Fabrication and Characterization of Ce-Doped LiCaAlF₆–CaF₂–Li₃AlF₆ and CaF₂–LiF–Li₃AlF₆ Scintillators for Thermal Neutron Detection

Abstract

In this study, we developed and characterized novel scintillators with Ce: LiCaAlF₆–CaF₂–Li₃AlF₆ and Ce: CaF₂–LiF–Li₃AlF₆ ternary systems for thermal neutron detectors. The eutectics were grown by the vertical Stochbarger-Bridgman (VB) technique, and their constituent phases were identified using powder X-ray diffraction and scanning electron microscopy. Radioluminescence spectra irradiated under an Ag-target X-ray tube and confirmed the 5d-4f and self-trapped exciton luminescence derived from Ce³⁺. Scintillation decay and pulse height measurements were performed using ²⁵²Cf and ⁶⁰Co sources. The Ce: CaF₂–LiF–Li₃AlF₆ sample exhibited approximately 5.6 times higher effective neutron sensitivity compared with a Ce: LiCaAlF₆ single crystal. A favorable decrease in the neutron discrimination threshold level (Q_th) due to reduced γ-ray emission was observed. ⁶Li-enriched Ce: CaF-based scintillators hold potential for nuclear decommissioning applications.

1. Introduction

In recent years, the global focus on nuclear reactor decommissioning has intensified, with significant progress made in dismantling aging nuclear infrastructures [1,2]. According to the World Nuclear Industry Status Report 2024, 23 units have been fully decommissioned, and decommissioning activities are underway for 213 additional units worldwide. The number of reactors undergoing decommissioning has a steady upward trend and is expected to continue increasing. In Japan, large-scale efforts are being directed toward the Fukushima Daiichi Nuclear Power Plant [3,4,5,6,7].

During nuclear decommissioning, it is essential to detect and monitor the fuel debris, radioactive waste, and other radioactive materials [8]. Passive neutron assay [9] is a key technique employed in nuclear material accounting and is particularly useful for the in situ assessment of dispersed nuclear fuel and for monitoring potential criticality. However, accurate neutron detection is challenging in practical applications. This is because the neutron emission rate tends to be low when only small quantities of nuclear fuel are present, whereas γ-ray backgrounds often exceed 10 Gy/h due to strong radioactive contamination. In environments with strong γ-ray backgrounds, distinguishing neutrons from γ-rays—commonly referred to as n/γ discrimination—is essential.

Traditionally, ³He proportional counters have been widely used as standard neutron detectors [10,11,12]. However, because of the global shortage of He gas, its cost has increased significantly. This has led to the development of alternative neutron detection technologies, such as ⁶Li, ¹⁰B, and ¹⁵⁷Gd [10]. In general, ⁶Li-doped inorganic scintillators exhibit a superior α/β ratio, which corresponds to the n/γ ratio, compared to organic scintillators. Thermal neutron detection using ⁶Li is founded on its nuclear reaction with incident neutrons.

$${}^{6}Li+n \to \alpha \left(2.05 \mathrm{M}\mathrm{e}\mathrm{V}\right)+{}^{3}H\left(2.73 \mathrm{M}\mathrm{e}\mathrm{V}\right)\hspace{1em}\hspace{1em}\left(Q=4.78 \mathrm{M}\mathrm{e}\mathrm{V}\right).$$

(1)

Tritons and α-particles release an energy of 4.78 MeV, known as the Q-value [13], which leads to a high scintillation light yield. These ⁶Li-containing inorganic scintillators allow for neutron detection without relying on pulse-shape discrimination (PSD) methods; instead, they employ pulse-charge discrimination techniques [14].

Among various ⁶Li-doped inorganic scintillators, Ce: LiCaAlF₆ (LiCAF) exhibits favorable properties, including fast decay time, low γ-ray sensitivity, low γ-ray light yield, and excellent chemical stability with non-hygroscopicity [15,16,17]. Our group has previously developed neutron detectors using ultra-thick (~100 μm) Ce: LiCAF and Li glass (KG2) and evaluated their n/γ discrimination performance under high γ-ray dose rate conditions [18]. As a result, while the effective neutron count rate (n_eff) based on pulse charges above the neutron threshold significantly decreased for the KG2 detector at γ-ray air kerma rates up to 2.97 Gy/h, the Ce: LiCAF detector maintained its n_eff, demonstrating its advantage for neutron detection in intense γ-ray fields. However, for further practical use, several technical challenges remain, including further thinning of the scintillator to reduce γ-ray sensitivity, enhancing neutron sensitivity by increasing ⁶Li content, and achieving a shorter decay time.

Over the last 20 years, our research has centered on the development of eutectic scintillators incorporating ⁶Li for the purpose of thermal neutron detection [19,20,21]. Figure 1 presents a schematic diagram of these materials, which are composed of two distinct phases: a neutron-capture phase, where ⁶Li undergoes nuclear transmutation, and a scintillation phase, where the resulting α-particles and tritons are absorbed, leading to the emission of visible light. Unlike conventional single-crystal scintillators, whose ⁶Li concentration is limited by their inherent chemical composition, the eutectic structure allows for a substantially higher lithium content.

In this study, the fabrication of binary eutectic scintillators using Ce: LiCAF and LiF was investigated. Here, LiF was selected as the neutron capture phase, and Ce: LiCAF was used as the scintillator phase, noting that Ce: LiCAF itself also contains ⁶Li. This allows for an increased ⁶Li concentration, enabling the fabrication of thinner samples while maintaining sufficient neutron sensitivity.

2. Materials and Methods

2.1. Eutectic Growth

The eutectics were grown using the VB method. Ce1%-doped LiCAF single crystals (C&A Corp.) were used as the starting materials for evaluating the scintillation performance. A piece of a single crystal was crushed into a powder. Based on the phase diagrams (Figure 2), powder mixtures were prepared with molar ratios of LiF (>99.9% purity): Ce: LiCAF = 48x: 52, where x = 1 corresponds to the reported eutectic composition, and x = 3 was investigated to increase the Li concentration. The mixed powders, with a total mass of approximately 0.5 g, were fed into a carbon crucible with an inner diameter of 14 mm and a height of 20 mm. To minimize the hydrolysis of the sensitive fluorides, the following fluorine pretreatment and graphite environment were selected. Baking was carried out at 150 °C under a vacuum of 10⁻⁴ Pa for at least 3 h to remove the trace amounts of moisture contained in the original powders. Subsequently, the atmosphere was purged with a gas mixture of Ar and CF₄ at a partial pressure ratio of 9:1, and heat treatment was conducted under a slightly negative pressure environment. The furnace output was maintained at 1.8 kW for 3 h and then promptly shut off to create a temperature gradient inside the crucible. The fabricated eutectics were sectioned and polished along the cross-sectional plane using a wire saw. Subsequently, the crystalline phases were characterized, and their scintillation performance was assessed.

2.2. Phase Identification

Phase identification was carried out via powder X-ray diffraction (XRD) using a D8 DISCOVER diffractometer (Bruker, Billerica, MA, USA). Measurements were conducted over a 2θ range of 10° to 60°, where 2θ denotes the diffraction angle between the incident X-ray beam and the detector. A Cu-Kα radiation source was employed with a tube voltage of 40 kV and a current of 40 mA. To examine the microstructure of the eutectic crystals, backscattered electron (BSE) imaging was performed using scanning electron microscopes (SEM), specifically the S-3400N (Hitachi Ltd., Tokyo, Japan) and JSM-7800F (JEOL Ltd., Tokyo, Japan).

2.3. Scintillator Performance

To evaluate the emission wavelength via radioluminescence (RL), an Ag-source X-ray tube operating at 30 kV and 130 mA was employed. The emitted light was analyzed using a CCD camera (iDus420-OE, Andor Technology, Belfast, UK) integrated with a spectrometer (SR-163, Andor Technology). The light yields and decay times of the prepared eutectics and a Ce: LiCAF single crystal as a reference sample with the same volume and thickness were evaluated individually under irradiation with ²⁵²Cf and ⁶⁰Co sources. The samples were mounted onto a photomultiplier tube (PMT; R7600-200, Hamamatsu, Shizuoka, Japan) using optical grease (KF-96H-60000CS, SHINETSU, Tokyo, Japan), and the resulting signals were recorded. The PMT was operated at a bias voltage of 700 V. For neutron measurements using ²⁵²Cf, a 5 cm thick polyethylene (PE) moderator was placed between the source and the sample, with a total separation of approximately 10 cm. For gamma-ray measurements with ⁶⁰Co, the source was positioned about 5 cm from the sample without a moderator. To evaluate the light yield and effective neutron sensitivity, pulse height spectra were acquired using a two-channel USB Wave Catcher module [23], and the spectra were fitted with a Gaussian function. The effective neutron sensitivity is defined in Equation (2)

$$n_{eff}=\sum n_c\left(Q\right)\left(Q>Q_{th}\right),$$

(2)

where n_c (Q) is a count rate per channel of neutron pulse height spectrum, and Q_th is the threshold of the neutron discrimination level for the pulse charge, defined as the maximum channel of the γ-ray pulse height spectrum.

Decay time measurements were performed by directly connecting the photomultiplier tube (PMT) to an oscilloscope (TDS3035B, TEKTRONIX, Beaverton, OR, USA). The decay curve was obtained by averaging 50 individual waveforms, from which the decay time was calculated.

3. Results and Discussion

3.1. Eutectic Growth

The grown samples are shown in Figure 3a,c, respectively. No mass loss was observed before and after the heat treatment, and the mass increased by approximately 0.5 wt.% in all cases. This slight increase is believed to be mainly due to carbon adhering to the surface of the samples, and its effect can be considered negligible by the subsequent cutting process. Figure 3b,d show photographs of 460 µm-thick wafers of the respective prepared crystals. As can be observed from the photographs of the wafers, the samples were optically transparent, but a quantitative evaluation was difficult.

3.2. Phase Identification

Figure 4 shows the powder XRD patterns of the samples with a nominal molar composition of LiF: LiCAF = 48x: 52. For the sample with x = 1, three phases—LiCAF (trigonal, P-31c, No. 163, PDF 77-0939), CaF₂ (cubic, Fm-3m, No. 225, PDF 89-4794), and Li₃AlF₆ (orthorhombic, Pna2₁, No. 33, PDF 22-1137)—were identified. For the sample with x = 3, three phases—CaF₂ (cubic, Fm-3m, No. 225, PDF 89-4794), LiF (cubic, Fm-3m, No. 225, PDF 04-0857), and Li₃AlF₆ (orthorhombic, Pna2₁, No. 33, PDF 22-1137)—were confirmed. Contrary to the proposed phase diagram in Figure 2, the LiF–LiCAF binary eutectic phase was not observed for either composition. Figure 5 shows BSE images of the x = 1 and x = 3 samples in the cross-sectional directions. The differences in the atomic numbers provided a contrast between each phase, and multiple phases were identified. The results of BSE images and powder XRD measurements indicate that in Figure 5a, the white phase corresponds to CaF₂, the gray phase to LiCAF, and the black phase to Li₃AlF₆; whereas in Figure 5b, the white phase corresponds to CaF₂, the gray phase to Li₃AlF₆, and the black phase to LiF.

To evaluate the synthesized ternary-phase microstructures within the context of the LiF–CaF₂–AlF₃ ternary system, the phase compositions were estimated using image analysis with ImageJ software [24] running on 64-bit Java 8. First, the area fractions of each phase were calculated from 20 BSE images. The area fractions were approximated as volume fractions and converted into molar fractions. As a result, the estimated molar ratios (%) of the x = 1 sample were approximately LiCAF: CaF₂: Li₃AlF₆ = 40:17:43, and for x = 3, CaF₂: LiF: Li₃AlF₆ = 35:35:30. Figure 6 shows these compositions plotted on the LiF–CaF₂–AlF₃ ternary-phase diagram. Eutectic points of LiF–CaF₂, CaF₂–Li₃AlF₆, and LiF–Li₃AlF₆ have been reported previously. The x = 3 sample, whose nominal composition lies within the CaF₂–LiF–Li₃AlF₆ triangular region, was therefore formed through a ternary eutectic reaction among these phases. By contrast, although the nominal composition of the x = 1 sample was also within the same compositional region, a LiCAF–CaF₂–Li₃AlF₆ ternary phase was formed instead. This deviation is likely attributable to a slight LiF deficiency caused by a reaction between LiF and the carbon crucible, which shifted the actual composition toward the LiCAF–CaF₂–Li₃AlF₆ region. However, further investigations are required to confirm this hypothesis. These results allow us to infer that the nominal composition at point P–LiCAF yields a LiCAF–CaF₂–Li₃AlF₆ system, whereas the composition at point P–LiF produces a CaF₂–LiF–Li₃AlF₆ system. Therefore, the formation of LiF–LiCAF eutectic compounds was unlikely.

3.3. Scintillation Properties

Figure 7 shows the RL spectra obtained under irradiation from an Ag-source X-ray tube. The black curve represents the RL spectrum of the reference Ce: LiCAF single crystal, and the colored curves correspond to the RL spectra of the ternary-phase samples. In the single crystal sample, emission peaks at 282 and 300 nm originating from the Ce³⁺ 5d–4f transitions in LiCAF were observed. In the ternary-phase sample with x = 3, the characteristic peaks at 320 and 337 nm due to the Ce³⁺ 5d–4f transitions in CaF₂, along with a peak at 286 nm attributed to self-trapped exciton luminescence, were detected. The emission wavelengths for Ce: LiCAF and Ce: CaF₂ were consistent with those reported in earlier studies by Getin et al. [15] and Yanagida et al. [28]. In the x = 1 sample, where both Ce: CaF₂ and Ce: LiCAF phases were present, emission originating from self-trapped excitation was scarcely observed. According to Yanagida et al. [28], self-trapped excitation in the RL spectra becomes prominent at a Ce concentration of approximately 0.1% but diminishes at approximately 0.5–3%. Therefore, the Ce concentration in the x = 1 sample was considered to be relatively high, whereas that in the x = 3 sample appeared to be much lower (approximately 0.1%). In the x = 1 sample, the 300 nm emission peak typically associated with Ce: LiCAF was not observed. This is likely because the excitation wavelength of CaF₂ is located at approximately 300 nm, resulting in reabsorption or suppression of Ce: LiCAF luminescence.

The pulse-height spectra of Ce: LiCAF and the ternary-phase samples under ²⁵²Cf (neutron source) and ⁶⁰Co (γ-ray source) irradiation are shown in Figure 8a–c, respectively. A Ce: LiCAF single-crystal standard with a light yield of 5000 photons/neutron (ph/n) was used as a reference. The light yields of the ternary-phase samples with x = 1 and 3 were calculated to be 1806 and 3944 ph/n, respectively. Although the light yield decreased in both ternary-phase samples owing to the reduced optical transparency compared with that of the Ce: LiCAF single crystal, the light yield of the sample with x = 3 reached approximately 4000 ph/n.

Subsequently, n_eff was calculated based on Equation (2). The n_eff of the Ce: LiCAF single crystal was found to be 0.708 s⁻¹, whereas those of the ternary-phase samples with x = 1 and 3 were 1.35 and 4.00 s⁻¹, respectively. In particular, the x = 3 sample exhibited the greatest improvement, with a n_eff value approximately 5.6 times higher than that of a single crystal. Moreover, Ce: CaF₂ are known to exhibit a low light yield under γ-ray excitation, typically less than 1000 ph/MeV [29]. The significant enhancement in n_eff can be attributed to the increased Li content compared with that of the Ce: LiCAF single crystal, as well as a reduced Q_th presumably due to the low γ-ray light yield of Ce: CaF₂.

Figure 9 presents the scintillation decay profiles of the Ce:LiCAF single crystal and the ternary-phase samples, measured under excitation by 662 keV γ-rays from a ⁶⁰Co source and neutrons from a ²⁵²Cf source. The decay curves were fitted using a bi-exponential model, as described by Equation (3)

$$I=\sum A_{i}exp\left(-{\displaystyle \frac{t}{{\tau }_i}}\right), i-1, 2,$$

(3)

where A_i is a constant, t is time, τ is the decay time, and I is the scintillation intensity. The ratio was calculated as: A₁τ₁/(A₁τ₁ + A₂τ₂) × 100 (%).

Table 1. Comparison of the properties of the prepared Ce: LiCAF single crystal and ternary-phase samples.

	neff (Ce:LiCAF = 1)	Light Yield (ph/n)	Decay Time (ns)	Density (g/cm3)	Hygroscopicity
Ce: LiCAF single crystal	1	5000	n: 46 (100%) γ: 49 (100%)	2.98	No
Ce: (LiCAF)0.4(CaF2)0.17(Li3AlF6)0.43 (x = 1)	1.9	1806	n: 61 (28%), 304 (72%) γ: 76 (100%)	2.92	No
Ce: (CaF2)0.35(LiF)0.35(Li3AlF6)0.3 (x = 3)	5.6	3944	n: 103 (20%), 650 (80%) γ: 165 (100%)	2.89	No

summarizes the decay time results of the Ce: LiCAF single crystal and the ternary-phase samples under excitation by the γ-ray source ⁶⁰Co and the neutron source ²⁵²Cf, along with the corresponding effective neutron sensitivity and neutron-induced light yield. An increasing trend in the decay time was observed with increasing x values, which was attributed to the decrease in Ce concentration. The decay times of both ternary-phase samples were slower than that of the Ce: LiCAF single crystal, with the x = 3 sample, presumed to have a trace Ce concentration of approximately 0.1%, exhibiting the slowest decay. Although neutron- or γ-ray-induced decay characteristics have not been reported, a significant increase in scintillation decay time has been observed at Ce concentrations of approximately 0.1% under X-ray excitation. Optimization of the Ce concentration of the samples is left for future investigation.

4. Conclusions

In this study, a Ce-doped LiF–LiCAF eutectic scintillator was fabricated using the VB method. For the nominal molar composition of LiF: LiCAF = 48x: 52, a ternary phase consisting of Ce: LiCAF–CaF₂–Li₃AlF₆ was obtained at x = 1, which corresponds to the previously reported LiF–LiCAF eutectic composition. At a higher LiF content (x = 3), the resulting material comprised ternary phases of Ce: CaF₂–LiF–Li₃AlF₆. These findings were interpreted in the context of the LiF–CaF₂–AlF₃ ternary-phase diagram, leading to the conclusion that the formation of a binary LiF–LiCAF eutectic is unlikely.

Both ternary-phase samples exhibited emission peaks at 286, 320, and 337 nm, which are attributed to Ce³⁺ in the CaF₂ scintillation phase. Among them, the x = 3 sample exhibited the highest light yield and thermal neutron sensitivity, achieving 3944 ph/n and an effective sensitivity approximately 5.6 times greater than that of the Ce: LiCAF single crystal. This significant improvement can be ascribed to the increased lithium content, which enhances neutron capture efficiency, and the lower γ-ray emission from Ce: CaF₂, which in turn contributes to a reduced Q_th value. Although the scintillation decay times of sample x = 3 for both neutron and γ-ray excitations were slower than those of the Ce: LiCAF single crystal—exhibiting 650 ns (80%) and 103 ns (20%) for neutrons and 165 ns (100%) for γ-rays—there remains significant potential to improve the decay characteristics through optimization of the Ce concentration. In conclusion, it was confirmed that the newly discovered Ce: CaF₂–LiF–Li₃AlF₆ ternary eutectic exhibits higher neutron sensitivity and reduced γ-ray response compared with the Ce:LiCAF single crystal, which is used in the existing detector [18]. We plan to proceed with the development of a new prototype detector incorporating the ternary eutectic and evaluate its performance as a thermal neutron detector.