Light Inorganic Scintillation Materials for Neutron and Charge Particle Detection

Abstract

The technological aspects of the light inorganic crystalline compounds suitable to create scintillation materials to detect charged particles and neutrons in a wide energy range have been examined. Among them, Li₂CaSiO₄:Eu was found to be a prospective candidate to control the valent state of the Rare Earth (RE) and to provide a high intensity of luminescence. It was demonstrated that the material has room for future improvement; however, this requires precise engineering of its composition—an experimental search of compositions or additives that will provide the maximum Eu²⁺/Eu³⁺ ratio to achieve a high scintillation light yield. The benefits of light inorganic materials are disclosed through the modeling of the linear density of nonequilibrium carriers along secondary particle tracks created in scintillators utilized for neutron detection. It is shown that oxide matrices have a larger linear density in comparison with halide crystalline compounds under alpha-particles and tritons, whereas light oxides can provide smaller numbers under protons.

1. Introduction

The scintillation method for measuring ionizing radiation [1] is widely used in a variety of fields, ranging from scientific research to radiography and medical imaging [2]. Simultaneously, the scope of their application is broadening, such as the requirement to measure neutrons with a wide range of energies, including those that are signatures of thermonuclear reactions that will be used in the future energy industry.

In fusion experiments with deuterium d and tritium t fuel, the major fusion reactions (d + d, d + t, t + t) involve the following neutron production channels [3].

$$\begin{array}{c}d+d\to {}^3\mathit{He}+n\left(2.45 MeV\right)\\ d+t\to {}^4\mathit{He}+n\left(14.0 MeV\right)\\ t+t\to {}^4\mathit{He}+n+n\left(0-8.8 MeV\right)\end{array}$$

(1)

To date, a few inorganic scintillation detector media are used to detect neutrons based on isotopes of light elements (³He, ⁶Li ¹⁰B), as well as heavy Gd [4,5,6] and U [7,8,9]. With the exception of gaseous helium, other elements can form inorganic crystalline compounds containing large concentrations (>10²² cm⁻³) of nuclei interacting with neutrons. A common disadvantage of the above light nuclei is the rapid decrease in the interaction cross section according to the law ~1/vn, where vn is the neutron speed. Their cross sections are reduced to bn units in the energy range of epithermal neutrons [10,11]. With an increase in the energy of neutrons, new reaction channels open with the emission of both protons and α-particles [12]. The heavier nuclei of the elements Cl, O, Al, and Ga contribute the most to the interaction in this energy range.

Table 1. Integral cross-sections of the nuclei of some isotopes of the elements creating the scintillation compounds [13].

Isotope	Energy Range, MeV	Integral Cross-Section (n, p), bn	Integral Cross-Section (n, Total), bn
35Cl	1–20	1.66	45.7
	1–150	>1.66	212.4
37Cl	1–20	>0.26	47.0
	1–150	>0.26	220.0
69Ga	1–20	0.439	65
	1–200	0.638	437
71Ga	1–20	0.206	64.6
	1–200	0.387	437
27Al	1–20	0.966	37.7
	1–150	0.966	198
16O	1–20	0.302	30
	1–150	0.435	144

shows the integrals of the neutron cross sections of the nuclei of individual isotopes of the listed elements for processes that open after a neutron energy of more than 8 MeV.

Evidently, the proton production cross-section is quite small and weakly dependent on energy, and the interaction products are dominated by α-particles, deutrons, and tritons.

A class of gadolinium-based garnet structure compounds for detecting neutrons in the thermal and epithermal energy ranges has been developed [14,15]. Gadolinium-based scintillation materials find applications for neutron detection, primarily due to the high neutron cross section of the natural mixture of gadolinium, equal to 46,000 bn for thermal neutrons. At the same time, due to the compound’s high effective charge, these crystals have a relatively high sensitivity to γ-radiation [16]. There is an elpasolite-structured scintillation material family that has a unique set of parameters for neutron detection; however, mass production remains difficult [17,18].

Thus, there is a demand for light scintillation materials, where the detection of high-energy neutrons can be provided by the detection of the same kind of secondary particles, which are generated by the light nuclei under thermal and epithermal neutrons. Such a combination of atoms in a compound should ensure the detection of neutrons in a fairly wide energy range. Analyzing the genesis of the development of the technology of inorganic scintillation materials, we can say that the awareness of the advantages of light inorganic compounds appeared more than 20 years ago [19,20,21]. A noticeable interest in them grew with the start of research on a wider set of the light materials for the purpose of creating scintillators for neutron detection [22,23,24,25]. The range of activating additives was expanded, and work was also carried out to combine several elements (Li, B, O) containing nuclei that effectively interact with neutrons [26,27].

An ideal crystalline matrix to detect neutrons in a wide energy range would be ⁶LiCl. It combines ⁶Li to absorb thermal neutrons, whereas chlorine isotopes ^35,37Cl provide an interaction with fast neutrons. However, this material is quite hygroscopic and is useless for doping with rare-earth ions. Nevertheless, its attractiveness led to the development of a series of eutectic crystalline combinations, which include lithium chloride and a compound allowing doping with 2+ and 3+ rare-earth ions [28,29]. At the same time, light compounds with a scintillation yield comparable to the relatively heavy LiI(Eu) or elpasolites have not yet been obtained.

Recently, Li₂CaSiO₄:Eu (LCS) was demonstrated to be a bright scintillator to construct neutron sensitive screens [30]. It has a low Z_eff ~ 15, which predetermines a low sensitivity to background γ-quanta. It contains light and middle-light elements as well as a small amount of the element europium with a high neutron resonance integral [31]. Aside from its polycrystalline form, it can also be grown in a single-crystalline form [32], potentially broadening the application area of Li₂CaSiO₄:Eu. A performance of a screen sample with 180 μm thickness and an area of 160 mm² in comparison to the commercially available Scintacor ND screen (⁶LiF-ZnS:Ag) was evaluated under the neutron radiation of a Pu-Be source with an activity of 4.7 × 10⁴ neutron/s and a γ-radiation source ¹³⁷Cs with a total activity of 9 × 10³ s⁻¹ (7.6 × 10³ s⁻¹ of 662 keV γ-quanta) [33]. It was demonstrated that the new scintillator provides a high neutron detection efficiency with a reduced screen thickness. Therefore, the sensitivity to the high-energy γ-quanta, particularly those of 4.4 MeV, will be reduced.

Let us compare lithium-containing crystalline materials in terms of the following parameters: density, which is correlated with Z_eff; the volume concentration of lithium; and the possibility of isovalent localization of activator rare-earth (RE) ions. A comparison of the parameters of the typical compounds is given in

Table 2. Parameters of some light lithium-containing crystalline inorganic materials.

Compound	Structure, Spatial Symmetry	Li Atoms/cm3	Density, g/cm3
Li2CaSiO4	Tetragonal, $\mathrm{I}\overline{42}\mathrm{m}$. Consists of LiO4 trigonal pyramids, SiO4 tetrahedra, and CaO8 polyhedra.	2.3 × 1022	2.86
γ-LiAlO2	Tetragonal, P4_$\overline{12}$_$\overline{12}$. Consists of LiO4, and SiO4 tetrahedra.	2.4 × 1022	2.64
LiYO2	Monoclinic, P21/c. Li⁺ is bonded in a distorted T-shaped geometry to three O2⁻ atoms. Y³⁺ is bonded to six O2⁻ atoms to form a mixture of distorted corner and edge-sharing YO₆ octahedra.	1.9 × 1022	4.10
LiYSiO4	Monoclinic, P21/c. Consists of LiO5 trigonal bipyramids, YO7 pentagonal bipyramids, and SiO4 tetrahedra.	1.1 × 1022	3.65
LiAlSiO4	Trigonal, R3. Consists of LiO4, AlO4 и SiO4 tetrahedra.	1.2 × 1022	2.69
Li3AlMO5 (M = Al, Ge, Ga)	Orthorhombic, Pna21. Stannite structure. Consists of LiO4, AlO4, GeO4, and GaO4 tetrahedra.	2.7 × 1022	3.08–4.2
LiCaAlF6	Trigonal, $\mathrm{P}\overline{31}\mathrm{c}$. Consists of LiF6, CaF6, SiF6 octahedra	9 × 1021	2.86

. The structure data and densities of the compounds were taken from the database [34], and the amounts of lithium atoms were calculated.

The listed compounds can be doped with 3d-ions. Compared to such activator ions, RE ions provide a higher light yield in scintillators due to the participation of mixed electronic states of df configurations in an energy exchange with the exciton subsystem of nonequilibrium carriers and, as a rule, have faster scintillation kinetics (with the exception of the Tb³⁺ ion, which has scintillation kinetics of the order of milliseconds). However, RE ions preferentially localize in octahedral and higher oxygen coordinations; therefore, not every type of crystal lattice is suitable for doping a compound with RE ions.

Compounds having Li and Y ions in the matrices are the heaviest in the list. From the standpoint of a combination of physico-chemical properties, the compounds Li₂CaSiO₄ and Li₃AlMO₅ (M = Al, Ge, Ga) may be distinguished. The Li₃AlSiO₅ compound is a deep-ultraviolet nonlinear optical crystal [35], but the entire class of materials is only suitable for 3d-element doping.

In this article, we focused on studying the potential for the future improvement of the Li₂CaSiO₄:Eu scintillation properties. Of particular interest was a comparison of LCS neutron-sensitive material with other species for the detection of the secondary particles created by neutrons. The materials were examined for ionization density, which regulates the light yield, as invented by Birks [36].

The Li₂CaSiO₄ compound is suitable for doping with divalent ions, providing isovalent substitution of Ca ions. This is quite difficult to manage in compounds where no appropriate site for defect-less substitution exists [37]. Nevertheless, Li-based compounds have a tendency to defect formation in the anion sublattice, which contributes to the stabilization of heterovalent impurity ions in different valence states. That is, the stabilization of europium ions predominantly in the divalent state in lithium compounds is a separate and interesting problem of inorganic chemistry.

2. Results and Discussion

2.1. Photoluminescence Study of Eu-Doped Scintillator Li₂CaSiO₄ and Possibilities for Light Yield Enhacement

Europium, even under the condition of isovalent substitution, is localized in inorganic compounds both in 3+ and 2+ valence states. It is known that, for Li₂CaSiO₄:Eu²⁺, the luminescence intensity depends on the ratio of Eu²⁺ and Eu³⁺, which is formed during high-temperature synthesis due to the volatilization of a part of lithium, resulting in the formation of cation vacancies in the crystal lattice, the charge compensation of which occurs due to the appearance of the Schottky pair $V_{Li}^{\prime }+\raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$2$}\right.V_O^{..}$ [38]. Under such conditions, the probability of luminescence quenching of Eu²⁺ ions increases both due to the interaction with quenching centers based on electron traps and the nonradiative transfer to Eu³⁺ ions. We compared two methods of preferential stabilization of Eu ions in the divalent state in Li₂CaSiO₄. First, light cations were used to nonisovalently replace silicon ions. Some of the Si⁴⁺ ions in the compound were replaced by Al³⁺ ions, as described in [39]. This procedure creates negatively charged vacancies in the crystal structure, which prevent the oxidation of Eu²⁺ to Eu³⁺. Another method was to substitute a part of the SiO₂ tetrahedra with the Si₃N₄ polyhedra. This method was used to boost the brightness of Li₂MSiO₄:Eu²⁺ phosphors, where M stands for Ca, Sr, and Ba [40]. Figure 1 compares the luminescence intensity of LCS samples obtained by two methods. Samples with nonisovalent substitution of aluminum ions for silicon ions demonstrated a twofold increase in luminescence intensity, while the Li₂CaSiO_2.5N:Eu compound showed an almost threefold increase in intensity compared to Li₂CaSiO₄:Eu²⁺. At the same time, the photoluminescence spectra recorded at 395 nm excitation show the bands of both Eu²⁺ and Eu³⁺ ions. They depict a significant decrease in the luminescence intensity of Eu³⁺ ions at the nonisovalent substitution of Si⁴⁺ by Al³⁺ ions. In the nitrogen co-doped sample, no Eu³⁺ luminescence bands were observed at all.

It is worth noting that a significant increase in the luminescence intensity is not accompanied by a similar increase in the scintillation yield. For the Li₂CaSiO_2.5N:Eu samples, it turned out to be even smaller than for Li₂CaSiO₄:Eu²⁺ at the same activator concentrations. This indicates that the concentration of nonisovalent additives requires additional fine tuning to prevent an increase in lattice point defects, which are, as a rule, electron traps.

It is important to note that, even in a polycrystalline form, lithium calcium silicate forms sufficiently large transparent crystallites (Figure 2). It might be crucial in terms of light extracting when LCS will be used as a component of neutron-sensitive coatings.

2.2. Evaluation of Various Scintillators for Charged Particles Detection

The ability of the material to separate α-particles of different energies is the most important factor in detecting secondary heavy charged particles formed by neutrons of different energies. Since the charged particles under consideration are not relativistic, the major energy losses are due to ionization. At the same time, the energy resolution in a simplistic form, as in the case of γ-quanta, is formed by three factors: the scintillation yield, material properties that determine the internal resolution of the material, and the track fluctuations of both the secondary particle itself and ionized electrons [41]. Although fluctuations are smaller in denser materials, the increased sensitivity to background γ-quanta becomes a significant factor, limiting their use for secondaries’ detection. Internal resolution is also a significant factor. For the registration of α-particles, one of the best inorganic scintillation materials with a low internal resolution is YAlO₃:Ce (YAP) [42]. However, this material is not sensitive to low-kinetic-energy neutrons. The yield of material scintillations for neutron detection should be considered in relation to the α/γ factor. A high scintillation yield with a low α/γ factor of the scintillator will provide overlap between the α-particles and background γ-ray signals.

The light yield of the scintillator correlates to the density of pairs of nonequilibrium carriers formed in the track. The average linear density of nonequilibrium carriers can be estimated as <N_pairs>/dx = <dE/dx>/(βE_g). Using a computer simulation, we estimated ionization losses <dE/dx> of various charged particles: electrons, protons, α-particles, and tritium nuclei (tritons) in bulk of widespread scintillators such as LSO (Lu₂SiO₅), CLYC (CsLiYCl₆), NaI:Tl, YAP (YAlO₃:Ce), and the new promising material LCS (Li₂CaSiO₄:Eu²⁺). The results of the simulation are presented in Figure 3. The average specific energy losses <dE/dx> = E₀/<l_tr> were calculated using the primary energies E₀ and <l_tr>.

The positions of the specific energy losses’ curves follow the change of the density of the material in the case of electrons; the heavier a material is, higher ionization losses occur. As expected, the specific losses for the heavier charged particles are significantly high. However, the positions of their curves are reshuffled. Apparently, the result that oxides exhibit average specific energy losses greater than the halide compounds is determined by a high electron density in oxides.

The choice of the conversion efficiency parameter of the material (β) requires more detailed comment. This parameter is the ratio between the average energy E_eh required to create one electron-hole pair and the bandgap of the material, which is equal to the minimum energy of such a pair. The value of β characterizes the efficiency of the multiplication cascade of electronic excitations in the scintillation material [43,44]. When interacting with the electron subsystem, an electron can only create a new pair if the energy conservation law is fulfilled in the process e→2e + h, and thus the kinetic energy of an electron must be greater than the bandgap. Similarly, the creation of a new electron-hole pair in the Auger process for holes h→2h + e will occur only if the hole is at least E_g below the top of the valence band. When secondary electron-hole pairs are created, secondary electrons and holes will be distributed in energy in the first approximation proportionally to the density of states (DOS). Thus, it is necessary to estimate the average energy of electrons in the region where they cannot create secondary pairs, the so-called “passive” region, with kinetic energies from zero to the multiplication threshold. In addition, one more relaxation process has to be taken into account. Relaxation by phonons occurs in parallel with the multiplication process at electron energies slightly above the threshold energy. The phonon emission probability can be quite high near the multiplication threshold and leads to a loss of carrier energy with a further transition to the “passive” region. The competition of relaxation channels causes the multiplication threshold to shift to higher energies. As an example, Figure 4 shows the densities of states in an NaI and YAP crystal (calculated by the DFT method and given in the AFLOWLIB.ORG repository [45]). Note that YAP is interesting for consideration because its E_g value was found to be different on the simulation model applied [46]. The yellow curves correspond to the passive region, and the blue curves correspond to the extended passive region, taking into account the shift in the effective manifestation of the multiplication threshold by 2 eV. The average kinetic energy of electrons and holes calculated from these passive regions is also shown in the figures.

The average energy required to create one pair is E_eh = E_g + <E_e> + <E_h>. As seen, the value of β = E_eh/E_g depends on several factors: the electronic structure of the material, and the strength of the electron–phonon interaction, which in turn depends on the ionicity of the crystal. For ionic crystals, the rate of energy loss of carriers for the emission of polarization optical photons is much higher than the corresponding rate in covalent crystals with deformation optical phonons. Figure 5 depicts the estimated values of a few materials based on the relaxation processes taken into account. Note that for a number of materials, the additional features of the electronic structure can significantly change the β value compared to the given simple estimate. For instance, the proximity of the 5pCs core level to the Auger process threshold for holes in CsI can significantly reduce the estimate of β [47].

The following bandgap values E_g of the compounds were used to calculate the average linear density of nonequilibrium carriers: LSO—6.2 eV, YAP—7.8 eV, LCS—5.2 eV, NaI—5.9 eV, and CLYC—4.9 eV. In addition, taking into account considerations of the β factor by authors [48,49,50], the following values of β were used for evaluation: LSO—2.5, YAP—2.5, LCS—2.5, NaI—1.7, and CLYC—2.

Figure 6 depicts the estimated dependence of the average linear density of nonequilibrium carriers on the energy of secondary particles.

For electrons and the lightest secondary particles (protons), the trend in the arrangement of the curves is preserved, as shown in Figure 3. The average linear density of nonequilibrium carriers for alpha particles and tritons is nearly equal for NaI(Tl) and LCS, whereas it is significantly higher for the other materials. It can be noted that the transverse dimensions of the track will also be smaller in heavier materials. Apparently, a smaller track volume implies a higher density of exciton states and, consequently, a stronger effect of the mutual quenching of excitons. This process competes with the capture of excitons by activator ions and, as a consequence, leads to a decrease in the scintillation yield. In this context, europium-activated scintillators have an advantage for detecting secondary particles, since they allow a large concentration of the activator in the crystalline compound.

3. Materials and Methods

3.1. Samples Synthesis

The method for producing the polycrystalline compound Li₂CaSiO₄:Eu is described elsewhere [30,33]. Samples of LCS co-doped with aluminum or nitrogen were prepared by a variety of the sol-gel method. Acidic hydrolyzed tetraethyl orthosilicate was used as a source of SiO₂. In obtained silica-sol sociometric amounts of reagent grade Li₂CO₃, CaCO₃, Eu₂O₃, and Al(NO₃)₃ corresponding to a composition of Li₂Ca_0.998Eu_0.002Si_1−xAl_xO₄ (x = 0.05, 0.1) were added. Formed gel was dried at 100 °C, comminuted by mortar and pestle, and sieved through a 150-mesh sieve. The powder was calcined at 850 °C during 2 h to release all CO₂. Then, the powder was comminuted again and sieved. In a case of nitrogen co-doping, a necessary amount of Si₃N₄ for the composition Li₂Ca_0.998Eu_0.002SiO_2.5N was added at this step. The reduction of Eu³⁺ to Eu²⁺ and the final synthesis of LCS were carried out in a tube furnace under a 5% H₂ + 95% N₂ atmosphere at 900 °C for 2 h.

3.2. Photoluminescence Measurements

Photoluminescence spectra were recorded on a Fluorat-Panorama (LCC “Lumex”, St. Petersburg, Russia) instrument using an excitation with a high-pressure xenon lamp operating in the short pulse mode (1 μs) with a repetition rate of 25 Hz.

3.3. Modeling and Simulation Procedure

In modelling, we focused on the study of ionization losses in the track of various types of secondary particles (protons, alpha particles, and tritons) formed under the neutrons. These parameters were compared with similar ones for electrons that are generated in materials due to the photo-absorption of γ-quanta. The evaluation was made using the GEANT4.10.0 software package [51]. The following compounds were selected for evaluation: NaI(Tl,) CLYC (CsLiYCl₆), YAP, LSO (Lu₂SiO₅), and LCS. Sodium iodide, in which part of the sodium is substituted for the lithium, is an attractive material for detecting neutrons [52]. CLYC has a high neutron detection yield. YAP and LSO oxygen compounds were chosen as materials with significantly different α/γ factors [53] and effective charges.

The following geometrical model was chosen for the simulation: a sphere of material with a diameter of 100 mm, in the center of which particles are emitted isotropically. For the energy value of each type of particle, 10⁵ particles were launched. When the energy decreased below the threshold of 100 eV, the length of the traversed track l_tr was determined. The obtained value was plotted on the histogram of track lengths. An approximation of this curve yielded the average track length, <l_tr>.

4. Conclusions

In this work, in the consideration of scintillation materials, the emphasis is on those that make it possible to detect neutrons in a wide energy range. A particular focus is placed on the discussion of the properties of the light crystalline lithium-containing materials in order to select a crystal that combines high lithium content and the ability to stabilize rare earth activator ions in the valent state of the interest in this matrix. It has been found that in oxide light matrices of lithium-silica compounds, europium ions can be stabilized predominantly in the divalent state. Technological methods for such stabilization are considered; luminescent measurements confirm the possibility of minimizing the concentration of Eu³⁺ ions in the Li₂CaSiO₄ compound. Technological methods to increase the intensity of Eu²⁺ photoluminescence have been examined. Nonisovalent substitution in cation and anion sub-lattices was investigated. Both were found to be prospective; however, to increase the yield of scintillations, their certain refinement is required, especially from the point of view of establishing sufficiently accurate ratios between the concentrations of europium and nonisovalent dopants. The described technological methods for the stabilization of Eu²⁺ in this matrix open the door for the creation of the Li₂CaSiO⁴:Eu²⁺ compound in a single-crystal form.

The developed material was compared with other materials occasionally used to detect neutrons and not only gamma-rays and charged particles. To compare the materials for the registration of secondary particles, the energy dependence of the average linear density of nonequilibrium carriers were compared. Peculiarities of the formation of thermalized pairs in various types of compounds are considered. It is shown that the major uncertainty in the determination of the average linear density of nonequilibrium carriers is introduced by the uncertainty in the estimate of the conversion efficiency parameter of the material. The specific energy losses of heavy secondary particles were found to be significantly greater than those of electrons produced by photo-absorption in scintillators.

The average linear density of nonequilibrium carriers created in several scintillation materials under electrons was compared to the average linear density of secondary particles created in scintillation materials under neutrons, namely protons, α -particles, and tritons. Alkali-halides, such as NaI, show minimal losses for the heaviest secondary particles: alpha particles and tritons. From this point of view, they can provide a high yield of scintillations at the detection of thermal neutrons with a partial substitution of Na ions by Li ions in the matrix. For the lighter secondary particles, protons, it is shown that the lighter the complex oxides are, the lower the average linear density of nonequilibrium carriers, which might be caused by the appropriate change of electron density in the oxides considered. This provides better conditions for an effective energy transfer from the matrix excitonic states to activator ions at the latter’s relatively high concentrations in the compound.