1 Introduction

Scintillation is a phenomenon whereby a material absorbs ionizing radiation and rapidly converts the absorbed energy into low-energy visible photons. Scintillation materials (scintillators), capable of detecting ionizing radiation, find applications in various fields such as gamma-ray detection in positron emission tomography, X-ray detection in medical imaging and industrial inspection [1, 2], and electromagnetic calorimetry [3]. For gamma- and X-ray detection, scintillators are required to have higher scintillation light yields, higher energy resolutions, shorter decay times, higher density, lower deliquescence, and lower production costs [4]. However, scintillators fulfilling all these requirements are currently not available.

Many researchers have focused on developing scintillators based on various inorganic crystals (e.g., halides and oxide crystals) [5]. Among the inorganic crystalline scintillators recently developed, cerium-activated LaBr3 (LaBr3:Ce) [6] and CeBr3 [7] have been widely commercialized in many applications. LaBr3:Ce has a high light yield (61000 photons/MeV), excellent energy resolution (2.8% at 662 keV gamma-ray), and high density (5.29 g/cm3). CeBr3 has a high light yield (60000 photons/MeV), high energy resolution (4.1% at 662 keV gamma-ray), high density (5.18 g/cm3), and short decay time (17 ns). Although inorganic crystals have appropriate scintillation properties, some industrial challenges remain in terms of their high production costs, long fabrication times, and difficulty in production scalability [8, 9].

In contrast to conventional inorganic crystals, inorganic glasses offer several industrial advantages in terms of their easily controllable shape, production scalability, low production cost, and high chemical durability in atmospheric conditions [10,11,12,13]. Therefore, using scintillators based on inorganic glasses is the primary method for resolving the industrial problems associated with commercial inorganic crystals. To date, glass hosts based on various glasses (e.g., silicate, aluminosilicate, borosilicate, borate, and phosphate) have been used to develop scintillators, and Ce has been selected as the emission center [14]. This is because Ce3+ ions exhibit a high emission quantum efficiency and short decay time (approximately 30–60 ns) owing to parity- and spin-allowed transitions from the Ce3+ 5d excited state to the 4f ground states [15]. However, inorganic glasses mostly have lower scintillation light yields compared with commercial inorganic crystals, because the former’s low energy transfer efficiency from the host to the emission centers in the scintillation process. For example, a commercially available Ce-doped Ga3(Al,Ga)5O12 crystal has a light yield of 46000 photons/MeV with an absolute photoluminescence quantum yield (PL QY) of 89.1% [16], whereas Ce-doped Al2O3–B2O3–SrF2 [17], Ce-doped SiO2–B2O3–Al2O3–Gd2O3 [18], and Ce-doped BaF2–Al2O3–B2O3 [19], have light yields of 240, 1300, and 1800 photons/MeV and PL QYs of 83.5, 59.6, and 70.1%, respectively.

We previously reported the synthesis and properties of an inorganic glass scintillator with a high light yield (3200 photons/MeV) and good energy resolution (14.4% at 662 keV gamma-ray), using phosphate glass doped with Ce3+ ions [20]. Phosphate glasses have high optical transparency in the ultraviolet-visible-near-infrared region, low melting points, and high thermal durability, offering industrial advantages over other glass hosts [21,22,23]. The reported glass also has a density of 3.33 g/cm3 and a short scintillation decay time of about 35 ns owing to the transitions from the Ce3+ 5d excited state to the 4f ground states. Comparatively, this glass has a higher density than the commercially available Li-glass GS20, commonly used for thermal neutron detection (with a density of approximately 2.2 g/cm3) [24, 25], but falls below the densities of widely commercialized inorganic crystals and previously reported glasses containing heavy elements, such as Ba and Gd. For example, Tl-doped NaI (NaI:Tl+) crystal [26], Ce-doped SiO2–Al2O3–BaF2–Gd2O3 [27], and Ce-doped SiO2–LiF–GdBr3 glasses [28] exhibit densities of 3.67 g/cm3, 4.2 g/cm3, and approximately 4.4 g/cm3, respectively.

In this study, we synthesized Al(PO3)3–CsPO3–Ba(PO3)2–CeBr3 glasses using phosphate glasses including Cs and Ba doped with Ce3+ ions. We selected Cs, which can achieve a higher scintillation light yield and higher energy resolution than other alkali metals based on our previous investigation [29], and Ba, which has a higher atomic number than Cs and Sr used in our pervious study [20]. In addition, Al(PO3)3 was used to form a glass network and suppress hygroscopicity [30], and CeBr3 was used to suppress the oxidation of Ce3+ to Ce4+ via creating a reducing environment. We investigated the structural, photoluminescence, and scintillation properties of Ce-doped Al(PO3)3–CsPO3–Ba(PO3)2 glasses having different Ce concentrations, and the Ce concentrations were optimized.

2 Experimental methods

Al(PO3)3 (99.99%, High Purity Chemicals Co., Ltd.), CsH2PO4 (99%, Mitsuwa Chemicals Co., Ltd.), Ba(PO3)2 (99%, Mitsuwa Chemicals Co., Ltd.), and CeBr3 (99.99%, Sigma-Aldrich) were used as raw materials.

Glass scintillators with the formula 0.6Al(PO3)3–1.5CsPO3–0.4Ba(PO3)2–xCeBr3 (x = 0, 0.05, 0.1, 0.2, 0.3, and 0.4) were synthesized by a conventional melt-quenching method. Al(PO3)3, CsH2PO4, Ba(PO3)2, and CeBr3 powders were mixed in stoichiometric ratios at room temperature. The mixed materials were placed in quartz tubes and dried at 200 °C for about 30 min under vacuum in an electric furnace. Next, the dried materials in quartz tubes melted at 1000 °C for about 30 min under vacuum. The melt was quenched in water at 20 °C for about 1 s to partially cool down, then immediately removed from the water to avoid thermal shock, and left in air to gradually cool down to room temperature. The raw material- and oxide-based composition of the prepared glasses are summarized in Table 1. Hereinafter, we represented the prepared glass composition with oxide-based composition.

Table 1 Raw material- and oxide-based composition of prepared glasses having different Ce concentrations
Full size table

2.1 Structural measurements

We estimated the densities of the prepared Ce-doped P2O5–Al2O3–Cs2O–BaO glasses using an electronic densimeter (DH-300, CgoldenWall, JP). X-ray diffraction (XRD) analyses were recorded using an Ultima IV diffractometer (Rigaku Corporation) with Cu-Kα radiation.

2.2 Photoluminescence measurements

The transmission spectra were measured using a spectrophotometer (UV-2700, SHIMADZU). A spectrofluorometer (F-7000, Hitachi High-Tech Corporation) equipped with a Xe lamp was used to record the PL excitation and emission spectra at room temperature. A Quantaurus-QY instrument (C11347, Hamamatsu Photonics) was used to measure the absolute photoluminescence quantum yields (PL QYs) and PL excitation/emission maps. The PL decay time profiles (excitation wavelength: 255 nm, emission wavelength: 350 nm) were collected with a fluorescence lifetime spectrofluorometer (DeltaFlex 3000U-TMK2, Horiba). The fabricated samples were excited using a 255-nm LED. The absorption spectrum of 0.6Al(PO3)3–1.5CsPO3–0.4Ba(PO3)2 (undoped sample) was obtained using a spectrophotometer (UV-2700, SHIMADZU).

2.3 Scintillation measurements

We obtained X-ray-induced radioluminescence spectra at room temperature with a QE Pro spectrometer (Ocean Optics) and an X-ray generator (XGD2300-HK, Rigaku Corporation). The scintillation decay time profiles were recorded using the measurement system shown in our previous report [29]. Gamma-ray-induced scintillation pulse-height spectra were obtained at room temperature to estimate the scintillation light yields of the fabricated glasses using 662 keV gamma rays generated by a 137Cs source. The detail of the setup for the pulse-height spectra is provided in our previous report [29].

3 Results and discussion

3.1 Structural properties

Photographs of Ce-doped P2O5–Al2O3–Cs2O–BaO glass scintillators having different Ce concentrations, captured under visible light at room temperature, are shown in Fig. 1. All the glass scintillators were transparent and nondeliquescent without regard to the Ce concentration (even the high Ce concentration of 13.8 mol%). To confirm the amorphous nature of Ce-doped P2O5–Al2O3–Cs2O–BaO having different Ce concentrations, we measured the XRD patterns of the fabricated samples (Fig. 2). For all the fabricated samples, we observed halo features indicating long-range structural disorder of amorphous network and no sharp peaks indicating the presence of crystalline phases. Therefore, the fabricated samples are considered to be amorphous regardless of the Ce concentrations. Table 2 summarizes the densities of Ce-doped P2O5–Al2O3–Cs2O–BaO glass scintillators having different Ce concentrations. The densities of the fabricated glasses increased with increasing the Ce concentration, with the glass with the Ce concentration of 13.8 mol% having the highest density (3.57 g/cm3), which exceeds those of earlier studied glass scintillators having high light yields such as Ce-doped Al2O3–SrO–Cs2O–P2O5 (3.33 g/cm3) [20] and is comparable to the commercial NaI:Tl+ (3.67 g/cm3) [26]. Figure 3 shows the transmission spectra of Ce-doped P2O5–Al2O3–Cs2O–BaO glasses having different Ce concentrations. All the fabricated glasses have the high transmittance of approximately 80% in visible light range and exhibited intense absorption around 350 nm, which corresponds to the 4f–5d transition of Ce3+ ions [17].

Fig. 1
figure 1

Photographs of Ce-doped P2O5–Al2O3–Cs2O–BaO glass scintillators having different Ce concentrations under visible light at room temperature

Full size image
Fig. 2
figure 2

XRD patterns of Ce-doped P2O5–Al2O3–Cs2O–BaO glasses having different Ce concentrations (intervals, 0.02°; scan rate, 1° min−1). au: arbitrary units

Full size image
Table 2 Densities of Ce-doped P2O5–Al2O3–Cs2O–BaO glasses having different Ce concentrations
Full size table
Fig. 3
figure 3

Transmission spectra of Ce-doped P2O5–Al2O3–Cs2O–BaO glasses having different Ce concentrations

Full size image

3.2 Photoluminescence properties

The PL excitation and emission spectra of Ce-doped P2O5–Al2O3–Cs2O–BaO glass scintillators having different Ce concentrations are shown in Fig. 4. The PL excitation spectra were recorded for emission at 350 nm, with two excitation peaks observed in the ranges 230–260 and 280–320 nm (Fig. 4). Because the excitation peak wavelengths were approximately consistent with earlier1 studied Ce-doped glasses [31, 32], these excitation peaks were ascribed to the transitions from the Ce3+ 4f ground states to the 5d excited state. The emission peak of the PL emission spectra recorded under excitation at 240 nm, was found in the range 300–440 nm, with a peak centered around 350 nm (Fig. 4). This observation aligns with the emission peaks observed in previously studied Ce-doped glasses [33, 34]. These emission peaks can be clearly attributed to the transitions from the Ce3+ 5d excited state to the 4f ground states. In addition, the e mission peak slightly red-shifted with increasing Ce concentration. Similar trends were reported for other phosphate glasses doped with Ce with different Ce concentrations [34, 35], which can be explained by the lowering of the bottom of the 5d levels and the self-absorption of Ce3+ ions with the increase in Ce concentration.

The PL QYs (recorded three times for each fabricated glass under excitation at 320 nm) are shown in Fig. 5. The average PL QYs were almost constant for each fabricated glass, as indicated by the colored dots. These PL QY values (approximately 96.0%) are higher than those for earlier studied Ce-doped glass scintillators exhibiting high light yields (above 2000 photons/MeV), such as Al(PO3)3–CsPO3–CsBr–CeBr3 (approximately 91.0%) [35], Al(PO3)3–CsPO3–CeBr3 (approximately 92.0%) [29], and Ce-doped Al2O3–SrO–Cs2O–P2O5 (approximately 92.0%) [20]. We also measured PL QYs for excitation at 250–340 nm and PL excitation/emission maps of Ce-doped P2O5–Al2O3–Cs2O–BaO glasses having different Ce concentrations (Table 3 and Fig. 6). We observed an intense emission band at 340–380 nm under excitation wavelength at 300–340 nm for all the fabricated glasses, and no pronounced decrease of PL QYs was observed in the excitation-emission overlap region.

Figure 7 shows the PL decay time profile, in which the decay time constant was approximated using a single exponential function. The decay time profiles were similar for each fabricated glass. Table 4 summarizes the estimated decay time constants of Ce-doped P2O5–Al2O3–Cs2O–BaO glasses having different Ce concentrations. We observed short decay time constants (32.1–33.8 ns) for each fabricated glass. Because similar decay time constants have been observed in earlier studied phosphate glasses doped with Ce [20, 29, 35], they are clearly ascribed to the transitions from the Ce3+ 5d excited state to the 4f ground states.

Fig. 4
figure 4

PL excitation and emission spectra of Ce-doped P2O5–Al2O3–Cs2O–BaO glasses having different Ce concentrations for emission at 350 nm and for excitation at 240 nm

Full size image
Fig. 5
figure 5

PL QYs of Ce-doped P2O5–Al2O3–Cs2O–BaO glasses having different Ce concentrations for excitation at 320 nm. The colored dots indicate the average PL QYs

Full size image
Table 3 PL QYs for excitation at 250–340 nm
Full size table
Fig. 6
figure 6

PL excitation/emission maps of Ce-doped P2O5–Al2O3–Cs2O–BaO glasses having different Ce concentrations

Full size image
Fig. 7
figure 7

PL decay time profile of 1.96 mol% Ce-doped P2O5–Al2O3–Cs2O–BaO glasses on a semilogarithmic plot (excitation wavelength: 255 nm, emission wavelength: 350 nm). The black line is a fit using a single exponential function

Full size image
Table 4 PL decay time constants of Ce-doped P2O5–Al2O3–Cs2O–BaO glasses having different Ce concentrations
Full size table

3.3 Scintillation properties

Figure 8 shows the X-ray-induced radioluminescence spectra of Ce-doped P2O5–Al2O3–Cs2O–BaO glasses having different Ce concentrations. For each glass, we observed an emission peak located in the range 300–440 nm with a maximum around 360 nm, which agrees well with the PL emission spectra (Fig. 4) and clearly indicates the transitions from the Ce3+ 5d excited state to the 4f ground states. When the concentration of Ce increased from 1.96 to 13.8 mol%, the emission peak slightly red-shifted. Similar to the PL emission spectra in Fig. 4, the red-shift is attributed to the lowering of the bottom of the 5d levels and the self-absorption of Ce3+ ions occurring with increasing Ce concentration.

To estimate the scintillation light yields of the fabricated glasses for gamma- and X-ray detection, we acquired the 137Cs gamma-ray-irradiated pulse-height spectra for the prepared glasses and a commercially available lithium glass scintillator (GS20) as a reference sample (Fig. 9). The spectra for all the fabricated glasses revealed a clear photoabsorption peak. The energy resolutions, defined as the full width at half maximum of the photoabsorption peak, and the peak channels were approximated by fitting the peaks with a Gaussian function. Furthermore, we approximated the light yields by comparing the estimated peak channels of the fabricated glasses with that of commercial GS20, which is widely known to have the scintillation light yield of 4000 photons/MeV for gamma-ray [36, 37, 14]. Table 5 summarizes the estimated peak channels, light yields, and energy resolutions of the samples and commercial GS20. The light yield increased with the increase in Ce concentration from 1.96 to 7.41 mol%, but was nearly constant from 7.41 to 13.8 mol%. The light yield is obtained by the product of the number of generated electron–hole pairs, the energy transfer efficiency from the host to the emission centers, and the PL QY [38]. Here, the number of generated electron–hole pairs was considered constant for each sample because the fabricated glasses have the same host (same bandgap energy) regardless of the Ce concentration, and because the PL QYs were almost constant for each glass (Fig. 5). Based on the light yield formulas, the observed concentration-yield dependence is related to difference in the energy transfer efficiency, indicating that the efficiency increases with the increase in Ce concentration from 1.96 to 7.41 mol%, but is nearly constant from 7.41 to 13.8 mol%. The light yields with the Ce concentrations of 7.41, 10.7, and 13.8 mol% were approximately 2200 photons/MeV, which exceeds those of earlier studied glasses containing heavy elements, such as Eu-doped SiO2–Al2O3–HfO2 (630 photons/MeV) [39], Ce-doped P2O5–Li2O–GdI3–Al2O3–CaCO3 (1600 photons/MeV) [30], Ce-doped BaF2–Al2O3–B2O3 (1800 photons/MeV) [19], and Ce-doped SiO2–Al2O3–BaF2–Gd2O3 (2050 photons/MeV) [27]. We consider that the high scintillation light yields of the fabricated glasses in this study mainly result from suppression of the oxidation of Ce3+ into Ce4+ by using CeBr3. The estimated energy resolutions also improved with the increase in the Ce concentration from 1.96 to 7.41 mol%, but were nearly constant from 7.41 to 13.8 mol%. (Table 5). The energy resolutions with the Ce concentrations of 7.41, 10.7, and 13.8 mol% (15.2–15.8%) were comparable to those of GS20 (15.0%). In addition, considering the results of this study and our previous study, we suggest that energy transfer efficiency is different for different alkali earth metals of Sr and Ba. We fabricated 0.6Al(PO3)3–1.5CsPO3–0.4Ba(PO3)2–0.3CeBr3 (10.7 mol% Ce-doped 41P2O5–6Al2O3–15Cs2O–8BaO) glass in this study and 0.6Al(PO3)3–1.5CsPO3–0.4Sr(PO3)2–0.3CeBr3 (10.7 mol% Ce-doped 41P2O5–6Al2O3–15Cs2O–8SrO) glass in previous study [20]. To estimate bandgap energy of the former glass containing Ba, we obtained an absorption spectrum of 0.6Al(PO3)3–1.5CsPO3–0.4Ba(PO3)2 as an undoped sample (Fig. 10). The absorption edge of 0.6Al(PO3)3–1.5CsPO3–0.4Ba(PO3)2 was observed in the wavelength range of 310–320 nm, which is consistent with the absorption edge of 0.6Al(PO3)3–1.5CsPO3–0.4Sr(PO3)2. This result indicates that both glasses have similar bandgap energies and the number of generated electron–hole pairs is similar regardless of the different alkali earth metals of Ba and Sr. In addition, the average PL QYs were similar: the PL QYs of the 0.6Al(PO3)3–1.5CsPO3–0.4Ba(PO3)2–0.3CeBr3 and 0.6Al(PO3)3–1.5CsPO3–0.4Sr(PO3)2–0.3CeBr3 were approximately 95.0% and 92.0%. The scintillation light yields of the fabricated Al(PO3)3–CsPO3–Ba(PO3)2–CeBr3 and the previously reported Al(PO3)3–CsPO3–Sr(PO3)2–CeBr3 were 2150 and 3200 photons/MeV, respectively. Considering the formula of the light yield [38], the difference in the light yields can be explained by the difference in the energy transfer efficiency. Therefore, we suggest that the energy transfer efficiency of phosphate glass containing Ba is lower than that of phosphate glass containing Sr. Because approaches to improve energy transfer efficiency are unclear, this finding will facilitate the development of novel glass scintillators for gamma- and X-ray detection having high scintillation light yields.

We measured the scintillation decay time profiles (Fig. 11), then the scintillation decay time constants were estimated using the sum of two exponential functions. The decay time profiles were similar for each prepared glass. The scintillation decay time constants of the Ce-doped P2O5–Al2O3–Cs2O–BaO having different Ce concentrations are summarized in Table 6. We observed the first (faster) and second (slower) exponential components, whose contributions were approximately 70% and 30% for Ce concentrations of 1.96 and 3.85 mol% and approximately 80% and 20% for Ce concentrations of 7.41, 10.7, and 13.8 mol%. The decay time constants (28.5–34.0 ns) and PL decay time constants (32.1–33.8 ns; Table 4) of the faster components correspond to the transitions from the Ce3+ 5d excited state to the 4f ground states, which are comparable with those of earlier studied phosphate glasses doped with Ce having short decay time, such as Al(PO3)3–CsPO3–CsBr–CeBr3 (29.1 ns) [35], Al(PO3)3–CsPO3–CeBr3 (28.9 ns) [29], and Ce-doped Al2O3–SrO–Cs2O–P2O5 (38.0 ns) [20]. In addition, the decay times of the fabricated glasses are shorter than those of widely commercialized inorganic crystals having short decay time, such as Lu1.8Y0.2SiO5: Ce3+ (LYSO: Ce3+; about 40–50 ns) [14, 40]. The decay time constants of the slower components in the scintillation decay profiles were not observed in the PL decay profiles (Fig. 7). This observed trend is considered to be related to the difference between PL and scintillation process: in PL process, emission centers are photo-excited from their ground state to their excited states, from which they can return to the ground state by emitting light. Meanwhile, scintillation process involves energy-transfer process from the glass host to the emission centers in addition to excitation and emission at the emission centers. During the energy-transfer process, trapping and detrapping of electrons and holes at defect levels in the glass host may occur. Then, while occurring repeated trapping and detrapping, the electrons and holes are finally trapped by the emission centers, which results in slow energy transfer. Thus, we consider that the observed slower components in scintillation decay result from the slow energy transfer from the glass host to the Ce3+ ions acting as the emission centers. Similar faster and slower components have also been observed in previously studied Ce-doped phosphate glasses, such as Ce-doped P2O5–Li2O–GdI3–Al2O3–CaCO3 (faster components, 39.9 ns; slower components, 221.7 ns) [30] and Ce-doped P2O5–Li2CO3–GdBr3–Al2O3 (faster components, 44 ns; slower components, 471 ns) [28].

Fig. 8
figure 8

X-ray-induced radioluminescence spectra of Ce-doped P2O5–Al2O3–Cs2O–BaO glasses having different Ce concentrations at room temperature

Full size image
Fig. 9
figure 9

137Cs gamma-ray-irradiated pulse-height spectra of Ce-doped P2O5–Al2O3–Cs2O–BaO glasses having different Ce concentrations and commercial GS20 (bias voltage, 600 V; shaping time, 0.5 µs)

Full size image
Table 5 Peak channels, light yields, and energy resolutions of Ce-doped P2O5–Al2O3–Cs2O–BaO glasses having different Ce concentrations and commercial GS20
Full size table
Fig. 10
figure 10

Absorption spectrum of 0.6Al(PO3)3–1.5CsPO3–0.4Ba(PO3)2 as an undoped sample

Full size image
Fig. 11
figure 11

Scintillation decay time profile of 1.96 mol% Ce-doped P2O5–Al2O3–Cs2O–BaO glasses on a single logarithmic plot (accumulated total counts, 100,000). The black line is a fit with the exponential functions

Full size image
Table 6 Scintillation decay time constants of Ce-doped P2O5–Al2O3–Cs2O–BaO glasses having different Ce concentrations
Full size table

4 Conclusions

Ce-doped P2O5–Al2O3–Cs2O–BaO glass scintillators having different Ce concentrations were synthesized, and their structural, PL, and scintillation properties were investigated. The density of the prepared glass increased with the increase in Ce concentration, with the glass for Ce concentration of 13.8 mol% showing the highest density (3.57 g/cm3), which is higher than our earlier studied Ce-doped Al2O3–SrO–Cs2O–P2O5 (3.33 g/cm3) glass scintillator and comparable to the commercial NaI:Tl+ scintillator (3.67 g/cm3). The emission peak ascribed to the transitions from the Ce3+ 5d excited state to the 4f ground states was observed in the PL emission and X-ray-induced radioluminescence spectra for all the prepared glasses. The scintillation light yield increased with the increase in Ce concentration, from 1.96 to 7.41 mol% but remained nearly constant at 7.41–13.8 mol%. The light yields with the Ce concentrations of 7.41, 10.7, and 13.8 mol% were approximately 2200 photons/MeV, which exceeds earlier reported for glasses exhibiting high light yields, such as Ce-doped P2O5–Li2O–GdI3–Al2O3–CaCO3 (1600 photons/MeV) and Ce-doped BaF2–Al2O3–B2O3 (1800 photons/MeV). For all the fabricated glasses, we observed short scintillation decay times (28.5–34.0 ns), corresponding to the transitions from the Ce3+ 5d excited state to the 4f ground states. These decay times are shorter than those of widely commercialized inorganic crystals with short decay times. Considering these results, the optimal Ce concentration is 13.8 mol%, and corresponding glass scintillator has a suitable performance for gamma- and X-ray detection.