The Effect of Heat Treatment on the Emission Color of P-Doped Ca2SiO4 Phosphor

In a series of (Ca2–x/2–yEuy□x/2)(Si1–xPx)O4 (x = 0.06, 0.02 ≤ y ≤0.5), various color-emitting phosphors were successfully synthesized by a solid-state reaction. These phosphors were characterized by photoluminescence (PL) spectroscopy, X-ray powder diffractometry, transmission electron microscopy, and X-ray absorption fine structure spectroscopy. We evaluated the effect of heat treatment on PL properties with various annealing temperatures at 1373–1773 K for 4 h before/after reduction treatment from Eu3+ to Eu2+. In the red-emitting (Ca1.95Eu3+0.02□0.03)(Si0.94P0.06)O4+δ phosphor, the highest PL intensity exhibited when it was annealed at 1773 K. On the other hand, in the green-emitting (Ca1.95Eu2+0.02□0.03)(Si0.94P0.06)O4 phosphor, the highest PL intensity was realized when it was annealed at 1473 K and consequently treated under a reductive atmosphere. With increasing annealing temperature, the emission peak wavelength steadily decreased. Furthermore, with increasing Eu2+ content, the emission peak wavelength increased, with the color of emitting light becoming yellowish. Thus, the PL properties of the phosphors were affected by both the structural change from β to α’L, which occurred by heat treatment, and the amount of doped Eu ions.


Introduction
Rare-earth-doped dicalcium silicate (Ca 2 SiO 4 , C 2 S) phosphors have been investigated by several researchers for application to white light-emitting diodes (LEDs) [1][2][3][4][5]. Two types of phosphors have been reported so far: Eu 3+ -doped and Eu 2+ -doped C 2 S phosphors. The former phosphors are characterized by the emission of red light due to the transition of 5 D 0 -7 F 2 for the Eu 3+ ion [3]. Thus, the luminescence originates from the 4f-4f dipole transitions of the Eu 3+ ion, and hence the wavelength of the emission is almost the same among the various phosphors with different host materials. For example, a group of A 2 SiO 4 : Eu 3+ (A = Ca, Sr, Ba) phosphors showed similar emission spectra at around the maximum wavelength of 620 nm [4]. On the other hand, Eu 2+ -doped C 2 S phosphors are generally useful for a wide range of applications, because their luminescent colors, due to the f-d transition of the Eu 2+ ion, are tunable by the crystal structures and/or compositions of the host materials. Recently, a crystal-site engineering technique was reported, which enabled us to customize the luminescent colors of C 2 S:Eu 2+ phosphors [5].
A series of structural studies of P-doped C 2 S crystals has demonstrated that the incorporation of P most effectively stabilized the high-temperature phases of C 2 S [6]. Hence, the authors present the idea of utilizing the P-doped C 2 S as the host material of Eu 2+ -activated phosphors. In a previous study, the relationship between the photoluminescence (PL) intensities and crystal structures of P-doped C 2 S phosphors was reported [7]. The doping of P ions in the green-emitting C 2 S:Eu 2+ phosphors were found to effectively increase the PL intensities. The polymorphs of C 2 S established are, in the order of increasing temperature, γ (orthorhombic), β (monoclinic), α' L (orthorhombic), α' H (orthorhombic), and α (trigonal) [8]. The phase transition temperatures are, during the heating process, 963 K for β to α' L , 1433 K for α' L to α' H , and 1698 K for α' H to α. The β to α' L phase transition of the cooling of doped C 2 S has been reported to be thermoelastic martensitic [9], hence the stabilized phases at ambient temperature systematically changed from β, β+α' L to α' L with increasing concentrations of foreign ions such as Sr 2+ , Ba 2+ , and/or P 5+ in C 2 S [10,11]. Thus, the change of crystal structures by heat treatment must be potentially able to improve the photoluminescence (PL) properties of the phosphors. In fact, we have succeeded in the further improvement of the PL intensities by the annealing of the green-emitting P-doped C 2 S phosphor [12].
In the present study, we successfully synthesized P-doped C 2 S phosphors showing various emission colors (red, green, and yellow) by a solid-state reaction. The PL intensities were compared between the phosphors with different activators of Eu 3+ or Eu 2+ ions that were annealed at 1373-1773 K. We also discussed the relationship between the PL intensities and crystal structures.

Red-Emitting P-Doped C 2 S:Eu 3+ Phosphor
In a previous study, we prepared a series of (Ca 1.98-x/2 Eu 2+ 0.02 x/2 )(Si 1-x P x )O 4 (P-doped C 2 S:Eu 2+ ) phosphors with 0 ≤ x ≤ 0.20 (y = 0.02), and investigated the effect of P 5+ -ion doping on the PL properties. Here, the box in the chemical formula represents vacancies according to a previous paper [6]. The P-doped C 2 S:Eu 2+ phosphor with x = 0.06 exhibited the highest PL intensity among those with 0 ≤ x ≤ 0.20 [12].
In the present study, we prepared the Eu 3+ -activated red-emitting phosphor with a chemical formula of (Ca 1.95 Eu 3+ 0.02 0.03 )(Si 0.94 P 0.06 )O 4+δ , and focused on the effect of annealing on its PL property. Figure 1 shows the X-ray diffraction (XRD) patterns of the red-emitting phosphor annealed at various temperatures from 1373 K to 1773 K for 4 h. With increasing annealing temperature, the relative amount of the α' L phase with respect to the β phase increased. This is caused by the phase transition from β to α' L during annealing. Figure 2 shows the emission and excitation spectra of the red-emitting phosphors annealed at 1373 K, 1573 K, and 1773 K. Red light emission of the Eu 3+ -activated phosphors were observed at an excitation wavelength of 394 nm due to the intraconfigurational 7 F 0 -5 L 6 transition. We observed the highest PL intensity for the specimen that was annealed at 1773 K. The sharp and strong emission peaks, induced by the transitions in the Eu 3+ ion, appeared at around 594 nm ( 5 D 0 -7 F 1 , magnetic-dipole), 625 nm ( 5 D 0 -7 F 2 , electric-dipole), and 706 nm ( 5 D 0 -7 F 4 , electric-dipole). We plotted the emission intensity at 594 nm (=I 594 nm) and that at 706 nm (=I 706 nm) with excitation by 394 nm in Figure 3. The intensity ratios of I 706 nm/I 594 nm (= 5 D 0 -7 F 4 / 5 D 0 -7 F 1 ) with different annealing temperatures are also plotted in the figure. Both I 594 nm-and I 706 nm -values steadily increased with increasing annealing temperature, while there was no significant change in the intensity ratio of I 706 nm/I 594 nm. The electric-dipole transitions, which is the f-f transition from 5 D 0 to 7 F 2,4 , are, according to the theory of selection rule [13], affected by the site symmetry of the crystal fields around the Eu 3+ ion. From the present results, the Eu 3+ ion would be in the same Ca site, although the site environment of the Eu 3+ ion must be changed from the β phase to the α' L phase. Interestingly, in our previous paper, bright red emission due to hypersensitive 5 D 0 -7 F 2 (electric-dipole) transition and very small 5 D 0 -7 F 4 and 5 D 0 -7 F 1 transitions were observed under the ultraviolet irradiation of 541 nm or 398 nm to Li 1.11 Ta 0.89 Ti 0.11 O 3 phosphor [14]. The eccentricity of the Eu 3+ positions in [(Li, Eu)O 12 ] polyhedron thus seems to be closely related to the PL efficiency of the doped lithium tantalate phosphors [15].

Green-Emitting P-Doped C 2 S:Eu 2+ Phosphor
The phase compositions of green-emitting (Ca 1 each other. Figure 5 shows the relationship between the PL intensity and annealing temperature in green-emitting (Ca 1.95 Eu 2+ 0.02 0.03 )(Si 0.94 P 0.06 )O 4 phosphors. The broad excitation spectra were composed of the four bands at 310 nm, 340 nm, 360 nm, and 400 nm, each of which could be attributed to the location of an Eu 2+ ion in the crystal structures of the βand α' L -phases [16,17]. The differences in peak shapes of the excitation spectra are caused by the differences in crystal structures of the host materials. The emission or excitation wavelengths were measured by monitoring them at the maximum wavelengths, as shown in Table 1. The emission peak wavelength decreased with increasing annealing temperature. This is caused by the expansion of Eu-O bond lengths by phase transition [17], because the 4f-5d transition of the Eu 2+ ion must be closely related to the crystal field. We observed the highest PL intensity for the specimen annealed at 1473 K, of which the phase composition was both β and α' L .
Materials 2017, 10, 1000 4 of 9 four bands at 310 nm, 340 nm, 360 nm, and 400 nm, each of which could be attributed to the location of an Eu 2+ ion in the crystal structures of the β-and α'L-phases [16,17]. The differences in peak shapes of the excitation spectra are caused by the differences in crystal structures of the host materials. The emission or excitation wavelengths were measured by monitoring them at the maximum wavelengths, as shown in Table 1. The emission peak wavelength decreased with increasing annealing temperature. This is caused by the expansion of Eu-O bond lengths by phase transition [17], because the 4f-5d transition of the Eu 2+ ion must be closely related to the crystal field. We observed the highest PL intensity for the specimen annealed at 1473 K, of which the phase composition was both β and α'L.   We investigated the profile intensity data of the phosphor sample annealed at 1473 K by the Rietveld method, and determined the precise phase composition. The structural parameters were taken from those reported by Jost et al. [18] for β-C2S, Udagawa et al. [19] for α'L-C2S, and Dickens and Brown [20] for Ca5P2SiO12. The sample was found to be composed of both the β-and α'L-phases  four bands at 310 nm, 340 nm, 360 nm, and 400 nm, each of which could be attributed to the location of an Eu 2+ ion in the crystal structures of the β-and α'L-phases [16,17]. The differences in peak shapes of the excitation spectra are caused by the differences in crystal structures of the host materials. The emission or excitation wavelengths were measured by monitoring them at the maximum wavelengths, as shown in Table 1. The emission peak wavelength decreased with increasing annealing temperature. This is caused by the expansion of Eu-O bond lengths by phase transition [17], because the 4f-5d transition of the Eu 2+ ion must be closely related to the crystal field. We observed the highest PL intensity for the specimen annealed at 1473 K, of which the phase composition was both β and α'L.   We investigated the profile intensity data of the phosphor sample annealed at 1473 K by the Rietveld method, and determined the precise phase composition. The structural parameters were taken from those reported by Jost et al. [18] for β-C2S, Udagawa et al. [19] for α'L-C2S, and Dickens and Brown [20] for Ca5P2SiO12. The sample was found to be composed of both the β-and α'L-phases  We investigated the profile intensity data of the phosphor sample annealed at 1473 K by the Rietveld method, and determined the precise phase composition. The structural parameters were taken from those reported by Jost et al. [18] for β-C 2 S, Udagawa et al. [19] for α' L -C 2 S, and Dickens and Brown [20] for Ca 5 P 2 SiO 12 . The sample was found to be composed of both the βand α' L -phases with a small amount of Ca 5 P 2 SiO 12 as the impurity phase. The phase composition was determined to be 75.3 mol % β, 23.5 mol % α' L , and 1.2 mol % Ca 5 P 2 SiO 12 ( Figure 6), under the condition of each effective particle radii being 5 µm. Because the highest PL intensity for the present sample among the phosphors annealed at 1373-1673 K, the co-existence of the βand α' L -phases would be essentially important for the enhancement of the PL intensity of green-emitting P-doped C 2 S:Eu 2+ phosphors. This is because the coherent interphase boundaries between α' L and β would store the strain energy, which would distort the crystal lattices and provide a favorable environment for the efficient PL emission of the Eu 2+ ion. with a small amount of Ca5P2SiO12 as the impurity phase. The phase composition was determined to be 75.3 mol % β, 23.5 mol % α'L, and 1.2 mol % Ca5P2SiO12 (Figure 6), under the condition of each effective particle radii being 5 μm. Because the highest PL intensity for the present sample among the phosphors annealed at 1373-1673 K, the co-existence of the β-and α'L-phases would be essentially important for the enhancement of the PL intensity of green-emitting P-doped C2S:Eu 2+ phosphors. This is because the coherent interphase boundaries between α'L and β would store the strain energy, which would distort the crystal lattices and provide a favorable environment for the efficient PL emission of the Eu 2+ ion.  The β-phase region showed a pseudo-merohedral (polysynthetic) twin structure, with the splitting of reflections for the SAED pattern. The reflection indices are based on the β-phase lattice. The coherent grain boundary (depicted by G. B. in Figure 7) was parallel to the (011) plane of the β phase (a = 5.513, b = 6.758, c = 10.460 nm, β = 117.27 degrees) and the (001) plane of the α'L phase (a = 20.470, b = 9. 390, c = 5.437 nm). We examined the XRD profile intensity data by the Rietveld method to find that the geometrical difference in the boundaries between (011) and (001) was as small as 0.023 nm. The coherent interphase boundaries between α'L and β would store the strain energy, which would distort the crystal lattices and, consequently, provide a favorable environment for the efficient PL emission of the Eu 2+ ion in the green-emitting phosphor.   Figure 7) was parallel to the (011) plane of the β phase (a = 5.513, b = 6.758, c = 10.460 nm, β = 117.27 degrees) and the (001) plane of the α' L phase (a = 20.470, b = 9. 390, c = 5.437 nm). We examined the XRD profile intensity data by the Rietveld method to find that the geometrical difference in the boundaries between (011) and (001) was as small as 0.023 nm. The coherent interphase boundaries between α' L and β would store the strain energy, which would distort the crystal lattices and, consequently, provide a favorable environment for the efficient PL emission of the Eu 2+ ion in the green-emitting phosphor.
(a = 5.513, b = 6.758, c = 10.460 nm, β = 117.27 degrees) and the (001) plane of the α'L phase (a = 20.470, b = 9. 390, c = 5.437 nm). We examined the XRD profile intensity data by the Rietveld method to find that the geometrical difference in the boundaries between (011) and (001) was as small as 0.023 nm. The coherent interphase boundaries between α'L and β would store the strain energy, which would distort the crystal lattices and, consequently, provide a favorable environment for the efficient PL emission of the Eu 2+ ion in the green-emitting phosphor.  In order to determine the environment of the Eu ion, X-ray absorption fine structure (XAFS) spectra were obtained with fluorescence mode at room temperature. Figure 8 shows the radial structure function of (Ca 1.95 Eu 2+ 0.02 0.03 )(Si 0.94 P 0.06 )O 4 phosphor annealed at 1473 K. The simulation was performed using the software Artemis [21] (Figure 8b), in which the Eu ion is substituted in the Ca(1n) site with 10 coordination of β-Ca 2 SiO 4 (structural parameters by   [18]. Figure 8c shows the calculated peaks as a sum of each peak shown in Figure 8b. With radial distance less than 2 Å, a distinct doublet peak appeared in the simulation in Figure 8c, although the corresponding peak was single for the measurement data in Figure 8a. The doping of the Eu ion in the Ca site would equalize the Ca(Eu)-O lengths, and hence we observed the single peak for the radial structure function in Figure 8a. This result agreed well with that obtained by Sato et al. (2014) [4]. For the doping of a large amount of Eu ions, the Eu ions would preferentially occupy the Ca (2n) site, leading to the change in emission color from green to yellow, and then to red, because the PL property is closely related to the crystal field of the host material.
Materials 2017, 10, 1000 6 of 9 In order to determine the environment of the Eu ion, X-ray absorption fine structure (XAFS) spectra were obtained with fluorescence mode at room temperature. Figure 8 shows the radial structure function of (Ca1.95Eu 2+ 0.02□0.03)(Si0.94P0.06)O4 phosphor annealed at 1473 K. The simulation was performed using the software Artemis [21] (Figure 8b), in which the Eu ion is substituted in the Ca(1n) site with 10 coordination of β-Ca2SiO4 (structural parameters by   [18]. Figure 8c shows the calculated peaks as a sum of each peak shown in Figure 8b. With radial distance less than 2 Å, a distinct doublet peak appeared in the simulation in Figure 8c, although the corresponding peak was single for the measurement data in Figure 8a. The doping of the Eu ion in the Ca site would equalize the Ca(Eu)-O lengths, and hence we observed the single peak for the radial structure function in Figure 8a. This result agreed well with that obtained by Sato et al. (2014) [4]. For the doping of a large amount of Eu ions, the Eu ions would preferentially occupy the Ca (2n) site, leading to the change in emission color from green to yellow, and then to red, because the PL property is closely related to the crystal field of the host material.

Green-Yellow-Emitting P-Doped C2S Phosphor with a Large Amount of Eu 2+ Ions
In the formula of (Ca1.97-yEu 2+ y□0.03)(Si0.94P0.06)O4, the phosphors were synthesized with a large Eu 2+ content (y = 0.1, 0.3, and 0.5). Figure 9 shows XRD patterns of the phosphors that were annealed at 1473 K and consequently reduced under a 97% Ar-3% H2 reductive atmosphere. With increasing Eu content, it was obvious that the structural change occurred after the reduction process. Before the reduction treatment, the β and α'L phases, together with the Ca5(PO4)2(SiO4) phase, coexisted. However, after the reduction, the main constituent phase was the α'L phase. For the doping of Eu with a smaller amount, the phase compositions were almost the same between the phosphors before and after the reduction treatment, as shown in Figures 1 and 4. Therefore, the phase compositions

Green-Yellow-Emitting P-Doped C 2 S Phosphor with a Large Amount of Eu 2+ Ions
In the formula of (Ca 1.97-y Eu 2+ y 0.03 )(Si 0.94 P 0.06 )O 4 , the phosphors were synthesized with a large Eu 2+ content (y = 0.1, 0.3, and 0.5). Figure 9 shows XRD patterns of the phosphors that were annealed at 1473 K and consequently reduced under a 97% Ar-3% H 2 reductive atmosphere. With increasing Eu content, it was obvious that the structural change occurred after the reduction process. Before the reduction treatment, the β and α' L phases, together with the Ca 5 (PO 4 ) 2 (SiO 4 ) phase, coexisted. However, after the reduction, the main constituent phase was the α' L phase. For the doping of Eu with a smaller amount, the phase compositions were almost the same between the phosphors before and after the reduction treatment, as shown in Figures 1 and 4. Therefore, the phase compositions were effectively affected by the doping of a large amount of Eu 2+ , the ionic radius of which is larger than that of Eu 3+ . Figure 10 shows the PL intensities of phosphors with the Eu 2+ content of y = 0.02, 0.1, 0.3, and 0.5. With increasing Eu content, the PL intensity steadily decreased due to the effect of concentration quenching. The top peaks shifted to the longer wavelength sides; they are 512 nm with y = 0.02, 525 nm with y = 0.1, and 533 nm with y = 0.3. With the Eu 2+ content of y = 0.5, two peak-tops appeared at around 550 nm for one top and around 640 nm for the other. As described above, the emission peak wavelength decreased by the expansion of Eu-O bond lengths by the phase transition from β to α' L [18]. The emission color would change in accordance with the site preference of Eu 2+ ions between the distinct two sites of Ca(1n) and Ca(2n). Sato et al. reported that the emission color of Ca 2-x Eu x SiO 4 phosphor changed from green-yellow to deep-red with increasing Eu 2+ content [4], which was closely related to the peculiar coordination environments of Eu 2+ in the two Ca sites of the host C 2 S [22]. which was closely related to the peculiar coordination environments of Eu 2+ in the two Ca sites of the host C2S [22].

Materials and Methods
The phosphors were synthesized by a conventional solid-state reaction without flux regents because their compositions should be precisely controlled. The starting materials used were the chemicals of CaCO3, SiO2, CaHPO4•2H2O, and Eu2O3 (>99.9% grade) for the preparation of the Eu 2+   which was closely related to the peculiar coordination environments of Eu 2+ in the two Ca sites of the host C2S [22].

Materials and Methods
The phosphors were synthesized by a conventional solid-state reaction without flux regents because their compositions should be precisely controlled. The starting materials used were the chemicals of CaCO3, SiO2, CaHPO4•2H2O, and Eu2O3 (>99.9% grade) for the preparation of the Eu 2+