Enhanced thermal stability and afterglow performance in Sr₂Ga_2−xAl_xSiO₇:Ce³⁺ phosphors via band gap tailoring

Abstract

In this study, the regulation of Al³⁺/Ga³⁺ on Sr₂(Ga,Al)₂SiO₇:Ce³⁺ phosphors is demonstrated to enhance the photoluminescence and persistent luminescence (PersL) performances of Ce³⁺. With the replacement of Al³⁺, the emission band of Ce³⁺ shifts to the longer wavelength side, thermal stability of Ce³⁺ emission becomes better, and Ce³⁺ persistent luminescence shows stronger intensity and longer lifetime. To get insight into the structure-luminescence relationship, the influences of Al³⁺/Ga³⁺ on the structure of Sr₂(Ga,Al)₂SiO₇:Ce³⁺, 4f–5d_i (i = 1–5) transition energies of Ce³⁺ and trap distributions are discussed. With the construction of the vacuum referred binding energy scheme and analysis of a series of thermoluminescence (TL) spectra, the increase of band gap is found to tune the trap depths and change persistent luminescence performances of Ce³⁺ in Sr₂(Ga,Al)₂SiO₇. Finally, with the stimulation of a 1060 nm laser, the optimized Sr₂(Ga,Al)₂SiO₇:Ce³⁺ phosphors show strong persistent luminescence of Ce³⁺, suggesting the potential for application in optical storage.

---

1. Introduction

Ce³⁺-activated persistent phosphors have been widely applied in emergency lighting and decorations, and show promising applications in optical information storage, dermatological treatments and so on.^1–3 Thus far, the composition regulation of a well-selected matrix compound has been an effective method to tune the luminescence properties and to develop novel Ce³⁺-doped persistent luminescent materials with high efficiency and long-lasting luminescence performances. For example, after the introduction of Si⁴⁺ ions, Mg₃Y₂Ge_3−xSi_xO₁₂:Ce³⁺ phosphors have millisecond afterglow and reduce the light flicker in alternating-current light-emitting diodes (LED).⁴ When gradually replaced by Ga³⁺ in Y₃Al_5−xGa_xO₁₂:Ce³⁺, V³⁺ phosphors, the emission color optimizes from green to yellow, and the trap depths are tailored from 1.2 to 1.6 eV, which makes them suitable for optical storage media.⁵ With the co-substitution of Si⁴⁺ and N³⁻ in Gd₃Ga₂(Al_3−ySi_y)(O_12−yN_y):Ce³⁺ phosphors, the emission band of Ce³⁺ gets broader and is located at longer wavelength, which increases the red emission component in warm white LEDs.⁶ The emission intensity, emission wavelength and radiative lifetime of Ce³⁺ are tunable and depend on the host matrix.

In addition, Ce³⁺ is an important lanthanide (Ln) ion with 4f¹ electronic configuration. Due to the typical 4f–5d transition, Ce³⁺ is often used to probe the ligand polarization, crystal field strength and coupling effect between Ln activators and vibrations of the host lattice.⁷ Moreover, the generation of persistent luminescence is related to the capture and release of electrons from trap levels, which are already lived or created by co-doping ions in phosphors. It is commonly accepted that the electron as a charge carrier transports through the host conduction band (CB) in the trapping and detrapping processes.^8,9 Thus, the Ce³⁺ ion is a favorable activator to connect the 5d excited states, trap levels and host conduction band, so as to understand the relationship between the host matrix and persistent luminescence properties.

The A₂BT₂O₇ (A = Ca²⁺ and Sr²⁺; B = Mg²⁺ and Al³⁺; T = Al³⁺ and Si⁴⁺) compounds belong to the melilite structure family,¹⁰ which has various strategies of ion substituents. In addition, Ce³⁺/Eu²⁺ doped A₂BT₂O₇ materials generally exhibit excellent optical performances. Ca₂Al₂SiO₇:Ce³⁺,Tb³⁺ phosphors are reported to be suitable for white-light-emitting diodes.¹¹ The Ca₂MgSi₂O₇:Eu²⁺,Dy³⁺ phosphors show potential for application in the sensors of mechanical stress with naked-eye green mechanoluminescence.¹² The phosphors for the application of optical information storage are better possessed of deep traps, which can be stimulated by low-energy light to release electrons to generate PersL.^13–15 The composition regulation of Ce³⁺ doped A₂BT₂O₇ is expected to tune the afterglow properties and to discover novel phases. In this paper, we systematically study the evolutions of the host structure, photoluminescence and persistent luminescence of Ce³⁺ in Sr₂(Ga,Al)₂SiO₇ phosphors. The positions of the emission band, radiative lifetime, 4f–5d transition energies and thermal stability of Ce³⁺ are investigated with low-temperature VUV-UV spectra and temperature-dependent decay curves. The persistent luminescence of Ce³⁺ and distributions of trap levels are discussed using thermoluminescence spectra and decay curves. This work demonstrates that the regulation of Al³⁺/Ga³⁺ in the host matrix tunes the persistent luminescence property of Ce³⁺ mainly via the change of band gap rather than that of 4f and 5d states of Ce³⁺.

2. Experimental

The host and Ce³⁺-doped Sr₂Ga_2−xAl_xSiO₇ (x = 0, 0.8, 1.2, 1.6, 1.8, 1.9 and 2) samples were prepared by a high-temperature solid-state reaction method. Na⁺ ions were introduced as charge compensators in the Ce³⁺ and Eu³⁺ singly doped samples. The details of preparations and characterization are described in the ESI.†

3. Results and discussion

3.1. P-XRD patterns and structure

Rietveld refinement results based on laboratory XRD data of the prepared Sr₂Ga₂SiO₇ sample using Sr₂Al₂SiO₇ (ICSD 30698) as the initial model are depicted in Fig. 1a, indicating that Sr₂Ga₂SiO₇ belongs to the class of A₂BT₂O₇ compounds.¹⁶ The obtained factors R_wp = 6.71%, R_p = 4.55%, and R_B = 4.41% indicate a reliable refined quality. The synthesized Sr₂Ga₂SiO₇ compound is crystallized in a tetragonal structure with the P4̄2₁m (113) space group. The lattice parameters are a = b = 7.9541(4) Å, c = 5.3185(7) Å, V = 336.49(7) ų, α = β = γ = 90° and Z = 2. Table S1† lists the structural parameters of Sr₂Ga₂SiO₇. Sr₂Ga₂SiO₇ has one Sr²⁺, one Si⁴⁺ and two Ga³⁺ sites. The Ga³⁺ ion shares one half of its crystallographic sites with Si⁴⁺. The unit cell structure of Sr₂Ga₂SiO₇ is shown in Fig. 1b. The structure of Sr₂Ga₂SiO₇ is built by the GaO₄ and (Ga,Si)O₄ tetrahedra as well as SrO₈ polyhedra. The SrO₈ polyhedron is of C_s point symmetry. The bond lengths of Sr²⁺–O²⁻ are listed in Table S2.† The average bond length of Sr²⁺–O²⁻ is ∼2.705 Å. The nearest distance of Sr²⁺–Sr²⁺ is ∼3.640 Å.

Fig. 1 (a) Rietveld refinement of P-XRD data of the synthesized Sr₂Ga₂SiO₇ sample; (b) unit cell structure of Sr₂Ga₂SiO₇; (c) P-XRD data of synthesized Sr_1.98Ce_0.01Na_0.01Ga_2−xAl_xSiO₇ (x = 0.8, 1.2, 1.6, 1.8, 1.9 and 2) samples; (d) lattice parameters (a, c and V) of Sr_1.98Ce_0.01Na_0.01Ga_2−xAl_xSiO₇.

In view of similar effective ionic radii of eight-fold coordination [r(Ce³⁺) = 1.143 Å, r(Sr²⁺) = 1.26 Å, and r(Na⁺) = 1.18 Å],¹⁷ Ce³⁺ and Na⁺ ions prefer to occupy the Sr²⁺ site. XRD data of synthesized Sr_1.98Ce_0.01Na_0.01Ga_2−xAl_xSiO₇ (x = 0.8, 1.2, 1.6, 1.8, 1.9 and 2) samples are shown in Fig. 1c to check the phase purity. The diffraction peaks match well with the refined results, indicating that these samples are in single pure phase, and Ce³⁺ is incorporated into the host matrix. Besides, we notice that the strongest diffraction peak of the (1 2 1) crystallographic plane shifts to the large-angle side with gradual substitution of Ga³⁺ into Al³⁺. On the basis of these XRD data, their lattice parameters are further calculated by Rietveld refinement. As shown in Fig. 1d and Table S3,† the lattice parameters a (b), c and V decrease linearly with the replacement of Al³⁺. It indicates that the unit cell of Sr₂Ga₂SiO₇ undergoes shrinkage with the substitution of Al³⁺.

3.2. VUV-UV photoluminescence of Sr₂Ga_2−xAl_xSiO₇:Ce³⁺

Highest-height normalized VUV-UV excitation spectra of Sr_1.98Ce_0.01Na_0.01Ga_2−xAl_xSiO₇ (x = 0.8, 1.2, 1.6, 1.8, 1.9 and 2) samples at 10 K are collected in Fig. 2a to analyse the evolutions of Ce³⁺ luminescence. These excitation spectra can be divided into two parts at ∼5.7 eV to analyse. The high-lying excitation band above 5.7 eV is mainly attributed to excitonic absorption of the host. By picking the maximum of host absorption above, the exciton energy (E_x) of Sr_1.98Ce_0.01Na_0.01Ga_1.2Al_0.8SiO₇ is evaluated to be ∼6.22 eV. When the content (x) increases from 0.8 to 2, the E_x increases about 0.58 eV. The mobility band gap (E_VC) can be calculated by extra adding the exciton binding energy (0.008 × E_x²) to exciton energy.¹⁸ For the Sr_1.98Ce_0.01Na_0.01Ga_1.2Al_0.8SiO₇ sample, it is estimated to be 6.53 eV (6.22 + 0.008 × 6.22² = 6.53 eV). With the substitution of Al³⁺, the E_VC increases about 0.64 eV (from 6.53 to 6.56, 6.64, 6.75, 6.95 and 7.17 eV). These E_VC values fall between those of other aluminates and silicates, such as Y₃Al_5−xGa_xO₁₂ (x = 0–5, E_VC = 7.50–6.40 eV),³ SrAl₂O₄ (7.28 eV),¹⁹ Ba₂MgSi₂O₇ (7.04 eV),²⁰ and Na₃GdSi₂O₇ (6.32 + 0.008 × 6.32² = 6.64 eV).²¹ The band gap energy increases with the increase in Al³⁺ content, which coincides with the variations of Y₃(Ga,Al)₅O₁₂,²² Lu₃(Ga,Al)₅O₁₂,²³ and Gd₃(Ga,Al)₅O₁₂ compounds.²⁴

Fig. 2 (a) Synchrotron radiation VUV-UV excitation (λ_em = 380 nm, 10 K) spectra of Sr_1.98Ce_0.01Na_0.01Ga_2−xAl_xSiO₇ (x = 0.8, 1.2, 1.6, 1.8, 1.9 and 2) samples; (b) emission (λ_ex = 320 nm, 10 K) spectra of Sr_1.98Ce_0.01Na_0.01Ga_2−xAl_xSiO₇ samples and the corresponding Gaussian fitting results.

The excitation bands below 5.7 eV are ascribed to the 4f–5d transitions of Ce³⁺ in Sr₂Ga_2−xAl_xSiO₇ (x = 0.8, 1.2, 1.6, 1.8, 1.9 and 2). In consideration of the occupancy of the eight-fold coordinated Sr²⁺ site with low point symmetry, 5d orbitals of Ce³⁺ are expected to split into five non-degenerated orbitals. In the 3.5–5.6 eV region of Fig. 2a, four excitation bands (A, B, C and D) are observed, and one missed excitation band may hide in these four excitation bands or host-exciton band. Considering that the highest 5d₅ state is generally located near the conduction band of the host, the 4f–5d₅ excitation band is most likely to be the lost excitation band and hides in band D with very weak intensity. The VUV-UV excitation spectra are further normalized to the height of band A and shown in Fig. S1.† With the replacement of Al³⁺, the intensity of bands B, C and D gradually increases; meanwhile, that of the host-exciton band decreases. In terms of the energy distribution, when bands A and B are assigned to the 4f–5d_1,2 transitions of Ce³⁺, bands C and D should contain the excitation bands of 4f–5d_3,4,5 transitions. The excitation spectrum of the representative Sr_1.98Ce_0.01Na_0.01Al₂SiO₇ sample in the 4.7–5.8 eV range is enlarged in the inset of Fig. S1.† Bands C and D are well fitted with a sum of three Gaussian functions to estimate the positions of 4f–5d_3,4,5 excitation bands. Table 1 summarizes the energies of Ce³⁺ 4f–5d_i (i = 1–5) excitation bands and host-exciton bands of Sr_1.98Ce_0.01Na_0.01Ga_2−xAl_xSiO₇ (x = 0.8, 1.2, 1.6, 1.8, 1.9, and 2) samples. In order to confirm the assignments and positions of Ce³⁺ 4f–5d_i (i = 1–5) excitation bands, the energies of centroid shifting and crystal field splitting of Ce³⁺ in Sr_1.98Ce_0.01Na_0.01Ga_2−xAl_xSiO₇ are evaluated. First, the centroid energy of Ce³⁺ in Sr_1.98Ce_0.01Na_0.01Ga_1.2Al_0.8SiO₇ is estimated to be 4.72 eV by calculating the average energy of the 4f–5d_i (i = 1–5) excitation bands. Then the 5d centroid of Ce³⁺ downshifts about 1.63 eV in Sr_1.98Ce_0.01Na_0.01Ga_1.2Al_0.8SiO₇ with respect to its free gaseous state (6.35 eV). With Al³⁺ substituted content x increasing from 0.8 to 2, the down shifting energy (ε_c) decreases slightly from 1.63 to 1.59 eV. These values are between the range of Sr²⁺-aluminates and silicates such as SrAl₁₂O₁₉ (1.24 eV),²⁵ Sr₂Al₂SiO₇ (1.58 eV),²⁶ SrSiO₃ (1.91 eV),²⁵ and Li₂SrSiO₄ (1.45 eV).²⁷ Moreover, the crystal field splitting energy (ε_cfs) of Ce³⁺ in Sr_1.98Ce_0.01Na_0.01Ga_1.2Al_0.8SiO₇, viz the energy difference between the 4f–5d₁ excitation band and 4f–5d₅ excitation band, is estimated to be 1.68 eV. We notice that band A shows a mild shift to the lower-energy side with the replacement of Al³⁺, while bands C and D to the higher-energy side. Also, bond distances of Sr²⁺ (Ce³⁺)-O²⁻ in Sr₂Ga_2−xAl_xSiO₇:Ce³⁺ show a slight decrease with the increase in content x (Table S4†). The observations indicate an increase of crystal field splitting with the substitution of Ga³⁺ into Al³⁺. By further comparing the estimated crystal field splitting energies of Sr_1.98Ce_0.01Na_0.01Ga_2−xAl_xSiO₇ samples, they show a slight increase with the increase of Al³⁺/Ga³⁺ ratio. And they are close to that of SrSiO₃:Ce³⁺ (1.92 eV),^25,28 SrB₂O₄:Ce³⁺ (2.17 eV),^25,29 and Sr₉Lu(PO₄)₇:Ce³⁺ (Ce³⁺(1), 1.76 eV),³⁰ in which Ce³⁺ occupies an eight-fold Sr²⁺ site with low point symmetry. The above analyses support the assignments of Ce³⁺ 4f–5d_i excitation bands.

Table 1 Energies (eV) of Ce³⁺ 4f–5d_i (i = 1–5) excitation bands, centroid shift, crystal field splitting, 5d₁–²F_5/2,7/2 emission bands and host-related excitation band in Sr_1.98Ce_0.01Na_0.01Ga_2−xAl_xSiO₇ (x = 0.8, 1.2, 1.6, 1.8, 1.9, and 2) samples

Property	x = 0.8	x = 1.2	x = 1.6	x = 1.8	x = 1.9	x = 2
4f–5d1 excitation band (eV)	3.72	3.71	3.71	3.69	3.69	3.69
4f–5d2 excitation band (eV)	4.22	4.20	4.12	4.10	4.19	4.20
4f–5d3 excitation band (eV)	4.98	4.99	4.99	5.00	5.00	5.00
4f–5d4 excitation band (eV)	5.27	5.30	5.34	5.34	5.34	5.35
4f–5d5 excitation band (eV)	5.40	5.42	5.48	5.50	5.51	5.54
Centroid shift (eV)	1.63	1.63	1.62	1.62	1.60	1.59
Crystal field splitting (eV)	1.68	1.71	1.77	1.81	1.82	1.85
Emission bands (eV)	3.46, 3.24	3.46, 3.24	3.46, 3.24	3.44, 3.22	3.43, 3.21	3.41, 3.19
Host-related excitation band (eV)	6.22	6.25	6.32	6.42	6.60	6.80

---

Emission spectra of Ce³⁺ in Sr_1.98Ce_0.01Na_0.01Ga_2−xAl_xSiO₇ samples displayed in Fig. 2b also imply the increase of crystal field splitting with the replacement of Al³⁺. The emission band shifts to the long-wavelength side when the x value increases. The emission band originates from the 5d₁–²F_J (J = 5/2, 7/2) transitions of Ce³⁺, and is fitted with a sum of two Gaussian functions. The energy difference of the doublet emission bands is about 0.22 eV, which is in line with the theoretical energy difference (0.25 eV) between ²F_5/2 and ²F_7/2 levels.³¹ In addition, the intensity of the emission band is found to increase systematically with the gradual replacement of Al³⁺. The integral intensity of the Ce³⁺ emission band (350–480 nm) as a function of the substituted content (x) is plotted in Fig. S2.† It finally increases about four times with x increasing from 0.8 to 2. The excited 5d electrons of Ce³⁺ give rise to the radiative luminescence accompanied by some non-radiative processes such as electron-vibrational interaction, energy transfer, thermal ionization and so on.^32,33 These processes usually degenerate the excited 5d states to some extent. In the series of Sr₂Ga_2−xAl_xSiO₇ host compounds, the electron-vibrational coupling effect between the host lattice and Ce³⁺ activator is regarded to be similar. Because of low doping concentration of Ce³⁺, energy transfer among Ce³⁺ also shows less influence on the evolution of emission intensity. Based on the discussions above, both the band gap and crystal field splitting of Sr_1.98Ce_0.01Na_0.01Ga_2−xAl_xSiO₇ systematically increase, which directly affects the energy difference between the host conduction band and the excited 5d states of Ce³⁺. We think the change of energy gap should play important roles in the thermal stability of Ce³⁺ luminescence, leading to the difference of Ce³⁺ emission intensity. In the following, the influences of temperature and energy difference of 5d-CB on Ce³⁺ luminescence are analysed.

3.3. Temperature-dependent luminescence of Sr₂Ga_2−xAl_xSiO₇:Ce³⁺

Temperature dependent decay curves of the representative Sr_1.98Ce_0.01Na_0.01Ga_1.2Al_0.8SiO₇ sample were recorded and are displayed in Fig. 3a. The decay curve of the Xe lamp is also recorded to rule out the influence of the Xe lamp on the variation of decay curves. At 77 K, the decay curve has deviated obviously from single exponential behavior, indicating the existence of a non-radiative process to quench Ce³⁺ luminescence. Decay curves continue to deviate from that of 77 K with the increase in temperature, implying that temperature has a significant effect on the non-radiative process. The lifetime of Ce³⁺ as a function of temperature is shown in Fig. 3b. With the temperature raising from 77 to 500 K, the lifetime of Ce³⁺ decreases about 76% from 19.4 to 4.60 ns. As a comparison, Fig. S3a–c† show temperature-dependent decay curves of Sr_1.98Ce_0.01Na_0.01Ga_2−xAl_xSiO₇ (x = 1.8, 1.9 and 2) samples. Their decay curves also show decreasing behavior, but different magnitude of reductions. In the Sr_1.98Ce_0.01Na_0.01Ga_0.2Al_1.8SiO₇ sample, the decay time of Ce³⁺ reduces about 35% from 26.5 to 17.2 ns with the temperature increasing from 77 to 500 K. The decrease is smaller than that of the Sr_1.98Ce_0.01Na_0.01Ga_1.2Al_0.8SiO₇ sample. By further comparing the thermal behaviors of x = 0.8, 1.8, 1.9 and 2 samples, the decrease of decay curves becomes smaller and smaller with the larger substituted content (x) of Al³⁺. The thermal ionization of Ce³⁺ from the excited 5d electrons into host conduction band is probable the non-radiative channel to quench Ce³⁺ emission.³² To analyze qualitatively, the Mott formula is adopted to obtain the activation energy (ΔE) of thermal ionization of Ce³⁺ emission and expressed as follows:³⁴where τ(T) is the lifetime of Ce³⁺ at the temperature T, and denoted as τ(0) when T = 0 K; A is a pre-exponential factor; k_B is the Boltzmann constant (∼8.6173 × 10⁻⁵ eV K⁻¹). According to the best fitting results, the activation energy (ΔE₁) of the Sr_1.98Ce_0.01Na_0.01Ga_1.2Al_0.8SiO₇ sample is estimated to be about 0.058 eV, which is smaller than that (ΔE₂ = 0.072 eV) of the Sr_1.98Ce_0.01Na_0.01Ga_0.2Al_1.8SiO₇ sample. Temperature-dependent lifetime of Ce³⁺ in Sr_1.98Ce_0.01Na_0.01Ga_2−xAl_xSiO₇ (x = 1.9 and 2) samples and their fitting results are displayed in Fig. S3d.† Their ΔE values are shown in the inset of Fig. 3a. It can be found that the activation energy gets larger and larger with the gradual increase of Al³⁺/Ga³⁺ ratio. Similar observations can be found in the reported Gd₃Al_5−xGa_xO₁₂:1%Ce³⁺,²⁴ and Y₃Al_5−xGa_xO₁₂:0.5%Ce³⁺.³ Accordingly, thermal stability of Ce³⁺ luminescence gets better with the gradual replacement of Al³⁺ in Sr₂Ga_2−xAl_xSiO₇.

Fig. 3 (a) Temperature dependent decay curves (λ_ex = 320 nm, λ_em = 400 nm) of the Sr_1.98Ce_0.01Na_0.01Ga_1.2Al_0.8SiO₇ sample; (b) lifetimes of Ce³⁺ in Sr_1.98Ce_0.01Na_0.01Ga_2−xAl_xSiO₇ (x = 0.8 and 1.8) samples as a function of temperature; the inset shows the activation energy (ΔE) as a function of Al³⁺ content (x).

The vacuum referred binding energy (VRBE) scheme reveals the locations of the host conduction band, host valence band and Ln²⁺/Ln³⁺ 4f and 5d_1–5 states in a specific inorganic compound, which can directly denote the evolutions of the energy difference between the excited 5d states of Ce³⁺ and the bottom of the host conduction band. For the construction of the VRBE scheme, the energies of band gap, 5d-centroid shifting of Ce³⁺ and charge transfer of Eu³⁺–O²⁻ in Sr₂Ga_2−xAl_xSiO₇ (x = 0.8, 1.2, 1.6, 1.8, 1.9, and 2) are necessary. Fig. S4† shows the VUV-UV excitation spectra of Eu³⁺ in Sr_1.98Eu_0.01Na_0.01Ga_2−xAl_xSiO₇ samples. The broad excitation bands located at about 4.77 eV are mainly ascribed to charge transfer bands (CTBs) of Eu³⁺–O²⁻. The weak excitation bands at the higher-energy side are attributed to the host exciton absorptions of Sr_1.98Eu_0.01Na_0.01Ga_2−xAl_xSiO₇ samples, whose positions coincide with that of Ce³⁺ doped Sr₂Ga_2−xAl_xSiO₇ samples as discussed above. With the ready information, we begin to build the VRBE scheme. First, the Coulomb repulsion energies and 4f-VRBE for Eu²⁺ and Eu³⁺ in Sr₂Ga_2−xAl_xSiO₇ are calculated with the help of 5d-centroid energy shift of Ce³⁺ and their experiential relationships.³⁵ Then the tops of valence bands of Sr₂Ga_2−xAl_xSiO₇ are determined with the charge transfer energies of Eu³⁺–O²⁻ and Eu²⁺ 4f-VRBE. And the bottoms of conduction bands are estimated with the band gap energies of Sr_1.98Ce_0.01Na_0.01Ga_2−xAl_xSiO₇. Finally, the 4f- and 5d-VRBE of Ce³⁺ are calculated with the 4d–5d_i (i = 1–5) transition energies of Ce³⁺ and the updated zigzag curves.³⁶ The stacked VRBE diagram including 4f and 5d states of Ce³⁺ in Sr₂Ga_2−xAl_xSiO₇ compounds is shown in Fig. 4. When x = 0.8, the 5d states of Ce³⁺ are located into the CB of Sr₂Ga_1.2Al_0.8SiO₇. Because 5d electrons of Ce³⁺ easily show delocalization into the host conduction band, the observed weaker intensity at low temperature and worse thermal stability of Ce³⁺ luminescence in Sr₂Ga_1.2Al_0.8SiO₇ can be interpreted. With the increase in Al³⁺ content, the 5d₁ state of Ce³⁺ downshifts slightly; meanwhile, the bottom of the host CB shifts upward obviously. As a result, the energy difference between the 5d₁ state of Ce³⁺ and host CB becomes larger and larger with gradual Al³⁺ replacement. We notice that the energy difference is different from the activation energy (ΔE) obtained from the decay curves, which may be due to the systematic error. And the systematic error can be accounted for by the effect of lattice relaxation.³ This trend confirms that the thermal stability of Ce³⁺ luminescence gets better and better with the gradual substitution of Ga³⁺ into Al³⁺.

Fig. 4 Vacuum referred binding energy (VRBE) scheme for Ce³⁺/Eu²⁺ 4f and Ce³⁺ 5d_1–5 states in Sr₂Ga_2−xAl_xSiO₇ (x = 0.8, 1.2, 1.6, 1.8, 1.9, and 2). The red and grey rectangles represent the conduction band and the valence band, respectively.

3.4. Thermoluminescence of Sr₂Ga_2−xAl_xSiO₇:Ce³⁺

Each Sr_1.98Ce_0.01Na_0.01Ga_2−xAl_xSiO₇ (x = 0.8, 1.2, 1.6, 1.8, 1.9 and 2) sample was heated to 700 K to clear up the signals before the thermoluminescence measurements. Then the samples are preirradiated with a mercury lamp (254 nm) for five minutes at 100 K. TL curves of Sr_1.98Ce_0.01Na_0.01Ga_2−xAl_xSiO₇ were recorded and shown in Fig. 5a. The TL peaks (labeled as A) show mainly distributions in the 100–300 K range in Sr_1.98Ce_0.01Na_0.01Ga_1.2Al_0.8SiO₇. With the gradual increase of Al³⁺/Ga³⁺ ratio, the intensity of band A increases; meanwhile, band A shifts to the high temperature side. In addition, the weak TL peaks (labeled as B) in the region of 300–700 K also raise gradually, and become dominant in the x = 2 sample. Further, we select the Sr_1.98Ce_0.01Na_0.01Al₂SiO₇ sample for wavelength-resolved TL measurements under the same conditions as displayed in Fig. 5b. The inset i of Fig. 5b shows the TL curve with the monitoring emission wavelength at 393 ± 15 nm, whose profile coincides with that in Fig. 5a. In addition, when monitoring TL spectra at 200 ± 15 K and 450 ± 15 K, obtained emission spectra are shown in the inset ii of Fig. 5b. It can be found that the emission profiles are in line with that of Ce³⁺ emission. Therefore, the TL peaks A and B originate from Ce³⁺ luminescence.

Fig. 5 (a) TL curves (254 nm Hg lamp, T_exc = 100 K) of Sr_1.98Ce_0.01Na_0.01Ga_2−xAl_xSiO₇ (x = 0.8, 1.2, 1.6, 1.8, 1.9 and 2) samples; (b) wavelength-resolved TL spectra (254 nm Hg lamp, T_exc = 100 K) of the Sr_1.98Ce_0.01Na_0.01Al₂SiO₇ sample.

Under preirradiation with a mercury lamp, 5d excited electrons of Ce³⁺ are activated into the host CB, and trapped by the charge carriers nearby. With the help of thermal activation, the trapped electrons can release back to the excited 5d state of Ce³⁺ through the host conduction band. The distribution of TL bands reflects the depth of traps and the concentration of electrons stored in traps. TL peaks are mainly located in the low temperature range, indicating that most of the traps are shallow and distribute close to the bottom of the host CB in the present case. Because TL peaks systematically shift to the higher temperature side with the increase in Al³⁺/Ga³⁺ ratio, we infer that the trap levels probably originate from the Al³⁺/Ga³⁺-related defects rather than oxygen vacancies (V′′_O). Because Al³⁺ and Ga³⁺ ions share their crystallographic sites with Si⁴⁺ in the Sr₂Ga_2−xAl_xSiO₇ host compound, the second coordination sphere of Ce³⁺ centers is non-uniform, which leads to the continuous distributions of trap depths. Furthermore, the second coordination sphere is varied by different Al³⁺/Ga³⁺ ratios, which also results in the changes of trap depths and finally shows their differences in TL spectra.^3,37 Because the XPS characteristic peak of Ce⁴⁺ at near 918 eV is barely observed,³⁸ Ce⁴⁺ ions are mostly reduced into Ce³⁺ (Fig. S5†). Herein, it is reasonable to postulate that the Al³⁺/Ga³⁺-related traps are induced by the Ga³⁺/Si⁴⁺ and Al³⁺/Si⁴⁺ point defects, V_Ga–Ce³⁺–V_O and V_Al–Ce³⁺–V_O defect clusters or other unknown defects.^4,39 Because the band gap increases with the substitution of Al³⁺, these defect levels may move away from the bottom of the host conduction band, and the stored electrons require larger activation energy (higher temperature) for detrapping from them. Considering that Al³⁺ ions replace Ga³⁺ with equal valence, non-equivalent defects that may act as charge carriers are regarded to show negligible influence on the thermoluminescence of Ce³⁺.

In the charging phase, the trapping and detrapping of excited 5d electrons occur at the same time.⁴⁰ Because the band gap of the x = 0.8 sample is the smallest in the Sr_1.98Ce_0.01Na_0.01Ga_2−xAl_xSiO₇ series, the trap depth is shallow and difficult to store electrons at 100 K. With the increase of band gap, the trap depth becomes deeper, and the electrons are relatively not easy to detrap in the charging process. Therefore, the concentration of trapped electrons increases, and the TL intensity of band A increases with the replaced content x increasing from 0.8 to 1.9. However, shallow traps capture electrons easily from the host CB because they are closer to the bottom of the host CB than deep traps. As observed, the TL intensity of band B is less intense than that of band A in the x = 0.8–1.9 samples. Besides the trap depth and trapping preference as discussed above, we believe that the concentration of traps also shows an effect on the TL intensity.⁴¹ Because the band gap of x = 2 is the largest in this series of samples, the shallow traps close to the host CB turn to be deeper, and the concentration of deep traps might be larger than that of the shallow traps. It can explain the phenomena that the TL intensity of band B increases and finally becomes dominant.

To investigate the trap depths, TL curves of the Sr_1.98Ce_0.01Na_0.01Al₂SiO₇ sample at different excitation temperatures (T_exc = 100–310 K) are measured and shown in Fig. 6a. With the increase of excitation temperature, band A shows an intense decrease and disappears; meanwhile, band B shifts to a high temperature and decreases slightly. The shift of broad band B indicates the distributions of TL peaks at higher temperature. In consideration of detection limit, we changed to the device of high excitation temperature. Excitation temperature-dependent TL curves (T_exc = 325–465 K) are shown in the inset of Fig. 6a. With excitation temperature reaching 465 K, band B decreases obviously and its maximum shifts to about 500 K. At higher excitation temperature, trapped electrons more easily get enough energy to release back into the CB. Therefore, when the excitation temperature becomes high in the charging process, shallow traps hardly stored electrons, but deeper traps can. Though the excitation temperature is not enough to release the electrons from the deep trap, the concentration of captured electrons becomes relatively lower than that of low excitation temperature. Besides, when the detrapped electrons move in the midway to the excited 5d states of Ce³⁺, they are also captured by deeper traps.⁴ The TL kinetics can affect the intensity and the shape of the TL curve. Because the initial rise analysis on the TL curve is independent of TL kinetics, this method is adopted to estimate the trap depths with eqn (2).⁴²

where I(T) is the TL intensity at the temperature T; E is the trap depth; B is the pre-exponential factor. We take the logarithm of the above formula and plot ln(I(t)) as a function of 1000/T as shown in Fig. 6b. It can be found that TL curves show a straight section on the low temperature side. The E values are obtained by the slope −E/k_Bvia a linear fitting to the straight section. The estimated trap depths as a function of excitation temperature (T_exc = 100–350 K) are displayed in Fig. S6 and Table S5.† In the Sr_1.98Ce_0.01Na_0.01Al₂SiO₇ sample, trap depths range from 0.05 to 0.65 eV. Based on the difference of TL intensity in Fig. 6b, the density of trap depths is displayed in the inset. In the range of 100–310 K, the trap depths mainly distribute at about 0.066, 0.150 and 0.226 eV in the Sr_1.98Ce_0.01Na_0.01Al₂SiO₇ sample.

Fig. 6 (a) Excitation temperature-dependent TL curves (254 nm Hg lamp) of the Sr_1.98Ce_0.01Na_0.01Al₂SiO₇ sample; (b) initial rise analysis on TL curves as a function of excitation temperature (T_exc); the inset shows the trap depth distribution of the Sr_1.98Ce_0.01Na_0.01Al₂SiO₇ sample.

Fig. 7 shows the temperature-dependent decay curves of persistent luminescence of the Sr_1.98Ce_0.01Na_0.01Al₂SiO₇ sample after charging with the excitation with a Hg lamp for five min, and then monitoring at 400 nm. The decay of PersL intensity is relatively faster in the range of 0–500 s and finally becomes slow after 500 s. With the temperature increasing from 77 to 500 K, the initial PersL intensity [I(t) at t = 0 s] increases gradually, and is finally beyond fifty times compared to that of 77 K. It shows a large increasing magnitude in the low temperature range, but increases slower with the temperature raising to 300 K. In addition, the decay processes are different at different temperatures. The PersL intensities in the 0–500 s region keep increasing with the temperature increasing from 100 to 380 K. When the temperature reaches 430 and 500 K, PersL intensities show a faster decrease than that of 380 K after the decay time t = 500 s. As demonstrated above, the Sr_1.98Ce_0.01Na_0.01Al₂SiO₇ sample contains different traps and the trap distributions are different at different temperatures. Most of the shallow and deep traps can capture the electrons from the excited 5d states of Ce³⁺ at low temperature, and then the stored electrons are released to produce PersL. Because of not enough energy to thermally activate electrons from deep traps and because the detrapping rate of electrons via thermal stimulation is slow at 77 K, the electrons mainly originate from the tunneling effect and detrapping from very shallow traps. To confirm the tunneling effect, selected decay curves are plotted with the double-logarithmic model in Fig. S7a.† The decay curves in the 500–3000 s period satisfy the linear decaying process (I⁻¹ ∼ t), suggesting that the tunneling effect is involved in the PersL processes.⁴³ Because multiple traps show contributions to the decay processes, PersL intensity shows non-single exponential decaying processes and decreases relatively fast in the 0–500 s range. When the electrons of the shallower traps are almost cleared up, the decay curves turn back to a slow decrease. Further, with the temperature increasing to 280 K, deep traps and a small number of shallow traps turn to store the electrons in the preirradiation process. Then the electrons are liberated from the shallow traps and a portion of deeper traps to generate PersL. At the same time, the detrapping rate of electrons becomes higher. Therefore the PersL intensity increases relatively at higher temperature on the whole.

Fig. 7 Temperature-dependent PersL decay curves (λ_ex = 254 nm Hg lamp, λ_em = 400 nm) of the Sr_1.98Ce_0.01Na_0.01Al₂SiO₇ sample.

Following the thought above, PersL intensity should keep increasing with the increase of detrapping rate when the temperature continues to increase. However, PersL intensity of 500 K in the 500–3200 s range decreases slightly compared with that of 380 K. In order to compare the decaying rate, the PersL decay curves are further normalized and shown in Fig. S7b.† It can be found that the decrease of PersL intensity is slower and slower with the temperature increasing from 77 to 280 K. When the temperature increases from 280 to 500 K, the decrease of PersL intensity gets faster and faster. Because the detrapping rate of electrons keeps increasing with the temperature increasing from 77 to 280 K, abundant electrons transport back for recombination of electron–hole pairs to generate PersL of Ce³⁺, which shows contributions to the slow decrease of PersL intensity. With the temperature increasing gradually to 500 K, the detrapping rate continues to increase, and then the stored electrons are cleared faster and hardly maintain the supplement for a long time. The decreasing phenomenon is caused by two main reasons. On the one hand, because only a small fraction of the deep traps can effectively store electrons at high temperatures (380–500 K), the concentration of stored electrons becomes relatively smaller. On the other hand, because the detrapping rate gets faster at high temperatures, stored electrons are easily in short supply with the increase of decay time, leading to a fast decay in the PersL intensity. Accordingly, the excitation temperature, trap depths, detrapping rate and concentration of stored electrons cooperatively show influences on the PersL intensity and the PersL decay.

In order to investigate the influences of Al³⁺/Ga³⁺ regulation on the trap depth and concentration of stored electrons, Fig. 8a shows the TL curves of Sr_1.98Ce_0.01Na_0.01Ga_2−xAl_xSiO₇ (x = 0.8, 1.2, 1.6, 1.8, 1.9 and 2) samples at RT. A weak TL band is located at about 350 K for the x = 0.8 sample. With the increase of the x value, the intensity of the TL band obviously increases, which is about thirty times with the x value increasing from 0.8 to 2. The increase of TL intensity indicates that the concentration of stored electrons becomes larger with the substitution of Ga³⁺ into Al³⁺. Simultaneously, the TL band gradually shifts to the high temperature side. According to the results in section 3.2, the band gap becomes larger and larger with the increase of x content in Sr_1.98Ce_0.01Na_0.01Ga_2−xAl_xSiO₇ samples. The schematic mechanism of PersL displayed in Fig. 8b shows the influence of increasing Al³⁺/Ga³⁺ ratio on the band gap and trap distributions. Assuming that trap levels are independent of the variation of Al³⁺/Ga³⁺ ratio, the energy difference between the trap levels and the bottom of the host conduction band gets larger, and the depths of trap levels increase. As observed, the TL band distributes in the higher temperature side with the increase of Al³⁺ content. Furthermore, when the traps get deep, the stored electrons are not easy to release in the charging process, which also shows contributions to the increase of concentration of stored electrons. Therefore, the trap depth gets deep, and the concentration of stored electrons increases with the replacement of Al³⁺.

Fig. 8 (a) TL curves (λ_ex = 254 nm Hg lamp) of Sr_1.98Ce_0.01Na_0.01Ga_2−xAl_xSiO₇ (x = 0.8, 1.2, 1.6, 1.8, 1.9 and 2) samples with the excitation temperature at RT; (b) schematic mechanism of PersL; (c) PersL decay curves (λ_ex = 254 nm Hg lamp, RT) of Sr_1.98Ce_0.01Na_0.01Ga_2−xAl_xSiO₇ samples.

To investigate the influences of Al³⁺/Ga³⁺ regulation on detrapping rate, Sr_1.98Ce_0.01Na_0.01Ga_2−xAl_xSiO₇ samples are charged with a 254 nm Hg lamp for five min at RT and their PersL decay curves are recorded in Fig. 8c. As predicted, the PersL intensity of the x = 0.8 sample is the weakest and decreases quickly after a short decay time. With the gradual replacement of Ga³⁺ by Al³⁺, the initial PersL intensity raises and remains strong after decaying for 1 h. The observations coincide with the analysis of TL curves above. The PersL intensity is proportional to the recombination rate between the electrons and 5d excited states of Ce³⁺ which act as holes in the present case. Further, the recombination rate is related to many factors such as the temperature, trap depth, concentration of stored electrons, concentration of holes and fixed recombination probability.⁴¹ Considering that the PersL curves are all collected at RT, the temperature shows a negligible effect on the variations of PersL decay. Moreover, because the trap depth becomes deeper with the increase of Al³⁺/Ga³⁺ ratio, more traps become suitable for the storage of electrons at RT, which also results in the increase of concentration of electrons in the preirradiation process. In addition, when the trap depths get deep, the detrapping rate of electrons will become slower from the deep traps, which lengthens the PersL decay time. The increase of electron concentration and the decrease of detrapping rate cooperatively make Ce³⁺ PersL have stronger intensity and longer decay time. Accordingly, the Al³⁺/Ga³⁺ ratio changes the PersL intensity and decay time by regulating the trap distribution and trap depth.

3.5. NIR stimulated persistent luminescence of Sr₂Ga_2−xAl_xSiO₇:Ce³⁺

To investigate the influence of laser stimulation, PersL decay curves of Sr_1.98Ce_0.01Na_0.01Ga_2−xAl_xSiO₇ (x = 0.8, 1.2, 1.6, 1.8, 1.9 and 2) samples are measured with a 1060 nm laser alternately turned on and off, and displayed in Fig. 9a. The background is measured for comparison, in order to rule out the possibility that the 1060 nm laser may act as a light source to charge the Sr_1.98Ce_0.01Na_0.01Ga_2−xAl_xSiO₇ samples.³⁷ When turning on the laser at the decay time t = 50 s, decay curves show obvious platforms. The platforms disappear when the laser is off, and PersL intensity follows back to the decay process similar to that in Fig. 8b. In addition, each time the laser is turned on, the platform intensity gets weaker than that of the previous. After stimulation for two and more times with a laser, the PersL intensities of the x = 0.8 sample almost become close to that of the background, confirming that the x = 0.8 sample shows the worst afterglow performance. The PersL intensities of the other samples are much higher than the background, and increase with the increase of Al³⁺/Ga³⁺ ratio. As observed at t = 50 s, their platform intensities are about twice the PersL intensity of no stimulation. When stimulating (t = 50–100 s) with a laser of 1060 nm, the PersL intensity of x = 1.2 and 2 samples becomes approximately four and ten times higher than the laser background, respectively. These phenomena imply that the platforms do not originate from the 1060 nm laser but the stimulated afterglow.

Fig. 9 (a) PersL decay curves (λ_ex = 254 nm Hg lamp, RT) of Sr_1.98Ce_0.01Na_0.01Ga_2−xAl_xSiO₇ (x = 0.8, 1.2, 1.6, 1.8, 1.9 and 2) samples with and without 1060 nm laser stimulation; (b) emission spectra of Sr_1.98Ce_0.01Na_0.01Ga_2−xAl_xSiO₇ (x = 0.8, 1.2, 1.6, 1.8, 1.9 and 2) samples with the stimulation of a 1060 nm laser (solid curves) and no laser (dash dot curves) after a delay time (t_d) of 2 min or 30 s.

To confirm that the stimulated afterglow originates from Ce³⁺ luminescence, emission spectra of Sr_1.98Ce_0.01Na_0.01Ga_2−xAl_xSiO₇ samples after delay for two min are further collected in Fig. 9b. Because the PersL intensity of the Sr_1.98Ce_0.01Na_0.01Ga_1.2Al_0.8SiO₇ sample is weak, the emission spectrum at the delay time t = 30 s is measured only. The PersL emission bands are in line with that of excitation with VUV-UV light. With the stimulation of a 1060 nm laser, the PersL emission profile remains the same, but the PersL emission intensity increases. It can be found that the PersL emission intensity is about twice that of no laser stimulation. Accordingly, the 1060 nm laser can activate the stored electrons to generate Ce³⁺ afterglow in Sr_1.98Ce_0.01Na_0.01Ga_2−xAl_xSiO₇ (x = 1.2, 1.6, 1.8, 1.9 and 2). With the UV light write-in and 1060 nm laser readout mode, the modulated Sr₂Ga_2−xAl_xSiO₇:Ce³⁺ afterglow materials show potential for application in multidimensional optical data storage.

4. Conclusions

In summary, we have studied the effects of Al³⁺/Ga³⁺ on the crystal structure of Sr₂Ga₂SiO₇, photoluminescence and persistent luminescence of Ce³⁺via XRD data, VUV-UV, temperature-dependent PL and TL spectra. With the gradual substitution of Ga³⁺ into Al³⁺, the host lattice of Sr₂Ga₂SiO₇ undergoes linear shrinkage. The mobility band gap is estimated to be 6.53 eV for Sr₂Ga_1.2Al_0.8SiO₇, and finally increases about 0.64 eV after being totally replaced by Al³⁺. The crystal field splitting of Ce³⁺ increases about 0.17 eV, while the 5d centroid shifting energies decrease slightly. The VRBE schemes of Sr₂Ga_2−xAl_xSiO₇ are constructed to reveal the variations of the bottom of the host conduction band and 5d energy levels of Ce³⁺. Due to the changes of band gap and 4f–5d transition energies of Ce³⁺, thermal stability of Ce³⁺ luminescence gets better with Al³⁺ replacement.

The intensity of Ce³⁺ PersL increases and the lifetime of Ce³⁺ PersL prolongs with the substitution of Al³⁺ due to the deeper trap distributions. With the initial rise analysis on TL curves, the trap depths are estimated, and the traps mainly distribute at near 200 K. With the increasing ratio of Al³⁺/Ga³⁺, the PersL intensity of Ce³⁺ increases with trap depths getting deeper. The host band gap is the key factor influencing traps levels and Ce³⁺ PersL in Sr_1.98Ce_0.01Na_0.01Ga_2−xAl_xSiO₇. With the simulation of a 1060 nm laser, Sr_1.98Ce_0.01Na_0.01Ga_2−xAl_xSiO₇ (x = 1.9 and 2) samples show intense Ce³⁺ persistent luminescence. Our results prove that the regulation of metallic ions of anionic group in a well-selected host compound optimizes the persistent luminescence performances mainly through the adjustment of the host band gap.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

We greatly appreciate Dr Jinwei Li and Miss Yunlin Yang (Sun Yat-sen University) as well as Dr Litian Lin (Guangdong Academy of Sciences) for the assistance in measurements of luminescence spectra of samples. This work is financially supported by the National Natural Science Foundation of China (Grant No. 61905289 and 51902053), Natural Science Foundation of Guangdong Province (Grant No. 2019A1515011988), Young Innovative Talents Project of Guangdong Provincial (No. 2020KQNCX242) and the Innovative Team Program of Guangdong Province (No. 2020KCXTD057).

Notes and references

1. P. Feng, G. Li, H. J. Guo, D. W. Liu, Q. F. Ye and Y. H. Wang, Identifying a cyan ultralong persistent phosphorescence (Ba, Li) (Si, Ge, P)₂O₅: Eu²⁺, Pr³⁺ via solid solution strategy, J. Phys. Chem. C, 2019, 123, 3102–3109.
2. S. K. Sharma, M. Bettinelli, I. Carrasco, J. Beyer, R. Gloaguen and J. Heitmann, Dynamics of charges in superlong blacklight-emitting CaB₂O₄: Ce³⁺ persistent phosphor, J. Phys. Chem. C, 2019, 123, 14639–14646.
3. J. Ueda, P. Dorenbos, A. J. J. Bos, K. Kuroishi and S. Tanabe, Control of electron transfer between Ce³⁺ and Cr³⁺ in the Y₃Al_5−xGa_xO₁₂ host via conduction band engineering, J. Mater. Chem. C, 2015, 3, 5642–5651.
4. H. Lin, J. Xu, Q. M. Huang, B. Wang, H. Chen, Z. B. Lin and Y. S. Wang, Bandgap tailoring via Si doping in inverse-garnet Mg₃Y₂Ge₃O₁₂: Ce³⁺ persistent phosphor potentially applicable in AC-LED, ACS Appl. Mater. Interfaces, 2015, 7, 21835–21843.
5. W. H. Li, Y. X. Zhuang, P. Zheng, T. L. Zhou, J. Xu, J. Ueda, S. Tanabe, L. Wang and R. J. Xie, Tailoring trap depth and emission wavelength in Y₃Al_5−xGa_xO₁₂: Ce³⁺,V³⁺ phosphor-in-glass films for optical information storage, ACS Appl. Mater. Interfaces, 2018, 10, 27150–27159.
6. A. J. Mao, Z. Y. Zhao, J. T. Wang, C. X. Yang, J. J. Ren and Y. H. Wang, Crystal structure and photo-luminescence of Gd₃Ga₂(Al_3−ySi_y)(O_12−yN_y): Ce³⁺ phosphors for AC-warm LEDS, Chem. Eng. J., 2019, 368, 924–932.
7. Y. Y. Ou, W. J. Zhou, D. J. Hou, M. G. Brik, P. Dorenbos, Y. Huang and H. B. Liang, Impacts of 5d electron binding energy and electron–phonon coupling on luminescence of Ce³⁺ in Li₆Y(BO₃)₃, RSC Adv., 2019, 9, 7908–7915.
8. Y. Zhang, D. X. Chen, W. L. Wang, S. Yan, J. W. Liu and Y. J. Liang, Long-lasting ultraviolet-A persistent luminescence and photostimulated persistent luminescence in Bi³⁺-doped LiScGeO₄ phosphor, Inorg. Chem. Front., 2020, 7, 3063–3071.
9. Z. Song, J. Zhao and Q. L. Liu, Luminescent perovskites: recent advances in theory and experiments, Inorg. Chem. Front., 2019, 6, 2969–3011.
10. Z. G. Xia, C. G. Ma, M. S. Molokeev, Q. L. Liu, K. Rickert and K. R. Poeppelmeier, Chemical unit cosubstitution and tuning of photoluminescence in the Ca₂(Al_1−xMg_x)(Al_1−xSi_1+x)O₇: Eu²⁺ phosphor, J. Am. Chem. Soc., 2015, 137, 12494–12497.
11. H. Y. Jiao and Y. H. Wang, Ca₂Al₂SiO₇:Ce^{3 +}, Tb^{3 +}: A white-light phosphor suitable for white-light-emitting diodes, J. Electrochem. Soc., 2009, 156, J117–J120.
12. H. W. Zhang, H. Yamada, N. Terasaki and C. N. Xu, Green mechanoluminescence of Ca₂MgSi₂O₇: Eu and Ca₂MgSi₂O₇: Eu, Dy, J. Electrochem. Soc., 2008, 155, J55–J57.
13. C. L. Wang, Y. H. Jin, Y. Lv, G. F. Ju, D. Liu, L. Chen, Z. Z. Li and Y. H. Hu, Trap distribution tailoring guided design of a super-long persistent phosphor Ba₂SiO₄:Eu²⁺, Ho³⁺ and photostimulable luminescence for optical information storage, J. Mater. Chem. C, 2018, 6, 6058–6067.
14. Y. X. Zhuang, Y. Lv, L. Wang, W. W. Chen, T. L. Zhou, T. Takeda, N. Hirosaki and R. J. Xie, Trap depth engineering of SrSi₂O₂N₂: Ln²⁺, Ln³⁺ (Ln²⁺ = Yb, Eu; Ln³⁺ = Dy, Ho, Er) persistent luminescence materials for information storage applications, Appl. Mater. Interfaces, 2018, 10, 1854–1864.
15. L. T. Zhao, T. P. Liu, H. X. An, H. W. Wang, K. M. Lv, J. H. Huang, X. S. Li and J. C. Zhou, Design and characterization of Eu²⁺-doped Ba₄Si₆O₁₆ photostimulated phosphor for optical information storage, J. Lumin., 2020, 227, 117556.
16. Y. Gong, Y. H. Wang, Y. Q. Li and X. H. Xu, Ce³⁺, Dy³⁺ Co-doped white-light long-lasting phosphor: Sr₂Al₂SiO₇ through energy transfer, J. Electrochem. Soc., 2010, 157, J208–J211.
17. R. D. Shannon, Revised effective ionic radii and systematic studies of interatomic distances in halides and chalcogenides, Acta Crystallogr., Sect. A: Cryst. Phys., Diffr., Theor. Gen. Crystallogr., 1976, 32, 751–767.
18. R. F. Zhou, F. K. Ma, F. Su, Y. Y. Ou, Z. M. Qi, J. H. Zhang, Y. Huang, P. Dorenbos and H. B. Liang, Site Occupancies, VUV-UV–vis photoluminescence, and X-ray radioluminescence of Eu²⁺-doped RbBaPO₄, Inorg. Chem., 2020, 59, 17421–17429.
19. R. Shi, M. M. Qi, L. X. Ning, F. J. Pan, L. Zhou, W. J. Zhou, Y. C. Huang and H. B. Liang, Combined experimental and ab initio study of site preference of Ce³⁺ in SrAl₂O₄, J. Phys. Chem. C, 2015, 119, 19326–19332.
20. Z. Y. Zhang and Y. H. Wang, UV-VUV excitation luminescence properties of Eu²⁺-doped Ba₂MSi₂O₇ (M = Mg, Zn), J. Electrochem. Soc., 2007, 154, J62–J64.
21. H. Y. Ni, H. B. Liang, C. M. Liu, Q. Su, S. H. Sun and Y. Tao, Luminescence of Eu³⁺ in two different sites of Na₃GdSi₂O₇ and Gd³⁺-Eu³⁺ energy transfer, ECS J. Solid State Sci. Technol., 2012, 1, R27–R31.
22. Y. Y. Ou, W. J. Zhou, F. K. Ma, C. M. Liu, R. F. Zhou, F. Su, Y. Huang, P. Dorenbos and H. B. Liang, Luminescence tuning of Ce³⁺, Pr³⁺ activated (Y, Gd)AGG system by band gap engineering and energy transfer, J. Rare Earths, 2020, 38, 514–522.
23. M. Fasoli, A. Vedda, M. Nikl, C. Jiang, B. P. Uberuaga, D. A. Andersson, K. J. McClellan and C. R. Stanek, Band-gap engineering for removing shallow traps in rare-earth Lu₃Al₅O₁₂ garnet scintillators using Ga³⁺ doping, Phys. Rev. B: Condens. Matter Mater. Phys., 2011, 84, 081102.
24. J. M. Ogiegło, A. Katelnikovas, A. Zych, T. Jüstel, A. Meijerink and C. R. Ronda, Luminescence and luminescence quenching in Gd₃(Ga,Al)₅O₁₂ scintillators doped with Ce³⁺, J. Phys. Chem. A, 2013, 117, 2479–2484.
25. P. Dorenbos, Ce³⁺ 5d-centroid shift and vacuum referred 4f-electron binding energies of all lanthanide impurities in 150 different compounds, J. Lumin., 2013, 135, 93–104.
26. L. T. Lin, R. Shi, R. F. Zhou, Q. Peng, C. M. Liu, Y. Tao, Y. Huang, P. Dorenbos and H. B. Liang, The effect of Sr²⁺ on luminescence of Ce³⁺-doped (Ca, Sr)₂Al₂SiO₇, Inorg. Chem., 2017, 56, 12476–12484.
27. V. P. Dotsenko, S. M. Levshov, I. V. Berezovskaya, G. B. Stryganyuk, A. S. Voloshinovskii and N. P. Efryushina, Luminescent properties of Eu²⁺ and Ce³⁺ ions in strontium litho-silicate Li₂SrSiO₄, J. Lumin., 2011, 131, 310–315.
28. Z. G. Cui, R. G. Ye, D. G. Deng, Y. J. Hua, S. L. Zhao, G. H. Jia, C. X. Li and S. Q. Xu, Eu²⁺/Sm³⁺ ions co-doped white light luminescence SrSiO₃ glass-ceramics phosphor for white LED, J. Alloys Compd., 2011, 509, 3553–3558.
29. H. J. Seo, B. K. Moon, B. J. Kim, J. B. Kim and T. Tsuboi, Optical properties of europium ions in SrB₂O₄ crystal, J. Phys.: Condens. Matter, 1999, 11, 7635–7643.
30. Y. Q. Geng, Q. F. Shi, F. T. You, K. V. Ivanovskikh, I. I. Leonidove, P. Huang, L. Wang, Y. Tian, Y. Huang and C. Cui, Site occupancy and luminescence of Ce³⁺ ions in whitlockite-related strontium lutetium phosphate, Mater. Res. Bull., 2019, 116, 106–110.
31. R. F. Zhou, L. T. Lin, C. M. Liu, P. Dorenbos, Y. Tao, Y. Huang and H. B. Liang, Insight into Eu redox and Pr³⁺ 5d emission in KSrPO₄ by VRBE scheme construction, Dalton Trans., 2018, 47, 306–313.
32. Y. Y. Ou, W. J. Zhou, C. M. Liu, L. T. Lin, M. G. Brik, P. Dorenbos, Y. Tao and H. B. Liang, Vacuum referred binding energy scheme, electron–vibrational interaction, and energy transfer dynamics in BaMg₂Si₂O₇: Ln (Ce³⁺, Eu²⁺) phosphors, J. Phys. Chem. C, 2018, 122, 2959–2967.
33. Y. C. Lin, M. Bettinelli and M. Karlsson, Unraveling the mechanisms of thermal quenching of luminescence in Ce³⁺-doped garnet phosphors, Chem. Mater., 2019, 31, 3851–3862.
34. R. F. Zhou, L. T. Lin, H. T. Zhao, T. T. Deng and J. W. Li, Constructing sensitive luminescent thermometers via energy transfer in Ce³⁺ and Eu²⁺ co-doped Ca₈Mg₃Al₂Si₇O₂₈ phosphors, Mater. Chem. Front., 2021, 5, 6071–6081.
35. P. Dorenbos, Lanthanide 4f-electron binding energies and the nephelauxetic effect in wide band gap compounds, J. Lumin., 2013, 136, 122–129.
36. P. Dorenbos, Improved parameters for the lanthanide 4f^q and 4f^q−15d curves in HRBE and VRBE schemes that takes the nephelauxetic effect into account, J. Lumin., 2020, 222, 117164.
37. W. Xie, W. Jiang, R. F. Zhou, J. H. Li, J. H. Ding, H. Y. Ni, Q. H. Zhang, Q. Tang, J. X. Meng and L. T. Lin, Disorder-induced broadband near-infrared persistent and photostimulated luminescence in Mg₂SnO₄: Cr³⁺, Inorg. Chem., 2021, 60, 2219–2227.
38. F. Ehré, C. Labbé, C. Dufour, W. M. Jadwisienczak, J. Weimmerskirch-Aubatin, X. Portier, J.-L. Doualan, J. Cardin, A. L. Richard, D. C. Ingram, C. Labrugère and F. Gourbilleau, Nitrogen concentration effect on Ce doped SiO_xN_y emission: towards optimized Ce³⁺ for DEL applications, Nanoscale, 2018, 10, 3823–3837.
39. H. Lin, B. Wang, Q. M. Huang, F. Huang, J. Xu, H. Chen, Z. B. Lin, J. W. Wang, T. Hua and Y. S. Wang, Lu₂CaMg₂(Si_1−xGe_x)₃O₁₂: Ce³⁺ solid-solution phosphors: bandgap engineering for blue-light activated afterglow applicable to AC-LED, J. Mater. Chem. C, 2016, 4, 10329–10338.
40. K. V. Eeckhout, A. J. J. Bos, D. Poelman and P. F. Smet, Revealing trap depth distributions in persistent phosphors, Phys. Rev. B: Condens. Matter Mater. Phys., 2013, 87, 045126.
41. A. J. J. Bos, Theory of thermoluminescence, Radiat. Meas., 2007, 41, S45–S56.
42. S. X. Wang, X. L. Liu, B. Y. Qu, Z. Song, Z. Z. Wang, S. Y. Zhang, F. X. Wang, W. T. Geng and Q. L. Liu, Green persistent luminescence and the electronic structure of β-Sialon: Eu²⁺, J. Mater. Chem. C, 2019, 7, 12544–12551.
43. Y. M. Yang, Z. Y. Li, J. Y. Zhang, Y. Lu, S. Q. Guo, Q. Zhao, X. Wang, Z. J. Yong, H. Li, J. P. Ma, Y. Kuroiwa, C. Moriyoshi, L. L. Hu, L. Y. Zhang, L. R. Zheng and H. T. Sun, X-ray-activated long persistent phosphors featuring strong UVC afterglow emissions, Light: Sci. Appl., 2018, 7, 88.

---

Footnote

† Electronic supplementary information (ESI) available: Details of samples preparations and characterizations; final refined structural parameters of Sr₂Ga₂SiO₇ host for Table S1; bond lengths of Sr²⁺–O²⁻ in Sr₂Ga₂SiO₇ sample for Table S2; highest-height normalized synchrotron radiation VUV-UV excitation spectra of Sr_1.98Ce_0.01Na_0.01Ga_2−xAl_xSiO₇ samples for Fig. S1; integral intensity of Ce³⁺ emission in Sr_1.98Ce_0.01Na_0.01Ga_2−xAl_xSiO₇ for Fig. S2; temperature-dependent decay curves of Sr_1.98Ce_0.01Na_0.01Ga_2−xAl_xSiO₇ samples and the lifetimes of Ce³⁺ as a function of temperature for Fig. S3(a–d); VUV-UV excitation spectra of Eu³⁺ in Sr_1.98Eu_0.01Na_0.01Ga_2−xAl_xSiO₇ samples for Fig. S4; estimated trap depths as a function of excitation temperature in Sr_1.98Ce_0.01Na_0.01Al₂SiO₇ for Fig. S5; normalized temperature-dependent PL decay curves of Sr_1.98Ce_0.01Na_0.01Al₂SiO₇ sample for Fig. S6. See DOI: 10.1039/d1qi01152a

---