X-ray-activated, UVA persistent luminescent materials based on Bi-doped SrLaAlO₄ for deep-Seated photodynamic activation

Currently, owing to the efficient storage of excitation energy in energy traps, persistent luminescent (PersL) materials with long-lasting emission after cessation of excitation light have attracted great research interest and have been widely applied in glow-in-the-dark paints, optical information storage, light-guided diagnostic imaging, and therapeutics.^1–4 To date, the commercial blue (CaAl₂O₄:Eu²⁺,Nd³⁺) green (SrAl₂O₄:Eu²⁺,Dy³⁺) and red (Y₂O₂S:Eu³⁺,Mg²⁺,Ti⁴⁺) PersL phosphors have been well developed.⁵ Recently, the near-infrared (NIR) PersL materials, such as Cr-doped ZnGa₂O₄ and LiGa₅O₈, have captured the attention of researchers for applications in biomedical imaging and treatment,^6,7 which can avoid tissue autofluorescence interference and provide high tissue penetration. In stark contrast to the evolving progress in the visible-infrared spectral regions, the research at the other end of the spectrum—the shorter-wavelength UV region—remains very quiet.^2,8

The growing multifunctional applications set a high standard for ultraviolet (UV, 200–400 nm) emitting processes and materials, such as 3D-printing technology, photocatalysis in the dark,⁹ UV bioadhesives in clinical needs,¹⁰ pest-trapping of crops, long-term sterilization,⁵ and photodynamic therapy (PDT) of cancer. Taking into account the potential applications of UV light, UV PersL, as an alternative form of luminescence that can be used as an energy resource in the absence of external excitation, is interesting and significant to be investigated. For instance, in a typical UV light-mediated PDT application, the photosensitizers are activated by UV light and then reacted with oxygen molecules that are present in the cancer tissues, thereby generating singlet oxygen (¹O₂) that is responsible for tumor cell destruction. However, the drawback of conventional UV light-mediated PDT is its inability to treat tumors located deep under the skin due to the short penetration depth of UV light in tissues. Recently, two-photon or upconversion nanoparticle-mediated PDT, which aims to minimize tissue interference and improve penetration depth, has achieved exciting results.¹¹ Unfortunately, NIR sources still suffer from tissue penetration limitations (<15 mm) and an undesired overheating effect.^12,13 To overcome these limitations, some researchers have developed an x-ray-activated scintillator and photosensitizer (LiYF₄:Ce³⁺@ZnO, LaF₃:Tb³⁺@RB)^14,15 nanoparticle-mediated PDT for deep-seated cancer therapy. Nevertheless, high-dose x rays (>5.0 Gy) are usually needed for efficient cancer therapy.^6,16 Such a high dose is comparable or even higher than a fraction of doses used in clinical radiotherapy and inevitably causes damage to normal tissues. In this regard, reducing the x-ray dose is a major concern in x-ray-activated deep-tissue PDT practical applications. The second concern is PDT efficiency. Considering that PersL materials can store excitation energy and then slowly emit light without continuous excitation, it can be used as potential energy mediators for depth-independent and long-acting PDT treatment by making use of a low dose of x ray.^6,16 Inspired by the wide application of x-ray-activated nanoparticle-mediated nanoplatforms for depth-independent therapeutics,^14,17 we speculate that the combination of UV (<400 nm) PersL and suitable photosensitizers, such as g-C₃N₄,¹⁸ can provide a more ideal curing effect by low-dose high-penetrance x-ray excitation, because shorter wavelengths have sufficient energy to initiate a photodynamic reaction.^19,20

Apparently, the ability to achieve excellent UV PersL performance by external stimuli, especially x ray, is at the core of imminent PersL material research and their PDT application. However, a rational design of UV PersL materials remains elusive. Here, we demonstrate that a tunable PersL can be accomplished by Bi³⁺ ion substitution into the lattice of SrLaXO₄ (X = Al, Ga, In) compounds with random cation arrangement as the local incommensurate structure. For many years, these layered perovskite-related compounds have been extensively studied due to their possible applications as substrate materials for high-temperature superconductors as well as host materials for laser media and phosphors, leading to the occurrence of novel emitters in unconventional frameworks.²¹ We intensively illustrate this concept using Bi³⁺-doped SrLaAlO₄ as a model system. A short-range disorder is found in SrLaAlO₄, indicating a predominantly disordered distribution of cation pairs formed by Sr²⁺ and La³⁺ ions with different charges occupying the same crystallographic site. These disordered pairs form different casual configurations and contribute to the polar local features in SrLaAlO₄,²² resulting in a fluctuating crystal field, especially after Bi³⁺ ion doping. The experimental results imply that the UVA PersL of Bi³⁺-doped SrLaAlO₄ derived from a disorder-induced trap center can store excitation energy and then release it slowly to the emitters. To verify the potential PDT application of x-ray-activated UVA PersL SrLaAlO₄:Bi³⁺ nanoparticles, as a proof of concept, these nanoparticles are combined with a matched photosensitizer (g-C₃N₄), and they exhibit excellent ¹O₂ generation. This work offers a protocol for the UV emission-related application and will stimulate the development of novel UV PersL materials. Furthermore, our research also shows that the UVA PersL of SrLaAlO₄:Bi³⁺ can be easily transferred to NIR PersL by codoping Yb³⁺ ions, which has high prospects for real-time diagnosis and therapy.

Bi³⁺ dopants were expected to replace La³⁺ cations in SrLaAlO₄. Accordingly, a series of SrLa_xBi_1−xAlO₄ (0 < x < 2%) samples were synthesized by high temperature solid state reactions. Starting materials, SrCO₃ (A.R.), Al₂O₃ (A.R.), La₂O₃ (99.99%, pre-heated overnight at 1000 °C), and Bi₂O₃ (99.99%) were mixed and ground in an agate mortar under ethanol. The obtained mixtures were calcined for 6 h at 1100 °C and then 10 h at 1380 °C in air. For the synthesis of SrLa_xBi_1−xGaO₄ and SrLa_xBi_1−xInO₄ (0 < x < 2%), Ga₂O₃ (99.99%) and In₂O₃ (99.99%) were used, respectively, while other synthesis parameters were the same as that of SrLa_xBi_1−xAlO₄. Note: All starting materials were purchased from Aladdin-reagent Co., Ltd.

The g-C₃N₄ photosensitizer used in this study was prepared by heating guanidine hydrochloride (A.R., 99%) to 600 °C for 3 h under ambient atmospheric conditions according to the procedure described in the literature.²³ The typical preparation of the SrLaAlO₄:Bi@g-C₃N₄ platform was as follows: first, an appropriate amount of g-C₃N₄ was added into methanol and then the beaker was placed in an ultrasonic bath for 10 h to completely disperse g-C₃N₄. Then, the finely ground SrLaAlO₄:0.3%Bi particles were added into the above solution and stirred in a fume hood for 24 h. Finally, after the volatilization of methanol, the powder was obtained after drying at 100 °C for 24 h. The mass ratios of g-C₃N₄ to SrLaAlO₄:0.3%Bi at 4% were synthesized and named SrLaAlO₄:Bi@g-C₃N₄. Note: To reduce SrLaAlO₄:0.3%Bi dimensions, the sample was mechanically ground, followed by sedimentation, filtration, and centrifugation, to yield <500 nm particles.

The detection of ¹O₂ was measured by monitoring the fluorescence quenching of 1,3-Diphenylisobenzofuran (DPBF). SrLaAlO₄:Bi@g-C₃N₄ (10 mg/ml) was dispersed in anhydrous acetonitrile and was then irradiated using an x-ray light tube (50 kV, 79 μA) for 5 min. DPBF (5 μM, 200 μl) was added into the solution immediately. The fluorescence of DPBF was recorded at different time points.

The x-ray diffraction (XRD) spectrum of samples was recorded on a diffractometer (Rigaku D/MAX 2200 VPC) equipped with Cu K_α radiation (λ = 1.5405 Å) operated with 40 kV and 26 mA radiation. A Varian Cary 500 UV–vis spectrophotometer was used to investigate the light absorption properties of the samples. Photoluminescence excitation (PLE), emission (PL) spectra, and lifetime curves were measured with a high-resolution spectrofluorometer (Edinburgh Instruments FLS1000) equipped with a 450 W xenon lamp. The room-temperature Raman spectrum was recorded using a micro-Raman system (Renishaw InVia, UK) with an excitation wavelength of 514.5 nm. X-ray photoelectron spectroscopy (XPS) spectra were acquired using an ESCALAB 250 photoelectron spectrometer (Thermo Fisher Scientific) at 3.0 × 10⁻¹⁰ mbar with monochromatic Al K_α radiation (E = 1486.2 eV). Binding energies shown in the present work were corrected by the C1 s peak at 284.8 eV. Thermoluminescence (TL) curves were measured on an LTTL-3DS thermoluminescence spectrophotometer (Guangzhou Radiation Technology Co., Ltd.). The samples were irradiated by x ray (or 2 mW LED light at 310 nm) for 5 min before the TL measurement and finally measured at a linear heating rate of 4 K/s in the temperature range of RT to 600 K. The morphology and microstructure of the samples were characterized using a transmission electron microscope (TEM; FEI Tecnai G2 F30), operating at an accelerating voltage of 300 kV. The x-ray photons are generated from a mini-X x-ray tube (Amptek, Inc.). The x-ray tube was set at 50 kV and 79 μA for all the experiments in this study.

To obtain the x-ray-activated PersL phosphor, we synthesize a series of Bi-doped SrLaXO₄ (X = Al, Ga, In) materials. The XRD patterns of SrLaXO₄:0.3%Bi (X = Al, In, Ga) samples are shown in Fig. 1(a). As a representative, the XRD patterns of SrLaAlO₄:Bi samples with different doping concentrations are shown in Fig. S1 in the supplementary material. No diffraction peaks of impurity phases are observed in all samples with the dopant concentrations below 2 at. %. This result proves the formation of a pure SrLaXO₄ phase and the successful doping of Bi³⁺ ions into the host lattice. Although SrLaInO₄ (ICSD No. 169079, C1m1) has a different space group to SrLaAlO₄ (ICSD No. 161180, I4/mmm) and SrLaGaO₄ (ICSD No. 54130, I4/mmm), while only one crystallographic site is available for La³⁺ and Sr²⁺ and the cations (La³⁺, Sr²⁺_, and Bi³⁺) are supposed to be randomly distributed in this site throughout the three structures. Aforesaid SrLaXO₄ compounds, a subgroup of the A₂BO₄-based oxides with K₂NiF₄-type structure, are composed of XO₆ octahedra connected to form a two-dimensional network by corner-shared oxygen ions.²⁴ The structure is built from Al/Ga/InO₆ octahedra and Sr/LaO₉ dodecahedra sketched in Figs. 1(b)–1(e). The Bi³⁺ ions probably take the place of the Sr²⁺/La³⁺ site to form BiO₉ dodecahedra.

Any combination of Bi-doped SrLaXO₄ (X = Al, Ga, In) can produce PersL when activated by an x-ray source (Fig. 2). As shown in Fig. 2(a), the relative persistent emission peaks of SrLaXO₄:Bi are tuned from UV (380 nm) to cyan (460 nm) or yellow (565 nm) light by completely replacing Al³⁺ with Ga³⁺ or In³⁺_. The PersL signals are detected more than 400 min for SrLaAlO₄:Bi and SrLaGaO₄:Bi, and 50 min for SrLaInO₄:Bi after stopping the x-ray irradiation [Fig. 2(b)].

In comparison with x ray, for also evaluating the photoluminescence properties of SrLaXO₄:0.3%Bi³⁺ (X = Al, In, Ga) under UV excitation, the PLE and PL spectra of the samples with various host cation ions are recorded and shown in Fig. S2 in the supplementary material. All the samples show a typical broadband emission corresponding to the ³P₁→ ¹S₀ transition of Bi³⁺ ions. With the increase of the cation radius from Al³⁺, Ga³⁺ to In³⁺, the excitation/emission peak red-shifted from 302/380 nm to 308/460 nm and then to 312/565 nm, respectively. The energy position and the ¹S₀↔ ³P₁ optical transition are profoundly influenced by the covalence of Bi–O bonding in a solid (nephelauxetic effect). Increasing covalence results in a shift of this transition to lower energy.²⁵ Besides, the UV-activated PersL emissions of all three samples are detected (Fig. S3 in the supplementary material). As for SrLaAlO₄:0.3%Bi³⁺, the profiles of PersL spectra are almost identical for both x-ray and UV light excitation, indicating that the UVA PersL originates from the Bi³⁺ emitters and exists in the whole photoluminescence emission band. Interestingly, the x-ray-activated PersL decay performance seems to be significantly improved compared with that of UV-light excitation. The PersL signals (over one order of magnitude stronger than the background signal) can be detected only over ∼5 min for SrLaGaO₄:Bi³⁺ and SrLaInO₄:Bi³⁺, and ∼60 min for SrLaAlO₄:Bi³⁺ after stopping the UV irradiation (Fig. S3 in the supplementary material). Although it is difficult to assess the irradiation dose of x-ray and UV excitation light source, the PersL duration of SrLaGaO₄:0.3%Bi³⁺ is immediately increased by ∼100 times when changing the excitation light source from UV light to x ray [Fig. 2(b), Fig. S3 in the supplementary material]. The PersL durations of both SrLaAlO₄:0.3%Bi³⁺ and SrLaInO₄:0.3%Bi³⁺ after x-ray irradiation are also increased by ∼10 times longer than those under the UV light irradiation. These results suggest that x ray is a different excitation mode to UV light and probably induces remarkable PersL enhancement in a specific matrix material. The details will be discussed in the TL results below.

To further explore the relationship between the PersL performance and structure, we take SrLaAlO₄:Bi³⁺ as the following research object, which has a representative structure and excellent UVA PersL performance as a potential bioprobe under the x-ray irradiation. Next, the PL curves, PersL curves, lifetime curves, absorption spectra, Raman spectra, XPS spectra, and TL spectra are systematically recorded to reveal the role of Bi³⁺ ions in influencing the local structure of SrLaAlO₄:Bi³⁺.

The optimal doping concentration of Bi³⁺ for PL is determined to be about 1% (Fig. S4 in the supplementary material), which is much higher than that for PersL (0.3%), as shown in Fig. S5 in the supplementary material. As is known, the PL intensity is proportional to the excitation cross section and the content of optically active centers. Hence, the cross relaxation between neighboring Bi³⁺ ions substantially quenches the PersL center when the concentration exceeds 0.3%, which is consistent with the lifetime result (Fig. S6 in the supplementary material).

The UV–visible diffuse reflectance spectra were recorded and converted into the absorption spectra using the Kubelka–Munk function for the pure and Bi-doped SrLaAlO₄ powders [Fig. 3(a)]. The absorption spectra of SrLaAlO₄:Bi³⁺ reveal a broad strong UV absorption extending from the absorption edge of the SrLaAlO₄ host centered at about 296 nm beside a broad absorption tail ranging from 350 to 700 nm. The absorption band at 296 nm originating from the ¹S₀–³P₁ electron transition of the Bi³⁺ ion is increased with Bi³⁺ doping content, indicating that this absorption is related to the Bi impurity in SrLaAlO₄. The peaks of 4.20 and 5.00 eV (the small band that is merged into the absorption edge of the host) are attributed to the ¹S₀–³P₁ transition in Bi³⁺.²⁶ The XPS results further prove that Bi is present in +3 [Fig. 3(b), Fig. S7 in the supplementary material]. The bandgap values (Eg) are obtained based on the Tauc plot as the intercept value of the plot of (αhν)² against light energy hν, as shown in Fig. 3(a). The bandgap of the undoped SrLaAlO₄ crystal is estimated to be about 5.15 eV. The incorporation of Bi³⁺ shifts the absorption edge to a lower energy, by about 0.8 eV for 2% Bi³⁺ doping. As shown in the curve covered by the shaded portion, the structure disorder-induced band-tail (Urbach tail) characteristic is clearly observed. Therefore, there is reason to believe that the absorption edge properties discussed here are a result of a fundamental limitation of structure disordering.^27–29 When we examine the preparation process of samples, it is found that Bi³⁺ doping at a high concentration (>3%) strongly leads to component segregation, resulting in the generation of an impure phase. Combining the above results indicates that Bi³⁺ doping can cause local incommensurate phase deviation, which is further accompanied by local structure disorders.

Raman spectroscopy affords convenient probes of local rather than long-range structures, and it can even determine the structural disorder in some cases.^30,31 Here, we use Raman spectroscopy to examine structural changes after doping Bi³⁺ into SrLaAlO₄ [Fig. 4(a)]. According to the factor group analysis, there are four Raman-active modes with two A_1g and two E_g symmetries for a K₂NiF₄-type structure with the I4/mmm space group. Two Raman bands observed at 129 and 221 cm⁻¹ are assigned to the Eg (Sr/La)–O1 bond in the ab plane and the bridge (Sr/La)-O1 bond along the c axis in A_1g symmetry, respectively [Fig. 1(c)]. So, we can conclude that these two bands are due to the two internal vibrations of the “Sr/LaO₉” dodecahedron. The other two bands at 300 and 440 cm⁻¹ are assigned to the vibrations of the (Sr/La)-O2-(Sr/La) oxygen bond in the Eg symmetry and the bridge (Sr/La)-O2-Al oxygen bond along the c axis in A_1g symmetry, respectively. The band at 583 cm⁻¹ is not predicted by the factor group analysis, and this could be due to a local distortion of the AlO₆ octahedron.^32,33 SrLaAlO₄ doped with different Bi³⁺ concentrations demonstrate an almost identical Raman spectrum relative to undoped SrLaAlO₄. Beyond this, enlarged Raman spectra lead to an occurrence of visible differences. It is found that the ratios of the two main peaks (I₂₂₁/I₃₀₀) decrease stepwise with an increasing Bi³⁺doping concentration [Fig. 4(b), Table S1 in the supplementary material], which can be explained by the expansion of (Sr/La)-O1 along the c axis and attenuation of the band in the Sr/LaO plane. The above results demonstrate that the substitution of Bi³⁺ will induce internal stress, which leads to microstructural inhomogeneities and distortions.

As is well known, radiation-induced PersL performance is closely related to the gradual thermally stimulated release of charge carriers from traps. At this stage, it is necessary to reveal the reason for the intense PersL of SrLaAlO₄: Bi³⁺ and the role of the Bi³⁺ dopant. Figure 5(a) shows the TL curves of SrLaAlO₄: Bi³⁺ after the removal of the x-ray excitation source. As can be seen, all SrLaAlO₄:Bi³⁺ samples show one main TL band observed at ∼364 K, and the strongest TL intensity is reflected by the 0.3%Bi-doped sample. Moreover, the Bi-doped samples show almost identical features under either the x-ray or UV light irradiation, except for a small additional bulge in the high-temperature region under x-ray excitation (Fig. S8 in the supplementary material), from which we can intuitively find that the electrons in deep traps are created after irradiation by high energy x ray. As expected, compared with UV excitation, the stronger PersL under x-ray excitation is due to the effective filling of deep traps.

Figure 5(b) gives the results of TL fading experiments of the SrLaAlO₄:0.3%Bi³⁺ sample. With the increase in the decay time, the TL intensity gradually decreases and is accompanied by a simultaneous right shift of the TL peak to the high-temperature region. It also demonstrates that the assumption of a continuous trap distribution rather than one or more straightforward discrete traps is reasonable. To simplify the calculation of the trap depth, the TL glow curve can be roughly deconvoluted into two Gaussian peaks centered at 350 and 385 K, respectively (Fig. S9 in the supplementary material). Deep traps caused by x-ray excitation at high temperatures are not considered. Therefore, to evaluate the trap depth of the dominant defects induced by Bi³⁺ and the kinetics of TL, the TL spectra of SrLaAlO₄ at 350 to 385 K are theoretically fitted by a general-order TL kinetic expression³⁴ 

where n₀ is the concentration of trapped charges at t = 0, k is Boltzmann's constant, β is the heating rate, E is the activation energy (which means the trap depth), S is the frequency factor, and b is the order of kinetics. The calculated trap depths of SrLaAlO₄:0.3%Bi³⁺ are 0.91 and 0.96 eV, respectively.

As the multiple Sr/La cations with different charges occupy the same crystallographical position, these cations are not in complete order in the local structure. Moreover, the hybridization between Bi 6 s and oxygen 2p states stabilizes the off-center displacement of Bi³⁺ cations and produces disorder, and the fractional aliovalent substitution of Bi³⁺ for Sr^{2 +}/La³⁺ may further increase the number of defects, which has been confirmed by Raman and absorption results. All these results give information about the relationship between PersL and the local disorder structure resulting from Bi³⁺ doping. In summary, we speculate that the intense x-ray-activated UVA PersL resulted from the short-range disorders caused by Bi³⁺ doping.

Based on the above results, disorder-induced traps are demonstrated to be easily formed in ABCO₄ compounds. We envisage that the PersL phenomenon should not be limited to SrLaXO₄:Bi³⁺ (X = Al, Ga, In) as the change of C site ions in ABCO₄ compounds. When changing A/B site ions in ABCO₄, the obtained compounds, such as SrGdAlO₄:Bi, CaGdAlO₄:Bi, and CaYAlO₄:Bi, also exhibit UVA PersL after x-ray irradiation (Fig. S10 in the supplementary material). Analogously, PersL can also be realized by changing the doped ions in the ABCO₄ compounds, e.g., Tb³⁺ (SrLaGaO₄:Tb)³⁵ and Eu³⁺ (SrLaAlO₄:Eu, Fig. S11 in the supplementary material).

Based on the experimental design and test results presented above, the x-ray (or UV light) induced PersL mechanisms of SrLaAlO₄:Bi³⁺ are concluded and shown in Fig. 6(a). Under x-ray excitation, the incident photons create electron–hole pairs via x-ray fluorescence, Auger, and electron–electron inelastic scattering. The majority of excited electrons is subsequently captured by traps through the conduction band (CB); however, as the heaviest stable element in the periodic table, Bi³⁺ ions can strongly absorb x-ray photons. Besides, the Bi³⁺ ions are the primary absorption center and have energy interaction with traps, which can be proved by the PersL excitation spectrum (Fig. S12 in the supplementary material). Therefore, it is possible that a section of electrons generated after the absorption of the x rays is captured near the Bi³⁺ ions. The captured process is to occur via tunneling between the Bi³⁺ ions and the proximity disorder-induced traps in the crystalline matrix [Fig. 6(b)]. Finally, the de-trapping process occurs, and these captured electrons are gradually released from the disorder-induced traps and return to the Bi³⁺ emitters via both tunneling and the conduction band, followed by the recombination and emission of UVA PersL.

Additionally, it is found that the UVA PersL phenomenon of the Bi-doped SrLaAlO₄ structure can be transferred to NIR PersL. For instance, after co-doping with Yb³⁺, the PersL emission bands of Yb³⁺ (978 nm) in SrLaAlO₄:Bi³⁺,Yb³⁺ are observed (Fig. S13 in the supplementary material), which are derived from energy transfer between Bi³⁺ and Yb³⁺. The x-ray-activated PersL materials with simultaneous UV and NIR emission will probably find applications in multifunctional therapeutic systems.

Considering the excellent UVA PersL performance, SrLaAlO₄:Bi³⁺ is strongly expected as the medium for PDT. Hence, we constructed the UVA PersL-mediated PDT platform by a conjugation of the SrLaAlO₄:Bi³⁺ nanoparticles with g-C₃N₄ [Figs. 7(a) and 7(b)]. First, g-C₃N₄, an efficient photosensitizer that has a good absorption spectral overlap with the PersL spectrum of SrLaAlO₄:Bi³⁺ [Fig. 7(c)],³⁶ was coupled to the surface of SrLaAlO₄:Bi to form the UVA PersL-mediated PDT platform—SrLaAlO₄:Bi@g-C₃N₄. In addition, the phase identification of as-obtained g-C₃N₄ was checked by XRD, as shown in Fig. S14 in the supplementary material. Then, the analyses of size and morphology of as-obtained SrLaAlO₄:Bi@g-C₃N₄ were performed by TEM, as shown in Fig. 7(a), Fig. S15 in the supplementary material. The SrLaAlO₄:Bi³⁺ nanoparticles with a small size (<500 nm) were successfully loaded onto g-C₃N₄ after being finely ground and after ultrasonic dispersion. Thus, we hypothesize that SrLaAlO₄:Bi was able to relay energy in the form of UV photons to g-C₃N₄ under x-ray irradiation and induced g-C₃N₄ to produce ¹O₂, as shown in Fig. 7(b). As is well known, cytotoxic intracellular ¹O₂ can cause damage to DNA, mitochondria, and plasma membranes of live cells, resulting in cell death. The ability to generate extracellular and intracellular ¹O₂ of the photosensitizer is one of the crucial factors that determine the efficacy of PDT.^16,20

Therefore, as a proof of concept, the ¹O₂ generation by the SrLaAlO₄:Bi@g-C₃N₄ composite under x-ray excitation was tested by employing DPBF, which is a commercial probe for the detection of ¹O₂ production by measuring the fluorescence change. In the presence of ¹O₂, DPBF was oxidized and accompanied by the quenching of its fluorescence at 455 nm. Using DPBF, we studied the ¹O₂ generation of a SrLaAlO₄@g-C₃N₄ acetonitrile solution (10 mg/mL) after x-ray irradiation (50 kV, 79 μA, 5 min) [Fig. 7(d)]. The dose of x-ray irradiation here is estimated to be ∼10 Gy.⁴ The fluorescence of DPBF decreased by almost 100% percent during the 15 min of the observation period, suggesting a significant ¹O₂ generation of the SrLaAlO₄:Bi@g-C₃N₄ composite. Similar tests were performed on acetonitrile solutions (control), SrLaAlO₄:Bi³⁺ nanoparticles, and g-C₃N₄, respectively, all of which showed a smaller decrease (∼0, ∼10%, and ∼30%, respectively) of fluorescence after x-ray irradiation [Fig. 7(e)]. ¹O₂ generation test results indicated that SrLaAlO₄:Bi³⁺ has a potential application in PDT. Besides, the penetration of the x-ray and UV light was measured on the representative SrLaAlO₄:0.3%Bi samples using 6 mm of raw pork tissue to simulate deep-seated tumors [Figs. 7(f)–7(h)]. The optical signal was declined by 68% after ceasing x-ray charging, whereas the signal completely disappeared after UV light excitation. Our results proved that ¹O₂ can be mass-produced in the presence of all the three components (SrLaAlO₄, g-C₃N₄, and x-ray), corroborating that the x-ray-activated SrLaAlO₄:Bi-mediated PDT platform is a potential technique for tumor therapy in deep tissue.

In summary, a series of tunable x-ray-activated PersL materials with the chemical formula ABCO₄:Bi have been successfully designed by utilizing the special valence difference and the occupation rules of cation sites. Comprehensive test results demonstrate that creating a disorder defect in the ABCO₄ type structure is a promising strategy for the development of large quantities of x-ray-activated PersL materials, which can cover the broad wavelength ranging from UV to NIR. As a representative, Bi³⁺ doped SrLaAlO₄ UVA PersL phosphors are investigated in all directions, which have exhibited some potential biological applications in x-ray inducible PDT with the help of g-C₃N₄. Systematic characterization methods indicate that doping with Bi³⁺ ions can lead to a short-range disorder structure, simultaneously induce the formation of trap centers in the crystalline matrix, and then generate intense UV PersL after ceasing the excitation of x ray. Hopefully, these findings will pave the way for designing novel x-ray-activated PersL materials for potential applications.

See the supplementary material for XRD data, PL and PLE profiles, lifetime curves, XPS analysis, PersL excitation spectrum, TEM images, and Raman results of SrLaAlO₄:Bi, and also luminescence property and of other ABCO₄-type compositions.

This work was financially supported by the National Science Foundation of China (NSFC) (Nos. 51772336, 51702373, 81871417, and 51961145101), the Fundamental Research Funds for the Central Universities (No. 19ykpy19), the Joint Funds of the National Natural Science Foundation of China and Yunnan Province (U1902222), the Natural Science Foundation of Guangdong Province (No. 2018A030313919), Guangdong Science & Technology Project (No. 2017A020215024), Key-Area Research and Development Program of GuangDong Province (No. 2019B010926001), and Guangzhou Science & Technology Project (Nos. 202007020005, 201807010104 and 201802020033).

The data that support the findings of this study are available from the corresponding author upon reasonable request.