1 Introduction

Scintillators convert high-energy photons into visible or ultraviolet (UV) photons. X- and gamma-ray scintillators are widely used in diagnostic imaging [1], security inspection systems [2], astrophysics [3], and high-energy physics [4]. An ideal scintillator for X- and gamma-ray detection should have a high light yield, fast scintillation decay, and a high effective atomic number (Zeff). However, commercially available scintillators such as NaI: Tl [5], CsI: Tl [6], Bi4Ge3O12 (BGO) [7], and Lu2SiO5:Ce (LSO: Ce) [8] cannot simultaneously satisfy all these requirements. Therefore, users must choose an appropriate scintillator that fulfills their particular requirements, highlighting the necessity for new scintillators that meet al.l these criteria.

Scintillator materials are classified into two major types: organic and inorganic. Inorganic scintillators have a higher Zeff and density (ρ) compared to organic scintillators, which are composed of light elements like carbon and hydrogen. A high Zeff is crucial because the photoelectric interaction of scintillators with X- and gamma-rays is proportional to the fifth power of their Zeff [9]. In addition, inorganic scintillators show higher light yields than organic scintillators (e.g., inorganic scintillators: − 80,000 photons/MeV, organic scintillators: − 16,000 photons/MeV) [10]. Here, inorganic scintillators are more suitable for X- and gamma-ray detection. Inorganic scintillators are primarily classified into halide-based and oxide-based scintillators. Halide-based scintillators have relatively small bandgaps (4–7 eV), compared to oxide-based scintillators (7–9 eV) [11]. The light yield can be estimated using the following formula (1) [10]:

$$\:\text{L}\text{Y}={\left(\beta\:{E}_{g}\right)}^{-1},$$
(1)

where LY is the light yield, \(\:\beta\:\) is the ratio of the average energy required for incident radiation to make one e–h pair, to band gap, and \(\:{E}_{g}\) is the bandgap energy. The light yield is inversely proportional to the bandgap energy, and higher light yields and energy resolutions can be expected from halide-based scintillators compared to oxide-based scintillators. Thus, halide-based scintillators tend to exhibit higher light yield and energy resolutions. This has led to extensive development of halide-based scintillators [12,13,14]. However, a disadvantage of existing halide-based scintillators is their high deliquescence, making them difficult to handle. Commercially available NaI: Tl scintillators are generally enclosed in an aluminum case. Moreover, halide-based commercial scintillators (e.g., NaI: Tl (ρ = 3.6 g/cm3, Zeff = 50.8), CsI: Tl (ρ = 4.53 g/cm3, Zeff = 54.1)) have lower density and Zeff compared to oxide-based scintillators (e.g., BGO (ρ = 7.13 g/cm3, Zeff = 75.2), LSO: Ce (ρ = 7.40 g/cm3, Zeff = 66.4)) [10]. Therefore, new types of halide-based scintillators are being developed to overcome these limitations.

Extrinsic luminescence due to dopants and intrinsic luminescence, such as self-trapped excitons (STE) and lattice defects, have been proposed as luminescence mechanisms of halide-based scintillators. Extrinsic luminescence is generated by doping elements, which could serve as luminescent centers, into host materials to form impurity levels internal within the band gap. Most dopant candidates are rare-earth elements exhibiting 5d–4f transitions, such as Ce3+, Eu2+, and Pr3+ [15, 16], and ns2 elements exhibiting s–p transitions, such as Tl+, Pb2+, and Sb3+ [17, 18], due to their high luminescence efficiency. Intrinsic luminescence refers to the specific luminescence of host materials such as BaF2 [19], CsCu2I3 [20], CaF2 [21], SrI2 [22], Cs3Cu2I5 [23], Cs2HfCl6 and Cs2ZrCl6 [24].

Among scintillators that exhibit intrinsic luminescence, the first thallium-based chloride crystal scintillator was the Tl2LiGdCl6 crystal [25]. Since its discovery, various thallium-based chloride crystal scintillators have been developed, such as TlCaCl3 [26], Tl2LaCl5 [27], Tl2GdCl5 [28], TlGd2Cl7 [29], Tl2LiScCl6 [30], and Tl2LiYCl6 [31]. This is attributed to the high Zeff of thallium-based chloride crystal scintillators, which is due to the large atomic number of Tl (81). We also investigated thallium-based chloride crystals that exhibit intrinsic luminescence, as listed in Table 1. Among them, TlCdCl3 crystals have no deliquescence, indicating good stability in air. However, TlCdCl3 crystals readily crystallize due to their congruent composition, and their light yields are relatively inferior to those of other thallium-based chloride crystals. Therefore, if the light yield of undoped TlCdCl3 crystals can be improved, TlCdCl3 crystal scintillators could potentially replace existing halide-based scintillators.

Table 1 Zeff and light yield of thallium-based chloride crystals
Full size table

In this study, we selected Sb3+ as the dopant cation for the undoped TlCdCl3 crystal to improve the light yield. This choice was based on a previous report [38] that demonstrated that Sb3+ can serve as an efficient luminescence center in CsCdCl3 crystals. These crystals have the same octahedral symmetry in their crystal structure as undoped TlCdCl3. In addition, several studies have revealed the improvement in light yield by doping Sb3+ in halide-based scintillators [39,40,41]. Therefore, we fabricated both undoped TlCdCl3 crystals and TlCdCl3:Sb crystals and compared their luminescence and scintillation properties.

2 Experimental procedures

Undoped TlCdCl3 and TlCdCl3:Sb crystal samples were prepared using a self-seeding solidification method. They were grown using TlCl (99.9%, Mitsuwa Pure Chemicals), CdCl2 (99.99%, ALDRICH Chemistry), and SbCl3 (99.9%, FUJIFILM Wako Pure Chemical Corporation) powders. To determine the optimal dopant-concentration, the powders were mixed in a molar ratio of TlCl: CdCl2:SbCl3 = 1:1:0, 1:0.995:0.005, 1:0.99:0.01, and 1:0.985:0.015 to fabricate TlCdCl3:Sb crystals with 0, 0.5, 1.0, and 1.5 mol%, respectively. The mixtures were then placed in quartz tubes and dried under vacuum at 0.067 Pa and 453 K for 5 h to remove water. Subsequently, the dried mixtures were transferred into vacuum-sealed quartz ampoules. These ampoules were then placed in a tube furnace, heated to 1023 K, maintained at that temperature for 2 h, and then cooled to room temperature over 56 h. After the growth process, a portion of the as-grown crystal was cut and polished for photoluminescence and scintillation measurements. Figure 1 shows a photograph of the as-grown crystals, with the 1.5 mol% TlCdCl3:Sb crystal appearing opaque and yellow, while the 0, 0.5, and 1.0 mol% TlCdCl3:Sb crystals appeared translucent and colorless.

Fig. 1
figure 1

Photograph of undoped TlCdCl3 crystal and TlCdCl3:Sb crystals (the scale bar is in millimeters (mm))

Full size image

To identify the obtained phase, the X-ray diffraction (XRD) patterns were measured using a Rigaku Ultima IV-PXS diffractometer with Cu Kα radiation. The photoluminescence (PL) spectra were obtained using a fluorescence spectrophotometer (Hitachi High-Tech F-7000) equipped with a 150 W Xe lamp at room temperature. Absolute photoluminescence quantum yields (PL QYs) were measured using a Quantaurus-QY instrument (C11347, Hamamatsu Photonics). PL decay time profiles were obtained using a fluorescence lifetime spectrofluorometer (Horiba, DeltaFlex 3000U-TMK2) at room temperature. A light-emitting diode (Horiba, NanoLED-250) with an emission wavelength of 255 nm was used as the excitation source. An optical cut filter was used to remove the scattered excitation light. X-ray-excited radioluminescence (XRL) spectra were obtained using an X-ray generator (Rigaku, D2300-HK) operating at 40 mA and 40 kV with a CuKα target used as the excitation source. Scintillation from the samples was directed to a charge-coupled device (CCD)-based spectrometer (Ocean Insight, QEpro) via an optical fiber. Scintillation decay time profiles were measured using the delayed coincidence method [42], employing the original setup described in a previous report [43]. In this setup,22Na was used as the gamma-ray radiation source. Additionally, 137Cs-gamma-ray pulse height spectra were recorded using our original setup [43], with a shaping time of 10 µs for all samples, including a commercially available Gd2SiO5:Ce (GSO) scintillator used as the reference sample.

3 Results and discussion

All the crystals were stable in air and did not show any deliquescence characteristics. Figure 2 shows the XRD patterns of both undoped TlCdCl3 and TlCdCl3:Sb crystals. Almost all the peaks observed in the XRD patterns of both undoped TlCdCl3 crystal and TlCdCl3:Sb crystals matched those of the reference data for undoped TlCdCl3 crystal (ICSD 39807). This indicates that all the crystals had a single phase without the precipitation of secondary phases. However, the doping of Sb3+ did not cause a noticeable shift in the diffraction peaks of the XRD patterns. The substitution of cations with different valences and the deficiency of Tl+ or Cd2+ for charge compensation may distort the crystal lattice, making it difficult to establish a correlation with the Sb3+ dopant concentration.

Fig. 2
figure 2

XRD patterns of undoped TlCdCl3 crystal and TlCdCl3:Sb crystals

Full size image

Figure 3 shows the PL spectra of both undoped TlCdCl3 and TlCdCl3:Sb crystals measured at room temperature. The excitation bands for the undoped TlCdCl3 crystal and TlCdCl3:Sb 0.5, 1.0, and 1.5 mol% crystals were observed at 280, 300, 295, and 295 nm, respectively (with emission wavelengths of 460, 480, 470 and 480 nm, respectively). A broad emission band with a peak at 470 nm is observed in the emission spectrum of the undoped TlCdCl3 crystal under excitation at 280 nm. In both emission spectra of TlCdCl3:Sb 0.5 and 1.0 mol% crystals, an emission band was observed at 480 nm under excitation at 300 and 280 nm, respectively. In the spectrum of the TlCdCl3:Sb 1.5 mol% crystal, an emission band was observed at 500 nm under excitation at 290 nm. Compared to the measurement results of the undoped TlCdCl3 crystal (excitation peak: 280 nm, emission peak: 460 nm), the peak wavelengths shifted to longer wavelengths in both emission and excitation peaks, indicating the formation of new luminescent centers. The Stokes shifts were 14,000 cm−1 for the undoped TlCdCl3 crystal, 13,100 cm−1 for the TlCdCl3:Sb 0.5 and 1.0 mol% crystal, and 14,500 cm−1 for the TlCdCl3:Sb 1.5 mol% crystal. Thus, the Stokes shifts of the TlCdCl3:Sb crystals cannot be distinguished from those of undoped TlCdCl3 crystals. Examining the emission origin in the TlCdCl3:Sb crystal is beyond the scope of this study; however, we speculate that the emission centers are self-trapped excitons (STE) [44], F-centers [45], localized Sb3+ ions [46], defects due to charge compensation [47], or mixed emission centers. Further studies, including the temperature dependence of luminescence, band gap measurements, and chemical composition analysis, are necessary for a clear understanding of this phenomenon. In addition, the PL QYs of TlCdCl3:Sb 0.5, 1.0, and 1.5 mol% crystals were recorded in triplicate for excitation at 300, 300, and 295 nm, respectively. The average PL QYs of TlCdCl3:Sb 0.5, 1.0, and 1.5 mol% crystals were 15.6, 25.3, and 8.4%, respectively. The TlCdCl3:Sb 1.0 mol% crystal showed the highest PL QY among the TlCdCl3:Sb crystals, and the PL QY decreased in the TlCdCl3:Sb 1.5 mol% crystal due to concentration quenching.

Fig. 3
figure 3

PL excitation spectra (dashed lines) and emission spectra (solid lines) of undoped TlCdCl3 crystal (λex = 280 nm, λem = 460 nm), TlCdCl3:Sb 0.5 mol% crystal (λex = 300 nm, λem = 480 nm) TlCdCl3:Sb 1.0 mol% crystal (λex = 280 nm, λem = 470 nm) and TlCdCl3:Sb 1.5 mol% crystal (λex = 290 nm, λem = 480 nm)

Full size image

Figure 4 shows the PL decay time profiles of undoped TlCdCl3 and TlCdCl3:Sb crystals at room temperature, with excitation and measurement wavelengths of 255 and 480 nm, respectively. The PL decay curves were fitted to two exponential decay functions, and the estimated PL decay time constants are listed in Table 2. Percentage percentiles in the brackets contribution rates each PL decay time constants to the emission of undoped TlCdCl3 crystal and TlCdCl3:Sb crystals. For the undoped TlCdCl3 crystals, the PL decay time constants were 76 and 2490 ns. Similarly, for the TlCdCl3:Sb 0.5 mol% crystal, the PL decay time constants were 96 and 2430 ns; for the TlCdCl3:Sb 1.0 mol% crystal, 49 and 2780 ns; and for the TlCdCl3:Sb 1.5 mol% crystal, 59 and 2450 ns. Notably, the PL decay constants remained unaffected by the addition of Sb3+, indicating a consistent emission mechanism between the Sb-doped samples and the undoped TlCdCl3 sample.

Fig. 4
figure 4

PL decay time profiles of undoped TlCdCl3 crystal and TlCdCl3:Sb crystals  (λex  = 255 nm, λem  = 480 nm)

Full size image
Table 2 PL decay time constants of undoped TlCdCl3 crystal and TlCdCl3:Sb crystals
Full size table

Figure 5 shows the XRL spectra of the undoped TlCdCl3 and TlCdCl3:Sb crystals measured at room temperature. The spectrum of the undoped TlCdCl3 crystal exhibits a broad emission band at 485 nm. The spectra of the TlCdCl3:Sb crystals exhibit a broad emission band at approximately 510 nm. The peak wavelengths of the Sb3+-doped samples shift towards longer wavelengths compared to those of the undoped TlCdCl3 crystals. This behavior is similar to that of the emission band in the PL spectra shown in Fig. 3. Thus, this X-ray-induced scintillation may be attributed to the same origin as the photoluminescence.

Fig. 5
figure 5

XRL spectra of undoped TlCdCl3 crystal and TlCdCl3:Sb crystals

Full size image

Figure 6 shows the scintillation decay time profiles of undoped TlCdCl3 and TlCdCl3:Sb crystals at room temperature. These decay curves were fitted using two exponential decay functions, and the scintillation decay time constants are listed in Table 3. For the undoped TlCdCl3 crystal, the scintillation decay time constants were 51 and 2613 ns. Similarly, for the TlCdCl3:Sb 0.5 mol% crystal, they were 81 and 6310 ns; for the TlCdCl3:Sb 1.0 mol% crystal, 80 and 7350 ns; and for the TlCdCl3:Sb 1.5 mol% crystal, 65 and 4550 ns. In both undoped TlCdCl3 and TlCdCl3:Sb crystals, the dominant scintillation decay components were slower than those of the photoluminescence components. The scintillation process involves three major stages: conversion, transport, and luminescence. In the conversion stage, high-energy electrons and holes are produced in the material. Subsequently, in the second stage, these electrons and holes migrate towards luminescent centers, albeit not all are transferred due to trapping and non-radiative recombination. Finally, in the luminescence stage, the electrons and holes reaching the luminescence centers induce luminescence [48]. In contrast, PL decay does not consider the first two stages. Therefore, scintillation decay components are often longer than those of PL, as observed in this study.

Fig. 6
figure 6

Scintillation decay time profiles of undoped TlCdCl3 crystal and TlCdCl3:Sb crystals

Full size image
Table 3 Scintillation decay time constants of undoped TlCdCl3 crystal and TlCdCl3:Sb crystals
Full size table

Figure 7 shows the pulse height spectra of 137Cs-gamma-ray for an undoped TlCdCl3 crystal, TlCdCl3:Sb crystals, and the commercial GSO scintillator (light yield = 10,000 photons/MeV) at room temperature. The sample sizes range from 2.0 to 3.0 mm in length, 4.0 to 5.0 mm in width, and about 1.0 mm in thickness compared with the GSO reference with dimensions of 6.0 × 6.0 × 2.0 mm. The spectra of all samples exhibited the 662 keV gamma-ray photo-peak. The photo-peak channel numbers were 221 for the undoped TlCdCl3 crystal, 426 for the TlCdCl3:Sb 0.5 mol% crystal, 634 for the TlCdCl3:Sb 1.0 mol%, 328 for the TlCdCl3:Sb 1.5 mol% crystal, and 571 for GSO. The PMT quantum efficiency values were 27% (485 nm) for the undoped TlCdCl3 crystal, 22% (508 nm) for the TlCdCl3:Sb 0.5 mol% crystal, 20% (511 nm) for the TlCdCl3:Sb 1.0 mol% crystal, 19% (514 nm) for the TlCdCl3:Sb 1.5 mol% crystal, and 37% (450 nm) for GSO. Based on the photo-peak channel numbers and PMT quantum efficiency values, the light yields were calculated as follows: 2700 photons/MeV for the undoped TlCdCl3 crystal, 6,300 photons/MeV for the TlCdCl3:Sb 0.5 mol% crystal, 10,300 photons/MeV for the TlCdCl3:Sb 1.0 mol% crystal, and 5500 photons/MeV for the TlCdCl3:Sb 1.5 mol% crystals. Among the TlCdCl3:Sb crystals, the TlCdCl3:Sb 1.0 mol% crystal exhibited the highest light yield, surpassing both GSO and BGO (light yield = 8600 photons/MeV). The variation in light yield was proportional to the PL QY, as shown in Fig. 8, indicating that the increase in light yield is primarily due to changes in PL QYs values.

Fig. 7
figure 7

137Cs-gamma-ray pulse height spectra of the undoped TlCdCl3 crystal, TlCdCl3:Sb crystals and GSO commercial scintillator

Full size image
Fig. 8
figure 8

Light yield and PL QYs of TlCdCl3:Sb crystals as a function of Sb concentration

Full size image

4 Conclusion

Undoped TlCdCl3 crystal and TlCdCl3:Sb crystals were fabricated using a self-seeding solidification method to develop scintillators with high Zeff for X-ray and gamma-ray detection applications. The luminescence and scintillation properties of these crystals were investigated to confirm that Sb3+ acts as an efficient luminescence center in the undoped TlCdCl3 crystal. In the XRL spectra, the undoped TlCdCl3 crystal exhibited a broad emission band at 485 nm, while the TlCdCl3:Sb crystals showed emission bands at 510 nm. The peak wavelengths of the TlCdCl3:Sb crystals were shifted to longer wavelengths compared to the undoped TlCdCl3 crystal. Moreover, the light yield of the undoped TlCdCl3 crystal and the TlCdCl3:Sb 0.5, 1.0 and 1.5 mol% crystals were estimated to be 2700, 6300, 10,300 and 5500 photons/MeV, respectively. This result indicates that doping the undoped TlCdCl3 lattice with Sb3+ increased the scintillation light yield by up to 4 times that of the undoped TlCdCl3 crystal. Consequently, we successfully developed new halide-based scintillators with high light yields that surpass the properties of GSO.