3.1. Crystal structure
The valence charge is unbalanced when Mn+ ions are substituted by M(n+1). This is undesirable for the phosphor materials and decreases the luminescence intensity. To avoid the charge unbalance and the formation of vacancy in the sample, Li+ ions were employed as charge compensators and added along with Dy3+ ions. The mechanism of charge compensation is based on the fact that two Sr2+ ions are replaced by one Dy3+ ion and one Li+ ion. Therefore, the crystallinity of CaSr2(PO4)2:Dy3+,Li+ improves because the doping with Li+ lowers the crystallization temperature. The XRD patterns of the synthesized CaSr2-2x(PO4)2:xDy3+,xLi+ (x = 0.01, 0.03, 0.06, 0.12, 0.18, 0.24, 0.27 and 0.3) are shown in Fig. 1. The XRD patterns of the synthesized CaSr2-1.5x(PO4)2:xDy3+, (x = 0.01, 0.03, 0.05, 0.07, 0.09, 0.12, 0.15 and 0.18) and CaSr2(PO4)2 are shown in Fig. 2. It is clearly seen that all the XRD patterns can be well fitted with the standard patterns of the Ca3(PO4)2 (JCPDS no. 09-0169) belonging to the trigonal crystal system, with the space group of R3c (161). The introduction of Li+ did not have any significant influence on the structure of the as-prepared samples, which suggests that all samples were crystallized in a single phase. All diffraction peaks shifted to the smaller 2θ angle side (larger d-spacing, i.e. interplanar distances), which can be ascribed to the substitution of Ca2+ by larger Sr2+. Dy3+ and Li+ ions have been successfully embedded into the crystal lattice, and also the Dy3+ and Li+ dopant ions had negligible influence on the structure of the host CaSr2(PO4)2 in varying amounts of doping.
The crystal structure of CaSr2(PO4)2 and different coordination environments for Ca/Sr atoms have already been discussed in the literature. [32] The crystalline structure of CaSr2(PO4)2 is a result of the lattice deformation of Ca3(PO4)2 by partial substitution of Ca2+ with Sr2+. In the crystal structure of CaSr2(PO4)2, the Ca2+ (Sr2+) ions are distributed between five crystallographic sites, four of which can be occupied by both Ca2+ and Sr2+ ions in various ways defined by chemical composition, and one site only can be occupied by Ca2+ ions, which are too small to be occupied by Sr2+ ions. If the number of Sr2+ ions continues to increase in the Ca2−xSr1+x(PO4)2 structure, the replacement of Ca2+ by Sr2+ in all these compounds does not change the space group, and the crystal structure remains unchanged. If all crystallographic sites are occupied by Sr2+ ions, the structure of CaSr2(PO4)2 would change to Sr3(PO4)2, which is different from the former one. In the crystal structure of Sr3(PO4)2 there are two nonequivalent crystallographic sites for the Sr2+ ions (Sr1 and Sr2). The Sr1 atoms are located on the threefold axis, and show 10 coordination sites with 6 oxygen atoms as nearest neighbors belonging to the PO4 group and site symmetry C3v slightly closer than other coordination sites with 8 oxygen atoms. The other Sr2 atoms in the unit cell show an octahedral configuration, coordinated with eight oxygen atoms. The Sr2 site has 12 coordination numbers and is the largest site in this structure from the PO4 tetrahedral on the six-fold axis.
It was found that the formation of CaSr2(PO4)2 structure is more favorable than the formation of Sr3(PO4)2 [33]. This phenomenon can be explained by the fact that the formation energy of CaSr2(PO4)2 from Ca2Sr(PO4)2 is lower by 17.34 eV than the formation energy of CaSr2(PO4)2 from Sr3(PO4)2 [34]. In the crystal structure of CaSr2(PO4)2, the Ca2+/Sr2+ ions are distributed between five crystallographic sites - all of them are occupied by both Ca2+ and Sr2+ ions in various ways. The Sr1/ Ca1, Sr2/Ca2, Sr3/Ca3, Sr4/Ca4 and Sr5/Ca5 positions are coordinated with six, six, seven, three and six oxygen atoms, respectively. It is worth noting that the Ca4 is three0fold coordinated, suggesting weak bonding and the formation of crystal defects.
To investigate the composition and morphology of the material, the CaSr1.88(PO4)2: 0.06Dy3+, 0.06Li+ phosphor was selected as a representative example for measurements. Fig. 3(a) displays the results elemental analysis of the sample measured by the EDS method, and the inset shows SEM images of the CaSr1.88(PO4)2: 0.06Dy3+, 0.06Li+ sample with elemental mapping of CaSr1.88(PO4)2: 0.06Dy3+, 0.06Li+ phosphor provided in Fig. 3(b). The EDS results indicate that the synthesized phosphor is composed of Ca, Sr, P, O, and Dy, which is consistent with the composition of CaSr1.88(PO4)2: 0.06Dy3+, 0.06Li+ material. Besides, the results revealed the contents of each element in Table 1. The SEM image reveals that the as-prepared samples are well-crystallized. The substances synthesized by solid-state method are usually agglomerated, but the sample obtained consist of irregular crystal sizes, which meet well the requirements of phosphor used in w-LEDs. Moreover, the mapping results showed that the distribution of the elements in this material is very uniform.
3.2. Photoluminescence properties
Fig. 4 depicts the excitation spectrum (λem = 578 nm) of various Dy3+-doped CaSr2(PO4)2 phosphors. The excitation spectrum of the Dy3+ ion, monitored at 580 nm emission (corresponding to the 4F9/2 → 6H13/2 transition), consists of several sharp peaks centered at 299, 328, 351, 390, and 455 nm, which are assigned to the f-f transitions of Dy3+ from its ground state 6H15/2 to the excited states 4M17/2, 6P7/2, 6P3/2, 4F7/2 and 4G11/2, respectively. Excitation peaks between 320 and 400 nm indicate that CaSr2(PO4)2:Dy3+ phosphors can be effectively excited by near-UV LED-chips. Fig.5 shows the emission spectra of the CaSr2-2x(PO4)2:xDy3+,xLi+ (x = 0.01–0.3) phosphors, measured at room temperature, wavelength with the excitation of 403 nm (6H15/2 → 6P3/2); the inset shows the dependence of the emission intensity at 486 and 578 nm on the Dy3+ doping concentration. Different from other Dy3+- doped phosphors [35], as the Dy3+ amount increases, the emission intensity increases and reaches a maximum at 6 mol.% Dy3+ doping content, which is considered as the optimum concentration. A higher Dy3+ ion concentration results in a reduction of the luminescence intensity associated with concentration quenching phenomenon. This is because when the concentration of Dy3+ increases, the distance between the ions reduces and enhances the energy transfer cross-relaxation processes between the dopant ions.
Fig. 6(a) shows the emission spectra of the CaSr2-2x(PO4)2:xDy3+,xLi+ (x = 0.01 and 0.03) and CaSr2-2x(PO4)2:xDy3+,xNa+ (x = 0.01 and 0.03) phosphors, under 351 nm excitation. The emission spectra of CaSr2(PO4)2:0.03 Dy 3+,Li+ / Na+ samples include some peaks centered at 486 and 578 nm, which are similar to those mentioned in Fig. 5. This indicates that co-doping with Li+ enhances the luminescence intensity of phosphors and it is more effective than co-doping with Na+. The maximum luminescence intensity is achieved with the content of 3% Li+ doping. Fig. 6(b) illustrates the emission spectra of CaSr2-2x(PO4)2:xDy3+, xLi+ and CaSr2-1.5x(PO4)2:xDy3+ phosphors, showing the beneficial effect of Li+ co-doping on the luminescence intensity of the samples. The phosphors need to produce ion defects to maintain the charge balance because the replacement of Sr2+ with Dy3+ is not equal. However, too many defects produce crystal lattice distortions that reduce the luminescence intensity. Instead, the introduction of Li+ ions means that the charge compensation takes place according to the following formula: 2Sr2+ = Dy3+ + Li+. This limits the number of crystal defects and leads to the enhanced luminescence intensity. Moreover, Li+ co-doping changes symmetry of the local coordination environment and the related crystal field strength. The reduced site symmetry may also lead to the enhanced luminescence intensity.
3.3. Energy transfer mechanism
The quenching of the emission intensity depends on the critical distance Rc, which is the shortest average distance between the nearest dopant Dy3+ ions at a critical concentration xc. The critical distance Rc is described by the following equation [36]:

in which V stands for the volume of the unit cell, xc is the critical concentration of activator ion (Dy3+) beyond the concentration quenching, and N represents the number of host cations in one unit cell. In our case, N = 6 V was estimated to be 3744.14 Å, and xc is 0.06 according to the above discussions. According to the eq. (1), Rc was changed to 27.08 Å (xc = 0.06). It is well-known that exchange interactions play a crucial role in the energy transfer mechanism when the critical distance between the sensitizer and the activator ions is less than 4 Å. With a much higher Rc value, the energy transfer mechanism is considered to be an electric multipolar interaction. Based on the Dexter's theory, if the energy transfer occurs by electric multipolar interactions, then the relationship between the luminescent intensity (I) and the activator concentration (x) can be expressed by the following equation [37]:

where x is the activator concentration, K and β are constants for each interaction at the same excitation. θ is a multipolar interaction constant equal to 3, 6, 8 or 10, corresponding to the nearest-neighbor ions, i.e., dipole–dipole (d–d), dipole–quadrupole (d–q) and quadrupole–quadrupole (q–q) interactions, respectively. We chose the CaSr2-2x(PO4)2:xDy3+,xLi+ (x = 0.06, 0.12, 0.18, 0.24, 0.27 and 0.3) samples for the constant emission intensity measurements at 578 nm exceeding the quenching concentration. The relation of log(I/x) vs. log(x) for the CaSr2-2x(PO4)2:xDy3+,xLi+ (x = 0.06, 0.12, 0.18, 0.24, 0.27 and 0.3) peaks at 486 and 578 nm was plotted and depicted in Fig. 7. All q values are close to 6, hence we can conclude that the quenching mechanism between Dy3+ ions in the CaSr2(PO4)2 samples conforms to the dipole–dipole (d–d) interactions.
3.4. Temperature-dependent luminescent properties
It is well-known that the luminescence intensity for most phosphors decreases if the operating temperature exceeds a certain value (thermal quenching). [38] Thus, the thermal stability of the phosphor is a key issue for high-power w-LEDs. One of the key requirements for a good phosphor is to maintain the performance at the operating temperature of the device. Usually, the luminescence intensity of the phosphors at 423 K with respect to that at room temperature is used to assess the thermal stability. The reason for this has been reported as an increase in the non-radiative transition probability in the configurational coordinated diagram. [39] Fig. 8 shows the temperature dependent emission spectra of the CaSr1.88(PO4)2:0.06Dy3+,0.06Li+ phosphor from 298 to 523 K, at 351 nm excitation; the inset shows the relative emission intensities at 486 and 578 nm as a function of temperature. The shape of the emission bands remains unchanged with increasing temperature (Fig. 8), suggesting that the phosphor has excellent color stability, which is crucial in LEDs or high temperature LEDs. Upon heating the phosphor samples in the temperature range from 298K to 523K, the emission intensity decreased slightly, since the probability of nonradiation is increased and luminescent c enter is released through the crossing point between the excitation state and the ground state, causing the luminescence quenching. [40] Besides, Fig. 9 illustrated the PL intensity of CaSr1.88(PO4)2:0.06Dy3+,0.06Li+ phosphor with respect to time, monitored under 351 nm excitation continuously for 60 min at (a) 100˚C and (b) 150˚C for 30 min at each temperature with a time interval of 5 min, and the emission profile of phosphor maintained at 100˚C and 150˚C continuously for 30 min at each temperature showed that the emission intensity are very stable and remain unchanged, as well as without variation of the emission wavelength. In addition, the emission intensities of CaSr1.88(PO4)2:0.06Dy3+,0.06Li+ maintains 77% of the initial emission intensity corresponding to a temperature of 423K, revealing that the CaSr1.88(PO4)2:0.06Dy3+,0.06Li+ phosphors have good thermal stability, which confirms the stable chromaticity coordinates of CaSr1.88(PO4)2:0.06Dy3+,0.06Li+ phosphors.
3.5. Luminescence decay curves and chromaticity coordinates
Fig. 10 shows the decay curves of Dy3+ emission for CaSr2-2x(PO4)2:xDy3+,xLi+ (x = 0.03, 0.09, 0.15 and 0.18) samples excited at 351 nm and monitored at 578 nm. All decay curves can be well fitted to a bi-exponential decay equation as follows:
I = A1 exp(-x/τ1) + A2 exp(-x/τ2)
where I is the luminescence intensity at time t, A1 and A2 are amplitudes, and t1 and t2 are the luminescence lifetimes. The average emission lifetimes τ were calculated by following formula:

The calculated average emission lifetimes decrease with increasing Dy3+ concentration, i.e. 0.74, 0.47, 0.22 and 0.11 ms for the CaSr2-2x(PO4)2:xDy3+,xLi+ samples with x = 0.03, 0.09, 0.15 and 0.18, respectively. This is due to the decreasing distance between Dy3+–Dy3+ ions, resulting in the observed concentration quenching phenomenon (enhanced cross-relaxations processes), as well as enhanced probability of energy transfer to the luminescence killer sites. Thereby, the luminescence lifetimes of Dy3+ ions are shortened due to the favorable nonradiative energy transfer processes when the Dy3+ concentration increases. The observed bi-exponential character of the decay curves could be because of the non-equal occupation of the cation sites by the emitting activator Dy3+ that has a concentration-dependent preferential occupation in one of the sites as well.
The Commission International de I′Eclairage (CIE) chromaticity coordinates for the representative sample (Ca2Sr1.88(PO4)2:0.06Dy3+,0.06Li+) were calculated based on the corresponding emission spectra, and the results are shown in Fig. 11(a). It can be seen that the coordinates (x = 0.3450, y = 0.3787) are located in the near-white region. In addition, we also calculated the CIE chromaticity coordinates of Ca2Sr1.88(PO4)2:0.06Dy3+,0.06Li+ at different temperature values (Table 1), and we found that with the temperature rising the phosphor exhibits almost no change in the color of emission, which is showed in Fig. 11(b). Thus, the CaSr2-2x(PO4)2:xDy3+,xLi+ material can be potentially used as a white phosphor for w-LEDs in solid-state lighting applications.