murnaghan equation of state using the EOSFIT7 software.28 This analysis allowed the determination of the bulk modulus at different temperatures. Results and Discussion Site Disorder Table 1 summarizes the refined site occupancy factors for lanthanide (Ln3+) cations on the Ln site and calcium (Ca2+) cations on the M1 and M2 sites for all of the investigated samples at room temperature. The results reveal significant and reproducible variations in cation-site order across the LnCa4O(BO3)3. Table 1Occupancy of the Ln, M1, M2 Sites in Ln(M1)2(M2)2O(BO3)3 of LnCa4O(BO3)3 Determined from Single Crystal X-ray Diffraction Study at Room Temperature Are Compared with Previous Reportsa18,30Lnr(Ln3+)/r(Ca2+)LnM1(Ca1)M2(Ca2)R1%)Rσ (%)Rint (%)Pr0.990.985(2) [0.930]180.993(1) [0.984]181 [0.983]184.302.943.45Nd0.9830.981(2) [0.989]180.990(1) [0.995]181 [1]182.292.742.87Gd0.9380.959(1) [0.968]180.979(1) [0.972]181 [0.997]182.264.023.18Er0.890.800(1) [0.802]180.900(1) [0.901]181 [1]181.863.83.15Tm0.880.759(3) [0.78]300.879(1) [0.89]301 [1]303.226.106.32aR1, Rσ, and Rint are also reported for indication Occupancy of the Ln, M1, M2 Sites in Ln(M1)2(M2)2O(BO3)3 of LnCa4O(BO3)3 Determined from Single Crystal X-ray Diffraction Study at Room Temperature Are Compared with Previous Reportsa18,30 R1, Rσ, and Rint are also reported for indication Crystals with a ratio of ionic radii for Ln3+ and Ca2+ close to 1 exhibited a nearly fully ordered state, indicated by a high occupancy factor of the Ln site, Table 1. Conversely, LnCa4O(BO3)3 with a larger size mismatch between the cations tended to exhibit less ordering, never reaching a fully ordered state. This trend is further supported by data for ErCa4O(BO3)3, TmCa4O(BO3)3, where smaller Ln3+ cations lead to lower ordering. Furthermore, the size of the trivalent cation also affects the type of disorder observed as also recently found in other lanthanide borates.29 In LnCa4O(BO3)3 with smaller Ln3+ cations, cation exchange occurs only between the Ln and M1 sites. The present study is in agreement with previous reports18,30 as summarized on Table 1. Table 2 shows the Ln-, M1-, and M2-site structural polyhedral properties as a function of the Ln-cations. As already reported in the literature,13Ln and M1 sites exhibit a near-octahedral coordination with oxygen atoms, while M2 site is more irregular with a six- to 8-fold coordination sphere, i.e., in TmCa4O(BO3)3 7-fold environment can be observed for the M2 site. This favors the preferential ordering between Ln and M1-sites due to the similar coordination environment around each site. Interestingly, a decrease in the ionic radius of the lanthanide (Ln) cation leads to a significant reduction in the distortion (d) of the Ln site’s coordination polyhedron. This effect, along with a coordination number (CN) approaching 6 for ErCa4O(BO3)3 and TmCa4O(BO3)3, likely facilitates cation disorder between the Ln and M1 sites. Table 2Average Bond Length (Å), Polyhedral Volume V(Å3), Distortion Index (Bond Length), Bond Angle Variance σ()(°2) and Effective Coordination Number CN of the Ln-, M1- and M2-Site Ln-siteM1-siteM2-siteyy(Å)V(Å3)dσ() (°2)CN(Å)V(Å3)dσ() (°2)CN(Å)V(Å3)dσ() (°2)CNPr2.40617.350.033139.45.612.3617.020.01158.85.972.4316.830.04280.85.58Nd2.39317.070.032138.25.652.3616.980.0159.135.972.4316.840.039280.85.58Gd2.34616.180.027128.75.762.3516.830.00861.015.9852.43316.850.038282.85.58Er2.31515.670.022114.15.832.33916.600.00863.875.982.42916.790.037285.15.59Tm2.30915.550.020112.35.862.33716.520.00966.175.9772.48622.350.057 5.89 Average Bond Length (Å), Polyhedral Volume V(Å3), Distortion Index (Bond Length), Bond Angle Variance σ()(°2) and Effective Coordination Number CN of the Ln-, M1- and M2-Site Thermal Expansion Figure 3 presents the relative thermal expansion of LnCa4O(BO3)3 obtained from single-crystal X-ray diffraction between 100 and 300 K. As explained earlier, our study employs the standardized orthogonal description (a*sin(β), b, c) for thermal expansion, allowing direct comparison with dilatometry data that follows the same convention. Relative expansion of LnCa4O(BO3)3 obtained from single crystal X-ray diffraction for a) Pr-, b) Nd-, c) Gd-, d) Er-, and e) Tm- Ca4O(BO3)3. Consistent with previous observations, LnCa4O(BO3)3 exhibit anisotropic thermal expansion. This means that the material expands at different rates along different crystallographic axes. The largest expansion coefficient is observed along the c-axis (α33), followed by the a-axis (α11, considering a•sin(β) due to the monoclinic system). The b-axis (α22) exhibits the lowest expansion coefficient. As previously noted on Table 2, smaller lanthanide cations cause a notable decrease in the distortion of their coordination environment (coordination polyhedron) in the Ln site, which might explain the fact that α11 tends to equal α33 in TmCa4O(BO3)3 from our data only. Table 3 summarizes the thermal expansion coefficients and volume expansion coefficient determined in the linear regime between 200 and 300 K. Our results are in good agreement with previously reported values within 15% standard deviation, validating this technique as a reliable and efficient method for studying thermal expansion in crystals. Table 3Linear Thermal and Volume Expansion Compared to Dilatometry Literature Reports in Parentheses(Ln = Pr, Nd, Gd, Er, Tm)ijT-range (K)αij 10–6K–1Pr11200–3008.4 (7.9931)22 6.0 (4.931)33 11.2 (9.4631)V 25.7 (26.5331)Nd11200–3008.2 (8.68,18 8.125)22 5.15 (5.35,18 5.965)33 11.9 (11.67,18 10.35)V 24.8 (25.70,18 24.45)Gd11200–3006.95 (7.8618)22 5.54 (4.4318)33 11.6 (12.1118)V 24.6 (24.5218)Er11200–3006.8 (7.54,18 6.3432)22 3.6 (4.031,18 4.0532)33 11.7 (12.6518, 9.7632)V 22.1 (24.3,18 20.1532)Tm11200–30010.82 (7.9632)22 5.32 (4.8632)33 12.3 (10.3232)V 28.2 (23.1432) Linear Thermal and Volume Expansion Compared to Dilatometry Literature Reports in Parentheses High-Pressure Synchrotron X-ray Diffraction Study at Low Temperatures To evaluate the influence of the site disorder over the structural features, we selected two systems with different behavior: an ordered NdCa4O(BO3)3 structure and a ErCa4O(BO3)3 presenting site disorder. The compressibility behavior at 300, 100, and 50 K of ordered and site-disordered LnCa4O(BO3)3 crystals is shown on Figure 4, NdCa4O(BO3)3 and ErCa4O(BO3)3, respectively. Due to limitations in accessing the reciprocal space during the experiment due to the restrained geometry, the volume error bars were chosen to be as high as four times the estimated standard deviation. This increased uncertainty reflects the potential for systematic errors arising from incomplete data collection. Relative unit cell volume fitted to a second-order Birch–Murnaghan equation of state for a) NdCa4O(BO3)3 and b) ErCa4O(BO3)3. Volume error bars were chosen to be as high as 4 times the estimated standard deviation due to the restrained access to the limited space. Despite the limitations, both data sets exhibit expected trends: Unit cell volume (V0) decreases with decreasing temperature. Unit cell volume (V0) decreases with decreasing temperature. Bulk modulus (K0) increases with decreasing temperature. Bulk modulus (K0) increases with decreasing temperature. At room temperature, NdCa4O(BO3)3 displays a bigger unit cell volume compared to ErCa4O(BO3)3, which aligns well with the larger ionic radius of Nd3+ (as shown in Table 1). Consequently, the observed difference in compressibility at 300 K (K0 = 66(7) GPa for NdCa4O(BO3)3 and K0 = 103(10) GPa for ErCa4O(BO3)3) is consistent with these observations. This finding is also in agreement with the value of K0 = 76 GPa reported by Yu et al.5 Our results contradict those reported by Münchalfen et al.,13 who observed similar K0 values (around 100 GPa) for both NdCa4O(BO3)3 and ErCa4O(BO3)3 using resonant ultrasound spectroscopy. This discrepancy might be due to the different characterization techniques employed. Resonant ultrasound spectroscopy measures the complete electromechanical tensor, potentially introducing additional uncertainties for the low-symmetry space group (monoclinic Cm) compared to our X-ray diffraction method. From the knowledge of the bulk modulus Ko and of the volume thermal expansion αV at room temperature reported above, it is possible to determine the thermodynamic Grüneisen parameter Γ at room temperature from the relation:33 with V0 the unit cell volume at room temperature, CP the heat capacity of the LnCa4O(BO3)3 compounds determined by Yu et al. at room temperature (at 300 K the mass heat capacity is 0.55 J/g.K and 0.71 J/g.K for the Nd and Er compounds, respectively).34 One finds Γ = 0.806 and 0.828 are found for the Nd and Er compounds, respectively, at room temperature, which is a moderate value indicating rather low anharmonicity in LnCa4O(BO3)3 compounds. Figure 5 presents the hk0 reconstructed reciprocal space for NdCa4O(BO3)3 and ErCa4O(BO3)3 at 50 K across the applied pressure range. In addition to the main lattice reflections, distinct satellite reflections are clearly visible for NdCa4O(BO3)3 at all investigated pressures (Figure 5a,b). These satellites can be indexed using a tripled a-unit cell parameter, indicating a superstructure with potential modulation of the NdCa4O(BO3)3 structure as already reported in scheelite tungstates Ln2(WO4)3 at high pressure.35 hk0 reconstruction of the reciprocal space of NdCa4O(BO3)3 at a) 50 K, 0.1 MPa, b) 50 K, 2.5 GPa and of c) ErCa4O(BO3)3 at 50 K, 2.5 GPa. We’ve highlighted the satellite reflections with white circles. This observation is particularly interesting considering the previously reported absence of phase transitions in LnCa4O(BO3)3, suggesting high thermodynamic stability. Our findings hint at a possible low-temperature structure for NdCa4O(BO3)3 that appears to be stable at high pressures. This could have significant implications for the material’s physical properties and warrants further investigation. In contrast, no satellite reflections were observed for ErCa4O(BO3)3 across the entire pressure–temperature range. This could be attributed to the inherent site disorder in ErCa4O(BO3)3, which might hinder the formation of a long-range structural transition. Conclusion In conclusion, this study employed a combined approach of single-crystal X-ray diffraction and high-pressure synchrotron to investigate the relationship between site disorder, thermal expansion, and compressibility in LnCa4O(BO3)3 (Ln = Pr, Nd, Gd, Er, Tm). With a standard deviation of less than 15%, our thermal expansion results validate the single crystal X-ray diffraction technique as a reliable and time-saving approach for crystal analysis. The observed trends in compressibility align with the varying ionic radii of the lanthanide elements. We also found a moderate thermodynamic Grüneisen parameter similar for the Nd and Er compounds. Notably, the identification of satellite reflections and a tripled a-unit cell parameter in NdCa4O(BO3)3 at low temperature suggests a potential difference from the previously assumed ground state for this material. These findings warrant further investigation using complementary techniques to definitively characterize this low-temperature structural behavior and its potential implications for the entire LnCa4O(BO3)3 family. Acknowledgments The authors acknowledge Xpress beamline, proposal number 20225246, Elettra Sincrotrone Trieste. The authors declare no competing financial interest. References