Introduction

Scintillators are essential functional materials for radiation detection and play crucial roles in medical diagnostics, nondestructive testing, security screening, and high-energy physics. In the medical field, photon-counting computed tomography (PCCT) has emerged as a next-generation X-ray imaging technology that offers an improved signal-to-noise ratio1, material decomposition capability2, multi-energy imaging3, and reduced radiation dose4,5 compared with conventional energy-integrating CT systems. These advantages have created a strong demand for scintillator materials specifically optimized for the stringent performance requirements of PCCT detectors. PCCT imposes fundamentally four key requirements on scintillators for practical implementation: (ⅰ) high light yield to ensure adequate energy resolution, (ⅱ) fast decay and minimal afterglow to suppress pulse pile-up and preserve accurate energy-bin discrimination, (ⅲ) absence of K-edge absorption within the 20–100 keV diagnostic X-ray range, which otherwise produces escape-peak artifacts and spectral distortion, and (ⅳ) centimeter-scale scintillation uniformity and stable producibility, which are indispensable for mass production and integration into multi-channel Photon-counting detectors (PCDs)6,7,8,9.

The performance of scintillator-based PCDs depends primarily on the intrinsic properties of the scintillator crystal. Owing to their optical transparency, compositional tunability, and chemical stability, Ce-doped garnet-type oxides are among the most promising scintillator candidates. Among these, Gd₃Ga₃Al₂O₁₂:Ce (GAGG: Ce) has been extensively studied because of its high light yield (~ 56,000 photons/MeV), relatively fast decay constants (76 and 826 ns), and high stopping power, making it one of the most successful commercial scintillators in recent years10,11,12. Although Mg co-doping reduces the light yield to ~ 35,000 photons/MeV, it accelerates the decay components to 51 and 173 ns13. However, the Gd K-edge at ~ 50 keV, which lies within the diagnostic X-ray energy window (20–100 keV), produces strong escape peaks that critically limit energy discrimination in PCCT2,14,15. In addition, slow scintillation components and afterglow induce pulse pile-up, degrading the count-rate performance of PCDs15. Furthermore, Gd³⁺ ions introduce 4f–4f transitions that interfere with co-dopants such as Tb³⁺, complicating energy-transfer processes16,17,18,19,20.

These drawbacks have motivated the search for alternative host lattices, with yttrium-based garnets Y3(Ga, Al)5O12:Ce (YAGG: Ce) emerging as promising candidates. The Y K-edge at 17 keV, well below the diagnostic energy window, avoids the spectral interference inherent to Gd-based materials21. Additionally, the Ga/Al ratio in YAGG can be tuned to optimize the bandgap and scintillation efficiency. Previous studies have shown that Mg co-doping suppresses the negative influence of shallow trap states near the conduction band, increasing the light yield and accelerating the decay. Kamada et al. reported that Y₃Ga₃Al₂O₁₂:Ce, Mg (YAGG: Ce, Mg) crystals grown via the micro-pulling-down (µ-PD) method exhibited a light yield of ~ 38,800 photons/MeV and a decay time below 55 ns, although key PCCT-relevant properties such as energy resolution and afterglow were not evaluated22.

For practical applications, large-diameter, uniform bulk single crystals with stable scintillation properties are essential. Garnet-type oxide scintillators such as GAGG13, (Lu, Y)₂SiO₅ (LYSO)23, and (La, Gd)₂Si₂O₇ (La-GPS)24 are routinely produced as 2–4-inch boules on an industrial scale using the Czochralski (Cz) method. While µ-PD is valuable for compositional exploration, it is intrinsically limited to millimeter-scale crystal growth; thus, scaling promising compositions requires bulk-growth techniques such as Cz.

In garnet materials, the melting point strongly depends on the elemental composition: approximately 2010 °C for Lu₃Al₅O₁₂25, 1970 °C for Y₃Al₅O₁₂26, and 1740 °C for Gd₃Ga₅O₁₂27, increasing with higher Lu, Y, or Al content. Gallium-containing garnets exhibit substantial Ga₂O₃ evaporation at high temperatures. In µ-PD growth, crystals are grown within a few hours, minimizing Ga loss. In contrast, the Cz method requires 7–10 days to grow a bulk single crystal—during which the melt is continuously exposed to high temperatures—resulting in markedly increased Ga evaporation, composition drift, and severe oxidation and degradation of the iridium crucible.

These issues were clearly demonstrated in attempts to grow Lu₂Gd₁Ga₃Al₂O₁₂ (LGAGG) via the Cz method28. Although LGAGG was identified as a promising scintillator through µ-PD screening due to its high density and fast decay characteristics10,11, the high Lu/Gd content implies an even higher melting point than GAGG. During Cz growth, Ga evaporation led to composition changes, secondary-phase precipitation, and increased turbidity as crystallization proceeded, causing the light yield to deteriorate. Luminescence was almost completely lost once the solidification fraction reached 30%. Given that YAGG is expected to exhibit a similarly high melting point, analogous issues—such as detrimental effects on crystal growth and nonuniformity in scintillation properties—must be carefully considered.

Segregation behavior also differs significantly between µ-PD and Cz growth. In µ-PD, segregation coefficients along the growth direction are theoretically close to unity for oxide garnets29,30, resulting in nearly uniform composition. In contrast, Cz growth exhibits much smaller segregation coefficients for dopants such as Ce and Mg —reported as 0.42 and 0.021, respectively, in GAGG31—and even host elements show measurable segregation. These differences lead to compositional gradients and corresponding fluctuations in scintillation performance along the boule, posing major challenges for the development of new scintillator materials.

In this study, we investigate whether Ga₂O₃ evaporation, dopant segregation, and crucible degradation constitutes critical issues for the Cz growth of YAGG, and evaluates the extent to which these challenges can be mitigated through optimized growth conditions aimed at practical implementation and mass production. We further assess how composition fluctuations arising from dopant segregation and Ga evaporation influence the uniformity of scintillation properties—including light yield, decay time, energy resolution, radioluminescence (RL) spectra, and afterglow—along the growth direction of the produced crystal.

This work therefore provides the first experimental demonstration that YAGG: Ce, Mg can be grown as inch-scale, compositionally and functionally uniform single crystals by the Cz method, overcoming the limitations of previous µ-PD studies and establishing its feasibility as a practical scintillator for PCCT.

Materials and methods

A (Y3 − x−yCexMgy)Ga3Al2O12 crystal was grown using the Cz method (x = 0.01, y = 0.006). Hereafter, (Y2.984Ce0.01Mg0.006)Ga3Al2O12 is referred to as YAGG: Ce, Mg. High-purity (4 N) CeO2, MgO, Y2O3, b-Ga2O3, and a-Al2O3 powders were prepared as starting materials. The mixed raw materials were sintered at 1500 °C in air. Crystal growth was performed using an Ir crucible with an inner diameter of 48 mm and < 100 > Y3Al5O12 (YAG: Ce) single crystals as seeds. The pulling and rotation rates were set to 0.7 mm/h and 10.0 rpm, respectively, and the crystal shape was controlled using an automatic diameter control (ADC) system. The growth atmosphere was maintained under N2 with 25% CO2 and Ar with 2% O2 to suppress Ga evaporation. The grown crystals were sectioned into seven parts according to the crystallization ratio, cut into 5 × 5 × 1 mm specimens, and mirror polished prior to analysis.

Powder X-ray diffraction (XRD; D8 DISCOVER, Bruker) with Cu Kα radiation was used to examine the phases of the grown samples. The acceleration voltage and current were 40 kV and 40 mA, respectively. To investigate the segregation behavior along the crystal growth direction, electron probe microanalysis (EPMA; JXA-8530 F, JEOL) with wavelength dispersive spectrometry (WDS) was performed for four cations: Y, Ce, Ga, and Al. In addition, inductively coupled plasma atomic emission spectrometry (ICP-AES; Thermo Fisher Scientific Inc., Waltham, MA, United States) was conducted to determine the concentrations of the five cations, including Mg, at each crystallization ratio.

RL spectra were recorded using an SR-163 spectrometer (Andor Technology) equipped with a CCD detector DU-420OE (Andor Technology). The measurements were performed under X-ray irradiation (40 kV, 20 mA) at an acquisition rate of 10 spectra per second. Scintillation light was collected using a photomultiplier tube (PMT; R7600-200, Hamamatsu Photonics K.K., Hamamatsu, Japan) operated at 700 V. A 10 MBq 137Cs source was positioned at a distance of 5 cm. Signals from the PMT were amplified using a spectroscopy amplifier (MSCF-16, mesytec GmbH & Co. KG, Putzbrunn, Germany) and digitized using a peak digital converter (PDC; A3400, NIKI GLASS Co., Ltd., Tokyo, Japan). The afterglow was measured using a custom-made spectrofluorometer 5000 M (Horiba Jobin Yvon), with steady-state excitation provided by an X-ray tube with a Mo anode (40 kV, 15 mA, Seifert GmbH), which was cut off at approximately 72 ms on the X-axis of the graph. A single-grating monochromator and a PCD TBX-04 (IBH, Scotland) were used for detection. Scintillation decay curves were recorded using a digital oscilloscope (RTB2004, Rohde & Schwarz GmbH & Co. KG, Munich, Germany). Considering that the YAGG: Ce, Mg scintillation decay contains two components, the experimental data were fitted using the sum of two exponential functions, I(t), where τ1 and τ2 represent the decay times of the fast and slow components, respectively.

$$\:I\left(t\right)\:={A}_{1}exp\left(\frac{-t}{{\tau\:}_{1}}\right)+{A}_{2}exp\left(\frac{-t}{{\tau\:}_{2}}\right)+const.$$

Results and discussion

For the mass production of GAGG: Ce single crystals, a mixed atmosphere of Ar with 30% CO2 was employed to minimize gallium oxide evaporation and Ir crucible degradation. The same atmospheric conditions were initially applied to the growth of the YAGG: Ce, Mg crystals. The results of the initial trial, conducted under a mixed atmosphere of N2 and 25% CO2, are shown in Fig. 1a. A metallic precipitate is visible on the crystal surface, with EPMA analysis identifying it as Ir. The growth process was suspended because of instability in crystal diameter control by the ADC system, caused by Ir film deposition on the melt surface and its adhesion to the crystal. Based on prior experience with GAGG: Ce growth, Ir deposition was considered to result from oxide evaporation. Moreover, in the Cz growth of β-Ga₂O₃ single crystals, several parameters such as temperature distribution, growth rate, rotation speed, crucible geometry, and oxygen partial pressure determine the crystal quality. Among them, precise control of the oxygen partial pressure is essential to suppress the decomposition and evaporation of Ga₂O₃ and to prevent oxidation and degradation of the Ir crucible. Therefore, maintaining an optimal oxygen partial pressure is indispensable for achieving thermochemical stability during crystal growth32. To mitigate gallium oxide evaporation, a second trial was conducted under an Ar + 2% O2 mixed atmosphere. As shown in Fig. 1b, a 1-inch diameter YAGG: Ce, Mg crystal with a body length of 80 mm was successfully obtained. Although Ir film formation on the melt surface was suppressed compared with the first trial, it still led to seeding instability, requiring manual intervention for crystal shape control. After the initial shoulder growth was manually adjusted, the process was switched to ADC to continue the crystal growth. Despite a sudden fluctuation in crystal diameter caused by iridium deposition, a YAGG: Ce, Mg crystal with a 1-inch diameter and 80 mm body length was successfully grown. Figure 1c shows the powder XRD results for phase identification of the samples. Owing to the large sample size, phase identification was performed at three positions: the seed side, central region, and tail side. The results confirm that the sample consisted of a single cubic phase with space group Ia-3d, exhibiting high crystallinity and phase purity.

To evaluate cation segregation from the tail side to the seed side, line analysis was performed in the central region of the sample using EPMA at intervals of 0.2 mm for four cations: Ce, Y, Ga, and Al. To confirm the segregation behavior of Mg, seven points were selected at equal intervals from the seed side to the tail side, and their concentrations were examined using ICP-AES. Figure 2 shows the detailed compositional profiles obtained along the growth direction of the YAGG: Ce, Mg single crystal, where Cs and C0 represent the actual and nominal compositions, respectively. The average solidification fraction (g) was approximately 0.46, with g defined by the following equation:

$$\:g=\:\frac{\text{M}\text{a}\text{s}\text{s}\:\text{o}\text{f}\:\text{s}\text{o}\text{l}\text{i}\text{d}\text{i}\text{f}\text{i}\text{e}\text{d}\:\text{p}\text{a}\text{r}\text{t}}{\text{T}\text{o}\text{t}\text{a}\text{l}\:\text{m}\text{a}\text{s}\text{s}\:\text{o}\text{f}\:\text{s}\text{t}\text{a}\text{r}\text{t}\text{i}\text{n}\text{g}\:\text{r}\text{a}\text{w}\:\text{m}\text{a}\text{t}\text{e}\text{r}\text{i}\text{a}\text{l}\:\text{i}\text{n}\:\text{t}\text{h}\text{e}\:\text{c}\text{r}\text{u}\text{c}\text{i}\text{b}\text{l}\text{e}}$$

These results indicate that the cation distribution was homogeneous throughout the sample, with no significant deviation in chemical composition across the crystal. Therefore, the successful growth of a 1-inch diameter YAGG: Ce, Mg single crystal with uniform chemical composition was achieved. In addition, slight disturbances in the linearity of the compositional distribution can be observed at approximately g = 0.03, 0.15, and 0.3, corresponding to the positions in Fig. 1b where irregularities in the crystal diameter occurred. This suggests that deviations in the pulling rate and temperature control during crystal growth disrupted melt convection, leading to local fluctuations in the cation distribution. The effective segregation coefficients (keff) of Al, Y, Ga, Ce, and Mg were calculated using the following equation:

$$\:{C}_{s}/{C}_{0}={k}_{eff}\:{(1-g)}^{{k}_{eff}\:-\:1}$$

The keff values of Al, Y, Ga, Ce, and Mg are 1.53, 1.10, 0.62, 0.25, and 0.19, respectively. The effective segregation coefficient indicates the ease with which each cation is incorporated into the crystal. These results show that Al and Y, with keff > 1, were preferentially incorporated into the YAGG: Ce, Mg crystal, whereas Ga, Ce, and Mg, with keff < 1, were relatively less incorporated. The effective segregation coefficient of Ce (0.25) is comparable to that reported for GAGG: Ce, Mg grown using the Cz method (0.42)31, indicating that Ce incorporation into the crystal is similar between the two hosts. In contrast, the effective segregation coefficient of Mg for GAGG: Ce, Mg is 0.021, whereas that for YAGG: Ce, Mg is 0.19, revealing that Mg incorporation is approximately ten times higher in YAGG: Ce, Mg. This difference in effective segregation coefficients can be explained by the differences in the ionic radii of the cations. Considering that previous studies performed crystal growth under the assumption that Ce and Mg substitute the dodecahedral Y site, the present crystals were grown using the same approach20. Accordingly, if the dopants in the crystal are assumed to occupy the dodecahedral sites, the ionic radii of the respective cations at these sites are 89 pm (Mg2+), 105.3 pm (Ga3+), and 101.9 pm (Y3+). The ionic radius difference in GAGG: Ce, Mg is RGd – RMg = 16.3 pm, whereas that in YAGG: Ce, Mg is RY – RMg = 12.9 pm33. These results suggest that Mg2+ is more readily incorporated into YAGG: Ce, Mg, where the ionic radius difference with the host cation is smaller, explaining the observed difference in the effective segregation coefficient. Furthermore, given that the ionic radius of Ce3+ at the dodecahedral site is 114.3 pm, the ionic radius difference in GAGG: Ce, Mg is RCe – RGd = 9 pm, whereas that in YAGG: Ce, Mg is RCe – RY = 12.4 pm. The larger ionic radius difference at the Y site suggests that Ce is less likely to be incorporated. However, considering that Mg2+ is readily incorporated, Ce may be more easily incorporated when it coexists as Ce4+ near the Mg2+ sites, providing charge compensation.

Fig. 1
Fig. 1
Full size image

Photographs of YAGG: Ce, Mg crystals grown (a) under N2 + 25% CO2 and (b) Ar + 2% O2 atmospheres. (c) XRD patterns of YAGG: Ce, Mg at the tail, middle, and seed positions.

Fig. 2
Fig. 2
Full size image

Compositional distribution along the growth direction of the YAGG: Ce, Mg crystal.

Figure 3 shows the RL spectra of the YAGG: Ce, Mg crystal at room temperature at various positions from the seed to the tail side. Broad emission peaks can be observed at approximately 505–510 nm at all positions. This indicates the luminescence characteristics in response to radiation are uniform throughout the entire crystal along the crystal growth direction. However, previous studies on GAGG have found that the Ga concentration in the crystal affects the position of the Ce3+ 5d1 energy level relative to the conduction band minimum (CBM), and a decrease in Ga concentration shifts the 5d1 energy level toward lower energy34,35,36 while the substitution of Gd by Y shows the opposite shift37. In the present study, ICP-AES analysis revealed a composition of Y2.997Ce0.001Mg0.002Ga2.263Al2.737O12, indicating a deficiency of Ga compared with the nominal composition. This deficiency can be attributed to Ga evaporation and segregation behavior. Jointly, this reduced Ga content and complete substitution of Gd by Y, shifted the 5d1-4f radiative transition of Ce3+ towards 505–510 nm. Furthermore, compared with the seed side, the Ga concentration decreased toward the tail side of the crystal, which increased the ionization barrier at the 5d1 level and the intensity of the emission peak.

Fig. 3
Fig. 3
Full size image

RL spectra of YAGG: Ce, Mg at various positions.

Figure 4a shows the pulse-height spectra measured at various positions using a 137Cs excitation source (662 keV). The calculated light yields and energy resolutions based on this pulse-height spectra results are listed in Table 1. The light yield was determined by comparison with a mirror-polished 5 mm × 5 mm × 1 mm GAGG: Ce standard, which exhibits a light yield of 46,000 photons/MeV and an emission peak of 507 nm11. The YAGG: Ce, Mg samples exhibited a light yield of 46,700 photons/MeV, which is nearly identical to that of the standard GAGG crystal. Though the light yield slightly decreased with increasing solidification fraction, the uniformity of the light yield along with the growth direction was confirmed, demonstrating that the crystal is homogeneous in light yield throughout its entire length. Furthermore, the energy resolution was calculated using the following function based on the pulse-height spectra:

$$\:y=\:{y}_{0\:}+\:{y}_{1}x+\:{y}_{2}{x}^{2}+\:\frac{A}{{\omega\:}_{1}\sqrt{\frac{\pi\:}{2}}}\:exp\left\{{-2\left(\frac{(x-{x}_{c})}{{\omega\:}_{1}}\right)}^{2}\right\}\:+\:\frac{B}{{\omega\:}_{2}\sqrt{\frac{\pi\:}{2}}}\:exp\left\{{-2\left(\frac{(x-C*{x}_{c})}{{\omega\:}_{2}}\right)}^{2}\right\}$$

Because the pulse-height spectrum contains three components (Compton scattering, energy escape caused by the K-absorption edge, and photoelectric absorption), it was fitted using an approximated function expressed as the superposition of a second-order polynomial function and two Gaussian functions38,39,40. The function outputs y for each input x (MCA channel), where y1, y2, A, and B are constant, and xc represents the x value of the peak position. The constant C is a correction factor that ensures that the peak x value corresponds to the position of the K-absorption edge of the host atom. Given that the K-absorption edge of the host atom Y is 17 keV, the value of C was set to (662 keV − 17 keV)/662 keV = 0.974. Then, energy resolution was calculated with the parameter of ω1 obtained from Gaussian1 and the equation that \(\:\varDelta\:E/E=\left(FWHM/xc\right)\times\:100\%,\:(FWHM=\sqrt{2ln2}\omega\:1)\). Figure 4b shows the fitting results for the energy resolution at g = 31%. The energy resolution varied from 8.5% to 11.4% throughout the crystal. In addition, it degraded on the tail side (g = 37% and 45%), corresponding to a decrease in light yield.

Fig. 4
Fig. 4
Full size image

(a) 137Cs-induced pulse-height spectra of YAGG: Ce, Mg scintillators at various positions. Cz-grown GAGG: Ce was used as a reference. (b) Fitting result of the energy resolution for the sample at g = 31% (R2 = 0.992).

Figure 5a shows the scintillation decay time measurements of the sample under 137Cs excitation at various positions. The case of g = 31% is shown in Fig. 5b as a representative example of the spectral fitting. The scintillation decay times calculated from double exponential fitting are summarized in Table 1. At all positions, both the first and second decay components exhibit shorter scintillation decay times than those of GAGG: Ce21. In addition, the scintillation decay time on the tail side is shorter than that on the seed side. The ICP-AES results indicate that the Mg concentration was 0.176 at% on the seed side (g = 1%) and 0.220 at% on the tail side (g = 45%). The observed shortening of the decay time is considered correlated with the elevated Mg concentration41.

Fig. 5
Fig. 5
Full size image

(a) 137Cs-induced scintillation decay curves of YAGG: Ce, Mg at various positions. (b) Fitting results of the scintillation decay curve for the sample at g = 31% (R2 = 0.999).

Table 1 Calculated light yields, energy resolutions, and scintillation decay times at various positions for YAGG: Ce, Mg single crystal scintillators.
Full size table

Figure 6a shows the measured afterglow curves compared with the standard BGO scintillator. An extremely low afterglow level was achieved owing to the stabilization of Ce4+ by the Mg2+ co-dopant, which prevented the localization of electrons from the conduction band in traps, thereby decreasing the delayed luminescence phenomenon22,31,42. Figure 6b shows the comparison of afterglow levels among the YAGG: Ce, Mg seed (Fig. 6a) and commercial GAGG: Ce and YAG: Ce samples from CRYTUR company. It is clear that YAGG: Ce, Mg perform the best having an average value of afterglow within 10-20ms after X-ray cut-off around 280 and 640 ppm for g = 1% and 45%, respectively, while GAGG: Ce, Mg and YAG: Ce commercial standards from CRYTUR show values of 510 ppm and 1000 ppm, respectively. The increased presence of Ce4+ is further confirmed by the absorption spectra in Fig. 7 where the charge transfer absorption of Ce4+ is evident below 350 nm.

Fig. 6
Fig. 6
Full size image

(a) Afterglow curves of YAGG: Ce, Mg seed and tail samples compared with the BGO standard. (b) Afterglow curve of YAGG: Ce, Mg tail samples compared with GAGG: Ce, Mg and YAG: Ce, which are representative multicomponent garnet crystals.

Fig. 7
Fig. 7
Full size image

Absorbance spectra of YAGG: Ce, Mg samples.

Scintillation mechanism in YAGG: Ce, Mg crystal is somewhat different with respect to classical YAG: Ce (LuAG: Ce) crystals and multicomponent GAGG: Ce ones as well. In the former crystals grown from the melt the stabilization of Y(Lu) at the octahedral Al site (called an antisite defect) occurs which creates a shallow electron trap detrimental in transfer stage of scintillation mechanism which was studied mainly in LuAG: Ce43. Furthermore, the trapped exciton around this antisite defect gives rise to a UV emission band which competes with Ce3+ emission center and negatively influences scintillation response as UV emission is far slower compared to Ce3+ one44. The admixture of Ga into LuAG: Ce was shown to prevent the trapping of free carriers due to the mentioned shallow traps as it lowers the bottom of conduction bands such that the energy level of shallow traps is no longer in the forbidden gap45. This leads also to destabilization of exciton trapping at antisite defect, i.e. to diminishing of unwanted UV emission in the Ga-admixed LuAG crystal46. The present study of YAGG: Ce crystal shows the same effect also in the Ga-admixed YAG host: UV emission is not present at all (Fig. 3) while it is routinely observed in YAG: Ce and LuAG: Ce radioluminescence spectra at RT47. In GAGG: Ce, part of the energy of the incoming ionization radiation absorbed in the host lattice is transferred to the Ce3+ centers through the Gd-sublattice48 which causes slower components in the scintillation response. Such a negative feature is not present in YAGG: Ce crystal. Finally, due to favorable segregation coefficient of both the cerium and magnesium ions in YAGG lattice we achieved the acceleration of scintillation response below the photoluminescence decay time of Ce3+ (about 50-55ns in YAGG host) which is explained by the quenching effect in closely spaced Ce-Mg pairs analogously to heavily doped GAGG: Ce, Mg49. Summarising, given the above description the YAGG host appears the most favorable in comparison with YAG, LuAG and GAGG ones in terms of fast and efficient energy transfer towards cerium emission centers in scintillation mechanism which is further reflected also in its ultralow afterglow demonstrated in Fig. 6.

Conclusions

In this study, the Cz growth of YAGG: Ce, Mg single crystals were investigated. As initially anticipated, pronounced evaporation of gallium oxide occurred, leading to oxidation of the iridium crucible and subsequent deposition of iridium metallic films on the melt surface. These effects significantly hindered stable crystal growth in the early stages of this study. Based on these findings, optimization of the growth atmosphere and other growth conditions were undertaken. The mitigation of gallium oxide evaporation, which effectively eliminated iridium deposition on the melt surface, enabled the successful Cz growth of the YAGG: Ce, Mg single crystal up to solidification fraction of 45%.

The grown YAGG: Ce, Mg achieved a light yield of 46,700 photons/MeV, energy resolution of 8.5 ~ 11.4% at 662 keV, low afterglow, and a fast decay time of approximately 50 ns, which are comparable to or surpass those of GAGG: Ce, Mg. Furthermore, the variation in light yield up to a solidification fraction of 45% was limited to 8.5%, and the afterglow remained below the 640 ppm level. Our analysis indicates that the uniformity of scintillation properties within the grown crystal correlates with the compositional homogeneity of the constituent elements and dopants. In conclusion, compared with the currently used GAGG: Ce, Mg scintillator for PCCT, we have demonstrated that a scintillator with higher light yield, superior afterglow characteristics, and no adverse effects from the Gd escape peak can be grown as a YAGG: Ce, Mg bulk crystal by the Cz method while maintaining uniformity in its scintillation properties. In particular, the market price of iridium metal has surged by an order of magnitude in recent years, creating a major cost challenge for the mass production of oxide scintillators. Degradation of Ir crucibles has been a critical factor limiting the practical application and large-scale manufacturing of these crystals. In this study, we demonstrated that Cz growth using Ir crucibles can be conducted while suppressing crucible deterioration associated with gallium oxide evaporation. These achievements represent a significant step toward the practical realization and industrial deployment of YAGG: Ce, Mg scintillators.

Further improvement in scintillator performance, particularly in energy resolution, can be expected for YAGG: Ce, Mg through optimization of the Ce and Mg doping concentrations, as well as advancements in crystal scaling and quality. Moreover, for YAGG: Ce, Mg, increasing the oxygen partial pressure of the atmosphere during growth to further suppress gallium oxide evaporation is expected to improve mass producibility, enabling larger-diameter crystals and higher solidification yields while maintaining uniform scintillation properties. Notably, our research group has been developing a novel melt growth method, called Oxide Crystal Growth from a Cold Crucible (OCCC), to grow over 2 inch diameter bulk single crystals without using precious metal crucibles50. In the OCCC method, Ir crucibles are not required, making it possible to grow oxide crystals even under a 100% oxygen atmosphere. Future studies will focus on growing YAGG: Ce, Mg single crystals using the OCCC method.