New Series of Red-Light Phosphor Ca_9−xZn_xGd_0.9(PO₄)₇:0.1Eu³⁺ (x = 0–1)

Abstract

In this study, a new series of phosphors, Ca_9−xZn_xGd_0.9(PO₄)₇:0.1Eu³⁺ (x = 0.00–1.00, step dx 0.05), was synthesized, consisting of centro- and non-centrosymmetric phases with β-Ca₃(PO₄)₂-type structure. Crystal structures with space groups R3c (0.00 ≤ x < 0.35) and R$\overline{3}$c (x > 0.8) were determined using X-ray powder diffraction and the method of optical second harmonic generation. In the region 0.35 ≤ x ≤ 0.75, phases R3c and R$\overline{3}$c were present simultaneously. Refinement of the Ca₈ZnGd(PO₄)₇ crystal structure with the Rietveld method showed that 71% of Gd³⁺ ions are in M3 sites and 29% are in M1 sites. A luminescent spectroscopy study of Ca_9−xZn_xGd_0.9(PO₄)₇:0.1Eu³⁺ indicated the energy transfer from the crystalline host to the Gd³⁺ and Eu³⁺ luminescent centers. The maximum Eu³⁺ luminescence intensity corresponds to the composition with x = 1.

1. Introduction

Structural type β-Ca₃(PO₄)₂ (space group (SG): R3c, Z = 21) [1] is a promising base for the development of materials with various interesting properties: luminescent [2], laser [3], non-linear optical [4,5,6], ferroelectric [4,5,6], antiferroelectric [7,8], catalytic [9], bioregenerative [10] and antibacterial [11]. In particular, these compounds can be expected to be of practical interest as narrow-band red light phosphors for the creation of white-light emission diodes (WLEDs), or as bone substitute material with ability to use imaging techniques.

The creation of new synthetic bone substitutes is an important task due to the rapid growth of the aging population and the increasing number of people with diseases of the musculoskeletal system [12]. Ceramics, based on calcium phosphates, are the most common material for filling bone defects [13]. Of all possible calcium orthophosphates (hydroxyapatite Ca₁₀(PO₄)₆(OH)₂ [14], β-tricalcium phosphate β-Ca₃(PO₄)₂ [15], a-tricalcium phosphate a-Ca₃(PO₄)₂ [16], brushite CaHPO₄·2H₂O [17], octacalcium phosphate Ca₈(HPO₄)₂(PO₄)₄·5H₂O [18]), β-Ca₃(PO₄)₂ is one of the most applicable for bone tissue restoration [19].

The development of a highly efficient red phosphor is an urgent task because commercially available WLED based on (Ga,In)N (465 nm) blue-chip [20] and bright-yellow phosphor Y₃Al₂(AlO₄)₃:Ce³⁺ (YAG:Ce³⁺) [21,22] suffers from several disadvantages: low color rendering index (CRI) [23,24,25,26], high correlated color temperature [27], bad influence on a person’s psychological state and health of the eyes [28,29]. Using the UV-chip [30] and varying blue, red [31,32,33] and green phosphors, a WLED with the necessary characteristics can be obtained [34,35].

The unique structure of β-Ca₃(PO₄)₂ has wide possibilities for cationic substitution. For example, incorporation of Eu³⁺ in β-Ca₃(PO₄)₂ makes it possible to obtain red-light phosphors, while co-doping of Eu³⁺ and Gd³⁺ allows for the acquirement of even more superior luminescent properties. In [36] it was shown that the Ca₉Gd_0.1Eu_0.9(PO₄)₇ phosphor emitted red light by 4.13 times brighter than the commercially available Y₂O₃:Eu³⁺. In [37,38], the authors found that sample Ca₈MgGd(PO₄)₇:Eu³⁺ had the highest luminescent intensity in comparison with Ca₈MgY(PO₄)₇:Eu³⁺ and Ca₈MgLa(PO₄)₇:Eu³⁺. This may be related to the sensibilization effect of Gd³⁺ in the energy transfer processes to Eu³⁺. Thus, the simultaneous incorporation of Eu³⁺ and Gd³⁺ can improve the luminescent characteristics of material (due to energy transfer processes from the matrix to Gd³⁺ and further to Eu³⁺-centers), and also can act as a contrast agent for MRI (magnetic resonance imaging) and X-ray dual imaging, which consists of combining two radiographs acquired at two different lanthanides [39].

Further improvement of luminescent properties is possible due to isovalent cation substitutions within the β-Ca₃(PO₄)₂ host. This can also lead to a change in SG (supporting information in [40]), which should be taken into account. Sometimes, the authors did not define the crystal structure of Ca₈MGd(PO₄)₇:Eu³⁺ (M²⁺ = Zn, Mg, Cd) with SG R3c correctly. Substitution Ca²⁺ → Zn²⁺ in the M5 site of β-Ca₃(PO₄)₂ leads to the change in SG R3c → R$\overline{3}$c, the improvement of materials’ luminescent properties (the explanation of this phenomena is provided in detail in [2,41]) and antibacterial characteristics [42] (Figure 1). Given the highest luminescent characteristics of phosphor with M = Zn²⁺ in Ca₈MEu(PO₄)₇ [41] (M = Ca²⁺, Mg²⁺, Zn²⁺, Cd²⁺), it can be assumed that Ca₈ZnGd(PO₄)₇:Eu³⁺ will show improved luminescent properties.

In the present study, a number of solid solutions with the general formula Ca_9−xZn_xGd_0.9Eu_0.1(PO₄)₇ (x = 0.00–1.00, dx = 0.05) are synthesized and examined for the first time. The following questions were in the focus of the study: (1) the boundary of the two-phase region with the change of SG R3c → R$\overline{3}$c at gradual substitution Ca²⁺ → Zn²⁺; (2) the distribution of cations in the Ca₈ZnGd(PO₄)₇ structure; (3) the modification of luminescence properties, in particular the energy transfer to Eu³⁺ with crystal compositions.

2. Results and Discussions

2.1. Elemental Composition and Preliminary Characterization

The quantitative ratio of elements was determined by EDX analysis. The results for samples with x = 0.35, 0.50, 0.75 and 1.00 in Ca_9−xZn_xGd_0.9Eu_0.1(PO₄)₇ series show the insignificant deviation from the theoretical composition.

Table 1. EDX analysis results, estimated crystalline size and average full width at half maximum (FWHM) for 0 2 10 (hkl) reflection at 2θ~31° (CuK_a1K_a2 radiation) of the diffraction peaks for Ca_9−xZn_xGd_0.9Eu_0.1(PO₄)₇.

x	Ca, at.%	Zn, at.%	Gd, at.%	Eu, at.%	Zn:Ca ratio	Crystalline sizes, nm	FWHM, °
0.35	87.29 ± 0.29	3.19 ± 0.16	8.80 ± 0.19	0.72 ± 0.11	0.32 ± 0.03:8.73 ± 0.11	75 ± 10	0.130 ± 0.013
0.50	83.56 ± 0.50	5.74 ± 0.23	9.90 ± 0.18	0.80 ± 0.10	0.57 ± 0.02:8.36 ± 0.05	103 ± 8	0.104 ± 0.006
0.75	81.39 ± 0.46	8.34 ± 0.37	9.44 ± 0.20	0.83 ± 0.13	0.83 ± 0.04:8.14 ± 0.05	108 ± 10	0.101 ± 0.006
1.00	78.25 ± 0.14	11.12 ± 0.22	9.65 ± 0.18	0.98 ± 0.08	1.11 ± 0.02:7.83 ± 0.01	270 ± 20	0.072 ± 0.002

summarizes the results of the EDX analysis.

2.2. Scanning Electron Microscopy

Figure 2 shows the images obtained with the SEM method for x = 0.00, 0.35, 0.75 and 1.00 in the Ca_9−xZn_xGd_0.9Eu_0.1(PO₄)₇ series. The shape of the particles becomes sharper, and the particles form larger agglomerates under Ca²⁺→ Zn²⁺ substitution in Ca_9−xZn_xGd_0.9Eu_0.1(PO₄)₇. This is correlated with the transition from the non-centrosymmetric to centrosymmetric (R3c → R$\overline{3}$c) state.

2.3. PXRD Analysis

A number and peaks position of synthesized solid solutions in Ca_9−xZn_xGd(PO₄)₇:0.1Eu³⁺ PXRD patterns related to similar compounds with β-Ca₃(PO₄)₂-type structure [2]. The absence of impurity phases shows that Gd³⁺ (r_VI = 0.94 Å), Eu³⁺ (r_VI = 0.95 Å) and Zn²⁺ (r_VI = 0.74 Å [43]) ions are completely incorporated into the β-Ca₃(PO₄)₂-type structure.

The unit cells parameters decrease with increasing of Zn²⁺ concentration because the ionic radius of Zn²⁺ (r_VI = 0.74 Å) is smaller than that of Ca²⁺ (r_VI = 1.00 Å) (Table S1). However, this decrease in parameters is nonlinear. Figure 3a shows that the slope of the curve in the region of 0.00 ≤ x < 0.30 is less than in the region of 0.30 ≤ x ≤ 0.80, and there is a sharp jump in the change of parameters for the composition x = 0.50. Such comprehensive behavior of unit cell parameters on Zn²⁺ concentration can be explained by a change in the SG R3c → R$\overline{3}$c. In a routine laboratory experiment, PXRD patterns of compounds with these SGs are indistinguishable [44,45].

In the region x = 0.35–0.75, a decrease in the crystalline size and an increase in the FWHM in comparison with x = 1 is observed. This circumstance may also indicate the coexistence of two phases—R3c and R$\overline{3}$c—further confirmed by the SHG method.

2.4. SHG Study

The above-defined limitations on the existence of regions with different symmetries are consistent with the SHG data (Table S1). As Zn²⁺ ions concentration rises from 0 to 0.35, there is a slight decrease in SHG signal (Figure 3b). More rapid and nonmonotonic decrease of SHG is observed in an interphase region with 0.35 < x < 0.75, where centro- and non-centrosymmetry fragments of whitlockite-like structures are mixed. Very small SHG activity at 0.75 < x ≤ 1.00 corresponds to centrosymmetric phase (SG R$\overline{3}$c) distorted with defects near the smaller boundary of x. For x > 0.75, the SHG signal is null or at the background level in accordance with macroscopic center of symmetry in this phase.

2.5. Crystal Structure Refinement of Ca₈ZnGd(PO₄)₇

Compounds with a β-Ca₃(PO₄)₂-type structure have extended possibilities for cationic substitution. There are six crystallographic positions with different sizes M1–M5 and M6 for Ca²⁺ in β-Ca₃(PO₄)₂. The occupancy of M4 site can vary from 0 to 1, while M6 is always vacant. Large numbers of positions and vacancies suggests a wide opportunity for iso- and heterovalent cationic substitution, including the lanthanoids Ln³⁺. These substitutions may lead to changing of SG from polar to non-polar (Supplementary information of [40]).

Atomic coordinates of phosphate Ca₈ZnLa(PO₄)₇ (SG R$\overline{3}$c) [46] were used as an initial model for the structural refinement of Ca₈ZnGd(PO₄)₇. The M1 and M3 sites (36f) are jointly occupied by La³⁺ and Ca²⁺; M5 (6b) is occupied by Zn²⁺ in Ca₈ZnLa(PO₄)₇. There is no M2 site in this structure with R$\overline{3}$c SG, since M2 is equivalent to M1 (M1 = M2). In centrosymmetric model with R$\overline{3}$c SG, the name of the M3 site was left the same, as for the polar model with R3c SG of β-TCP-type structure. There are two phosphorus atoms in 12c and 36f Wyckoff positions. There is one oxygen atom in 12c Wyckoff position, and the other five are in 36f positions.

At the first step of the refinement, the f-curves for Ca²⁺ (in M1 and M3 sites) and Zn²⁺ (in M5) were used to form the determination of the atoms’ positions. This analysis (

Table 2. Crystallographic data for Ca₈ZnGd(PO₄)₇ (SG R$\overline{3}$ c, Z = 6, T = 293 K).

Sample Composition	Ca8ZnGd(PO4)7
Lattice parameters: a, Å	10.3796(6)
c, Å	37.1316(3)
Unit cell volume, Å3	3460.69(4)
Calculated density, g/cm3	3.452(2)
Data Collection
Diffractometer	Empyrean X-ray
Radiation/wavelength (λ, Å)	CuKα/1.540593, 1.544427
2θ range (o)	10.013–109.983
Step scan (2θ)	0.013
Imax	51281
Number of points	7691
Refinement
Refinement	Rietveld
Background function	Pseudo-Voigt, 16 terms
No. of reflections (all/observed)	488/473
No. of refined parameters/refined atomic parameters	64/54
R and Rw (%) for Bragg reflection (Rall/Robs)	8.81/9.07, 9.10/9.08
RP, RwP, Rexp (%)	4.73, 6.64, 1.96
Max/min residual density (e) (Å3)	1.58/−1.78
nf-CaM1	1.263(8)
nf-CaM3	0.605(5)
nf-ZnM5	0.864(8)

) shows that the Gd³⁺ ions are distributed between the positions M1, M3 (exceeding the maximum occupancy—1 for M1 and 0.5 for M3), while the M5 position is completely occupied by Zn²⁺ ions.

The M3 and P1 sites must be in special Wyckoff positions (18d) (0.5, 0, 0) and (6a) (0, 0, 0.25), respectively, in Ca₈ZnGd(PO₄)₇ (SG R$\overline{3}$c). The structure refinement of this model led to large parameters of atomic displacement, U_iso. = 0.162(2) Ų for Ca²⁺/Gd³⁺ in the M3 site, and U_iso. = 0.173(4) Ų for P1. For this reason, the refinement of the Ca₈ZnGd(PO₄)₇ structure was performed with a shift of the phosphorus atom P1 from a special (6a) position to a half-occupied (12c) position. The Ca²⁺ in M3 was shifted from a special (18d) position to a half-occupied (36f) site. Moreover, the positions of Ca²⁺ and Gd³⁺ in M1 and M3 sites were additionally split. This led to a significant decrease in M1 and M3 U_iso.

After refinement, there is a good agreement between the calculated and experimental X-ray diffraction patterns (Figure 4) with acceptable R-factors (Table 2). Fractional atomic coordinates, site symmetry, isotropic displacement of atomic parameters and site occupation for Ca₈ZnGd(PO₄)₇ are shown in Table S2. The main interatomic distances are shown in Table S3. The distribution of Gd³⁺ ions in Ca₈ZnGd(PO₄)₇ over crystal sites was found to be 71% in M3, and 29% in M1. In Ca₈MgGd(PO₄)₇ [47], the distribution of Gd³⁺ ions was 77% in M3, and 23% in M1 sites. Thus, the M3 site is a preferable location for relatively big Gd³⁺ (r_VI = 0.94 Å) ions.

The average length of the M-O bonds in the M1O₉ and M3O₈ polyhedra are d_<M1-O> = 2.347 Å and d_<M3-O> = 2.526 Å, respectively. At the same time, the deviations of M-O distances from the average values of d_<M1-O> and d_<M3-O> are significant and indicate a large distortion of the polyhedra. Calculated distortion indexes (DI [48]) for M1O₉ and M3O₈ polyhedra are DI_M1O = 0.0362 and DI_M3O = 0.0563. These values were calculated according to the Equation (1)

$$DI=\frac{1}{n}{\displaystyle \sum }_{i=1}^n\frac{d_i}{〈d〉}$$

(1)

where n—number of bonds, d_i—length of a bond and <d>—average bond length for polyhedra. In Ca₉Gd(PO₄)₇, DI’s for M1O₈, M2O₈ and M3O₈ are DI_M1O = 0.0486, DI_M2O = 0.0234 and DI_M3O = 0.0500 (M1 and M2 sites in SG R$\overline{3}$c are equivalent) [49]. It indicates that M3 site (77% Gd³⁺) in Ca₈ZnGd(PO₄)₇ (SG R$\overline{3}$c) is more distorted than M3 site in Ca₉Gd(PO₄)₇ (SG R3c).

2.6. Luminescent Properties of Ca_9−xZn_xGd(PO₄)₇:0.1Eu³⁺

The luminescence spectra of Ca_9−xZn_xGd_0.9Eu_0.1(PO₄)₇ (x = 0, 1) compounds are presented in Figure 5a,b at 6 and 300 K. The spectra consist of a set of narrow emission lines in the region of 305–317 nm and 575–720 nm, related to 4f-4f transitions in Gd³⁺ and Eu³⁺, respectively. The features of the structure of Eu³⁺ emission change with the incorporation of Zn²⁺. It can be shown for the ⁵D_0–⁷F₀ transition (Figure 6), which is not split by the crystal field and sensitive to the number of nonequivalent Eu³⁺ positions. Two peaks can be observed for the sample with x = 0, while only one for x = 1. The number of peaks correlate with the decrease of the number of cationic positions. However, presence of slightly different crystallographic sites with similar crystallographic properties may also result in a single peak consisting of several superimposed peaks. For x = 0 (SG R3c), there are three cationic positions—M1, M2 and M3, while for x = 1 (SG R$\overline{3}$c) there are only two cationic positions—M1 and M3. Analysis of the emission spectra in the 305–317 nm region demonstrates that the number of peaks related to Gd³⁺ also depends on the crystal composition. Presence of additional features such as the shoulder at 310.6 nm and peak at 314 nm depends on the Zn²⁺ content in the sample. This is especially demonstrative when the temperature drops to 6 K (Figure 5b)—the peak at 314 nm becomes more prominent for the compound with x = 0.

The relative intensities of Gd³⁺ and Eu³⁺ emissions depend on the Zn²⁺ content in the sample under VUV excitation (163 nm). The given wavelength is tentatively related to the fundamental absorption region [2], and observed modifications are connected with the features of energy transfer to competitive emission centers. It was found that in the sample with x = 0, the Gd³⁺ luminescence band dominated in the spectrum, while for x = 1, the Eu³⁺ luminescence band intensity increased (Figure 5). Therefore, the energy transfer from the host to Eu³⁺ is more efficient in the sample with Zn²⁺.

PLE spectra of Eu³⁺ and Gd³⁺ emissions are presented in Figure 7. In the excitation spectra of Eu³⁺ a set of narrow lines in the region of 320–500 nm is connected with 4f-4f Eu³⁺ transitions, while the broad band peaking at ~245 nm to charge transfer transitions from the valence band (VB) to 4f Eu states. A narrow excitation band at 273 nm is related to ⁸S_7/2–⁶I_J transitions in Gd³⁺, thus indicating the energy transfer from Gd³⁺ to Eu³⁺. The scheme of Eu³⁺ and Gd³⁺ energy levels position relative to the top of the valence and bottom of the conduction bands is presented in Figure 8. The position of Eu³⁺ and Gd³⁺ 4f ground states position were taken from [2], while 4f excited states were taken from a Dieke diagram [50]. In the excitation spectra of Gd³⁺ emissions, a set of narrow lines is observed at 246, 253 and 273 nm, which are connected with ⁸S_7/2–⁶D_J and ⁸S_7/2–⁶I_J Gd³⁺ transitions. A very weak broad band can be found in the region of 210–250 nm. The position of this band coincides with the position of the charge-transfer band (CTB) in the excitation spectra of Eu³⁺ and shows that the energy transfer from Eu³⁺ to Gd³⁺ is possible as well; however, its efficiency is low. Similarity of the PLE of Ca₉Gd_0.9Eu_0.1(PO₄)₇ and Ca₈ZnGd_0.9Eu_0.1(PO₄)₇ in the energy range up to the fundamental absorption region indicates that electronic states of Zn²⁺ do not form additional channels of energy transfer to the emission centers. In the region of the fundamental absorption edge, the excitation spectra of the studied samples differ considerably. In the sample with Zn²⁺, a broad band peaking at 157 nm (7.89 eV) is observed. The obtained value corresponds to the energy of the direct creation of excitons in Ca₈ZnLn(PO₄)₇ compounds—7.94 eV [2]. Therefore, we also attribute the peak at 7.89 eV to the direct exciton creation in Ca_9−xZn_xGd_0.9Eu_0.1(PO₄)₇ (x = 0, 1). A small hump can also be found at 172 nm (7.21 eV) in the excitation spectrum of Eu³⁺ for the sample with x = 1, while this peak dominates in the excitation spectra of both Eu³⁺ and Gd³⁺ emissions in the sample with x = 0. Previously, a set of sharp peaks related to 4f-5d transitions in Eu³⁺ were observed in the VUV spectral region in some wide bandgap compounds [51]. A similar origin could be supposed for the detected peak at 172 nm. However, this sharp peak is intensive in the excitation spectra of Gd³⁺ emission, thus indicating efficient energy transfer from Eu³⁺ to Gd³⁺. However, according to the analysis of the excitation spectra in UV spectral region such kind of energy transfer is inefficient (CTB is barely observed in the excitation spectra of Gd³⁺ emission) and therefore the attribution of the peak at 172 nm to 4f-5d transitions in Eu³⁺ is low probable. This peak can be tentatively ascribed to excitons localized near Gd³⁺. This may be the reason for the enhanced energy transfer from the host to Gd³⁺ in this compound. In the sample with Zn²⁺, localization of excitons near Gd³⁺ is less effective, which results in the increase in intensity of Eu³⁺ luminescence increases.

3. Experimental Section

3.1. Sample Preparation

Ca_9−xZn_xGd_0.9Eu_0.1(PO₄)₇ (x = 0.00–1.00, dx = 0.05) compounds were synthesized by solid-state method from stoichiometric mixtures of CaHPO₄·2H₂O (99.9%), CaCO₃ (99.9%), ZnO (99.9%) and Ln₂O₃ (Ln³⁺ = Gd, Eu) (99.9%), purchased from Sigma-Aldrich, according to the reaction:

$$\begin{gathered} 7CaHP{O}_4·2H_{2}O+\left(2-x\right)CaC{O}_3+xZnO+0.45G{d}_2{O}_3+0.05E{u}_2{O}_3 \\ \hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}=C{a}_{9-x}Z{n}_{x}G{d}_{0.9}E{u}_{0.1}{\left(P{O}_4\right)}_7+17.5H_{2}O+\left(2-x\right)C{O}_2 \end{gathered}$$

All reagents were controlled by PXRD for the absence of impurity phases. The stoichiometric mixtures were carefully grounded and very slowly heated up to 600 K for 9 h and then annealed at 1273 K for 10 h with several intermediate grindings followed by slow cooling (10 h) to room temperature (T_R).

3.2. Experimental Description

The powder X-ray diffraction (PXRD) patterns were collected on a Thermo ARL X’TRA powder diffractometer (CuK_α radiation, λ = 1.5418 Å, Bragg–Brentano geometry, scintillator detector) at T_R in 2θ range of 5–65° with steps of 0.02°. The phase analysis of the obtained samples was carried out using the Crystallographica Search-March program (Version 2.0.3.1) and JCPDS PDF-2 database. Le Bail analysis was performed using JANA2006 program package [52]. The Debye–Scherrer equation [53] was implemented to count coherent scattering regions (crystalline sizes). LaB₆ (SRM 660c) as a line shape standard was applied to determine instrumental broadening.

Scanning electron microscopy (SEM) study was performed using Tescan VEGA3 microscope equipped with LaB₆ cathode. The SEM images were obtained using secondary electron detector. The analysis of the quantitative of elements concentration was determined by energy-dispersive X-ray (EDX) analysis. Ca_K, Eu_L, Gd_L and Zn_L lines in the EDX spectra were used for the element content determination.

An Empyrean X-ray diffractometer (PANalytical, Almelo, Netherlands) equipped with a PIXcel^3D 2D solid-state hybrid detector providing for counting photons with a high spatial resolution and a high dynamic range was used for registration a powder diffraction patterns. Each pixel is 55 μm × 55 μm and detector array is 256 × 256 pixels. Bragg–Brentano geometry was realized with a Bragg—Brentano^HD X-ray optical module (parabolic multilayer mirror), which monochromatized the primary X-ray beam and provided it with high intensities compare to the commonly used divergence slits and beta filters as well as increase the peak-to-background ratio and minimize excitation of fluorescent radiation from the sample. The X-ray generator (CuK_α-radiation) was operated at 40 kV and 40 mA. Diffraction patterns were recorded in the range of 10÷110⁰ (2θ) with the step size of 0.0131⁰ using continuous scan mode. PIXcel^3D operated in a scanning line detector (1D) mode over its total active length (14 mm), which corresponds 3.3⁰ (2θ) on the Empyrean goniometer with the radius of 240 mm. To avoid an influence of transparency effect of a material, we used a zero-background sample holder consisting of an obliquely cut silicon single crystal with a 32 mm diameter and 2 mm thickness. Rietveld analysis [54] was performed using Jana2006 program package. Illustrations were made using the VESTA program.

The second harmonic generation signal was measured with a Q-switched YAG:Nd laser at λ_ω = 1064 nm in reflection mode for powder series with particle sizes of 40–60 μm. The laser operated with a repetition rate of 10 impulses/s and an impulse duration of about 5 ns. The laser beam was split into two beams to excite the radiation at the halved wavelength, λ_2ω = 532 nm, simultaneously in samples to be measured and in reference sample polycrystalline α-SiO₂. The incident beam peak power was about 0.1 mW on a spot 3 mm in diameter on the surface of the sample.

Photoluminescence emission (PL) and excitation (PLE) spectra were measured using specialized setups for ultraviolet (UV) and vacuum ultraviolet (VUV) luminescence spectroscopy. A DDS-400 deuterium lamp was used as an excitation source for measurements in UV-visible spectral regions (200–500 nm). Excitation wavelengths were monochromatized using primary prism monochromator DMR-4. PLE spectra were measured with 5 nm spectral resolution. The sample was installed in a Janis VPF-800 nitrogen cryostat, which allows for changing the temperature in the range of 80–700 K. The luminescence signal was registered using an ARC SpectraPro-300i monochromator and a Hamamatsu H28259-01 photon counting head used for the measurements of luminescence.

The measurements in UV-VUV spectral region (130–400 nm) were performed using deuterium lamp Hamamatsu L11798 with MgF₂ output windows as an excitation source. The lamp radiation was monochromatized using a McPherson 234/302 vacuum primary monochromator. The PLE spectra were measured with 14 nm resolution. The samples were placed in an ARS vacuum cryostat, which allows for measurements in the temperature range of 5–300 K. Luminescence was registered using Shamrock 303i monochromator equipped with ANDOR iDUS CCD detector with a spectral resolution of 0.3 nm.

4. Conclusions

The as-synthesized Ca_9-xZn_xGd_0.9Eu_0.1(PO₄)₇ solid solutions with β-Ca₃(PO₄)₂-type structures do not contain impurity phases and all crystallize in R3c or R$\overline{3}$c SGs depending on the content of Zn²⁺ ions. Centrosymmetry in Ca_9−xZn_xGd_0.9Eu_0.1(PO₄)₇ phosphors with x > 0.75 was reliably established based on the absence of the SHG effect. Together with data from X-ray analysis, this proves the existence of a pure centrosymmetric variant of the β-Ca₃(PO₄)₂-type compounds with R$\overline{3}$c SG and the interphase region between this phase and compounds with polar SG R3c at lower x. It is worth noting that SGs R3c and R$\overline{3}$c are practically indistinguishable in routine laboratory X-ray diffraction experiments because the regions of centro- and non-centrosymmetric phases can only be definitely determined using additional methods such as SHG. Namely, in a Ca_9−xZn_xGd_0.9Eu_0.1(PO₄)₇ system, the change of symmetry was confirmed by this method. Concentration phase boundaries in this system were established according to zero (or very small) SHG signal for the R$\overline{3}$c phase in interval 0.80 < x ≤ 1.00. Then, at 0.35 ≤ x ≤ 0.75 a transient two-phase state was observed, and at 0.00 ≤ x < 0.35 a non-centrosymmetric SG R3c was fully stabilized according to nearly constant non-zero SHG signal.

In the single-phase phosphate Ca₈ZnGd(PO₄)₇, it was shown using the Rietveld method that Zn²⁺ ions occupy the M5 position, thus relieving a geometric stress in the structure. The M3 position is shifted from the special position (18d) to the semi-occupied position (36f), and P1 position is shifted from the special position (6a) to the semi-occupied general position (12c). The rationale for these shifts was the large values of the U_iso parameters (atomic displacement) in refining the structure under the assumption of locating M3 and P1 in special positions. Thus, the M3 position is shifted from the third-order axis, and its occupancy is equal to 0.5. The distribution of Gd³⁺ ions in Ca₈ZnGd(PO₄)₇ by position in the structure turned out to be 71% in M3 and 29% in M1. This is characteristic of large cations such as Gd³⁺. Thus, we can distinguish the following factors, which probably positively affect the intensity of Eu³⁺ luminescence:

(1)

Shifting of the M3 position from the third-order axis;

(2)

Distortion of M3O₈ polyhedra (local decrease of the symmetry);

(3)

General increase in symmetry of the structure (R3c → R$\overline{3}$c).

The relative intensity of Eu³⁺/Gd³⁺ emissions depend on their composition. It is shown that in compounds with x = 1 the energy transfer from the host to Eu³⁺ is improved, which results in the increase in Eu³⁺ luminescent intensity.

Two nonequivalent positions of Ln³⁺ ions were deduced from the emission spectra of Eu³⁺ as well as Gd³⁺ in the sample with x = 0. Therefore, Gd³⁺ ions could also be used as a luminescent marker to study the crystal structure of the compound.