Luminescence Efficiency and Spectral Compatibility of Cerium Fluoride (CeF₃) Inorganic Scintillator with Various Optical Sensors in the Diagnostic Radiology X-ray Energy Range

Abstract

The aim of this study was to experimentally assess the luminescence efficiency of a cerium fluoride (CeF₃) inorganic scintillator in crystal form as a possible alternative to high-luminescence but hygroscopic cerium bromide (CeBr₃). The experiments were performed under typical diagnostic radiology X-rays (50–140 kVp). Parameters such as the crystal’s absolute luminescence efficiency (AE) and the spectral matching with a series of optical detectors were examined. The replacement of bromine with fluorine appeared to drastically reduce the AE of CeF₃ compared to CeBr₃ and other commercially available inorganic scintillators such as bismuth germanate (Bi₄Ge₃O₁₂-BGO). CeF₃ reaches a maximum luminescence efficiency value of only 0.8334 efficiency units (EUs) at 140 kVp, whereas the corresponding values for CeBr₃ and BGO were 29.49 and 3.41, respectively. Furthermore, the emission maximum (at around 313 nm) moved towards the lower part of the visible spectrum, making CeF₃ suitable for spectral coupling with various photocathodes and photomultipliers applied in nuclear medicine detectors, but completely unsuitable for spectral matching with CCDs and CMOS. The obtained luminescence efficiency results denote that CeF₃ cannot be applied in medical imaging applications covering the range 50–140 kVp; however, examination of its luminescence output in the nuclear medicine energy range (~70 to 511 keV) could reveal possible applicability in these modalities.

1. Introduction

Research on novel inorganic scintillators for a variety of applications, including medical applications, is an ongoing task for research groups worldwide [1,2,3,4,5,6,7,8,9,10,11]. An efficient single-crystal scintillator with a light yield of the order of 6 × 10⁴ photons/MeV is cerium bromide (CeBr₃). It also has a decay time of only 19 ns [12] and a spectrum lying in the lower part of the visible range, making it suitable for coupling with various optical detectors used in nuclear medicine [13,14,15,16,17,18,19,20,21].

However, CeBr₃ is hygroscopic and should be encapsulated in order to maintain its structural integrity, leading to a reduction in the luminescence signal [22]. A material with a similar structure with the bromine (Br) atom replaced by fluorine (F) is cerium fluoride (CeF₃) [13,23,24,25,26,27,28].

CeF₃ has been in the spotlight in the last 30 years due to experiments including the Large Hadron Collider (LHC) [29,30,31,32,33,34,35,36,37,38]. Since CeF₃ single crystals have attractive optical, mechanical and luminescence characteristics, they are suitable for a variety of applications [25,29,39,40,41,42]. CeF₃ crystals are rare-earth trifluorides with a tysonite structure [43]. The crystalline structure can be hexagonal, as per the crystal sample used in our study, at high temperatures and trigonal at low ones [43]. CeF₃’s hexagonal-shape crystalline structure includes layers of cerium ions situated very close together. At the same time, fluorine ions occupy three distinct positions inside the lattice [43,44,45]. Previous X-ray diffraction (XRD) analysis exhibits characteristic peaks that agree with the hexagonal structure of the crystal [43,44,45]. Optical and structural properties, such as a large Verdet constant (247 rad/Tm at 450 nm), high transmittance of the crystal, about 92% compared to terbium gallium garnet (TGG) and a great cutoff absorption edge (at 270 nm compared to 400 nm for the TGG crystal), render it a good scintillator candidate for magneto-optical materials and mostly for various optical applications in the region of high-energy physics [46,47].

Furthermore, a CeF₃ crystal’s mechanical properties reveal an important anisotropy. Young’s modulus anisotropy (E_max/E_min = 195 GPa/100 GPa = 1.95), shear modulus anisotropy (G_max/G_min = 65.7 GPa/34.2 GPa = 1.92), Poisson’s ratio (0.16–0.48) and Vickers microhardness (2.3–2.9 GPa) make CeF₃ crystals very useful for technologies that need flexible and durable materials [45].

It has been discovered in earlier studies that CeF₃ responds non-proportionally to gamma rays up to 5.1 MeV [48,49]. Nonetheless, the analysis of CeF₃ monocrystals demonstrates that they are appropriate for high-rate calorimetry. Tests utilizing CeF₃-based prototype calorimeters (such as the high-luminosity Large Hadron Collider, HL-LHC) demonstrated encouraging results in terms of energy resolution [36,37,38].

According to recent studies, a CeF₃ crystal is an excellent Faraday rotator for broadband high-precision spectroscopy, particularly suitable for high-power laser applications [40,50,51,52]. This material appears with better characteristics than the commercial TGG crystal in terms of transparency in the UV-NIR region [46]. Additionally, a CeF₃ mono-crystal as an optical isolator is an advantageous alternative for protecting laser devices from back reflection, particularly in the UV-NIR wavelength ranges [40,46,50].

An essential disadvantage of a CeF₃ inorganic scintillator is that it tends strongly to overlap the luminescence band with the absorption cutoff, which is due to the fact that the Stokes shift is small and the material is a self-activated scintillator [39,40,41,43,44,46,47,48,49,53,54].

The aim of this work was to experimentally assess the general performance (emission efficiency and spectral compatibility with optical sensors) of a CeF₃ crystal under conditions found in medical laboratories as well as in various other applications such as industrial, geological, archeological, in security, etc. Most of these applications use mainly continuous spectrum X-rays produced by suitable X-ray tubes.

The findings of this work were compared with a CeBr₃ crystal that was recently examined by our group as well as to a crystal that has been incorporated in commercially available nuclear medicine scanners, i.e., bismuth germanate (Bi₄Ge₃O₁₂-BGO) [1,55]. The properties of the three materials are shown in

Table 1. Properties of CeF₃, CeBr₃ and BGO single crystals [1,18,19,26,39,40,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75].

	Units	CeF3	CeBr3	BGO
Wavelength of max emission	Nm	340 (slow), 300–310 (fast) [40,68]	380 [1,18]	480 [19,74]
Emission wavelength range	Nm	250–425 [68]	325–475 [1]	375–650 [1]
Decay times	Ns	30 (slow), 8 (fast) [40]	19 [1,18]	300 [19,56]
Light yield	photons/MeV	4.4 × 103 [40]	6 × 104 [1,18,26]	8.9 × 103 [19]
Photoelectron yield	% of NaI:Tl	8 [68]	122 [1]	15–20 [74]
Radiation length	Cm	1.654 [68]	1.96 [1]	1.12 [76]
Refractive index @ max nm		1.62 @ 440 nm [40]	2.09 @ 340 nm [1]	2.15 @ 480 nm [77]
Density	g/cm³	6.16 [40]	5.1 [1]	7.13 [19]
Melting point	°C	1443–1450 [39,40,71]	722 [1]	1050 [74]
Mechanical hardness	Mohs	4.5–5 [39]	5–6 [1]	5 [78]
Radiation hardness	Rad	>106 [73]	2 × 103 [1]	107 [79]
Hygroscopic		No [40]	Yes [1,18]	No [19]

.

2. Results

Figure 1 shows the signal reading produced by the CeF₃ crystal, after X-ray irradiation, compared to the X-ray exposure rate. In the same Figure can be seen the output signal of the CeBr₃ and BGO crystals, under similar irradiation conditions. All the crystal samples had a purity of >99.99% (personal communication with Advatech (London, UK). From Figure 1, it can be depicted that the output signal and the exposure rate of the medical X-ray tube can be considered linear in the examined energy range (0 to 372 mR/s), except for CeBr₃ where linearity is lost for less than 50 mR/s. However, the replacement of bromine with fluorine in CeBr₃ results in a significant reduction in the signal output curve (from 29.54 in CeBr₃ to 0.834 in CeF₃). The output signal of BGO, which is a commercially available crystal in medical imaging applications, is also greater than that of CeF₃, denoting that CeF₃ has a low light yield for medical applications in the energy range 50–140 kVp.

Figure 2 shows the experimental results regarding the absolute luminescence efficiency of CeF₃ compared to the CeBr₃ and Bi₄Ge₃O₁₂ crystals. The results are shown for the available kVp of the X-ray tube, with steps of 10 kVp. A common finding, for the three examined crystals, is that in the examined X-ray range, the luminescence efficiency increases when the tube voltage increases. The increase in the luminescence efficiency is more evident in the range 50–80 kVp, due to the fact that the K-edges of all three materials are in this energy range. CeF₃ reaches a maximum value of only 0.8334 EU (EU is the S.I. equivalent μWm⁻²/(mGy/s) at 140 kVp), whereas the corresponding values for CeBr₃ and BGO are 29.49 and 3.41, respectively, despite the higher density (6.16 compared to 5.1 g/cm³) of CeF₃ compared with CeBr₃ (Table 1) [1]. Luminescence efficiency measurements in higher available X-ray energies could reveal a more efficient performance that possibly fulfils the requirements for nuclear medicine applications.

Figure 3 shows the spectral matching values of CeF₃ with a broad variety of optical sensors. The emission spectrum of CeF₃ in the range of 260–400 nm and with an emission maximum at around 313 nm, which shifted to the shortwave region by 67 nm compared to CeBr₃, limits the available sensors commonly used in medical imaging applications that could be coupled with, such as charge coupled devices (CCDs) or complementary metal oxide semiconductors (CMOS) [80].

The main luminescence mechanism, observed in the CeF₃ crystal, is mainly due to 5d-4f transitions of the cerium (Ce³⁺) ions [25]. Electron transitions from the 4f to the 5d energy level occur when these ions (CeF₃) are excited. The return of electrons to the 4f state produces photons. Due to this mechanism, the two primary luminescence bands of the crystal are created. These bands are located at around 290 nm and 340 nm, respectively [25,81,82]. Previous studies have demonstrated an association between the emission at 290 nm and the 5d-4f transitions in cerium (Ce³⁺) ions. This band is produced by the electrons that radiatively decay from the excited 5d ground state to the 4f ground state [25,81,82]. Frenkel self-trapped excitons are produced as a result of the presence of sub bands in the conduction band, which are obtained from the 5d states of the ions and provide the procedure of the luminescence [25,43,44,45]. The electronic structure and the luminescence mechanism of CeF₃ crystal are valuable in evolving scintillators for radiological applications and for other applications in high-energy physics [40,46,47,50,51,52].

As can be depicted from Figure 3, the spectral matching factors between CeF₃ and most of the CCDs and CMOS are close to zero. However, CeF₃ can be matched with a variety of optical sensors, such as silicon (SiPMs), position-sensitive photomultipliers, etc., which are used in nuclear medicine applications [83,84]. For example, the emission spectrum of CeF₃ has a 97% match with the spectral sensitivity of multialkali photocathodes and flat-panel position-sensitive (PS) photomultipliers (H8500D-03) (for both, the spectral matching factor is 0.97). Lower compatibility values can be found with gallium arsenide (GaAs) photocathodes (α_s = 0.8) and the extended (E-S20) photocathode used to measure the luminescence efficiency signal in this work (α_s = 0.76). This value is also used for the calculation of the AE.

The spectral responses that were used for the calculation of the α_s between CeF₃ with a series of light sensors are shown in Figure 4 and Figure 5, along with the emitted spectrum of CeF₃ [68]. Figure 4 shows the grouped spectral responses of the photocathodes (Figure 4a) and silicon photomultipliers (Figure 4b), whereas Figure 5 shows the grouped spectral responses of the CCDs (Figure 5a) and CMOS (Figure 5b).

The CeF₃ spectral matching results, presented in the spectral matching factors data of Figure 3, were used to calculate the EE that could be produced upon coupling the specific crystal with the examined optical sensors. Figure 6 and Figure 7 show the calculated effective luminescence efficiency values with the X-ray tube voltage for the above-mentioned optical sensors. Figure 6 shows the grouped spectral responses of the photocathodes (Figure 6a) and silicon photomultipliers (Figure 6b), whereas Figure 7 shows the grouped spectral responses of the CCDs (Figure 7a) and CMOS (Figure 7b).

Following the spectral matching factor results, the effective efficiency is maximized for multialkali photocathodes and flat-panel position-sensitive photomultipliers, since in these cases, most of the emitted light from the CeF₃ scintillator can be adequately collected by the detectors.

The combination of the multialkali photocathode with CeF₃ results in the effective efficiency values being reduced by 2.58% compared to the absolute efficiency values at 140 kVp. In the case of the combination of CeF₃ with the gallium arsenide photocathode, the corresponding absolute efficiency reduction is 24.26% at 140 kVp.

Typical EE values for single-crystal scintillators, already integrated in medical imaging systems, range from 3 EE for BGO (used, for example, in the GE Discovery IQ scanner, GE Healthcare, Milwaukee, WI, USA) [85], 5 EE for GSO:Ce (Philips Gemini GXL PET/CT, Philips Medical Systems, Eindhoven, The Netherlands) [86], 8 for LYSO:Ce (Philips Gemini TF PET/CT, Philips Medical Systems, Eindhoven, The Netherlands) [87] to 12 for LSO:Ce (Siemens Biograph TruePoint PET/CT, Siemens Healthineers, Forchheim, Germany) [88,89].

Figure 8 shows the X-ray luminescence efficiency of the CeF₃ crystal compared to the data for CeBr₃ and BGO, in the examined X-ray tube voltage range. The XLE is an important parameter for detectors that are operated as energy integrators. This type of sensor is used in medical imaging systems, in which the signal output is related to the amount of energy that is captured by the crystal. The CeF₃ XLE values are below the corresponding ones for CeBr₃ and BGO as was expected from the previous findings (Figure 1 and Figure 2). They start from 3.2 × 10⁻⁴ at 50 kVp and drop down to 2.3 × 10⁻⁴ at 140 kVp. The corresponding values for CeBr₃ and BGO were 7.6 × 10⁻³ and 7.3 × 10⁻⁴ at 50 kVp, reaching a maximum of 8.9 × 10⁻³ (CeBr₃ at 90 kVp) and 1.1 × 10⁻³ (BGO at 80 kVp) and afterwards, both drop down to 8.2 × 10⁻³ and 9.4 × 10⁻⁴ at 140 kVp, respectively. In the case of CeF₃, even the characteristic K-edge (at ~40 keV as shown in Figure 9) does not contribute to an XLE increase, like in the case of CeBr₃, with a K-edge at almost the same energy with CeF_{3 (}~40 keV) or BGO (~90.5 keV).

Figure 9 shows the calculated attenuation coefficients (μ_att) and the energy absorption (μ_en) coefficients of CeF_3, CeBr₃ and BGO, in the energy range 8 to 150 keV. The elemental coefficients (cerium, fluorine, bromine, bismuth (Bi), germanium (Ge) and oxygen (O)) that were used to calculate the CeF_3, CeBr₃ and BGO mixtures were obtained from the NIST XmuDat database [90,91]. The decrease in the coefficients, except from the sudden increase due to the characteristic K-absorption, for every mixture point, where the probability of photoelectric interaction maximizes, can also be seen in Figure 9.

Figure 10 shows the EAE values of CeF₃ compared to CeBr₃ and BGO. The energy absorption efficiency strongly depends on the μ_en and μ_att coefficients, shown in Figure 9, as well as on their ratio (μ_en/μ_att, Equation (4)), shown as an inset in Figure 10. Initially, for energies up to 39 keV, CeF₃ maintained a higher attenuation coefficient ratio than both CeBr₃ and BGO, leading to a higher energy absorption efficiency value at 40 kVp. However, when the energy increases, the energy absorption of BGO showed superior energy absorption efficiency values across the examined energy range, due to the combined effects of the density, higher μ_en/μ_att ratio (from 39 to 90 keV) and the increase in absorption due to the K-edge. Both CeBr₃ and CeF₃ have close attenuation coefficient values due to the presence of cerium (gram atomic mass = 140.116 g/mol) in both crystals; however, the influence of the three atoms of bromine (gram atomic mass = 79.904 g/mol) is stronger than the corresponding three atoms of fluorine (gram atomic mass = 18.998 g/mol), leading to a gram molecular mass for CeF₃ equal to 197.111 g/mol compared to 172.549 g/mol for CeBr₃. After the K-edge absorption, CeF₃ exhibits a lower μ_en/μ_att ratio and consequently lower EAE values up to 129 keV across the examined energy range.

3. Materials and Methods

The cubic-shaped (edge of 10 mm) and polished CeF₃ crystal was obtained from Advatech (London, UK) [68]. For the irradiation of the sample, a typical radiological X-ray tube (CPI Inc., Georgetown, ON, Canada, model CMP 200 DR high-frequency X-ray generator and an X-ray tube (IAE SpA, Milano, Italy, model SpA-RTM90HS) was used, with voltage settings of 50–140 kVp, tube current fixed at 63 mAs, large focus (1.2 mm), a source to detector distance of 72.5 cm and an additional aluminum (Al) filter with a thickness of 20 mm.

The emitted light energy produced by the crystals per unit area $\left({\Psi }_{\Lambda }\right)$ was determined as the output signal in the energy range that diagnostic medical X-ray tubes allow (50–140 kVp). The methodology is described in detail in [1].

The CeF₃ sample was wrapped with Teflon tape (only one face was free) and inserted in the input port of an integrating sphere ((Newport Corp., Irvine, CA, USA, Oriel model 70451). The light that was produced by the sample, after X-ray excitation, was collected by a photomultiplier tube (PMT, (EMI, London, UK, EMI model 9798). The PMT’s photocathode was the extended sensitivity (S20) type. The dose was monitored using an RTI Piranha P100B dosimeter (Mölndal, Sweden). In our experiments, the PMT works with a 22 volts difference between the initial dynode (all dynodes shorted) and the photocathode. The produced current is recorded by a sub-femtoamp remote source meter (Keithley, Cleveland, OH, USA, model 6430).

The ratio of the light energy flux emitted by the examined sample, normalized by the X-ray exposure rate, can be expressed as the absolute luminescence efficiency:

$$AE={\eta }_A={\displaystyle \frac{{\Psi }_{\Lambda }}{X}}=\left({\displaystyle \frac{i_{elec}}{S{\eta }_p{\alpha }_s{c}_g}}\right){\dot{X}}^{-1},$$

(1)

In Equation (1), ${\Psi }_{\Lambda }$ is the light energy flux (output signal) in units of μW m⁻². $X$ is the exposure rate (mR s⁻¹). i_elec is the current produced by the electrometer in pA and S denotes the surface of the crystal, excited by X-rays (mm²). The peak sensitivity of the photocathode (η_p) is expressed in units of pA/W. ${\alpha }_s$ is the spectral matching between the light source (in this case, crystal) to the spectral response of the optical sensor (in this case, the photocathode). Finally, the percentage of the sample’s emitted light escaping the integrating sphere and reacting with the PMT is the geometric light collection efficiency ($c_g$) with a measured value of 15.6. The units of the luminescence efficiency are EU = (μW m⁻²)/(mR s⁻¹). The experimental error is within ±3.4% due to the combined effects of the reproducibility of the X-ray tube output and optical measurement setups.

If the AE is reduced by the spectral matching factor (${\alpha }_s$) (denoting the percent of the light produced by the scintillator, which is within the same wavelength range as the optical response of a light sensor), then the effective efficiency is derived [1]:

$$EE={\eta }_{eff}={\eta }_A{\alpha }_s$$

(2)

The effective efficiency was calculated for various optical detectors, often used in medical applications, while the spectrum of CeF₃ was obtained from manufacturer data [68].

The X-ray luminescence efficiency (XLE) is the ratio of the crystal’s emitted light energy flux over the X-ray energy flux (${\eta }_{\Psi }={\Psi }_{\Lambda }/{\Psi }_0$) [1].

Calculation of the XLE can be performed by the conversion of the X-ray exposure (in Equation (1)) to the X-ray energy flux $({\Psi }_o$), using ${\Psi }_o=X\widehat{\Psi }$ where $\widehat{\Psi }$ is the X-ray energy flux per exposure rate:

$$\widehat{\Psi }={\displaystyle \frac{{\displaystyle \int }{\Psi }_0\left(E\right)dE}{{\displaystyle \int }{\Psi }_0\left(E\right)\left[\frac{{\rm X}}{{\Psi }_0\left(E\right)}\right]dE}}={\displaystyle \frac{{\displaystyle \int }{\Psi }_0\left(E\right)dE}{{\displaystyle \int }{\Psi }_0\left(E\right)\left[{\left(\frac{{\mu }_{en}\left(E\right)}{\rho }\right)}_{air}·{\left(\frac{W_A}{e}\right)}^{-1}\right]dE}}$$

(3)

${({\mu }_{en}\left(E\right)/\rho )}_{air}$ is the X-ray mass energy absorption coefficient of the air and $W_A/e$ is the average energy/unit of charge, for the production of an electron/ion pair in the air.

The energy absorption efficiency is used to account for the energy locally absorbed at the points of X-ray interaction, within the crystal [1]:

$$EAE\left(E\right)={\displaystyle \frac{{{\displaystyle \int }}_0^{E_0}{\Phi }_0\left(E\right)E\left(\frac{{\mu }_{en}\left(E\right)/\rho }{{\mu }_{att}\left(E\right)/\rho }\right)\left(1-e^{-\left({\mu }_{att}\left(E\right)/\rho \right)\rho W}\right)dE}{{{\displaystyle \int }}_0^{E_0}{\Phi }_0\left(E\right)EdE}}$$

(4)

In Equation (4), the X-ray photon fluence is depicted by Φ₀(E) in units of photons/unit of area and is multiplied by the energy of the photons (E), in order to reproduce the X-ray energy fluence. The X-ray mass attenuation (μ_att(E)/ρ) and the energy absorption coefficients (μ_en(E)/ρ) were calculated from the NIST XmuDat database [90,91]. The crystal thickness is denoted by W and the density by $\rho$, in units of g/cm³ [92].

4. Conclusions

This study aimed to investigate the luminescence properties of a CeF₃ single-crystal inorganic scintillator as a possible alternative to high-luminous but hygroscopic CeBr₃. The experiments were performed under the exposure of a typical X-ray unit, with X-rays spanning from 50 to 140 kVp. The fluorine in CeF₃ appeared to drastically reduce the luminescence efficiency, compared to the CeBr₃ crystal and other commercially used crystals in medical applications such as BGO. Furthermore, the emission maximum moved even lower than CeBr₃, towards the lower part of the visible spectrum and beyond, making CeF₃ suitable for spectral coupling with some multialkali photocathodes and flat-panel position-sensitive photomultipliers applied in nuclear medicine detectors, but completely unsuitable for spectral matching with CCDs and CMOS, used in a variety of medical imaging sensors. The obtained luminescence efficiency results denote that CeF₃ cannot be applied in medical imaging applications, covering the range 50–140 kVp; however, examination of its luminescence efficiency at higher energies could reveal a possible applicability for nuclear medicine detectors.