Elemental Fluctuation in Gd₃Al₂Ga₃O₁₂:Ce Crystals Imposed by Li⁺ and Mg²⁺ Co-Doping: The Impact on Defects, Luminescence, and Scintillation Properties

Abstract

This research revealed the response of Ga and Al sublattices to the incorporation of mismatching substituents in Gd₃Al₂Ga₃O₁₂:Ce single crystals. Incompatible in size and charge, Li⁺ and Mg²⁺ substituents violated configurational entropy. This led to lattice distortion and triggered structural rearrangements. The radial fluctuation of the Ga and Al elements was proven by multi-elemental energy-dispersive X-ray spectroscopy mapping and elemental composition analysis. Further evidence was observed by the shift of the exciton creation energy toward higher energy in the vacuum ultraviolet excitation spectra recorded with synchrotron radiation. In the Li⁺ and Mg²⁺ co-doped samples, the crystal core was depleted with Ga atoms and enriched with Al elements. The crystal rim showed the opposite behavior. The change in thermoluminescence peak positions revealed a different mechanism for the formation of localized traps. As a result, Li⁺ co-doping slightly improved the light yield value, but at the same time decelerated the scintillation decay time. On the contrary, Mg²⁺ co-doping markedly diminished scintillation parameters.

1. Introduction

Investigation of the structure–property relationship plays an important role in the development of new functional materials and provides new physical insights. Crystal engineering strategies are the basis for the tuning and understanding of material properties, as well as the discovery and structural design of new luminescent materials for emerging applications [1]. RE₃M₅O₁₂ (RE = Lu, Y, Gd, Tb; M = Al, Ga, Sc) simple or complex garnets form a wide range of inorganic compounds. Their physical and chemical properties are closely related to the structure, composition, and synthesis methods [2,3,4,5,6,7]. Composition flexibility is a unique feature of RE₃M₅O₁₂ garnets, as it allows extreme variations in the design of structure and composition in relation to applications. Recently, complex Ce³⁺ doped garnets have received a lot of attention as promising scintillators for medical imaging [4], phosphors for light-emitting laser/diode applications [6], and persistent phosphors [8]. Such intense research on aluminum garnets resulted in the invention of a complex Gd₃Al₂Ga₃O₁₂:Ce (GAGG:Ce) single crystal with an excellent light yield (LY) of 46,000 ph/MeV [9]. The further development of GAGG:Ce crystals resulted in achieving an extremely high LY of around 55,000 ph/MeV with a scintillation decay time of about 90–120 ns [10]. The GAGG host lattice, on the other hand, features several structural defects that degrade scintillation performance and lead to the rise time, the slow component of scintillation decay time, and afterglow. These drawbacks exclude the use of the GAGG:Ce single crystal in medical imaging applications [11,12]. Further research efforts have focused on improving the scintillation timing performance for the GAGG:Ce crystal. This was achieved by aliovalent co-doping with Mg²⁺ or Ca²⁺ ions. Owing to the charge imbalance, the Ce⁴⁺ centers are formed to maintain the charge neutrality in the crystal. Stable Ce⁴⁺ ions are characterized by the presence of a distinctive wide charge transfer absorption band in the 320–200 nm spectral region [13]. Furthermore, stable Ce⁴⁺ ions are also an excellent electron trap that prevents electron capture in various intrinsic defects, creating an alternative channel for fast radiative recombination. This significantly reduces the slow component in scintillation decay and suppresses afterglow [11]. However, at the same time, a significant decrease in light yield is observed [11]. Another optimization based on monovalent Li⁺ co-doping was applied to improve both timing parameters and LY [14,15,16]. However, Li⁺ co-doping, similar to Mg²⁺, imposed the formation of stable Ce⁴⁺ ions [14,15,17]. The appearance of additional Ce⁴⁺ related broad charge transfer absorption below 320 nm and the reduction in Ce³⁺ absorption strength with the increasing Li⁺ content indicated the change in the Ce valence state from +3 to +4 [14,15]. Furthermore, the acceleration in the scintillation decay time and improvement in timing resolution along with simultaneous reduction in the light yield were also reported. However, this effect was much weaker compared with Mg²⁺ co-doping. Recent research [17] revealed that the incorporation of aliovalent and mismatched Mg²⁺ ions into the GAGG:Ce lattice imposed the formation of specific types of defects that significantly changed radial elemental homogeneity in the crystal. This significantly influenced the luminescence and scintillation properties of GAGG:Ce,Mg crystals. Similarly, the recent EPR and NMR study [15] revealed a detailed mechanism of charge imbalance compensation in Li⁺ co-doped YAG:Ce single crystals.

The spectral and temporal properties of emission centers strongly depend on the host lattice composition. Much research is devoted to how doping and co-doping concentration levels change optical and luminescence parameters and impose asymmetric crystal field perturbation and defect formation. In such research, the investigation of how doping or co-doping imposes structural rearrangement and a change in elemental homogeneity in the crystal is quite often omitted. The elemental distribution significantly influences the spectral properties of the emission centers in the crystals. Therefore, this research is devoted to studying the influence of the co-doping of Li⁺ and Mg²⁺ ions on the radial distribution of the Gd, Ga, Al, and O atoms in Gd₃Ga₃Al₂O₁₂:Ce³⁺ single crystals grown using the μ-PD method. The changes in atom distributions imposed by Li⁺ and Mg²⁺ co-doping are compared and discussed. The optical, luminescence, and scintillation features are compared and discussed on the elemental inhomogeneity imposed by co-doping. The change in elemental homogeneity is additionally confirmed by photoluminescence properties measured with synchrotron radiation. Thermally stimulated luminescence (TSL) measurements between 10 and 670 K revealed the formation of different types of defects in Li⁺ and Mg²⁺ co-doped crystals.

2. Methodology

2.1. Crystal Growth

The powders of Gd₂O₃, α-Al₂O₃, Ga₂O₃, CeO₂, Li₂CO₃, and MgCO₃ (99.99%, Iwatani Corporation, Tokyo, Japan) were mixed stoichiometrically and added directly to the Ir crucible according to the nominal melt compositions: Gd_2.9925Ce_0.0075Al₂Ga₃O₁₂, Gd_2.9925Ce_0.0075Al₂Ga_2.995Mg_0.005O₁₂, and Gd_2.9895CeLi_0.003Al₂Ga₃O₁₂. Therefore, the nominal concentration for Ce³⁺ was 0.25 at. % (with respect to Gd, at. %—the abbreviation ‘at.’ is now skipped), and Li⁺ and Mg²⁺ concentrations were 1000 ppm with respect to Gd and Ga, respectively. The concentrations of Ce, Li, and Mg atoms were chosen to provide a good compromise between a fast scintillation response and sufficient light yield value. The crystals were grown at the Institute for Materials Research, Tohoku University, Japan. The crystals were grown in an iridium crucible (Furuya Metal Co., Ltd. Tokyo, Japan) of ~2.5 mm in diameter using the micro-pulling down method (μ-PD) [18]. Growth was carried out in an atmosphere of argon mixed with 2% of oxygen. The <100> orientated GAGG:Ce single crystal was used as a seed with a 0.1 mm/min pulling down speed. To compensate for Ga evaporation, a 1.0 wt.% excess of Ga₂O₃ oxide was optimized and added to the stoichiometric compositions.

2.2. Powder X-ray Diffraction and Energy-Dispersive X-ray Spectroscopy Analysis

In a mortar, a small portion of the grown crystal was crushed and ground to a powder. The Bruker D8 DISCOVER-HS diffractometer (Billerica, MA, United States) was used to perform powder X-ray diffraction (PXRD) investigations in the 2θ 15–65° range. CuK-α-radiation with a wavelength of about 1.54 Å and a photon energy of E = 8.04 keV was used for diffraction measurements. A field emission scanning electron microscope, type FEI Nova NanoSEM 230 (Field Electron and Ion Company, Hillsboro, OR, United States), with an Apollo X Silicon Drift Detector and EDAX Genesis Software (Version 4.5) was used for the morphology and chemical composition analysis. The conditions of sample preparation were described in [17]. The Ce, Li, and Mg concentrations in the crystals were beyond the EDAX detector’s detection limit. As a result, neither the EDS mapping nor the analysis of EDS elemental compositions was measured or discussed.

2.3. Optical, Luminescence, and Scintillation Characteristics

The optical absorption spectra at 300 K were measured using a Shimadzu 3101PC spectrometer (Shimadzu Corporation, Kyoto, Japan). Photoluminescence excitation and emission spectra in the vacuum ultraviolet (VUV) spectral range were recorded at 300 K using synchrotron radiation (SR) at the National Institutes of Natural Sciences, Institute for Molecular Science, Okazaki, Japan. The liquid helium cryostat was used for temperature dependence measurements. Correction of these spectra for the wavelength-dependent excitation intensity was performed with the use of sodium salicylate as a standard. Excitation spectra in the UV/Vis range were measured at 300 K using Edinburgh Instruments FLS920 (Edinburgh Instruments Ltd, 2 Bain Square, Kirkton Campus, United Kingdom) equipped with a xenon lamp as an excitation source. Pulsed Diode Lasers (Edinburgh Instruments Ltd, 2 Bain Square, Kirkton Campus, United Kingdom) was used for fast decay kinetic measurements using the same spectrofluorometer. X-ray excited measurements were performed under soft X-ray excitation (40 mA and kV, model RINT2000, Rigaku Corporation, Tokyo, Japan). The scintillation measurements were described in [17].

2.4. Thermoluminescence Properties

Thermally stimulated luminescence was recorded between 10 and 350 Kelvins at the National Institutes of Natural Sciences, Institute for Molecular Science, Okazaki, Japan. The liquid helium cryostat was used. Crystals were irradiated at the liquid helium temperature or room temperature with vacuum ultraviolet radiation with an energy of 12.4 eV. The heating rate was 1 K per second. In the thermally stimulated luminescence measurements between 350 and 760 Kelvins, the crystals were irradiated with vacuum ultraviolet radiation with an energy of 12.4 eV. The heating rate was 1 Kelvin per second. The measurements were performed with a Risø DA-15 TL reader (Risø DTU National Laboratory, Kongens Lyngby, Denmark).

3. Results and Discussion

3.1. Impact of Li⁺ and Mg²⁺ on Atom Distribution and Crystal Phase by XRD and SEM-EDS Analysis

Figure 1 shows as-grown rods and polished cross-sectional plates of Li⁺ (1000 ppm) and Mg²⁺ (1000 ppm) co-doped Gd₃Al₂Ga₃O₁₂:Ce0.25% crystals. Each crystal shows a regular shape with a plain surface. This can indicate that the crystallization process was not disturbed by Li⁺ and Mg²⁺ co-doping. To minimize any concentration discrepancies resulting from atom segregation during crystal growth, the cross-section plates were cut from the same position in each crystal rod. Cross-sectional plates show good transparency throughout the surface.

Figure 2 shows the powder X-ray diffraction (PXRD) patterns for the Li⁺ and Mg²⁺ co-doped GAGG:Ce single crystals. The crystals show a pure garnet phase with an Ia-3d space group (No. 230). This confirms that both Li⁺ and Mg²⁺ ions enter the garnet structure and do not influence the thermodynamical stability of the garnet phase. The SEM-EDS technique is used to specify the changes in the cross-sectional distribution of atoms imposed by the co-doping with Li⁺ and Mg²⁺ ions.

Multielemental EDS mapping images and EDS elemental analysis for atom distribution analyses allow for a precise study of the impact of co-doping with Mg²⁺ and Li⁺ ions on the radial distribution of atoms in GAGG:Ce crystals. The EDS mapping images were recorded for Ga, Al, Gd, and O atoms. Whereas, EDS elemental analysis was performed for Ga, Al, and Gd atoms. The sensitivity of the EDAX detector is not sufficient for the precise detection of Mg²⁺, Li⁺, and Ce³⁺ ions. Hence, the EDS mapping images and EDS elemental analysis are not discussed for those elements. Figure 3a shows the EDS mapping images for Ce³⁺, Ce³⁺:Li⁺, and Ce³⁺:Mg²⁺ doped GAGG single crystals. The mapping of Ga, Al, Gd, and O atoms in the Ce³⁺ doped GAGG crystal confirms the elemental homogeneity in the entire cross section. The Li⁺ and Mg²⁺ co-doping significantly disturbed the distribution of the Ga and Al atoms, while the distribution of the Gd and O elements is unchanged. In both crystals, the EDS mapping images for the Al element show a dark shade on the crystal rim and a brighter shade on the crystal core, while the EDS mapping images for the Ga element show the opposite behavior (see Figure 3a). The brightening reflects an increase in the concentration of atoms, whereas the darkening is related to a decrease in the concentration of the element. Therefore, in both Li⁺ and Mg²⁺ co-doped GAGG:Ce crystals, the Ga atoms enrich the rim of the crystal, while the Al elements are concentrated in the core of the crystal. Furthermore, the EDS elemental mapping images show a distinct impact of the Ga and Al distribution imposed by the Li⁺ and Mg²⁺ co-doping. Specifically, in the Mg²⁺ co-doped sample, the crystal core enriched with Al has a smaller area and is brighter than in the Li⁺ co-doped crystal. This may suggest that Mg²⁺ co-doping reduces crystal uniformity to a greater extent than Li⁺ co-doping. To deeply understand such differences, it is necessary to compare and discuss the EDS elemental distribution of Al, Ga, and Gd atoms in all crystals; see Figure 3b and

Table 1. EDS elemental analysis of the crystal rim and core for Li⁺:Ce³⁺ and Mg²⁺:Ce³⁺ doped GAGG crystals, 300 K.

Crystal	Al at.% (Kα X-ray)/Number of Atoms		Ga at.% (Kα X-ray)/Number of Atoms		Gd at.% (Lα X-ray)/Number of Atoms
	Rim	Core	Rim	Core	Rim	Core
GAGG:Ce3+ 0.25%	25.89/2.07	26.21/2.10	37.13/2.97	36.89/2.95	36.98/2.96	36.90/2.95
GAGG:Ce3+ 0.25%, Li+ 1000 ppm	24.93/1.99	28.05/2.24	38.07/3.05	35.03/2.80	37.00/2.96	36.92/2.95
GAGG:Ce3+ 0.25%, Mg2+ 1000 ppm	23.25/1.86	28.83/2.41	39.86/3.19	34.11/2.73	36.89/2.95	37.06/2.96

. The EDS elemental distribution is consistent with the EDS mapping images. In the GAGG:Ce:Li crystal, there is a slight gradient between the content of Ga and Al atoms in the crystal core and rim. The Mg²⁺ co-doping imposed a significantly stronger gradient between the content of Ga and Al atoms in the crystal core and rim. Furthermore, the variation in Ga and Al elements across the cross section is greater in Li⁺ co-doped samples than in the Mg²⁺ crystal. The radial variation in the Ga and Al atoms is complex and can be caused by a few effects. The first effect can be due to the process of the charge compensation mechanism. The crystal lattice to incorporate charge mismatching Li⁺ and Mg²⁺ ions needs to compensate for the charge imbalance to maintain crystal neutrality. As a result, Li⁺ and Mg²⁺ ions impose the formation of specific types of neutral Li⁺-based defects and neutral Mg²⁺-based defects [14,15,17,19]. These defect clusters violated configurational entropy [20]. This led to lattice distortion and triggered structural rearrangements [20]. Considering Al and Ga atoms in the tetrahedral and octahedral coordination, it is obvious that the smaller Al atoms prefer to segregate to the core and the larger Ga atoms to the crystal rim. It is believed that both Li⁺ and Mg²⁺ ions show low solubility because charge compensation is required. Consequently, segregations of Li⁺-based and Mg²⁺-based defect complexes to the crystal rim are inevitable. This enhances the perturbation of the host lattice and violates configurational entropy. As a result, the radial distribution of Al and Ga atoms changes significantly [21]. Additionally, the Marangoni melt flow, which happens at the melt’s outer surface in the molten zone, can strengthen this effect. Photoluminescence and scintillation analyses provide evidence of host lattice disturbance.

3.2. Optical and Photoluminescence Characterization under Synchrotron Radiation

Figure 4 shows the effect of Li⁺ and Mg²⁺ co-doping on the concentration of stable Ce³⁺ centers and the formation of stable Ce⁴⁺ ions. The absorption bands centered at 440 and 340 nm are attributed to the inter-configurational 4f→5d₁ and 4f→5d₂ transitions of stable Ce³⁺ ions [9,10]. The strong absorption intensity below 320 nm is related to the O²⁻→Ce⁴⁺ charge transfer transition (CTT), that is, electron transfer between the energy levels of the O 2p and Ce 4f energy levels [14]. Furthermore, the decrease in the absorption intensity of the 4f→5d₁ transition is due to the reduction in the stable content of Ce³⁺ ions as a result of the Ce³⁺→Ce⁴⁺ conversion process [14]. The Li⁺ co-doping does not increase the absorbance intensity below 300 nm, but the opposite does, markedly decreasing its intensity. It should be noted that absorbance below 300 nm is frequently complex owing to the overlap of Ce³⁺-related absorption bands and various bands related to color and impurity centers [22,23]. Furthermore, Li⁺ co-doping reduces the absorbance intensities of the inter-configurational 4f→5d₁ and 4f→5d₂ transitions. This observation leads to two important conclusions: Li⁺ co-doping negligibly imposes the change in the valence state of Ce from +3 to +4 and changes the distribution of Ce³⁺ ions between the crystal core and rim. The sharp lines at 276–282 nm and 310–315 nm are attributed to the ⁸S_7/2→⁶I_J and ⁸S_7/2→⁶P_J absorption transitions within the Gd³⁺ ions, respectively [24].

Synchrotron radiation is a kind of spectroscopy that allows to study in the vacuum ultraviolet (VUV) spectral range. A special focus of VUV spectroscopy is the investigation of wide band-gap crystals for which the edge of the intrinsic absorption lies in the VUV spectral range. Therefore, SR spectroscopy is an excellent tool for the determination of host band gap energies (i.e., exciton creation energy). Figure 5a compares normalized VUV excitation spectra for Ce³⁺ emission centered at 540 nm in Ce³⁺, Ce³⁺:Li⁺, and Ce³⁺:Mg²⁺ doped GAGG single crystals, 300 K. Spectra were recorded using synchrotron radiation between 160 and 330 nm to precisely reveal the change in the edge of the host lattice imposed by co-doping with Li⁺ and Mg²⁺ ions. The spectra show different profiles depending on the co-doping ions. The excitation spectra for GAGG:Ce and GAGG:Ce,Li crystals show very similar shapes. The spectra consist of well-separated sets of absorption lines at 276–282 nm and 310–315 nm attributed to the Gd^{3+ 8}S_7/2→⁶I_J and ⁸S_7/2→⁶P_J intra-configurational transitions, respectively [24]. The presence of Gd³⁺ absorption lines on the excitation spectra of Ce³⁺ luminescence points to efficient non-radiative energy transfers from the Gd sublattice toward the Ce³⁺ ions [24]. The broad absorption bands centered at 230 nm are most probably due to the inter-configurational 4f→5d₃ absorption transition of Ce³⁺ ions [25]. The Mg²⁺ co-doped GAGG:Ce crystal shows significantly different excitation characteristics. In particular, the Gd³⁺ lines at 276–282 nm related to the ⁸S_7/2→⁶I_J excitation transition are very weak. Furthermore, the ⁸S_7/2→⁶P_J excitation transition at 310–315 nm is at the level of noise. The 4f→5d₃ absorption transition of Ce³⁺ ions shows a very low intensity at 230 nm. The VUV excitation spectra revealed a different position of the fundamental valence-to-conduction band absorption of the host lattice (i.e., exciton creation energy). The fundamental edge of host lattice absorption is at (i) GAGG:Ce ~210 nm (5.90 eV); (ii) GAGG:Ce:Li ~207 nm (5.98 eV); and (iii) GAGG:Ce:Mg ~203 nm (6.01 eV). This shift in the host lattice edge towards higher energy can be due to the variation in the Al and Ga content in the crystal imposed by Li⁺ and Mg²⁺ co-doping [26]. It is worth mentioning that the substitution of Ga for Al significantly reduces the band gap [4]. In the Li⁺ and Mg²⁺ GAGG:Ce samples, the crystal core is enriched with Al elements (depleted with Ga atoms); therefore, the band gap energy is wider than that of the GAGG:Ce crystal. The Mg²⁺ co-doped GAGG:Ce sample shows a stronger gradient between the Ga and Al atoms in the crystal core area than in the Li⁺ co-doped GAGG:Ce sample. Therefore, the Mg²⁺ co-doped sample has the widest band-gap energy. The shift in the fundamental valence-to-conduction band absorption of the host lattice toward higher energy in Li⁺ and Mg²⁺ co-doped crystal is in contrast with elemental EDS mapping images and EDS elemental composition analysis; see Figure 3a,b and Table 1. The reduced intensity of the excitation lines of the Gd³⁺ ions and Ce³⁺ bands in the Mg²⁺ co-doped sample can be explained by considering crystal chemistry and the charge compensation mechanism. Specifically, Mg²⁺ co-doping efficiently transforms Ce³⁺ into Ce⁴⁺. As a consequence, the content of stable Ce³⁺ ions decreases significantly. This reduces the intensity of the Ce³⁺ 4f→5d₃ absorption transition and hampers Gd³⁺→Ce³⁺ energy transfer [25]. Furthermore, the strong CT of stable Ce⁴⁺ below 350 nm overlaps with the energy levels of ⁶I (~275 nm) and ⁶P (310 nm) in Gd³⁺ ions. This can enable efficient energy transfer from Gd³⁺ ions to Ce⁴⁺, where energy is lost [25].

Figure 5b shows the excitation spectra for three single crystals. Excitation spectra were recorded for Ce³⁺ emission at 540 nm related to the inter-configurational 5d₁→4f transition. The bands centered at 345 and 440 nm correspond to the inter-configurational 4f→5d₂ and 4f→5d₁ transitions of Ce³⁺ ions, respectively [4]. The excitation features below 330 nm are consistent with Figure 5a. The band located at 265 nm corresponds to the inter-configurational 4f→5d₁ transition of Tb³⁺ ions [7]. The intensity of Tb³⁺ inter-configurational 4f→5d₁ transition in Mg²⁺ and Li⁺ co-doped crystals is reduced, most probably because of a change in the Tb valence state from +3 to +4 [17]. The absorption strength of the Ce³⁺ 4f→5d₂ excitation transition is significantly reduced in Li⁺ and Mg²⁺ co-doped crystals. The decrease in Ce³⁺ 4f→5d₂ absorption strength can be due to two reasons. The first is the decrease in the concentration of stable Ce³⁺ ions. In Mg²⁺ co-doped crystal, this decrease is due to Ce³⁺→Ce⁴⁺ conversion [17,26]. In the case of the Li⁺ co-doped sample, this decrease is mainly due to the change in Ce³⁺ ions’ distribution between the crystal core and rim imposed by Li co-doping. The second reason could be due to the local distortion of CeO₈ dodecahedra imposed by the formation of defect clusters to maintain crystal neutrality after co-doping with aliovalent Li⁺ and Mg²⁺ ions [15,17].

Figure 6a compares the emission spectra of GAGG:Ce and Li⁺ and Mg²⁺ co-doped GAGG:Ce single crystals excited at 200 nm, 300 K. The broad emission bands centered at 540 nm belong to the Ce³⁺ inter-configurational 5d₁→4f transition. The lack of Gd³⁺ emission is due to the efficient energy transfer from the Gd sublattice to Ce³⁺ ions. Furthermore, owing to the higher temperature and stronger oscillations of the host lattice phonons, the Ce³⁺ inter-configurational 5d₁→4f emission bands broaden. Therefore, the impact of Li⁺ and Mg²⁺ co-doping cannot be observed.

The impact of the Mg²⁺ and Li⁺ co-dopants on Ce³⁺ energy levels can be seen in the decay times of Ce³⁺ emission; see Figure 6b. The single exponential function was used to fit the experimental data according to Equation (1):

I(t) = Σ _iI_i exp[−t/τ_i] + B

(1)

where I—luminescence intensity, I_i—intensity at 0 ns, t—time, τ_i—decay time, and B—time-independent background intensity. Each decay curve was recorded under Ce³⁺ inter-configurational 4f→5d₁ excitation at 450 nm. The decay constants vary depending on the co-doping ion. Li⁺ co-doping slightly decelerated the decay time to τ_Li+ = 58 ns, while Mg²⁺ co-doping slightly accelerated the decay constant to τ_Mg2+ = 53 ns compared with the non-co-doped sample τ_Ce3+ = 55 ns. This small change in the decay constants can be related to the radial fluctuation of the Ga and Al atoms (see Figure 3a,b and Table 1) as well as different defect structures (see Figure 7a,b).

3.3. Thermally Stimulated Luminescence Characteristics

The interband irradiation of the crystals at 10 K and 300 K allows the trapping of electrons and holes at different traps, leading to the creation of various electron and hole-type centers. The recombination of electrons and holes, thermally released from different electron traps, with the Ce³⁺ and Ce⁴⁺ centers results in the appearance of different peaks of the TSL glow curve. Figure 7a shows the TSL glow curves measured after irradiation at 10 K. Between 25 and 150 K, there is a complex structure of the most intense TSL peaks for the GAGG:Ce and GAGG:Ce,Li crystals. The intensity of this low-temperature peak is reduced by three and sixty times in Li⁺ and Mg²⁺ co-doped samples, respectively, compared with the GAGG:Ce crystal. Moreover, their variation in position and complex structure might be related to the radial fluctuation of the Al and Ga atoms [26]. This observation is in contrast with EDS elemental mapping and elemental composition analysis, VUV excitation spectra, and previous research in [26]. The decrease in intensity of the low-temperature glow peaks in GAGG:Ce³⁺ co-doped with Li⁺ and Mg²⁺ crystals is not related to the reduction in the content of shallow traps. Rather, Li⁺ and Mg²⁺ imposed defects, including Ce⁴⁺ centers, which compete more efficiently to capture electrons and holes with shallow trapping centers [15,17,26]. Furthermore, Li⁺ co-doping imposed the formation of a new broad peak around 310 K [14]. This trap can be responsible for the deceleration of the scintillation decay time and a slight increase in LY. The charge carriers captured in this trap can be released around room temperature and radiatively recombine with the Ce³⁺/Ce⁴⁺ centers, increasing the LY value. The co-doping of Mg²⁺ ions formed a new broad TSL glow peak around 280 K [14].

The TSL glow curves obtained after interband irradiation at 300 K are presented in Figure 7b. The TSL glow curves show very similar profiles, with different intensities and slightly different positions. Li⁺ co-doping slightly reduces the intensity of the peaks at ~370 K; however, it significantly increases the intensities of the peaks at 430 and 530 K, corresponding to the non-co-doped GAGG:Ce crystal. In contrast, Mg²⁺ co-doping markedly decreases the intensities of these TSL glow peaks. The increase in TSL intensity in the higher temperature range can be imposed by the formation of oxygen vacancies (V_O^••) for the excess negative (−2) charge compensation induced by Li⁺ [14,15]. This indicates a similar origin of the corresponding traps in all of the crystals. The TSL glow peaks <200 K can be related to shallow antisite defects [26]. Their slightly different positions in different crystals can be caused by different Ga³⁺ contents, as SEM-EDS multi-elemental analysis (Figure 3) and excitation spectra (Figure 5a), as well as previous research, suggest [14,17,26]. Peaks centered at 305 K and 265 K, which are absent in the GAGG:Ce crystal, could arise from the intrinsic defects associated with Li⁺ and Mg²⁺ ions, respectively [15,26]. The lower-temperature part of the 265 K peak in the Mg²⁺ co-doped sample can arise from a hole trapped at an oxygen ion located close to the Mg²⁺ ion (Mg_GaO_O)^x-type hole center [17,26]. The peak around 305 K in the Li⁺ co-doped sample may originate from the thermal release of holes from the (Li_GdO_O)′ centers [15]. The peak around 370 K, presented in all samples, is most probably connected to a single oxygen vacancy [26]. Peaks above 370 K can arise from both intrinsic defects and the electron and hole centers perturbed by Li⁺ and Mg²⁺ ions (for example, antisite defects and O⁻ type hole centers [14,15,26]). The strong decrease in the intensity of TSL in the Mg²⁺ co-doped GAGG:Ce is because of electron recombination with Ce⁴⁺ centers, which results in the release of photons, and recombination with Mg²⁺ defect clusters, which results in the release of phonons (i.e., Mg²⁺ defect clusters operate as quenching centers) [17,26]. These data indicate that the distribution of electrons and holes between traps in GAGG:Ce crystals depends on the Li⁺ and Mg²⁺ co-dopants [15,26].

3.4. Scintillation Characteristics

Ionization radiation (X-ray or γ-ray) can penetrate matter deeply thanks to its high penetration power. An important byproduct of ionizing radiation in solids is that electrons and holes are frequently trapped at lattice sites. Depending on the depth of the trap sites, the electron or hole may either recombine rapidly, giving a prompt luminescence or be trapped for an extended period, even many years. The release of these electrons and holes can yield delayed luminescence recombination. Moreover, the radiative recombination of electrons and holes can reveal all luminescence centers in the host lattice. The radioluminescence spectra of Ce³⁺, Ce³⁺:Li⁺, and Ce³⁺:Mg²⁺ doped GAGG single crystals are shown in Figure 8. The bands centered around 550 nm correspond to the inter-configurational 5d₁→4f transition in Ce³⁺ ions. After co-doping with Li⁺ and Mg²⁺, no extra luminescence is seen. This result shows that both Li⁺ and Mg²⁺ ions do not impose additional defect-related centers that can radiatively recombine.

The kinetics of scintillation decay offers important information on defect centers and their impact on scintillation parameters. The high dynamic ranges and timing of the decay experiments allow for precise convolution of the experimental curve and simultaneous replication of the rapid and slow components. The decay curves for Ce³⁺, Ce³⁺:Li⁺, and Ce³⁺:Mg²⁺ doped GAGG single crystals are shown in Figure 9a. The decay curves are approximated by a double exponential function, which corresponds to radiative recombination at the Ce³⁺ and Ce⁴⁺ centers.

Table 2. Scintillation decay constants and light yield values (shaping time 2 μs) for Ce³⁺, Ce³⁺:Li⁺, and Ce³⁺:Mg²⁺ doped GAGG single crystals under γ-ray excitation from ¹³⁷Cs radioisotope (662 keV), 300 K.

Sample	Scintillation Decay Constants		Light Yield (Phot/1 MeV) [2 μs]
	τ1 [ns]/%	τ2 [ns]/%
GAGG:Ce3+ 0.25%	50/61	225/39	31,500 ± 1100
GAGG:Ce3+ 0.25%, Li+ 1000 ppm	58/89	240/11	33,200 ± 1400
GAGG:Ce3+ 0.25%, Mg2+ 1000 ppm	19/42	80/58	21,500 ± 800

summarizes the decay constants calculated from an experimental data fit. The scintillation decays and pulse-height spectra (see Figure 9b) show the distinct influence of Li⁺ and Mg²⁺ co-doping on the scintillation properties. The Li⁺ co-doping markedly decelerates the scintillation response and increases the light yield values; see Table 2. The Li⁺-related defect can rather slowdown the e–h transport toward the Ce³⁺ and Ce⁴⁺ emission centers than provide quenching channels. This slowdown can be related to the appearance of the defect structure around 310 K; see Figure 7a. As a result, the charge carriers can temporarily be captured in this trap and, after some time, released back to the conduction band and delivered to the Ce³⁺/Ce⁴⁺ luminescence centers (i.e., delayed radiative recombination). Furthermore, the small content of Ce⁴⁺ luminescence centers could increase LY. On the contrary, Mg²⁺ co-doping significantly accelerated the scintillation response and reduced LY. This can be due to the formation of Mg²⁺-based defect clusters [17]. Another reason is the decrease in the efficiency of Gd³⁺→Ce³⁺ energy transfer [25]. This is caused by the decrease in the stable content of Ce³⁺ ions due to the Ce³⁺→Ce⁴⁺ conversion and the energy loss imposed by energy transfer from the Gd³⁺ sublattice to the non-luminescent charge transfer absorption transition of Ce⁴⁺ [25].

4. Conclusions

The experimental data presented in this investigation revealed that Ga and Al sublattices are very sensitive to the local distortion of the host lattice imposed by the incorporation of incompatible substituents. Incompatible in size and charge, Li⁺ and Mg²⁺ substituents violated configurational entropy. This led to lattice distortion and triggered structural rearrangements. As a result, both Li⁺ and Mg²⁺ co-dopants significantly depleted the crystal core in Ga atoms, at the same time enriching it with Al elements. The radial fluctuation of the Ga and Al elements was revealed by EDS elemental mapping and elemental composition analysis. Further evidence was observed by the shift of the fundamental host lattice edge (exciton creation energy) toward a higher energy in the VUV excitation spectra recorded with synchrotron radiation. The local host lattice disturbance was shown by the PL data. Scintillation and TSL data showed that Mg²⁺ ions imposed rather defect/defect clusters that acted as quenching centers, while Li⁺ defect clusters increased the content of deep host lattice defects. These observations showed the distinct impact of the co-doping of Li⁺ and Mg²⁺ ions on defect formation processes, atom distribution, and changing the cerium valence state from +3 to +4.