1 Introduction

Inorganic rare-earth borates have received intensive interests owing to their wide range of applications in nonlinear optics (NLO), laser hosts and photoluminescence materials [1,2,3]. Among the most attractive features of borate crystals, the key characteristics are their transparency in the wide range from mid-infrared to deep-ultraviolet [4], high optical damage threshold [5] and suitable physico-chemical stability [6]. Hitherto, Na3La9O3(BO3)8 [7], La2CaB10O19 [8], and YAl3(BO3)4 [9] have been widely used as NLO crystals for second harmonic generation (SHG). Since the ionic radii of the rare-earth elements are close to each other, these rare-earth borates are favorable to host optical active cations such as Ce3+, Nd3+ and Yb3+ [10]. Therefore, the rare-earth borates can further serve as an array of functional materials in optics: Ce3+:Li6Y(BO3)3 [11] as scintillator materials for neutron detection, Nd3+:YAl3(BO3)4 [12] as crystals for green laser emission, and Yb3+:LiGd6O5(BO3)3 [13] as ultrashort-pulse lasers. Rare-earth orthoborate YBO3 doped with Eu3+ and Tb3+ [14] are appropriate phosphors in plasma-display panels due to their high luminous efficacy and thermally stable luminescence yield.

Recently, Pr3+ cation-doping materials have drawn an intense research attention due to their promising applications in white-light emitting diodes [15,16,17]. The Pr3+ cation is able to release rich fluorescence spectral lines in the red, orange, green, and blue regions, which originate from the electronic transition between the 4f inner-shell configurations. For instance, in the alkali-metal borates steady and high-efficiency red emission have been detected in Pr3+:KSr4(BO3)3 powder [18], Pr3+:Na3La9O3(BO3)8 crystal [19] and Pr3+ doped lithium borate glasses [20] when activated by blue or ultraviolet light due to 3P0 → 3H6 and 1D2 → 3H4 transitions. Despite much works on Pr3+-doped borates, purely praseodymium borates have not been extensively studied. In search of novel rare-earth borate compounds the present investigation focus on A2O–Pr2O3–B2O3 (A = alkali metal) quasi-ternary systems and their crystal-, chemico-physical properties. To the best of our knowledge, only Li3Pr2(BO3)3 [21], LiPr6O5(BO3)3 [22], Na3Pr(BO3)2 [23] and K9Li3Pr3(BO3)7 [24] were reported in these ternary systems. The alkali-metal praseodymium borate K2Pr2O(BO3)2 has been found to be a new member of the K2La2(BO3)2O family [25]. The report particularly emphasizes on the synthesis, crystal structure, vibrational, optical and thermal properties of K2Pr2O(BO3)2.

2 Experimental

2.1 Synthesis

Single crystals of K2Pr2O(BO3)2 were obtained by means of the flux-assisted solid state reaction method. The starting materials K2CO3, PrO2 and H3BO3 with a molar ratio of 1.5: 2: 3, respectively, were thoroughly ground in an agate mortar, and put into a platinum crucible. The mixture of the starting chemicals can be regarded as the target product K2Pr2O(BO3)2 together with the assisted-flux KBO2 in the molar ratio of 1:1. The crucible was heated up to 1173 K in a muffle furnace with a heating rate of 100 K/h and maintained at that temperature for 12 h. Afterward, the crucible was cooled down to room temperature with a cooling rate of 300 K/h. The obtained greenish single crystals are found to be chunk-shaped with a dimension of tens of micrometer.

The polycrystalline samples of K2Pr2O(BO3)2 were prepared using a solid-state synthesis method in a platinum crucible. First, a stoichiometric batch of intimately mixed K2CO3, PrO2 and H3BO3 (analytical grade) was heated at 773 K for 12 h to completely decompose K2CO3 and H3BO3. The resulting samples were then ground in agate mortar and heated further at 1123 K for 48 h. A 10% excess molar ratio of K2CO3 was added into the samples during the sintering process to compensate the evaporation of potassium. The mixtures were repeatedly heated at 1123 K with intermediate grindings until their powder X-ray diffraction pattern confirms a pure phase.

2.2 X-ray single crystal diffraction

The small crystals of K2Pr2O(BO3)2 were isolated by mechanical fragmentation and picked using a polarization microscope. Single-crystal diffraction data were collected on a Bruker D8-Venture diffractometer in Kappa geometry with \( {\text{Mo}}_{{{\text{k}}{\overline{\upalpha }}}}\) radiation (\(\uplambda _{{{\text{k}}{\overline{\upalpha }}}}\) = 71.0747(6) pm) at 297(2) K. A numerical absorption correction was applied to the intensity data sets. The systematic extinctions and |E2 − 1| statistics suggested the monoclinic space group P21/c; therefore the structure determination of K2Pr2O(BO3)2 was performed in this space group. The structure was finely solved via the intrinsic phasing method and successfully refined on F2 by full-matrix least-square methods with the ShelxT and Shelxle program packages [26, 27]. All atoms were refined with anisotropic displacement parameters and the final difference Fourier synthesis did not reveal any significant residual electron density. All relevant details of the data collection and the refinement are listed in Table 1, atomic coordinates and displacement parameters in Table 2 and the bond valence sums (BVSs) in Table 3.

Table 1 Crystal structure determination of K2Pr2O(BO3)2 in the space group P21/c (No. 14) from single crystal and powder X-ray diffraction data
Full size table
Table 2 Crystal structural data of K2Pr2O(BO3)2
Full size table
Table 3 Interatomic bond distance/pm and bond valence (BV/v.u.) and bond valence sum (BVS/v.u.) of the PrO8, KO6,7 and BO3 polyhedra in K2Pr2O(BO3)2
Full size table

Further details of the crystal structure information can be obtained from FIZ Karlsruhe, 76,344 Eggenstein Leopoldshafen, Germany (fax: +49-7247-808-666; e-mail: crysdata@fiz-karlsruhe.de) on quoting the deposition number CSD-1978814.

2.3 X-ray powder diffraction

The powder X-ray diffraction (PXRD) pattern was recorded on a Panalytical X’Pert Pro powder diffractometer using Bragg–Brentano geometry with CuKα1,2 (λkα1 = 154.05929(5) pm, λkα2 = 154.4414(2) pm) radiation. The measurement was carried out at ambient condition in a range between 5° and 130° 2θ with a step size of 0.0167° and a data collection time of 30 s/step. The fundamental parameter approach, where the fundamental parameters were fitted against a LaB6 standard material, was applied for the Rietveld refinement using “DiffracPlus Topas 6” software (Bruker AXS GmbH, Karlsruhe, Germany). The starting lattice parameters and atomic coordinates were taken from the results of the single crystal structure determination. The average crystallite size was calculated from all observed X-ray reflections, which is described as LVol(IB) by the TOPAS suite. Of notes, LVol(IB) refers to the volume-weighted mean of the coherently diffracted domain size using the integral breadth for the description of the reflection profile.

2.4 UV/Vis spectroscopy

The UV/Vis diffuse reflectance measurement was collected from 200 to 850 nm with a step of 1 nm on a UV-2700 spectrophotometer (Shimadzu, Japan) equipped with an ISR-2600 plus two-detector integrating sphere (Pike Technologies, USA). The baseline correction was carried out against BaSO4 powder.

2.5 Vibration spectroscopy

The Fourier transform infrared (FTIR) spectrum was recorded on a Bruker IFS66v/S spectrometer using the standard KBr method between 370 and 4000 cm−1. KBr pallets consist of 2 mg sample mixed with 200 mg KBr (sample) and 200 mg KBr (reference), pressed at 100 kN, forming disks of 12 mm in diameter.

Due to laser induced susceptibility of the 4f-electronic transitions the Raman spectra were measured at ambient condition using at least three lasers (532 nm, 633 nm and 785 nm), and found that the band numbers and the respective frequency positions are laser-independent. To avoid the absorption areas (see UV–Vis reflectance spectrum below), the 785 nm laser was chosen for better spectral resolution. Temperature-dependent Raman spectra were recorded on a LabRam ARAMIS (Horiba Jobin Yvon) Micro-Raman spectrometer equipped with a laser working at 785 nm and less than 5 mW power on the sample surface. The use of a 50 × long working distance objective (Olympus) with a numerical aperture of 0.55 provides a focus spot of about 2 µm diameter when closing the confocal hole to 200 μm. Raman spectra were collected in the range 85–1700 cm−1 with a spectral resolution of approximately 1.1 cm−1 using a moving grating of 1800 grooves/mm and a thermoelectrically cooled CCD detector (Synapse, 1024 × 256 pixels). The spectral positions were calibrated against the Raman mode of Si before and after the sample measurements. The position of the Si peak was repeatedly measured against the Rayleigh line (0.0 cm−1) yielding a value of 520.7 ± 0.1 cm−1. The linearity of the spectrometer was calibrated against the emission lines of a neon lamp. For the low-temperature measurements, a pressed pellet of powder sample was placed on a Linkam cooling stage (THMS600) attached to a pump (LNP95 Cooling Pump) that provides a continuous flow of liquid nitrogen. The measurements were carried out between 78 and 300 K. A ramp rate of 5 K/min and a holding time of 5 min were followed to properly equilibrate the temperature. For the spectrum at ambient condition the baseline was linearly corrected, bands were fitted with Pseudo-Voigt lineshape.

2.6 Thermal analysis

Simultaneous thermogravimetric analysis/differential scanning calorimetry (TGA/DSC) measurements were performed on TGA/DSC 3+ STARe system of Mettler Toledo. The sample was measured with a heating rate of 10 K/min and a continuous N2 flow of 20 mL/min from 300 to 1473 K. Afterward, the data were normalized to their respective mass. Approximately 16.159 mg of K2Pr2O(BO3)2 was measured relative to an empty corundum crucible as the reference. A drift correction was applied based on empty crucible data.

3 Results and discussion

3.1 Crystal structure of K2Pr2O(BO3)2

K2Pr2O(BO3)2 crystallizes in a new structure type in the monoclinic space group P21/c (no. 14) with lattice parameters of a = 1133.77(3) pm, b = 660.47(2) pm, c = 1071.30(3) pm, β = 117.072(1)°, and Z = 4. The structural building units of K2Pr2O(BO3)2, as shown in Fig. 1, can be regarded as PrO8 dodecahedra, KO6 octahedra and KO7 pentagonal bipyramid together with isolated trigonal planar BO3 groups. Both Pr(1) and Pr(2) form distorted PrO8 dodecahedra, where the Pr–O bond lengths vary from 235.9(2) pm to 260.1(2) pm in Pr(1)O8, and from 234.6(2) to 274.6(2) pm in Pr(2)O8. The Pr(1)O8 dodecahedra are axially connected with its neighbors via sharing two triangular faces to construct an infinite zig-zag chain along the b-axis (Fig. 1b). These zig-zag chains are further joined together by edge-sharing to form infinite two dimensional Pr(1)O(3+3+2)/2 layers with elongated hexagonal voids in the bc-plane. Such a 2D REOn layer with “honeycomb-like” voids was also found in other rare-earth borates, for instance, in Na3La2(BO3)3 [28]. By contrast, two Pr(2)O8 dodecahedra share their O(5)-O(5) edge to construct Pr(2)2O14 dimers that fill into those “honeycomb-like” voids (Fig. 1c). It is interesting to mention that O(5), the isolated oxygen anion in K2Pr2O(BO3)2, connects only to Pr-atoms and plays an important role in the formation of a Pr(2)2O14 dimer. Therefore, [Pr2O(BO3)2]2− sheets are composed of Pr(1)O(3+3+2)/2 layers, Pr(2)2O14 dimers, isolated B(1)O3 and B(2)O3 planar groups via corner and edge sharing in the ac-plane. Two adjacent [Pr2O(BO3)2]2− sheets in the structure are regulated by the inversion center (Fig. 1d). Two types of potassium K(1)and K(2) are found to occupy the inter-sheet space and separate the [Pr2O(BO3)2]2− sheets along the a-axis. Similar stacking modes were also observed in other layered alkali-metal rare-earth borates such as K9Li3Nd3(BO3)9 [24] and Rb2LiNd(BO3)2 [24]. K(1) is six-fold coordinated to form a distorted K(1)O6 octahedron while K(2) connects to seven oxygens in exhibition of a distorted pentagonal bipyramid K(2)O7. The K–O distances range from 265.9(2) to 282.4(2) pm with mean values of 273.6(2) pm and 282.3(2) pm for K(1)O6 and K(2)O7 polyhedra, respectively. These values fit well with the sum of the ionic radii of oxygen and potassium in sixfold and sevenfold coordination [29]. In the unit cell, both B(1) and B(2) atoms are three-coordinated to oxygen atoms, forming isolated planar BO3 groups. The B-O distances range from 134.2(5) to 141.0(6) pm with an averaged bond length of 137.0(3) pm for B(1)O3 and 137.9(3) pm for B(2)O3, respectively. These values are in agreement with those of other alkali-metal rare-earth borate-containing isolated BO3 groups, for instance 137.1 pm in K3Sm(BO3)2 [30], 137.4 pm in Li3K3Y7(BO3)9 [31] and 137.6 pm in K9Li3Nd3(BO3)7 [24]. The bond valence sums (BVS) calculation was performed for K2Pr2O(BO3)2 using the Bondstr software of the FullProf suite [32], which are listed in Table 3. The structural BVSs of the atoms correspond well with the formal integer charge of the respective atoms. To confirm the single crystal structure of its bulk representative as well the purity of the as-synthesized K2Pr2O(BO3)2 polycrystalline powder sample, X-ray powder data Rietveld refinements were performed. The corresponding Rietveld plot is shown in Fig. 2. The metric parameters (a = 1135.00(3) pm, b = 660.641(15) pm, c = 1072.03(3) pm, β = 117.128(2)° and V = 715.41(3) × 106 pm3) are in excellent agreement with those obtained from the single crystal structure determination (Tables 1, 2).

Fig. 1
figure 1

Crystal structure of K2Pr2O(BO3)2, showing polyhedral structural units (a) connectivity of Pr(1)O8 dodecahedra (b), Pr(2)2O14 polyhedral dimer (c), and view of the structure of K2Pr2O(BO3)2 along the b-axis (d)

Full size image
Fig. 2
figure 2

X-ray powder data Rietveld plots of K2Pr2O(BO3)2 and of the decomposition products as inset

Full size image

3.2 UV/Vis spectrum

The UV/Vis reflectance spectrum of K2Pr2O(BO3)2 in the range of 200–850 nm is shown in Fig. 3. The optical absorption in the visible range accounts for the green color of K2Pr2O(BO3)2 and also corresponds well to the presence of Pr3+ cation. All visible absorption bands in K2Pr2O(BO3)2 result from the inner-shell 4f2-configuration electronic transitions from the ground state (3H4) to various excited states. According to the energy level scheme proposed by Dieke and Crosswhite [33], these bands from 400 to 520 nm can be assigned, respectively to 3H4 → 3P2, 3P1, and 3P0, while the band in the range of 550–650 nm ascribed to 3H4 → 1D2. Of particular notes, we clearly observe the Stark splitting of 3P2, 3P1, and 3P0 and 1D2 multiplets, which may help calculate the crystal field levels of Pr3+ cation in K2Pr2O(BO3)2. The monotonic drop of the reflectance between 290 and 270 nm (Fig. 3) corresponds to the valence-to-conduction-band absorption edge. The Kubelka–Munk function [34] treatment followed by the Tauc method [35] are often used to estimate the corresponding bandgap energy by finding the intercept of the abscissa from the following relations:

$$ \begin{aligned} & F\left( R \right) = \frac{{\left( {1 - R} \right)^{2} }}{2R} \\ & F\left( R \right)\left( {h\nu } \right) = B\left( {h\nu - E_{g} } \right)^{n} \\ \end{aligned} $$

where R is the reflectance [%] in the UV/Vis spectra, h the Planck’s constant, ν the frequency of light, Eg the bandgap in eV, and n-the type of optical transition. That is, n = 2 for an indirect transition (plotted as [F(R)·(hν)]1/2 vs. hν) and n = 1/2 for a direct transition (plotted as [F(R)·(hν)]2 vs. hν). The intercepts of the abscissa (Fig. 3) demonstrates bandgap values of 4.40(1) eV and 4.50(1) eV for an indirect and direct transition, respectively. These bandgap values of K2Pr2O(BO3)2 are smaller than those of 4.96 eV for λ-PrBO3 [36] and 6.32 eV for Rb2LiLaB2O6 [37]. Recently, the derivation of absorption spectrum fitting (DASF) method was proposed by Souri et al. [38] for thin films to calculate the bandgap energies without any presumption of the nature of the transition. Notably, the DASF method can also be expressed as follows proposed by Kirsch et al. [39, 40] for powder samples:

$$ \frac{{{\text{dInF}}\left( {\text{R}} \right)}}{{{\text{d}}h\nu }} = \frac{{\text{n}}}{h\nu - Eg} $$
Fig. 3
figure 3

Reflectance spectrum of K2Pr2O(BO3)2 (top left), Tauc plots for indirect (bottom left) and direct (top right) optical transitions, and DASF plot (bottom right)

Full size image

The obtained bandgap of 4.49(1) eV using the DASF method is similar to the determined bandgap of 4.50(1) eV for a direct transition within the estimated uncertainty. Therefore, the combined approach suggests a direct bandgap transition for K2Pr2O(BO3)2.

3.3 FTIR and Raman spectra

The observed FTIR and Raman spectra along with the fitted compound and component models are shown in Fig. 4. Since the crystal structure of K2Pr2O(BO3)2 adopts P21/c space group, factor group analysis predicts 156 vibrational modes at the zone center (39Ag + 39Au + 39Bg + 39Bu), where 75 modes (38Au + 37Bu) are IR active, 78 modes (39Ag + 39Bg) are Raman active and 3 are acoustic modes (Au + 2Bu). To fit the observed IR spectrum, it requires 41 bands. The bands between 400 and 550 cm−1 can be attributed to bending of Pr–O, which was explicitly characterized in other praseodymium oxides [41, 42] The bands between 600 and 1600 cm−1 owing to vibration of planar BO33− groups, can be categorized into four different types [37, 43, 44]: the in-plane bending (ν4; 550–700 cm−1), out-of-plane bending (ν2; 740–780 cm−1) of BO3, B–O symmetric stretching (ν1; ~ 903 cm−1) and B–O asymmetric stretching (ν3; 1000 cm−1 and 1500 cm−1). Clearly, the ν2 and ν3 modes have much stronger absorption than that of ν3 and ν1. Group analysis of an ideal planar BO3 group possesses D3h symmetry, where v2 and v3 are IR active, whereas the ν1 is IR inactive. The clear appearance of ν1 in the IR spectrum indicates that the BO3 groups in K2Pr2O(BO3)2 are distorted from an ideal symmetry.

Fig. 4
figure 4

FTIR (left top) and Raman (left bottom) spectra of K2Pr2O(BO3)2 at ambient condition along with fitted compound and component models. Temperature-dependent Raman spectra (right); temperature increases upward from 78 to 300 K

Full size image

The observed Raman spectrum of K2Pr2O(BO3)2 at ambient condition could be fitted with 56 component peaks (Fig. 4). Several intense bands below 550 cm−1 can be ascribed to the bending and stretching vibrations of K–O and Pr–O bonds as well as the lattice vibrations. The bands observed in the range of 600–700 cm−1 and 700–800 cm−1 correspond to the ν4 and ν2 modes, respectively. The most intense bands occur at 915 cm−1 resulting from the ν1 mode. The symmetric stretching mode (ν1) of BO33− group is a strong Raman active vibration as known from other Pr3+-containing orthoborates such as λ-PrBO3 [37, 43, 44] and KCaPr(BO3)2 [45]. Above 1000 cm−1, one peak locating at 1050 cm−1 and several border bands are found and can be ascribed to the ν3 mode of BO33− groups. Of notes, the Raman spectrum also clearly confirms the vibrational features of BO3 and PrOx groups in K2Pr2O(BO3)2. The temperature-dependent Raman spectra of K2Pr2O(BO3)2 from 78 to 300 K (Fig. 4) show that the overall global quasi-harmonic change is not significant within the investigated low-temperature range. For instance, the intense band at 916.2(1) cm−1 at 78 K shifts only to 914.6(1) cm−1 at 300 K. Any phase transition driven by any optical soft-mode was not observed during the sample cooling from 300 to 78 K. Extrapolation of the frequency of this mode down to 0 K suggests that the compound K2Pr2O(BO3)2 would be stable, a guideline for the phonon calculation using density functional theory calculation.

3.4 Thermal analysis

For the bulk single-crystal growth of inorganic compounds, determination of the melting point or the decomposition temperature is indispensable. The thermal behavior of K2Pr2O(BO3)2 was investigated by using simultaneous TGA and DSC methods. As shown in Fig. 5, the DSC curve demonstrates one sharp endothermic signal peak at 1318(1) K as well as one border and tiny peaks after 1400 K. The distinct endotherm in the range of 1273(3) K and 1330(3) K corresponds to melting or thermal decomposition. From the TG curve a weight-loss of 21.96(1)% is found to be at 1273(3) K. After the TGA /DSC experiment, X-ray powder data Rietveld refinement (inset Fig. 2) demonstrates that the residual consists of Pr26O27(BO3)8 and Pr6O11 phases. It is worthwhile to note that we used Nd26O27(BO3)8 [46] as the starting model and replaced Nd -atom with Pr-atom in the same Wyckoff position during the Rietveld refinement. For the obtained isotypic Pr26O27(BO3)8 lattice parameters of a = 676.160(17) pm, b = 1269.56(4) pm, c = 1432.28(3) pm, α = 89.9979(17)°, β = 99.9008(17)°, γ = 89.9995(19)° were calculated. From the distinct weight loss during the heating process one can safely assume that K2Pr2O(BO3)2 decomposes into Pr26O27(BO3)8, Pr6O11 and potassium borates which eventually evaporates at the high-temperature regime. The TGA /DSC also confirms that K2Pr2O(BO3)2 possess an incongruent melting point, suggesting that the flux method would be necessary for the growth of its bulk crystal.

Fig. 5
figure 5

Thermogravimetry (TG) and differential scanning calorimetry (DSC) curves of K2Pr2O(BO3)2

Full size image

4 Conclusion

The detailed structural, spectroscopic and thermal analysis reveal that the synthesized alkali metal praseodymium borate K2Pr2O(BO3)2 belongs to the new member of the K2La2(BO3)2O family [25] providing also a new structure type. The crystal structure consists of 2D [Pr2O(BO3)2]2− sheets and K+ cations in a layered stacking along the a-axis. The greenish color stems from the electronic transitions from the ground state to various excited states within the 4f2-configuration rather than from the fundamental absorption edge that lies in the UV-region. Since Pr3+ cation is of interest due to its photoluminescence property, K2Pr2O(BO3)2 may be a prospective red phosphor under the radiation of blue light [15,16,17]. Density functional theory (DFT) calculation would be necessary to distinguish the observed Raman and infrared bands between phonon at the zone-center and photoluminescence as well as for a fuller assignment of the modes. However, the 4f-electrons of the rare-earth elements are over delocalized, sometimes leading to unrealistic/unphysical potentials and may provide inaccurate phonon calculations. Therefore, additional cares must be taken of in calculating the phonon spectra of the compound, which may be computationally very expensive. Given that K2Pr2O(BO3)2 is an incongruent melting compound, the bulk crystals may be grown via high-temperature flux method, which would help elucidate possible anisotropic properties.