Identification of defects in pure and Al/Ga-doped ZnO to improve X-ray detector performance: experimental and simulation methods

Abstract

ZnO is an essential material used in various devices, but its performance can be significantly enhanced by introducing or removing defects, such as oxygen and zinc vacancies. In this article, we explore the relationship between different types of defects in pure and Al/Ga-doped ZnO and their corresponding optical and electronic properties, which are vital for X-ray detection. The nanoparticles were initially synthesized using the sol–gel auto-combustion method, with calcination temperatures of 400 °C, 500 °C, and 600 °C. The samples were then analyzed using multiple techniques, including XRD, FESEM, FTIR, photoluminescence (PL), electron paramagnetic resonance (EPR), and positron annihilation lifetime spectroscopy (PALS). Subsequently, DFT+U calculations were conducted to examine the electrical, optical, and theoretical EPR properties, as well as to identify defects that occur at different calcination temperatures. Analyzing the connection between defects and PL spectra in X-ray scintillation detectors, along with exploring the relationship between electron and hole concentrations in X-ray semiconductor detectors, provides valuable insights into the fundamental properties of ZnO. These insights pave the way for utilizing ZnO in developing next-generation scintillation and semiconductor X-ray detection technologies.

---

1. Introduction

In the past decade, zinc oxide (ZnO) nanostructures have been extensively used in the fabrication of various electronic devices and functional materials.^1–4 Furthermore, ZnO is well-suited for detecting X-rays and other ionizing radiation at room temperature due to its wide band gap, strong exciton binding energy, rapid scintillation response, and high resistance to radiation damage.^5–7 When compared to detectors made from materials such as silicon (Si), gallium arsenide (GaAs), cadmium sulfide (CdS), and gallium nitride (GaN),^8,9 ZnO demonstrates superior resistance to nuclear radiation. It can withstand electron radiation of up to 1.6 MeV and proton radiation of 10¹⁵ cm².¹⁰ Inorganic crystals, such as NaI:Tl, CsI:Tl, BGO, and CdWO₄, are utilized for radiation detection; however, they have some drawbacks, including inefficiency, which necessitates the use of specialized equipment and expensive methods.^1,11,12 ZnO, particularly when doped with gallium, exhibits ultrafast scintillation response times below 1 ns,^13–15 making it a promising material for high-speed X-ray and alpha-particle detection. This rapid response capability can be harnessed to develop ultrafast X-ray imaging detectors that could be employed in materials science and advanced medical diagnostics. However, it should be noted that ZnO also has certain limitations compared to other semiconductors and scintillation materials. For example, achieving stable p-type conductivity remains challenging, which limits the fabrication of p–n junction detectors.¹⁶ Furthermore, intrinsic point defects can introduce visible luminescence and leakage current, reducing energy resolution. Although ZnO offers superior radiation hardness and ultrafast response compared to Si and GaAs, materials such as CdTe and perovskites still provide higher X-ray absorption efficiency due to their higher atomic numbers.^17,18 Therefore, further optimization of defect control and doping strategies is necessary for ZnO to fully compete with existing detector materials.

An efficient X-ray detector based on ZnO requires intense luminescence under ultraviolet excitation and the ability to sustain an ultra-low leakage current, typically within a few tens of nanoamperes, under reverse-bias conditions.^19,20 Furthermore, in the absence of incident radiation, ZnO should exhibit a low dark current density (typically on the order of a few microamperes per square centimeter) to ensure high signal-to-noise performance in low-background environments.²¹ Native defects, including vacancies and interstitials, mainly influence the performance of detectors based on ZnO. Such intrinsic defects can, however, result in increased leakage current as well as the generation of visible and infrared light, which, in turn, decreases the response current of a detector. The type, concentration, and distribution of these defects can be associated with annealing temperature and the reaction of the doping elements.^22,23 In this context, extrinsic impurities can significantly modify the defects and electronic properties of ZnO.²⁴ Therefore, understanding how dopants such as Al and Ga interact with the defects in ZnO can significantly enhance the performance of ZnO-based radiation detectors.

This study explores how impurities and thermal treatment influence defect formation in pure and Al- or Ga-doped ZnO (3% and 5%), using a combination of PL, EPR, and PALS measurements along with DFT+U calculations. It also analyzes how these factors affect the performance of scintillation and semiconductor detectors, particularly in relation to defects in the materials. Nanoparticles with the chemical formula Zn_1−xAl_xO and Zn_1−xGa_xO (x = 0.00, 0.03, and 0.05) are synthesized at calcination temperatures of 400, 500, and 600 °C. Additional discussion on the rationale behind the chosen calcination temperatures and doping concentrations can be found in Section S1 of the SI. XRD, FESEM, FTIR, PL, EPR, and PALS are employed to elucidate the relationships between defect states, excitonic effects, and electronic transitions.

DFT+U calculations were also performed to investigate the electrical and optical properties, with a focus on identifying defects. The origins of typical emissions are discussed, providing new insights into the properties and transition mechanisms of defect luminescence. Moreover, calculations involving the g-tensor and hyperfine coupling constants (HFCCs) are performed to obtain theoretical EPR spectra,²⁵ further facilitating the analysis of defect-related electronic structures. By analyzing the relationships between defects and PL spectra in scintillation detectors and electron–hole concentrations in semiconductor detectors, this study provides insights into the properties of pure and doped ZnO and its potential applications in next-generation scintillation and semiconductor X-ray detectors. This article will be divided into three parts, beginning with an overview of the experimental and computational methods employed. The Results section examines the material's structure, spectroscopic findings (PL, EPR, and PALS), defect stability, electronic properties, and impact of defects on charge carriers, as well as their potential applications in X-ray detectors. The article ends with a conclusion that ties together the main insights and their significance.

2. Experimental and computational methods

2.1. Experimental methods

The sol–gel auto combustion method is employed to synthesize the nanoparticles. Stoichiometric amounts of Zn(CH₃COO)₂·2H₂O, Al(NO₃)₃·9H₂O, and Ga(NO₃)₃·nH₂O were added to 100 cc of deionized water in a beaker. The solution was stirred for 10 minutes, and citric acid was added as a chelating agent. Ammonia solution was gradually added to the mixture until the pH reached approximately 7–8. The mixture was stirred continuously and kept at a temperature of 300 °C for 60 minutes to ensure that the combustion process was fully completed. Nanoparticles of pure zinc oxide and doped with aluminum and gallium impurities at concentrations of 3 and 5% were prepared at equilibrium temperatures of 400, 500, and 600 °C.²⁶ Characterization techniques were utilized to analyze the properties of the synthesized nanomaterials, including XRD, FESEM, EDS mapping, EPR, PL, and PALS. Experimental EPR spectra were obtained by using an X-band (ν = 9.776 GHz) EMS 104 EPR spectrometer (Bruker Co., Germany) with the following parameters: a microwave power of 1.99 mW, a sweep width of 200 G, a modulation amplitude of 1.01 G, a receiver gain of 30 dB, and a scan range of 338–358 G at room temperature.

2.2. Computational methods

The structure of pure and Al and Ga-doped ZnO superlattices with defects was optimized using DFT calculations. The calculations employed PBE methods, a DZVP Gaussian basis set, GPW, and a Rydberg cutoff energy of 320 Ry in the CP2K software.²⁷ The electronic structures and parameters associated with EPR, such as the components of the g-tensor and HFCCs, were subsequently determined using the optimized structures calculated within the QUANTUM ESPRESSO package.²⁸ The theoretical background for calculating the EPR parameters is provided in Section S2.1 of the SI. A comprehensive investigation was conducted on energy cutoffs for the plane-wave basis set, in conjunction with different Monkhorst–Pack grids. This process continued until a convergence in total energy was obtained, as illustrated in Fig. S1(a) and (b) of the SI. To provide an accurate characterization of the electronic properties, the DFT+U approach was utilized, incorporating the Hubbard U potential for the Zn-3d, O-2p, Ga-3d, and Al-2p states.

Various values for U p(O) and U d(Zn), ranging from 0 to 12 eV, were investigated to ensure precise calculations of the band gap energy, as illustrated in Fig. S1(c). It was found that a cutoff energy of 40 Ry, k-points with dimensions 2 × 3 × 3, a charge density of 320 Ry, and Hubbard potential corrections of U 3d(Zn) = 10 eV, U 2p(O) = 11 eV, U 2p(Al) = 5 eV, and U 3d(Ga) = 3 eV provide optimal convergence.²⁹ The dominant defects were identified based on a comprehensive evaluation of the formation energies, complemented by EPR spectra obtained using the GIPAW method.²⁸ The computational EPR spectra were simulated using the MATLAB-based EasySpin package,³⁰ specifically the pepper subroutine. The simulations incorporated parameters including the g-factor, HFCCs of aluminum and gallium atoms, a microwave frequency of 9.776 GHz, a magnetic field sweep range of 320–390 mT, and a linewidth of 0.9 mT.

3. Results and analysis

3.1. Structural and microstructural characterization

The structural and microstructural properties are examined through characterization methods, including XRD, FTIR, and FESEM, and the results for a calcination temperature of 500 °C are shown in Fig. 1. XRD analysis indicates a wurtzite crystal structure for the pure and doped materials, confirmed by the presence of characteristic diffraction peaks compatible with the standard JCPDS card no. 00-005-0664. A slight shift to the left is observed, which is caused by the change in lattice constants. The lattice constants for a hexagonal lattice can be obtained using the following equation:and

nλ = 2d_hkl sin θ (2)

The obtained results are presented in Fig. S2a and b (SI). The a-axis lattice parameter increases very negligibly, while the c parameter initially decreases and then increases very slightly. The observed lattice parameter changes in ZnO doped with Al and Ga can be attributed to the ionic size mismatch between the dopants and Zn. The initial decline in c with Al and Ga doping can be attributed to the smaller ionic radius of the dopants compared to that of Zn. However, with an increase in the dopant concentration, the lattice is distorted due to the repulsive interaction between the dopant and Zn ions, leading to a slight increase in the c parameter.

Fig. 1 (a) The XRD diagrams of the pure and doped samples, (b) the FTIR spectrum of pure ZnO, and the FESEM images of (c) pure ZnO, (d) ZnO:Al (3%), and (e) ZnO:Ga (3%) samples.

The crystallite size (D_c) can be evaluated using Scherrer's relation:

where β is the FWHM, k is a constant number, and λ is the X-ray wavelength. The crystallite size ranges from 8 to 24 nm and generally decreases as impurities are added (see Fig. S2c (SI)). Increased nucleation processes, which are more pronounced for Al doping due to its smaller ionic radius, are responsible for the decrease in the crystallite size. However, relaxation mechanisms may prevent crystallite formation at higher Ga concentrations.³¹

FTIR spectroscopy was used to study the chemical bonding and defects (see Fig. 1b and Fig. S2d (SI)). The hydroxyl groups on the nanoparticles are responsible for the peak at approximately 3400 cm⁻¹. The C–H stretching vibrations of alkane functional groups result in peaks between 2830 and 3000 cm⁻¹. The peaks at around 1630 and 1384 cm⁻¹ are due to the asymmetric and symmetric stretching vibrations of the zinc carboxylate functional group, respectively.³² Usually, Zn–O stretching vibrational modes fall between 400 and 500 cm⁻¹. The vacancies in the bulk of the ZnO material are responsible for the peak at 1100 cm⁻¹.³³ The extra peaks in the FTIR spectrum of the sample synthesized at 500 °C can be due to the change in ZnO lattice vibrational modes because of vacancies. These vacancies can create new vibrational modes that are not present in a perfect ZnO lattice.³⁴ The shift of the peak positions from 466 to 433 and 489 cm⁻¹ can also be attributed to changes in the Zn–O bond environment due to vacancies. In particular, the peak at 534 cm⁻¹ for the sample calcined at 500 °C is due to vacancies.^34,35 However, it would be preferable to use EPR, PALS, and DFT+U studies to further validate the existence of vacancies.

The FESEM micrographs are also presented in Fig. 1(c)–(e). The average size of pure ZnO is approximately 40 nm and decreases with increasing dopant concentrations. The abundance percentages of oxygen, zinc, aluminum, and gallium atoms were obtained using EDS analysis. Fig. S3 and S4 display the EDS elemental maps obtained from ZnO:Al(3%) and ZnO:Ga (3%), respectively. The atomic abundance percentages for all samples are presented in Fig. S5 (SI). It is clear that the dopant addition process is well executed. As will be discussed in the following sections, pure zinc oxide primarily exhibits ZnOV_O, ZnOV_Zn, and defects. Regarding the defect formation energy (see Section 3.3), the formation energy of the ZnOVO defect increases with rising temperature, while the formation energy of the ZnOV_Zn defect decreases. As a result, the abundance percentages of oxygen atoms rise, whereas those of zinc atoms fall. It can be noted that for zinc oxide doped with 3% aluminum, the predominant defects are ZnOV_O:Al and ZnOV_Zn:Al. Regarding defect formation energy, as the temperature rises, the formation energy of the ZnOV_O:Al defect remains constant, while that of the ZnOV_Zn:Al defect decreases (see Section 3.3). Consequently, the abundance percentages of oxygen increase, while those of zinc decrease. The same pattern is observed for other doped samples as well.

3.2. PL, EPR, and PALS

PL spectroscopy is an effective non-invasive technique used to analyze defects in semiconductors. It provides essential details regarding the energy levels of the defects (even at low densities), which helps detect structural defects in semiconductors.³⁶ The PL spectra obtained with an excitation wavelength of 325 nm were fitted using an improved fitting algorithm. This method utilized Gaussian functions and a weighted non-linear least squares approach to determine multiple bands.³⁷ The corresponding findings for the samples calcined at 500 °C are presented in Fig. 2, and the results for the ones calcined at 400 °C and 600 °C are also presented in Fig. S6 and S7 (SI), respectively. The peaks in the PL spectra for all modes of emitted photon energy are in the ranges of 3.14–3.22 eV (high intensity), 2.74–2.94 eV (medium intensity), and 2.52–2.66 eV (low intensity), which correspond to near band edge (NBE) states, shallow donor states, and deep central states, respectively. Strong UV emissions near the band-gap energy arise from free exciton recombination, while visible emissions stem from shallow or deep-level defects.³⁸ It can be observed from Fig. 2 that for ZnO:Ga (3%) calcined at 500 °C, the intensities of both the shallow donor states and the deep central states are lower than those of the near-band-edge (NBE) state found in other synthesized samples. This is because the crystal structure quality of ZnO doped with 3% Ga is better than that of other samples, resulting in a lower defect concentration.

Fig. 2 The deconvolution of the PL spectra for (a) pure, (b) ZnO:Al (3%), (c) ZnO:Al (5%), (d) ZnO:Ga (3%), and (e) ZnO:Ga (5%) samples calcined at 500 °C.

PALS of the pure and 3% doped samples calcined at 500 °C was performed using a digital PALS spectrometer at room temperature. The source was sandwiched between two identical samples. The spectrometer was equipped with a 500 MHz CAEN waveform digitizer. The digitizer samples directly from the anode signals of the two plastic detectors. The timing resolution was measured to be 160 ps employing the coincidence gamma lines of ⁶⁰Co (1173 and 1332 keV). The channel width was adjusted to 28 ps. The details of the PALS spectrometer are published in our previous study.³⁹ The statistics of 1 million positron annihilation events were compiled over 7 days for each sample. The PALS spectra of the samples were analyzed using the LT-10 code.⁴⁰Fig. 3 shows the PALS spectra of the pure sample. The results of the data analysis are presented in Table 1. Three parameters were considered for data analysis. τ_i (ns) is the lifetime of the ith component, and I(i) is the corresponding intensity. The shortest lifetime and its related intensity are attributed to the para-positronium self-annihilation. The intermediate component is due to the positron annihilation with free electrons. The longest-lived component and its associated intensity are related to positron annihilation in free-volume sites.

Fig. 3 The PALS spectra of the ZnO sample.

Table 1 The results of the PALS experiment

Sample	τ 1 (ns)	I 1 (%)	τ 2 (ns)	I 2 (%)	τ 3 (ns)	I 3 (%)	κ (s^−1)
ZnO	0.259 ± 0.002	65.59 ± 0.13	0.714 ± 0.019	31.98 ± 0.21	1.841 ± 0.031	2.43 ± 0.10	(7.87 ± 0.02) × 10^8
ZnO:Al (3%)	0.368 ± 0.001	62.41 ± 0.18	0.669 ± 0.008	29.34 ± 0.17	1.732 ± 0.020	8.25 ± 0.24	(3.59 ± 0.06) × 10^8
ZnO:Ga (3%)	0.262 ± 0.001	63.79 ± 0.07	0.743 ± 0.013	33.95 ± 0.08	1.974 ± 0.026	2.26 ± 0.05	(8.39 ± 0.10) × 10^8

---

The third-lifetime component (τ₃) and its intensity (I₃) obtained from PALS provide insights into the presence and nature of large open-volume defects, such as voids or vacancy clusters, where ortho-positronium (o-Ps) formation can occur. In this study, the ZnO:Al (3%) exhibited the highest I₃ value (8.25%), indicating a significantly higher concentration of such defects compared to the pure ZnO and ZnO:Ga (3%). This suggests that Al doping induces extensive structural disorder, likely through the formation of vacancy clusters and charge-compensating defects associated with the substitution of Zn by ZnO:Al (3%). Although the τ₃ value of ZnO:Al (3%) is slightly shorter (1.732 ns), it still falls within the expected range for o-Ps annihilation in large but possibly interconnected or partially confined voids. In contrast, the ZnO:Ga (3%) shows the longest τ₃ (1.974 ns), indicating the presence of larger and more isolated free volumes; however, its I₃ parameter is low (2.26%), suggesting a much lower density of such sites. The pure ZnO exhibits intermediate behavior, with τ₃ = 1.84 ns and I₃ = 2.43%, indicating a relatively compact microstructure with fewer and smaller open-volume defects. These observations highlight the distinct impact of the dopant type on the defect structure and porosity of ZnO at the atomic scale.

To provide a more quantitative interpretation of the PALS results, the standard two-state positron trapping model was applied. The positron trapping rate κ was calculated using the following relationship:

Based on the experimental values shown in Table 1, the calculated trapping rates are also listed in Table 1. These κ values represent the positron capture probability per second and directly measure the trapping strength of vacancy-type defects. The highest κ value obtained for the Ga-doped sample indicates more effective positron trapping by vacancy-related centers, consistent with the defect energetics predicted by DFT and with EPR observations. In contrast, the Al-doped sample shows a lower κ and a higher ortho-positronium intensity, suggesting that larger open-volume voids are dominant, which favor positronium formation rather than direct positron trapping.

EPR spectroscopy has uncovered significant insights into the defects in the crystal structure of the materials. This technique, suitable for samples with unpaired electrons, was applied to pure and 3% and 5% Al- and Ga-doped zinc oxide powders synthesized at 500 °C. The EPR spectra of the pure and 5% Al- and Ga-doped samples showed that they are diamagnetic, and the experimental spectra only contained noise signals. Fig. 4 shows the EPR spectrum of pure ZnO, and almost the same spectrum is observed for 5% Al/Ga-doped samples. The EPR spectra of 3% doped zinc oxide are shown in Fig. 4, which exhibit resonance signals at magnetic fields of 356 and 357 mT for aluminum and gallium, respectively. The EPR spectrum of zinc oxide doped with 3% gallium has a narrower width compared to that of zinc oxide doped with 3% aluminum. A wider EPR spectrum suggests a greater presence of defects with varying resonance field positions for the 3% Al-doped sample. For more details, refer to Section 3.4 on theoretical EPR spectroscopy.

Fig. 4 The experimental EPR spectra of ZnO, ZnO:Al (3%), and ZnO:Ga (3%) samples calcined at 500 °C.

3.3. Thermodynamic stability of defects

DFT calculations in the GGA+U framework have been used to investigate the presence of various defects in the crystal structures of the specimens. The simulated defects are the oxygen vacancy (V_O), zinc vacancy (V_Zn), zinc interstitial (Zn_i), oxygen interstitial (O_i), nearby Zn–O vacancy pair (V_OZn), and distant Zn–O vacancy pair . To examine the effect of impurities and intrinsic defects on the electronic properties, the structure of pure and doped samples with intrinsic defects was carefully optimized. The a and c lattice parameters of pure zinc oxide were calculated to be 3.27 and 5.22 Å, respectively, with an insignificant deviation from the experimental value obtained earlier, indicating the accuracy and validity of the model. Besides, the spin-polarized band structure of pure zinc oxide and the total and partial densities of states of the oxygen and zinc atoms are presented in Fig. 5. The pure zinc oxide semiconductor has a direct band gap of 3.39 eV, which aligns well with the experimental results (3.37 eV).⁴¹ The partial density of states reveals that, for all structures, the conduction band minimum of zinc oxide is formed by the overlap of Zn-4s, Zn-3d, and O-2p orbitals. Furthermore, the 2p orbitals of most oxygen atoms essentially form the valence band maximum (VBM).

Fig. 5 The spin-polarized band structure, total and partial densities of states, and electron density distribution (0.25 a.u.) in the crystal structure for pure zinc oxide. Magenta/cyan bands show spin-up/spin-down states, and gray/red spheres denote Zn/O atoms.

We calculated the formation energy of the defects to understand the stability of different defects. The focus is on native defects when they are neutral, and we used the following equation to determine the formation energies:

In this equation, E_tot(ZnO) and E_tot(D) are the total energies for the pure zinc oxide and zinc oxide with defects, respectively. n_i is the number of atoms of type i added or detached from the superlattice, which we consider negative and positive, respectively. The chemical potential of atom i (oxygen, zinc, aluminum, and gallium) is μ_i, which is affected by the conditions used during the material synthesis. To evaluate the chemical potential of oxygen and zinc at equilibrium, we used the relationship μ_Zn + μ_O = μ_ZnO. In addition, the chemical potentials must obey the boundary conditions of and . These limits ensure that ZnO remains thermodynamically stable with respect to metallic Zn and molecular O₂, in complete agreement with both experimental growth conditions and previous theoretical studies.^42–44 As a result, this range offers a relevant and meaningful description of realistic synthesis environments for both pure and doped ZnO. Under the O-rich synthesis conditions, and and under the Zn-rich synthesis conditions, and . Besides, μ_Al = E_Al(bulk) and , where E_Al(bulk) and E_Ga(bulk) are the total energies of bulk Al and Ga, respectively.

Synthesis conditions at various calcination temperatures can influence the formation energy and the occurrence of different defects. Low oxygen flow rates at low temperatures are considered O-poor conditions, while high oxygen flow rates at high temperatures are deemed O-rich conditions. Given that nanoparticles were prepared within the equilibrium temperature range (400, 500, and 600 °C), the formation energy for all defects was calculated under equilibrium conditions. To assess the effect of the supercell size on defect formation energies, ZnO:Al, ZnO:Ga, ZnOV_O, ZnOV_O:Al, and ZnOV_O:Ga configurations were modeled in 2 × 2 × 2, 2 × 2 × 3, 2 × 3 × 3, and 3 × 3 × 3 supercells (32, 48, 72, and 108 atoms, respectively) and fully optimized. The corresponding formation energies were calculated using the GGA functional under equilibrium conditions. Additionally, the formation energies were calculated using GGA+U approximations under equilibrium conditions for all states and defects of a 2 × 3 × 3 supercell to investigate the effect of the Hubbard potential (see Fig. S8).

As shown in Fig. S8, the size of the superlattice has little effect on the formation energy of the ZnO:Al and ZnO:Ga states, as well as the ZnOV_O, ZnOV_O:Al, and ZnOV_O:Ga defects. The formation energies of the 2 × 3 × 3 and 3 × 3 × 3 superlattices are nearly identical, with the maximum difference of about 0.38 eV occurring for the ZnO:Al and ZnO:Ga states. The minimal impact of the superlattice size on formation energy results from the neutral charge of the superlattices, which prevents Coulomb interactions between defect charges and their periodic images. In contrast, the Hubbard potential significantly lowers the defects' formation energy relative to the superlattice size. Nonetheless, the trend of decreasing formation energy across all states and defects is consistent, with defect energies being similar in both GGA and GGA+U calculations.

The intrinsic defect formation energies for pure zinc oxide and zinc oxide doped with 3% and 5% aluminum and gallium were calculated under equilibrium conditions, and the results are shown in Fig. 6 and Fig. S9 (SI). To examine the effect of interactions between the dominant defects (oxygen and zinc vacancies), nearby Zn–O vacancy pairs (V_OZn) and distant Zn–O vacancy pairs were also simulated in pure and Al- and Ga-doped samples. Their formation energies are also reported in Fig. 6. As shown in Fig. S9, when aluminum replaces oxygen, the formation energies for defects are significantly higher than when it replaces zinc. Therefore, the probability of aluminum substituting for zinc is higher than that of substituting for oxygen. Under equilibrium temperature conditions, the formation probabilities of ZnOV_O and ZnOV_Zn defects in pure zinc oxide; ZnOO_i:Al, ZnOV_O:Al, and ZnOV_Zn:Al defects in 3% Al-doped ZnO; ZnO:Ga, ZnOV_O:Ga, and ZnOV_Zn:Ga defects in 3% Ga-doped ZnO; ZnOO_i:Al, ZnOV_O:Al, and ZnOV_Zn:Al defects in 5% Al-doped ZnO; and ZnOV_O:Ga and ZnOV_Zn:Ga defects in 5% Ga-doped zinc oxide are higher than those of other defects. The highest error corresponds to the ZnOV_Zn:Al (5%) defect, while the lowest error is associated with the ZnOV_O:Ga (3%) defect.

Fig. 6 Intrinsic defect formation energies for pure and 3% and 5% Al/Ga doped ZnO, with error bars under equilibrium conditions.

3.4. Theoretical EPR spectroscopy

In addition to the defect formation energy, EPR spectra were simulated for all the optimized configurations of intrinsic defects in ZnO doped with 3% Al/Ga to identify the nature of the defects. In a paramagnetic material, the unpaired electrons may be naturally occurring or generated by defects, and they can be identified by EPR. The structure and defects in pure zinc oxide and zinc oxide doped with 5% aluminum and gallium are diamagnetic due to the lack of unpaired electrons. The EPR spectrum is generated by the interaction between an external magnetic field and the spins of unpaired electrons, along with the influence of nearby nuclear spins, referred to as the HFCC. The resonance field was determined by calculating the EPR parameters (g-factor and HFCC(A)) and laboratory parameters, including the range of the magnetic field sweep and microwave frequency, as given below:In this equation, m_l is the nuclear spin quantum number and B₀ represents the magnitude of the magnetic field in mT. Nevertheless, since ¹⁷O and ⁶⁷Zn isotopes have a nuclear spin of 5/2, but their natural abundances are negligible (0.038% and 1.4%, respectively), they make insignificant contributions to the EPR spectrum and the resonance field position. The Al atom with a nuclear spin of 5/2 has a natural abundance of 100%, and the ⁶⁹Ga and ⁷¹Ga isotopes with nuclear spins of 1/2 and 3/2 have natural abundances of 60.11% and 38.89%, respectively. Therefore, their HFCCs can affect the EPR spectrum. Here, the EPR parameters are Δg = g_e − g and the HFCC(A) for aluminum and gallium atoms, where g_e is the Zeeman energy for a free electron (g_e = 2.002319304). The HFCC of each atom is proportional to that atom's spin density.⁴⁵ The spin density around the aluminum atom is greater when it substitutes for the oxygen atom than it is when it replaces the zinc atom. Consequently, the HFCC for aluminum is significantly larger in the case of substituting for the oxygen atom compared to its substitution for the zinc atom. Fig. 7 shows the calculated values of Δg and A for 3% Al/Ga-doped ZnO and also for 3% Al-doped ZnO when the Al atom has replaced the O atom (ZnO:Al_O).

Fig. 7 EPR spectrum of defects in zinc oxide doped with 3% aluminum and gallium.

The HFCCs for Al and Ga atoms when they replace zinc atoms are negligible, so for each defect, excluding ZnOV_OZn:Al and ZnOV_OZn:Ga defects, there is a resonant magnetic field whose position is determined only by the value of the g factor. However, when aluminum replaces oxygen and ZnOV_OZn:Al and ZnOV_OZn:Ga defects, they exhibit a large HFCC, resulting in six resonant magnetic fields (2m_l + 1 in number) that are equally spaced, with their separation determined by the hyperfine coupling constant of the aluminum atom. By comparing the position of the resonant field in the theoretical EPR with the experimental spectrum, it is possible to determine whether defects are present or present in negligible concentrations. A comparison of the simulated EPR spectrum of aluminum replacing oxygen with the experimental EPR spectrum reveals that, similar to the defect formation energy results, aluminum cannot replace oxygen. Comparison of the simulated resonance field position with experimental results shows that the concentration of ZnOZn_i:Al and ZnOV_OZn:Al defects for Al-doped materials and the concentration of ZnOZn_i:Ga, ZnOO_i:Ga, ZnOV_O:Ga, and ZnOV_OZn:Ga defects for Ga-doped samples are insignificant and practically zero. Besides, the concentration of ZnOO_i:Al and ZnOV_O:Al defects is very low and causes a broadening of the EPR spectrum for the Al-doped samples. A high intensity of the EPR spectrum was observed at the resonance fields of 356 and 357 mT for Al- and Ga-doped zinc oxide, respectively. These peaks indicate that 3% Al-doped ZnO contains one, two, or all three types of states: ZnO:Al, ZnOV_Zn:Al, and . At the same time, the 3% Ga-doped sample exhibits ZnO:Ga, ZnOV_Zn:Ga, and defects.

3.5. Electronic structure and defect-induced carrier concentration

The photoluminescence properties of the specimens are directly influenced by the band structure characteristics, which, in turn, are influenced by intrinsic defects. The intrinsic defects cause the emission of photons at different energies by creating shallow and deep traps. As observed in the previous section on defect formation energy, the concentration and type of defects change with the calcination temperature, resulting in a change in the electronic structure and, consequently, the emission of photons with different energies in the PL spectrum. We will explain how defects, optical emission, and calcination temperature are interconnected to optimize the properties of pure and doped ZnO for X-ray detectors.

From the previous two sections, we found that ZnOV_Zn, ZnOV_O, and defects are present in pure zinc oxide. For 3% Al-doped ZnO, the ZnOV_Zn:Al is the dominant defect; however, ZnOO_i:Al, ZnOV_O:Al, and defects are less prevalent. ZnOV_Zn:Ga is the most prevailing defect, and is present at a lower concentration in 3% Ga-doped ZnO. Therefore, zinc oxide doped with 3% aluminum exhibits the highest defects compared to pure zinc oxide and the 3% Ga-doped sample, which is consistent with PALS results. For the 5% Ga-doped ZnO, ZnOV_Zn:Ga and are the most dominant defects, and the same condition is also observed for the 5% Al-doped ZnO, although the ZnOO_i:Al defect is present in 5% Al-doped ZnO at a lower concentration.

Fig. 8 presents the spin-dependent band structures and electron density distributions of pure ZnO for the dominant defects, while Fig. S10–S13 (SI) display the results for doped ZnO, along with the spin density distribution. The corresponding figures for the less dominant defects are presented in Fig. S14–S18 (SI). Since the aluminum and gallium density of states do not contribute to the VBM and CBM, they are not shown in the corresponding figures. As can be seen, the electron distribution is centered around most of the oxygen atoms. For zinc oxide doped with 3% Al and Ga, we observe an asymmetry between the spin-up and spin-down states. The asymmetry in the total density of states arises from the 2p orbitals of the oxygen atoms, and as can be seen, the spin density is located around these oxygen atoms.

Fig. 8 Spin-polarized band structures, total and partial densities of states and electron density distribution (0.25 a.u.) for ZnO with dominant defects. Magenta/cyan bands show spin-up/spin-down states, green/violet circles indicate electron/hole traps, and gray/red spheres denote Zn/O atoms.

The presence of defects induced new energy states in the band gap. It changes the optical band splitting, which causes the emission of photons with different energies (under laser light excitation) and also changes the concentration of electrons and holes in the zinc oxide semiconductor. In Fig. 9, the energy of the emitted photons in the presence of defects is compared with the energy of the experimentally emitted photons at synthesis temperatures of 400 to 600 °C. In the experimental PL spectra for all modes, the emitted photon energy is in the ranges of 3.14–3.22 eV (brown triangle symbols), 2.74–2.94 eV (red diamond symbols), and 2.52–2.66 eV (blue plus symbols), which correspond to emissions near the band edge, shallow donor, and deep central states, respectively. The purple cross symbols also correspond to the very deep central states, which are absent in the experimental PL spectra.

Fig. 9 Comparison of the energy of photons emitted from defects in pure and doped zinc oxide with the energy of experimentally emitted photons at the calcination temperatures of 400 to 600 °C.

Comparison of the photon energies emitted from the experimental PL spectra with the photon energies calculated from the band structures shows that the emitted photon energy corresponding to the near band edge (brown triangle) is obtained from pure ZnO. Besides, the two emitted photon energies corresponding to the shallow donor states (red rhombus) and the deep central states (blue sum) are obtained from the ZnOV_Zn or , and three ZnOV_O, ZnOV_Zn and defects, respectively. For ZnO doped with 3% Al, the energies of the emitted photons corresponding to the shallow donor state, the deep central state, and the one near the band edge are obtained from ZnOV_Zn:Al, ZnOV_O:Al defects and mainly from the ZnO:Al state (with a negligible amount from ZnOO_i:Al and defects), respectively. For zinc oxide doped with 5% aluminum, the energy of the emitted photons corresponding to the shallow donor level, the deep central level and the one near the band edge comes from ZnOV_Zn:Al defects, defects and mainly the ZnO:Al state (with a small amount of ZnOO_i:Al defect), respectively. For 3% Ga-doped zinc oxide, the energies of the emitted photons corresponding to the shallow donor states, the deep central state, and the one near the band edge are obtained from ZnOV_Zn:Ga defects, defects, and ZnO:Ga state, respectively. Finally, for 5% gallium-doped zinc oxide, the energy of the emitted photons corresponding to the shallow donor level and the deep central level originates from the defects, and the photon energy corresponding to NBE originates from the ZnO:Ga state.

The carrier concentrations and free volumes at room temperature for pure and doped zinc oxide with dominant defects are reported in Table S1 of the SI. Section S2.2 of the SI provides the theoretical framework and the procedure used to calculate the charge carrier density. It is found that zinc vacancy defects have minimal impact, while oxygen vacancy defects lead to an increase in electron concentration. According to Table S1, zinc oxide doped with 3% aluminum has higher electron and hole concentrations than pure zinc oxide and zinc oxide doped with 3% gallium, which is in good agreement with the PALS results. Table S1 shows that adding aluminum and gallium does not significantly affect the free volumes, while the free volumes for zinc vacancies are greater than those for oxygen vacancies. The presence of oxygen and zinc vacancies simultaneously and at a distance from each other increases the free space relative to the zinc vacancy. Therefore, the free volumes for zinc oxide doped with 3% gallium are increased compared to those for pure and 3% Al-doped ZnO and are in good agreement with the PALS results. Increasing the aluminum and gallium doping concentrations results in a significant increase in both the electron concentration and the free volume.

The comprehensive identification of intrinsic defects through PL, EPR, PALS, and DFT+U analyses allows us to link microscopic defect structures with macroscopic detector performance. In scintillation detectors, strong luminescence near the band edge and weak luminescence in defects, especially those related to deep-level energy emission, are required to achieve fast-response X-ray detection.^46,47 However, defects that induce shallow trap-to-NBE transitions may also suppress visible emissions and enhance near-band-edge (NBE) emission, which improves fast response and timing resolution. Conversely, defects that induce deep trap-to-NBE transitions act as radiative recombination centers, increasing visible luminescence but introducing slow decay components that reduce temporal resolution in scintillation detectors. For pure and doped zinc oxide, the PL spectrum exhibited three photon emission peaks in the energy ranges of 3.14–3.22 eV (NBE), 2.74–2.94 eV (shallow donor states), and 2.52–2.66 eV (deep central states). The lower the intensity of shallow donor state and deep central state peaks relative to the NBE, the faster the detector response. For scintillator applications, Fig. 2 shows that the shallow trap-to-NBE and deep trap-to-NBE intensity ratios are 0.19 and 0.08, respectively, which are considerably lower than those of the other samples. This indicates lower defect concentrations and thus a more perfect crystal lattice, consistent with the higher defect formation energies of ZnOV_Zn:Ga (3%) (1.16 eV) and (6.64 eV) in the 3% Ga-doped sample (Fig. 6). Therefore, a scintillation detector made of zinc oxide doped with 3% gallium can have a faster response than the other samples.

Detecting X-rays with semiconductor devices involves optimizing two key aspects: achieving rapid device responses and detecting weak signals. To do this, researchers often rely on optical conductors that cover a wide area and can keep unwanted currents, especially dark current, to an absolute minimum. ZnO-based detectors should exhibit low dark current, which enhances the signal-to-noise ratio and improves energy resolution. However, to keep the dark current low, the material needs to have very few free carriers before it is exposed to X-rays. Additionally, one should also consider the issue of trap states, which are imperfections or defects within the material that can trap electrons or holes. Sometimes, they release visible light, and sometimes they do not; however, what they do create is interference. They boost noise and reduce sensitivity. For semiconductor-type X-ray detectors, reducing the oxygen vacancy concentration minimizes trap-assisted leakage currents, while controlled doping with gallium effectively decreases defect density and stabilizes the electron concentration, thereby enhancing sensitivity and reducing noise. Therefore, optimizing synthesis temperature and dopant concentration to minimize the oxygen vacancy while maintaining structural integrity is the key pathway for achieving both high-speed scintillation and high-sensitivity semiconductor X-ray detection performance. According to DFT calculations and PALS, the 3% Ga-doped sample exhibits the lowest carrier concentration (6.22 × 10⁹ cm⁻³) compared with 1.82 × 10¹³ cm⁻³ for pure ZnO, 5.14 × 10¹² cm⁻³ for 3% Al, 8.39 × 10¹⁹ cm⁻³ for 5% Al, and 4.77 × 10¹⁶ cm⁻³ for 5% Ga, suggesting reduced leakage current and noise. It therefore exhibits reduced leakage current and noise, making 3% Ga-doped zinc oxide an ideal material for highly sensitive semiconductor X-ray detectors. The large amount of free volume causes the electrons and holes to remain there for a long time after the X-ray radiation is produced, allowing them to recombine later and thus slowing down the detector response. Although the free volume for 3% Ga-doped ZnO is larger than that for 3% Al-doped ZnO and pure ZnO, the number of significant free volumes created by defects is small. Furthermore, this is less significant compared to the sensitivity of the semiconductor detectors.

4. Conclusions

We have explored the promising capabilities of Al/Ga-doped ZnO as both an X-ray scintillator and a semiconductor detector. For ZnO to be an ideal X-ray detector, it must emit intense light under UV exposure, be transparent to visible and infrared light, and maintain a very low leakage current of just a few tens of nanoamperes in reverse bias. Native defects, including vacancies and interstitials, mainly influence the performance of detectors based on ZnO. We utilized PL, EPR, and PALS, along with DFT+U calculations, to identify defects in pure zinc oxide and zinc oxide doped with aluminum and gallium, which were synthesized at calcination temperatures of 400, 500, and 600 °C. Our results can be outlined in five major points:

(i) We found that at three calcination temperatures, ZnOV_O, ZnOV_Zn, and defects are present in pure zinc oxide. In samples doped with 3% Al, the ZnOV_Zn:Al defect can be identified at higher concentrations, while ZnOV_O:Al, ZnOO_i:Al, and defects occur at very low concentrations. ZnOV_Zn:Ga is the most prevailing defect, and is present at a lower concentration in 3% Ga-doped ZnO. For the 5% Ga-doped ZnO, the most dominant defect is . A similar pattern is observed in the 5% Al-doped ZnO, although the ZnOO_i:Al and ZnOV_Zn:Al defects appear at a lower concentration.

(ii) 3% Al-doped ZnO exhibits higher electron and hole concentrations compared to both pure zinc oxide and 3% Ga-doped ZnO.

(iii) Increasing the aluminum and gallium doping concentrations results in a significant increase in both the electron concentration and the free volume.

(iv) The 3% Ga-doped ZnO sample exhibits the lowest shallow and deep trap-to-NBE intensity ratios (0.19 and 0.08), consistent with higher defect formation energies and a reduced defect density, indicating superior crystal quality and a faster response for scintillator applications.

(v) It also exhibits the lowest carrier concentration (6.22 × 10⁹ cm⁻³) among all studied samples, which minimizes leakage current and noise, making 3% Ga-doped ZnO a highly promising candidate for sensitive semiconductor X-ray detectors.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request. Data for this article including calculation data are available at https://drive.google.com/file/d/1Fj2rllYao1At4Wfr4-6s04hc91fhH-oN/view?usp=drive_web.

Supplementary information (SI) is available. See DOI: https://doi.org/10.1039/d5cp03107a.

Acknowledgements

The financial support from Nuclear Science and Technology Research Institute (NSTRI) and Kharazmi University is gratefully acknowledged.

References

1. Y. Zhong, X. Yu and W. Li, Nanoscale scintillating materials for X-ray imaging: fundamentals and applications, J. Mater. Chem. C, 2025, 13, 1036–1062.
2. A. K. Singh, V. Singh and P. I. Anand, An innovative laser decal transfer of ZnO ceramic in LIG for advanced hybrid nanogenerator applications, Ceram. Int., 2025, 51, 4957–4970.
3. Z. Chen, et al., Enhanced photocatalytic degradation of tetracycline hydrochloride by hollow nanofiber Ag@ZnGa₂O₄/ZnO with synergistic effects of LSPR and S-scheme interface engineering, J. Mater. Chem. C, 2024, 12, 17448–17457.
4. Y. Zou, S. Yu, S. Cao, X. Lu and Y. Zhou, Ni-CdS composited with ZnO for improved surface reaction and charge efficiency for photocatalytic hydrogen production from formic acid, Phys. Chem. Chem. Phys., 2025 10.1039/D5CP01271F.
5. F. T. Z. Toma, M. S. Rahman and K. H. Maria, A review of recent advances in ZnO nanostructured thin films by various deposition techniques, Discovery Mater., 2025, 5, 60.
6. J. Han, H. Li, J. Cheng, X. Ma and Y. Fu, Advances in metal oxide semiconductor gas sensor arrays based on machine learning algorithms, J. Mater. Chem. C, 2025, 13, 4285–4303.
7. J. Chen, et al., Preparation and application of nanostructured ZnO in radiation detection, Materials, 2024, 17, 3549.
8. R. T. Williams, W. W. Wolszczak, X. Yan and D. L. Carroll, Perovskite quantum-dot-in-host for detection of ionizing radiation, ACS Nano, 2020, 14, 5161–5169.
9. X.-Y. Liu, G. Pilania, A. A. Talapatra, C. R. Stanek and B. P. Uberuaga, Band-edge engineering to eliminate radiation-induced defect states in perovskite scintillators, ACS Appl. Mater. Interfaces, 2020, 12, 46296–46305.
10. C. E. Lee, E. Lee, S. Lee and W. Lee, Radiation damage in proton-irradiated ZnO single crystal, J. Korean Phys. Soc., 2010, 56, 2108–2111.
11. J. Fan, H. Li, W. Liu and G. Ouyang, Fabrication Strategies of Mn²⁺-Based Scintillation Screens for X-Ray Detection and Imaging, Angew. Chemie, 2025, 137, e202425661.
12. T. Wang, et al., Advances in Metal Halide Perovskite Scintillators for X-Ray Detection, Nano-Micro Lett., 2025, 17, 275.
13. M. Kano, et al., Response-time-improved ZnO scintillator by impurity doping, J. Cryst. Growth, 2011, 318, 788–790.
14. M. Xu, et al., Transient radiation imaging based on a ZnO:Ga single-crystal image converter, Sci. Rep., 2018, 8, 4178.
15. R. Lin, et al., Ultrafast (600 ps) α-ray scintillators, PhotoniX, 2022, 3, 9.
16. R. Yang, et al., ZnO with p-type doping: Recent approaches and applications, ACS Appl. Electron. Mater., 2023, 5, 4014–4034.
17. M. Fiederle and T. Baumbach, Perspective of perovskite-based X-ray hybrid pixel array detectors, Front. Phys., 2024, 12, 1395690.
18. F. Yao, K. Dong, W. Ke and G. Fang, Micro/Nano perovskite materials for advanced x-ray detection and imaging, ACS Nano, 2024, 18, 6095–6110.
19. X. Zhao, et al., High sensitivity X-ray detector based on a 25 µm-thick ZnO film, Sens. Actuators, A, 2022, 334, 113310.
20. A. Datta, Z. Zhong and S. Motakef, A new generation of direct X-ray detectors for medical and synchrotron imaging applications, Sci. Rep., 2020, 10, 20097.
21. P. Jin, et al., Realizing nearly-zero dark current and ultrahigh signal-to-noise ratio perovskite X-ray detector and image array by dark-current-shunting strategy, Nat. Commun., 2023, 14, 626.
22. I. Ayoub, et al., Advances in ZnO: Manipulation of defects for enhancing their technological potentials, Nanotechnol. Rev., 2022, 11, 575–619.
23. J. Ji, et al., ZnO Luminescence and scintillation studied via photoexcitation, X-ray excitation and gamma-induced positron spectroscopy, Sci. Rep., 2016, 6, 31238.
24. R. R. Jalolov and B. N. Rustamova, Defect formation in hydrothermally grown ZnO and S-doped ZnO nanostructures, Ceram. Int., 2025, 51, 40833–40840.
25. N. B. Bidmeshki, Y. T. Azar, F. Ziaie and M. Janbazi, Unravelling the nature of the α-keratin EPR signal: an ab initio study, Phys. Chem. Chem. Phys., 2021, 23, 6815–6822.
26. A. Saidani, et al., Effect of calcination temperature on the photocatalytic activity of precipitated ZnO nanoparticles for the degradation of rhodamine B under different light sources, Water, 2024, 17, 32.
27. J. VandeVondele, et al., Quickstep: fast and accurate density functional calculations using a mixed Gaussian and plane waves approach, Comput. Phys. Commun., 2005, 167, 103–128.
28. P. Giannozzi, et al., QUANTUM ESPRESSO: a modular and open-source software project for quantumsimulations of materials, J. Phys.: Condens. Matter, 2009, 21, 395502.
29. M. Sharbatdaran and M. Janbazi, Effect of temperature on the structure, catalyst and magnetic properties of un-doped zinc oxide nanoparticles: experimental and DFT calculation, RSC Adv., 2024, 14, 31153–31164.
30. S. Stoll and A. Schweiger, EasySpin, a comprehensive software package for spectral simulation and analysis in EPR, J. Magn. Reson., 2006, 178, 42–55.
31. B. Choudhury and A. Choudhury, Local structure modification and phase transformation of TiO₂ nanoparticles initiated by oxygen defects, grain size, and annealing temperature, Int. Nano Lett., 2013, 3, 1–9.
32. G. Xiong, U. Pal, J. G. Serrano, K. B. Ucer and R. T. Williams, Photoluminesence and FTIR study of ZnO nanoparticles: the impurity and defect perspective, Phys. Status Solidi C, 2006, 3, 3577–3581.
33. M. D. McCluskey and S. J. Jokela, Defects in ZnO, J. Appl. Phys., 2009, 106, 71101.
34. G. Xiong, U. Pal, J. G. Serrano, K. B. Ucer and R. T. Williams, Photoluminescence and FTIR study of ZnO nanoparticles: the impurity and defect perspective, Phys. Status Solidi C, 2006, 3581, 3577–3581.
35. G. R. Dillip, et al., Oxygen Vacancy-Induced Structural, Optical, and Enhanced Supercapacitive Performance of Zinc Oxide Anchored Graphitic Carbon Nanofiber Hybrid Electrodes, ACS Appl. Mater. Interfaces, 2016, 8, 5025–5039.
36. M. J. F. Empizo, et al., Structural and optical characterization and scintillator application of hydrothermal-grown ZnO microrods, Opt. Mater., 2017, 65, 82–87.
37. H. Zhao and H. Kalt, Energy-dependent Huang-Rhys factor of free excitons, Phys. Rev. B: Condensed Matter Mater. Phys., 2003, 68, 125309.
38. A. Sahoo, et al., Facile control of giant green-emission in multifunctional ZnO quantum dots produced in a single-step process: femtosecond pulse ablation, Nanoscale Adv., 2024, 7, 524–535.
39. A. Biganeh, O. Kakuee and H. Rafi-Kheiri, Positron Annihilation Spectroscopy of KCl (Zn) crystals, Appl. Radiat. Isot., 2020, 166, 109330.
40. D. Giebel and J. Kansy, LT10 program for solving basic problems connected with defect detection, Phys. Proc., 2012, 35, 122–127.
41. A. M. Toufiq, et al., The influence of Mn doping on the structural and optical properties of ZnO nanostructures, Phys. B, 2021, 604, 412731.
42. A. Janotti and C. G. de Walle, Fundamentals of zinc oxide as a semiconductor, Rep. Prog. Phys., 2009, 72, 126501.
43. S. B. Zhang and J. E. Northrup, Chemical potential dependence of defect formation energies in GaAs: Application to Ga self-diffusion, Phys. Rev. Lett., 1991, 67, 2339.
44. F. Oba, A. Togo, I. Tanaka, J. Paier and G. Kresse, Defect energetics in ZnO: A hybrid Hartree-Fock density functional study, Phys. Rev. B: Condensed Matter Mater. Phys., 2008, 77, 245202.
45. M. Janbazi, Radiation-Induced N-methylacetamide and β-Alanine radicals in aqueous solutions at different pH Levels: EPR spectroscopy and First-Principles calculations, J. Mol. Liq., 2025, 423, 127035.
46. A. Spustaka, D. Millers, E. Einbergs, V. Vitola and E. Vanags, Luminescence of ZnO:Ga ceramics under sub-threshold electron irradiation, Opt. Mater., 2022, 126, 112181.
47. L. Chen, et al., Comparative study on fluorescence decay time of doped ZnO crystals under α and β excitation, Nucl. Instrum. Methods Phys. Res., Sect. A, 2019, 933, 71–74.

---