3.1. Crystal Structure Analysis and Micro-Morphology
Figure 1a presents the XRD patterns of Lu
2.94Ga
xAl
5−xO
12: 0.06Ce
3+ (x = 0, 1, 2, 3, 4) synthesized at 1450 °C. All of these samples exhibited almost the same XRD patterns to that of the Lu
3Al
5O
12 (LuAG, PDF#73-1368) pure phase. Since no impurities with other diffraction peaks were found in the figure, this indicated that the pure phase of these compounds could be successfully prepared due to the combined effect of the sintering process at 1450 °C/3 h (3 h) and co-solvents. With the increase in the Ga
3+ content, the LuAGG: Ce
3+ structure gradually evolved from Lu
3Al
5O
12 to Lu
3Ga
5O
12 (LuGG, Ga-PDF#73-1372).
The detailed XRD patterns from 32.8° to 33.8° are depicted in
Figure 1b. It is clear that the enlarged diffraction peaks belong to the (4 2 0) crystal face of the sample shifted to lower angles in the process of the increasing replacement of Ga
3+ for Al
3+ in the host, demonstrating that Ga
(1)3+ (62 pm) and Ga
(2)3+ (47 pm) had a bigger cationic radius instead of Al
(1)3+ (53.5 pm), and Al
(2)3+ (39 pm) entered the matrix of LuAG, resulting in lattice expansion.
In order to further explore the possibilities of low energy consumption synthetic processes, the Lu
2.94Ga
xAl
5−xO
12: 0.06Ce
3+ samples with different Ga
3+ concentrations (x = 0, 1, 2, 3, 4) were synthesized at 1350 °C for 3 h, whose XRD patterns are shown in
Figure 2.
As x ≥ 2, there were no diffraction peaks of the impurity phase in the XRD patterns. However, as x is less than 2, diffraction peaks of Al2O3 (PDF#46-1215, 74-1081), AlLuO3 (PDF#24-0690), and LuBO3 (PDF#74-1938) were found in the Lu2.94Al5O12: 0.06Ce3+ and Lu2.94GaAl4O12: 0.06Ce3+ phosphors. It is clearly indicated in the figure that the higher the Ga3+ concentration, the lower the formation temperature of the pure phase LuAGG becomes. The reason for this phenomenon can be explained by the bond dissociation energies. The value of the Ga-O pair (285 kJ/mol) was obviously less than that of the Al-O pair (512 kJ/mol). It can be concluded that Ga3+ ions are easier to combine with O2− ions, promoting the formation of a garnet structure at lower temperature. In contrast, when Al3+ ions completely replace Ga3+ in the raw material, a higher temperature is required to synthesize the Lu3Al5O12 pure phase.
It is well-known that the size, crystallinity, and morphology of phosphor can affect its emission intensity. In this case, the SEM images of the Lu
2.94Ga
2Al
3O
12: 0.06Ce
3+ phosphors sintered at 1350 °C, 1450 °C, and 1550 °C were exhibited in
Figure 3. It can be observed from the figure that, as the temperature rises, the diameter range of the aggregated particle increased from 2–5 to 20–30 μm, and the morphology of the grain changed from an oval shape to an irregular shape. The reason for this phenomenon can be attributed to the increasing concentration of Ga
3+ ions contributing to the decrease in the phase formation temperature. Furthermore, the size of the phosphor particles gradually grew with the increasing sintering temperature, and the sample shrank more seriously. Finally, it resulted in an increased hardness of the samples and led to fragmented crystal grains after grinding.
3.2. Photoluminescent Spectra
In order to reveal the relationship between the luminescence properties, the concentration of Ga
3+ ions, and synthesis temperature, the PL and PLE spectra of the samples synthesized at different temperatures are recorded in
Figure 4a–e.
Figure 4e shows the relationship between the maximum emission intensity and sintering temperature at different Ga
3+ concentrations. It can obviously be observed that the variation trend in the maximum emission intensity of samples with different Ga
3+ concentrations with the increasing temperature was different. Furthermore, when the synthetic temperature was at 1550 °C, the emission intensity of sample with x = 0 was the strongest one relative to other Ga
3+ concentrations of the samples; x = 1 was strongest at 1450 °C, and x = 2 was strongest at 1350 °C.
This phenomenon is consistent with the analysis on the crystal structure and micro-morphology. Generally, after exceeding the phase-forming temperatures, the emission intensity becomes stronger with the increasing particle size and sintering temperature. In general, the defect-free particles of phosphors with spherical or nearly spherical shape exhibit a superior emission intensity. The emission intensity of the as-prepared phosphors reached its maximum at the just phase forming temperature. Subsequently, with the increase in Ga3+ ion concentration, the phase formation temperature point decreases, resulting in the increase in particle aggregation at the same sintering temperature. Finally, it causes the emission intensities to decrease gradually with the increase in the synthesis temperature.
As shown in
Figure 4a–c, the PL spectra of the Lu
3Ga
xAl
5−xO
12: Ce
3+ phosphors excited at 450 nm displayed a gradual blue-shift with the increasing Ga
3+ content. Although similar phenomena have also been found in previous studies of the luminescent properties of YAG: Ce
3+ [
13,
14], few have been systematically researched for the effect of Ga
3+ doping on the luminescent properties in LuAG: Ce
3+ at different temperatures. As exhibited in
Figure 4b, the variety of peak position in the emission spectra display from 508 nm to 482 nm was accompanied by the change of x from 0 to 3. The blue-shift offset was 26 nm. Based on the above results, it can be predicted that a series of Ga
3+-substituted lutetium aluminum garnet will be excellent stuffing into the blue-green cavity of the emission spectrum around 490 nm, which is conducive to a high CRI. Furthermore, the full-width at half-maximum (FWHM) of the emission profile with the introduction of Ga
3+ ions broadened a small amount from 75 nm (x = 0) to 79 nm (x = 3), which will also be helpful in obtaining a high quality illumination source in w-LED applications.
As illustrated in
Figure 4d, monitored at the optimal emission, all of the PLE spectra with similar morphological characteristics included two broad absorption bands. These two broad bands, peaking around 350 nm and 430 nm, were assigned to the 4f→5d2,1 spin-allowed transition of the Ce
3+ ions. By comparing these curves, it can be found that with the increase in the Ga
3+ ion concentration, the higher energy excitation band (4f-5d2) and the lower energy excitation band (4f-5d1) moved toward each other gradually. This phenomenon for the spectral shift of PL and PLE can be determined by two possible factors: the crystal field splitting (CFS) and the nephelauxetic effect (NE) [
13,
21,
22].
According to reports by Robertson et al., the degree of crystal field splitting (
Dq) can be defined as follows [
23]:
where
Dq represents the degree of energy level separation;
Z is the charge of anion;
e is the charge of electron;
r is the radius of the
d wavelength; and
R is the bond length between the activator and coordination ion. As we know, a garnet-type structure with the chemical formula A
3B
2C
3O
12 consists of three different structures: AO
8 (distorted dodecahedral), BO
6 (octahedral), and CO
4 (tetrahedral) framework. Ascribed to the larger cationic radius of Ga
3+ ions than Al
3+ ions, the cell parameters and cell volume (V) will increase with the replacement of Al
3+ by Ga
3+ at the tetrahedral and octahedral sites. This means that the bond distance (
RCe−O) will become longer when Ce
3+ ions enter the Lu
3+ site. It can be inferred from the inverse proportional formula between
Dq and
R that the decrease in the crystal field intensity will cause the emission wavelength to move to the high-energy region, resulting in a continuous blue-shift of the PL spectra.
In addition, the positive correlation exists between the centroid shift (ε
c) of the Ce
3+ 5d levels and the anion polarizability (α
sp) based on the reports of Morrison and Dorenbos [
19,
20]. The anion polarizability in Ce
3+ doped oxide compounds can be estimated by the following equation [
24,
25],
Meanwhile, the
αsp is affected by the joint action of the average electronegativity
χav of the cations in the host:
where
χi denotes the electronegativity of cation
i with formal charge
zi;
Nc represents the summation over all cations in the compound;
Na and
γi stand for the number and the formal negative charge of the anion in the formula, respectively. By Plugging Pauling-type electronegativity values in Equation (3), the values of χ
av were calculated to be 1.483, 1.508, 1.533, 1.558, and 1.583 in Lu
2.94Ga
xAl
5−xO
12: 0.06Ce
3+ for x = 0, 1, 2, 3, and 4, respectively. Based on Equation (2), the increasing χ
av will lead to the decrease in the anion polarizability α
sp. As a result, the centroid of the Ce
3+ 5d levels shift to a higher-energy position, resulting in a gradual blue-shift of the emission and excitation bands with the increase in the Ga
3+ ion concentration from 0 to 4.
Figure 4d also shows that the 4f–5d1 excitation band of Lu
2.94Ga
3Al
2O
12: 0.06Ce
3+ exhibited a larger blue-shift (20 nm), while the 4f–5d2 excitation band showed a smaller redshift (9 nm) compared to the excitation spectral band of Lu
2.94Al
5O
12: 0.06Ce
3+. Similar phenomena have been found in other research, but have been rarely mentioned [
11,
19]. This phenomenon of spectral shift caused by Ga
3+ doping in Re
3Al
5O
12:Ce
3+ host (Re = Lu, Y) can be explained by the combined effects of CFS and NE. The experimental results combined with previous studies are listed in
Table 1 [
13,
18,
19,
20]. It can be seen that a larger shift (represented by ↑↑) of the excitation band, which was attributed to the elevated energy level, was co-promoted by CFS and NE. The contradictory role of CFS and NE on the energy level led to a slighter shift (represented by ↓) of the excitation band. Meanwhile, this result also illustrates that CFS plays a more significant role than that of the increase in NE in the phosphors that are activated by Ce
3+.
In addition, a clear evolutionary trend could also be observed in
Figure 4 where the PL and PLE intensities of the samples with the increase in x gradually weakened and even emitted no radiation (as x = 4). The thermoelectric ionization model can be used to explain this phenomenon. According to previous studies [
26], the valence band energy E
V rose from −9.6 eV for Lu
3Al
5O
12 to −9.0 eV for Lu
3Ga
5O
12, but the conduction band energy E
C declined from −1.7 eV to −2.5 eV. As a result, with the increase in the Ga
3+ ion content, the 5d1 level will gradually approach the conduction band, and the 5d2 level will even enter the conduction band. Namely, the energy gap between the conduction band and the lowest energy of the 5d excited state of the host lattice will be reduced. This result enhances the probability of thermal ionization, and ultimately leads to a reduction in the emission intensity. The schematic diagram is described in
Figure 5.
In summary, the luminescence intensity of samples was determined by the synthesis temperature and energy gap. For one thing, the increasing sintering temperature will result in the aggregation of particles in Lu3GaxAl5−xO12: Ce3+, which will worsen the emission strength after grinding; for another, the energy gap between the conduction band and the lowest energy of the 5d excited state of host lattice is reduced with the increase in the Ga3+ concentration as it forces enormous excited electrons into entering the conduction band and leads to the poor or even non-luminous intensity.
3.3. Temperature Dependence Luminescence Intensity
When w-LEDs are in service, ambient temperatures of phosphors can typically reach 150 °C. In order to ensure the high energy efficiency and color stability of w-LEDs, the excellent thermal stability of phosphor is a crucial performance parameter. The temperature dependence of the normalized emission intensity for the Lu
2.94Ga
xAl
5−xO
12: Ce
3+0.06 (x = 0, 1, 2, 3) phosphors at T = 30–225 °C are shown in
Figure 6. Lu
2.94Ga
4AlO
12: Ce
3+0.06 is not shown in the figure as it had almost no luminous properties. Statistical error of the thermal quenching measurement equipment was estimated as error bars.
As can be observed in the figure, the integrated normalized emission intensity of the phosphors exhibited a quite monotonic decrease with the increase in the ambient temperature. Meanwhile, as the Ga
3+ ion concentration increased, the normalized emission intensity tended to decrease, indicating that the thermal stability gradually deteriorated. It can also be observed that when the temperature was measured at 150 °C, the emission intensities of these samples dropped to 83%, 81%, and 79% of the initial value (25 °C), demonstrating an excellent stability against the thermal quenching of x = 0, 1, and 2, respectively. However, the emission intensity declined sharply as x = 3 with the rising temperature. The reason for this phenomenon is the gradual intensification of the interaction between the phonons and electrons with increasing temperature, leading to an increase in the probability of non-radiative transition. As a widely accepted theory, the thermally activated crossover mechanism is used to explain the weakening in thermal stability of Ga
3+-doped phosphor. As shown in
Figure 7a, excited electrons with sufficient thermal activation energy(ΔE) can pass through the cross-relaxation point between the 5d1 level and 4f level, and return relatively easily to the ground state with the increasing temperature in a non-radiative relaxation manner. However, according to the spectral analysis results, with the increase in the Ga
3+ ion concentration, the blue-shift phenomenon of the emission spectrum indicates that the Stokes shift will decrease, which will lead to the increase in the ΔE value. In general, the thermal stability of phosphor with a larger value of ΔE will improve rather than deteriorate. However, this theory and its conjectures are contrary to the experimental results of thermal quenching. Herein, it can be inferred that the thermally activated crossover mechanism is not the dominant factor influenced on the thermal quenching of the system. Therefore, the thermal ionization mechanism was used to elaborate this abnormal thermal quenching phenomenon.
As shown in
Figure 7b, the energy gap (ΔE
T) for the thermal ionization process, which is between the conduction band and the lowest energy of the 5d excited state, will narrow with the increase in the Ga
3+ ion concentration. It can be inferred that the probability of 5d electrons enter into the conduction band based on thermal fluctuation will become more obvious with the increase in temperature. Hence, compared to the thermally activated crossover process, thermal ionization is more likely to cause non-radiative transition of 5d1 level electrons in the excited state and explains the reason quite well as to why there was a sudden deterioration phenomenon of thermal stability from x = 2 to 3. Therefore, the thermal ionization process, instead of the thermally activated crossover process, becomes the main reason for the strong thermal quenching for Lu
2.94Ga
xAl
5−xO
12: Ce
3+0.06 with the increase in the Ga
3+ content.
3.4. Packaging Test
Aiming to evaluate the effectiveness of the cyan-green phosphor LuAGG: Ce
3+(x = 0, 1, 2, 3) in practical application, a specimen of the w-LED device was obtained by mixing the as-prepared Lu
2.94Ga
2Al
3O
12: Ce
3+0.06 sample with a certain amount of commercial red phosphor Sr
2Si
5N
8: Eu
2+ and encapsulating them on the blue GaN chip (450 nm). The emission spectrum of the as-fabricated w-LED device driven by a current of 20 mA is present in
Figure 8a. Obviously, compared with the YAG: Ce
3+ and LuAG: Ce
3+ spectra [
11,
25], it can be conspicuously observed that the blue-green cavity of the emission spectrum of the as-prepared phosphor was filled and the spectrum covered the whole visible region smoothly.
The experimental LED parameters of the specimen LED device marked as a, b c, and d including the CIE chromaticity coordinates, CRI, CCT, and luminous efficiency are listed in
Table 2 and drawn in
Figure 8b.
The results in
Table 2 show that with the increase in the Ga
3+ content in the as-prepared cyan-green phosphor, the CRI of the specimen LED device increased at first, and then dropped, which reached its maximum value at x = 2; the color temperature and the light efficiency decreased gradually. The reason for the increase in the CRI value is that the blue-green cavity of the emission spectrum was filled by an appropriate substitution of Ga
3+ for Al
3+, which could significantly improve the color index properties of the white LED device. However, due to the significant reduction in the emission intensity of the samples and the spectral absorption of red phosphors to the samples, these may be the reasons as to why the luminous efficiency of the w-LED device decreased when x = 3.
The CIE chromaticity diagram of these samples is drawn in
Figure 8b, which demonstrates a succession of transitions from x = 0.236, y = 0.608 (green region) to x = 0.162, y = 0.464 (cyan region) with the Al
3+ sites gradually occupied by Ga
3+ ions. The results indicate that the as-prepared phosphors are hopeful in making a tunable blue-green emission under the excitation of blue light in order to compensate for the lack of spectrum in the 470–510 nm region.