A universal strategy toward two-component organic-inorganic metal halide luminescent glasses and glass-crystal composites

Abstract

The development of melt-quenched organic-inorganic metal halide (OIMH) glasses is hampered by the scarcity of suitable organic molten salts and low luminescence efficiency. Herein, we developed a series of two-component OIMH amorphous glasses consisting of (TPG)₂MnBr₄ (TPG⁺, triphenylguanidium) and A₂MnBr₄ (A, organic molten cation), named α_G(A_xTPG_y). The high glass-formation ability (GFA) in (TPG)₂MnBr₄ provides a platform to modulate the crystallization of another molten A₂MnBr₄ by homogeneous melting. Moreover, the GFA modulation allows controlled in situ crystallization of α_G(A_xTPG_y) and the formation of transparent glass-crystal composites with higher luminescence efficiency. For instance, the light yield of α_G(PTP₉₉TPG₁) (PTP⁺, propyltriphenylphosphonium) is improved from 18,800 to 35,140 photons per mega–electron volt after annealing at 55°C, showing huge application potentials in radiation detection and high-resolution x-ray imaging. The present research would inspire further exploration of high-performance OIMH glasses and facilitate multiple applications in advanced photonics such as scintillators, photoconductive fibers, light-emitting diodes, and laser crystals.

A universal co-melting strategy broadens the exploration of highly luminescent organic-inorganic metal halide glasses.

INTRODUCTION

Scintillation-based x-ray detectors with high sensitivity and efficiency play an important role in various fields such as medical diagnosis, security screening, and non-destructive inspection (1–3). Currently, most commercial radiation detectors are based on all-inorganic materials including Bi₄Ge₃O₁₂ (BGO), cerium-doped Lu₃Al₅O₁₂ (LuAG: Ce), and thallium-activated CsI (CsI: TI), which have bright scintillation and high radiation hardness (4, 5). However, their sustainability is hampered by energy-intensive production processes and the high demand for toxic or rare-earth metals. Toward these issues, organic-inorganic hybrid metal halides (OIMHs) are considered a promising category of scintillators (6–8). The presence of metal-halogen polyhedrons (e.g., Sb-Cl, Mn-Br, and Cu-I) gives them strong x-ray absorption capabilities. The versatility in assembling organic and inorganic modules allows the physical and chemical properties to be tailored to practical requirements. Specifically, the eco-friendly Mn(II)-based OIMHs have been consistently reported as high-performance scintillators (9). For instance, (C₃₈H₃₄P₂)MnBr₄, (C₂₄H₂₀P)₂MnBr₄, and (C₈H₂₀N)₂MnBr₄ crystals all exhibit high photoluminescence quantum yields (PLQYs) of >90% and excellent light yields of >30,000 photons per mega–electron volt (MeV⁻¹) when exposed to x-ray irradiation (10–12).

The fabrication of large-area scintillation screens is a crucial step in practical applications (13). The prevailing approach for OIMH scintillation screens is to incorporate crystalline powders into organic polymers, whose thickness must be on the order of a few hundred micrometers or even millimeters to ensure sufficient radiation attenuation and scintillation brightness (14, 15). However, such kind of scintillation screens often exhibits poor imaging spatial resolutions below 5 line pairs per millimeter (lp mm⁻¹) due to insufficient phase compatibility, particle agglomeration, or uneven crystal distribution (15–17). Large-area transparent single crystals are another option to achieve high-quality x-ray imaging. Some centimeter-sized OIMH single crystals, (C₈H₂₀N)₂MnBr₄ and (C₇H₁₀N₂)₂MnBr₄, have shown higher spatial resolutions than 25 lp mm⁻¹, surpassing the commercial standard (2 to 8 lp mm⁻¹) (18, 19). Although achieving effective light management, the growth of large-size transparent single crystals remains a challenge (20). The discovery of melt-quenched OIMH glass allows the mass production of large-area transparent scintillation screens (21). For now, Mn(II)-based OIMH glasses have achieved high spatial resolutions of 6 to 18 lp mm⁻¹, demonstrating promising potentials in the scintillation field (22–25). However, compared to the tens of thousands of single crystals published in the Cambridge Crystallographic Data Center (CCDC), glassy OIMH scintillators are rare and confined to triphenylphosphonium-based and guanidinium-based materials. The presence of rigid organic cations with larger steric hindrance can effectively suppress crystallization in supercooled melts, but further development of OIMH glasses is hampered by the limitation of suitable organic molten salts, which act as the organic cation in OIMHs (22, 26). Therefore, it is important to develop a straightforward, efficient, and universal method for the preparation of OIMH glasses.

The challenge of preparing OIMH glasses can be divided into two distinct areas. The first is to achieve a stable melt, and this can be solved by using a low melting temperature (T_m) organic salt to construct OIMHs. The second is to ensure a successful phase transition from melt to glass, whereas most OIMH melts tend to crystallize upon quenching, hindering the expansion of novel OIMH glasses (27, 28). Previous studies on amorphous alloys have provided a suitable strategy to improve crystallization resistance of supercooled metal melts, that is, the introduction of a second component (or multiple components) (29, 30). The crystallization in the melt can be modulated by regulating the structural “order” or “entropy,” which helps to maintain the long-range disorder upon quenching. The “Confusion principle” proposed by Greer (31) also suggests that the melt comprising multiple metals tends to be highly disordered, facilitating the formation of amorphous alloys with a large mixing entropy. Recently, Bennett et al. first synthesized the zeolitic imidazolate framework (ZIF) glass comprising ZIF-4 and ZIF-62. The two-component ZIF glass has a single glass transition temperature (T_g) of 306°C that falls between the T_g values of ZIF-4 and ZIF-62, indicating the homogeneous melting of parent materials by melt-quenching (32). Also, another two-component ZIF glass comprising ZIF-8 and ZIF-62 was prepared by flux melting (33). It is conceivable that if this strategy to modulate crystallization is applied to the OIMH materials, it would offer rich opportunities to tailor their physical and chemical properties, thereby stimulating the potential applications in advanced photonics (34).

In this work, we developed a universal strategy toward two-component OIMH glass, which can expand the material systems of OIMH glasses. Specifically, the (TPG)₂MnBr₄ (TPG⁺, triphenylguanidium) with a high T_g/T_m ratio of 0.80 and a large viscosity at T_m of 1820 mPa s was firstly synthesized, showing the good glass-forming ability (GFA) and being suitable to modulate the crystallization of two-component OIMH melts. To illustrate the universality and scalability of this strategy, a series of A₂MnBr₄ crystals (A represents the organic molten cation, including triphenylphosphonium, guanidinium, imidazolium, pyridinium, and quaternary phosphonium/ammonium) were co-melted with the (TPG)₂MnBr₄, respectively, resulting in the formation of stable two-component amorphous glasses, named α_G(A_xTPG_y). Systematic analysis including thermodynamics, rheology, and structure confirmed the homogeneous melting and structural characteristics of two-component glasses. Meanwhile, the tunable GFA enables the amorphous α_G(A_xTPG_y) glass to be annealed into a transparent glass-crystal composite, named α_G–C(A_xTPG_y), with an improved luminescence efficiency. As an example, the light yield of α_G–C(PTP₉₉TPG₁) (PTP⁺, propyltriphenylphosphonium) is increased from 18,800 to 35,140 photons MeV⁻¹ after being annealed at 55°C for 1 day, as well as a low limit of detection (LoD) of 43.6 nGy s⁻¹.

RESULTS

Synthesis of high-GFA (TPG)₂MnBr₄ glass

The (TPG)₂MnBr₄ single crystals were grown by dissolving TPG·HBr and MnBr₂·4H₂O in ethanol at a molar ratio of 2:1 and then evaporating the solvent at 80°C for 2 days. As illustrated in Fig. 1A and table S1, the (TPG)₂MnBr₄ crystallizes in the monoclinic space group of C2/c, in which isolated [MnBr₄]²⁻ tetrahedrons are surrounded by TPG⁺ cations, forming a zero-dimensional crystal structure. Thermogravimetric analysis (TGA) gives the decomposition temperature (T_d) of 241°C. Differential scanning calorimetry (DSC) proves the stable melting with a T_m of 158.5°C, and the glass formation with a T_g of 72.6°C through melt-quenching (Fig. 1B). The structural difference between crystalline and glassy states was declared through powder x-ray diffraction (PXRD), where the pattern of long-range disordered glass shows a diffuse feature (Fig. 1C). Meanwhile, crystallographic measurements including single-crystal XRD (SCXRD) and PXRD were performed to support the successful synthesis of TPG·HBr (fig. S1). The T_g/T_m ratio (assessing the GFA) of (TPG)₂MnBr₄ is 0.80, which is larger than those of reported Mn-based OIMH glasses, indicating the strong resistance to crystallization (Fig. 1D and table S2) (21–24, 35, 36). Rheological properties including a large viscosity of 1820 mPa s and a high flow activation energy (E_a) of 83.6 kJ mol⁻¹ further suggest its remarkable GFA (fig. S2). As illustrated in fig. S3, the glassy (TPG)₂MnBr₄ remains amorphous without crystallization even after being annealed at temperatures of 80° to 140°C (between T_g and T_m) for 100 hours. The excellent GFA of (TPG)₂MnBr₄ glass makes it an ideal parental material for preparing two-component OIMH glass, which can be used to modulate the crystallization of another parent OIMH material during annealing. Meanwhile, the identical Fourier transform infrared (FTIR) spectra between (TPG)₂MnBr₄ crystal and glass indicate the complete retention of TPG⁺ cations upon melt-quenching (fig. S4). The local electronic and coordination features around Mn ions were studied using x-ray absorption fine structure (XAFS) analysis. The identical normalized x-ray absorption near edge structure (XANES) spectra at Mn K-edge show similar local structures around Mn ions for (TPG)₂MnBr₄ crystal and glass. Compared with the Mn foil, their absorption edges shift to higher energy regions and approach the MnO, indicating that the Mn valence state in the glassy state is still +2 (fig. S5). Also, as shown in Fig. 1E, the extended XAFS (EXAFS) data were fitted by Fourier transform in Artemis, in which the main peak is attributed to the local coordination environment of one Mn ion to four Br ions (37). The detailed fitting results are summarized in table S3. The presence of a small amount of Mn-O coordination is probably ascribed to the slight hygroscopicity of some powder samples (38).

Fig. 1. Structural, thermodynamic, and photophysical properties of (TPG)₂MnBr₄ crystal and glass.

(A) The single crystal structure of (TPG)₂MnBr₄. (B) TGA and DSC plots of (TPG)₂MnBr₄ crystal. Inset images are photographs of (TPG)₂MnBr₄ crystal and glass under natural light and ultraviolet light excitation. (C) PXRD patterns of (TPG)₂MnBr₄ crystal and glass and the simulated one derived from SCXRD result. (D) T_g/T_m ratios of (TPG)₂MnBr₄ and other reported Mn-based OIMH glasses [P₄₄₄₄⁺, tetrahexylphosphonium; MTP⁺, methyltriphenylphosphonium; ETP⁺, ethyltriphenylphosphonium; BuTP⁺, (n-butyl)triphenylphosphonium; HTPP⁺, hexyltriphenylphosphonium; BTP⁺, benzyltriphenylphosphonium; DPG⁺, N,N′-diphenylguanidinium; DOTG⁺, 1,3-di-o-tolylguanidine]. (E) Fitting Fourier-transformed EXAFS spectra at Mn K-edge of (TPG)₂MnBr₄ crystal and glass. (F) PLE and PL spectra of (TPG)₂MnBr₄ crystal and glass. a.u., arbitrary units.

Crystalline and glassy (TPG)₂MnBr₄ both exhibit the green emission and photoluminescent (PL) decay lifetimes of hundreds of microseconds under 365-nm excitation, originating from the ⁴T₁(G)–⁶A₁(S) transition of tetra-coordinated Mn²⁺ ions (Fig. 1F and fig. S6A) (39). However, because of the long-range disorder feature, there is severe nonradiative recombination in the (TPG)₂MnBr₄ glass compared with the crystal, resulting in an emission redshift to 540 nm, a shortened decay lifetime to 193.19 μs, and a photoluminescence quantum yield (PLQY) of 5.1% (fig. S6B). Temperature-dependent PL spectra showed the emission behaviors during a whole melt-quenching process. The complete melting of (TPG)₂MnBr₄ crystals occurs within the temperature range of 420 to 440 K, resulting in severe PL quenching. Then, the quenched glass shows a weaker PL intensity compared with the crystalline state due to the higher disorder degree (fig. S7). In addition, the (TPG)₂MnBr₄ glass exhibits green emission under x-ray irradiation with a light yield of 6400 photons MeV⁻¹ and an LoD of 247.1 nGy s⁻¹ (fig. S8). Combined with high transmittance of up to 82% and a strong attenuation efficiency of 92.15% (1 mm thick), the (TPG)₂MnBr₄ glass displays a high x-ray imaging spatial resolution of 18 lp mm⁻¹ and huge potential for practical scintillation application (figs. S9 and S10).

Preparation of single-component OIMH glasses

Despite the important application of glassy materials in various scenarios, their exploration is deficient compared with the crystalline materials. Considering that most reported OIMHs decompose at 250° to 400°C, halogenated ionic liquids with low T_m are a good choice. On the basis of the above conditions, as depicted in fig. S11, five classes of low-T_m organic salts (triphenylphosphonium, guanidinium, imidazolium, pyridinium, and quaternary phosphonium/ammonium) were chosen to synthesize A₂MnBr₄ single crystals. Detail preparation conditions and crystallographic information are summarized in tables S4 and S5 (22, 24, 35, 40, 41). As shown in Fig. 2A and table S6, these crystals exhibit a T_m of <200°C and a ∆T of >98°C (ΔT = T_d − T_m), thus achieving a stable melt (TGA and DSC results are illustrated in fig. S12). Also, all of them can be synthesized by the solid-phase melting method, which avoids the use of organic solvents and has the advantages of economy, environmental friendliness, and being suitable for mass production. Specifically, the brominated organic salt and MnBr₂·4H₂O are mixed in a 2:1 molar ratio and then heated in a muffle furnace at 200°C for 10 min. The crystals are grown by cooling and stirring the melt. As shown in fig. S13, the identical PXRD pattern with the simulated one supports the phase purity.

Fig. 2. Preparation, structure, and thermodynamics of single-component and two-component OIMH glass.

(A) T_d and T_m values of single-component A₂MnBr₄ materials. (B) Preparation process and photographs for two-component OIMH glasses. (C) PXRD patterns of two-component OIMH glasses. (D) The second up-scan heating DSC curves of two-component OIMH glasses.

Notably, most of these melts tend to crystallize upon quenching, while only the (MTP)₂MnBr₄, (PTP)₂MnBr₄^, and (DPG)₂MnBr₄ melts could form amorphous glasses with diffuse PXRD patterns (fig. S14). The difference in glass formation can be attributed to the presence of organic cations with large rigid groups, leading to an elevation in melting enthalpy and viscosity, as well as hindering diffusion and rearrangement of atoms in supercooled melts (42). To further investigate the relationship between structure and glass formation, we calculated the radius of guanidinium cations including TMG⁺ (4.307 Å), DPG⁺ (6.575 Å), and TPG⁺ (7.148 Å), which increases with the number of benzene groups (43, 44). The larger DPG⁺ and TPG⁺ favor the formation of OIMH glasses, while the molten (TMG)₂MnBr₄ crystallizes during cooling (fig. S15, A to C). Meanwhile, although the radius of TBP⁺ (6.709 Å) is larger than that of MTP⁺ (5.902 Å) and PTP⁺ (6.042 Å), it is prone to crystallize during quenching because of the lack of a rigid group, which results in a T_g below the room temperature (fig. S15, D to F). The DSC curves confirm the above phenomenon (fig. S16). The T_g values of these glasses are affected by the rigidity of organic cations, where (MTP)₂MnBr₄ and (TPG)₂MnBr₄ glasses have a high T_g value of 57.7° and 72.6°C, respectively. A relatively high T_g allows glassy OIMH to be processed and used over a wider temperature interval. Nevertheless, the synthesis and modification of suitable low-T_m rigid organic salts involve complex organic reactions and experimental trials, rendering it challenging to screen a suitable candidate for preparing novel OIMH glasses with optimal T_d, T_m, and T_g.

Preparation of two-component OIMH glasses

The preparation of two-component OIMH glasses is considered a potential solution to the above challenge. The introduction of high-GFA (TPG)₂MnBr₄ melt can effectively regulate the crystallization tendency of A₂MnBr₄ melt, thereby forming a stable two-component OIMH glass. To demonstrate the universality of this method, as shown in Fig. 2B, all A₂MnBr₄ crystals were co-melted with (TPG)₂MnBr₄ at 220°C at a mass ratio of 1:1, respectively, and then cooled to room temperature to obtain transparent two-component OIMH glasses, named α_G(A_xTPG_y) (x:y is the mass ratio of A₂MnBr₄ and (TPG)₂MnBr₄). PXRD patterns verify their amorphous state and further demonstrate the feasibility of this method (Fig. 2C). The DSC plots in Fig. 2D exhibit that all amorphous α_G(A₁TPG₁) have a singlet T_g located between A₂MnBr₄ and (TPG)₂MnBr₄, suggesting that the parental materials underwent a homogeneous melting process, instead of forming the separate domains (32). Such kind of phenomena is also observed in the fields of metallic glass, organic polymer, or metal-organic framework glasses (45–47). Notably, the different thermodynamic performances between A₂MnBr₄ and α_G(A₁TPG₁) highlight the GFA modulation effect of high-GFA (TPG)₂MnBr₄ on the two-component OIMH glass. As shown in Fig. 3A, the increased T_g value of α_G(A₁TPG₁) affords an increased T_g/T_m ratio and enhanced GFA compared with single-component A₂MnBr₄, thereby forming the amorphous glass. The T_g values of low-GFA A₂MnBr₄ are predicted in table S7 using a simplified Fox model

$$T_{\mathrm{g}}={\mathrm{\phi }}_1{T}_{\mathrm{g},1}+{\mathrm{\phi }}_2{T}_{\mathrm{g},2}$$

where φ₁, φ₂ is the weight fraction of the parent materials. It can be used to predict the vitrification tendency of binary polymer mixtures, especially for those mixtures where the T_g is linearly dependent on the composition (45, 48). Notably, although there are currently no reported examples demonstrating the applicability of the Fox model to OIMH materials, the simulated T_g values of A₂MnBr₄ using the Fox model (fig. S17) are very close to the DSC test results (an error is less than 5%).

Fig. 3. Thermodynamic, rheological, and structural analysis of two-component OIMH glasses.

(A) T_g/T_m ratios of A₂MnBr₄ and α_G(A₁TPG₁). (B) Viscosities and flow activation energies of A₂MnBr₄ and α_G(A₁TPG₁). (C) Fitting Fourier-transformed Mn K-edge EXAFS spectra of α_G(PTP₁TPG₁) glass. (D) PDFs for two-component α_G(PTP₁TPG₁) and α_G(PTP₉TPG₁) glasses.

Viscosity (η), as an indicator describing the migration ability of liquids, is closely related to the intermolecular forces and microstructure of the melt. According to the classical nucleation theory, the nucleation and growth of crystals are inversely proportional to the viscosity, where migration of atoms in high-viscosity melts is effectively inhibited and benefits the formation of glass. The η(T_m) value of molten A₂MnBr₄ markedly increases after co-melting with (TPG)₂MnBr₄, indicating stronger resistance to crystallization (Fig. 3B). At the same time, the η values of all A₂MnBr₄ are close to that of (TPG)₂MnBr₄ at the preparation temperature (220°C), which improves the flow and diffusion of melts and allows for homogeneous dispersion of two OIMH-based parent materials (fig. S18A). The smooth viscosity curves of α_G(A₁TPG₁) as a function of temperature also prove the uniformity of the co-melting process (fig. S18B). Further analysis of the temperature-dependent viscosity data provides insight into the melt dynamics. As shown in fig. S19, the flow activation energy (E_a), which represents the energy barrier encountered by each component in the melt to diffuse at a specific temperature, can be calculated using the Arrhenius model (0.9 < T_m/T < 1.0)

$$\mathrm{\eta }\left(T\right)=A\times \text{exp}\left(\frac{E_{\mathrm{a}}}{\mathit{RT}}\right)$$

where A is a constant, R is the molar gas constant, and E_a is correlated with the evolution of melt structure in the cooling process (49). The high E_a value of two-component OIMH melt displays a rapid increase in the η value as the temperature decreases, which hampers the migration of molecules and then facilitates the formation of amorphous α_G(A₁TPG₁) (Fig. 3B and table S8). Although the E_a values are temperature-dependent because of the limitations of the Arrhenius model when extrapolated over a broader temperature range, their comparative evolution still provides valuable insights. The temperature range of 0.9 < T_m/T < 1.0 was carefully selected for fitting, which corresponds to the duration of the complete flowing homogeneous melt. At this condition, the structure of the material is relatively stable, and the molecular movement is mainly affected by temperature and the intrinsic properties of the molecular chains, thereby minimizing the impact of structural heterogeneity on the measurements. To further understand the dynamics of supercooled melts, we adopted the improved Mauro-Yue-Ellison-Gupta-Allan model as shown below (0.7 < T_g/T < 1.0)

$$\begin{array}{c}\text{log}\mathrm{\eta }\left(T\right)=\text{log}{\mathrm{\eta }}_{\infty }+(12-\text{log}{\mathrm{\eta }}_{\infty })\frac{T_{\mathrm{g}}}{T}\\ \text{exp}\left[(\frac{m}{12-\text{log}{\mathrm{\eta }}_{\infty }}-1)(\frac{T_{\mathrm{g}}}{T}-1)\right]\end{array}$$

where η_∞ is the η value at an infinitely high temperature, and m is the fragility index, which describes changes in the viscosity of a melt as a function of temperature (50). As shown in fig. S20, the m values of αG (A₁TPG₁) glasses span from 85.2 to142.1, exceeding those of commercial inorganic glasses such as soda-lime-silica glass (15–25), borosilicate glass (10–21), and lead glass (18–28). Relatively high m values observed in α_G(A₁TPG₁) are attributed to weaker intermolecular interactions compared with inorganic glass, indicating a smoother transition in the η value from the high-temperature flow state to the supercooled state. This property facilitates the processing of glasses over a wide temperature range and prevents the formation of irregularities, defects, or stress concentrations.

The XAFS experiment was carried out to explore the structural features of two-component OIMH glass. As shown in fig. S21, the normalized Mn K-edge XANES spectrum of α_G(PTP₁TPG₁) is nearly identical to that of (TPG)₂MnBr₄, indicating that there are similar local structures around Mn ions. Their absorption edges are close to that of MnO, which proves that the Mn valence of two-component glasses remains +2. The EXAFS data of two-component α_G(PTP₁TPG₁) were fitted by Fourier transform in Artemis, where the main peak demonstrates the local four-coordination environment of Mn-Br (Fig. 3C and table S9). Complementary pair distribution function (PDF) analysis was used to further probe the local atomic arrangements in two-component amorphous OIMH glasses. As shown in Fig. 3D, both α_G(PTP₁TPG₁) and α_G(PTP₉TPG₁) glasses exhibit distinct flat features between 6 and 30 Å, indicating the absence of long-range ordered structure. The preservation of identical PDF oscillation profiles below 6 Å reveals remarkable short-range structural consistency across different compositions. The prominent peaks at 2.5 and 4.1 Å are attributed to Mn-Br and Br-Br atomic pairs due to the larger x-ray scattering strength of Mn and Br relative to lighter elements. In addition, these peaks ranging from 0 to 2 Å are assigned to C/N–H, C–C/N, and C–P atomic pairs within isolated organic cations. These structural characterizations confirm the invariance of organic cations (PTP⁺ and TPG⁺) and inorganic [MnBr₄]²⁻ units in the glass state, as well as the retention of a zero-dimensional structure at the molecular level.

Energy-dispersive spectroscopy (EDS) elemental mapping verified uniform dispersion of two parent materials during co-melting. For the single-component (TPG)₂MnBr₄ glass, Mn and N elements are uniformly dispersed (fig. S22). Similarly, in the two-component α_G(PTP₁TPG₁) glass, Mn, N [belong to (TPG)₂MnBr₄), and P [belongs to (PTP)₂MnBr₄] elements are homogeneously dispersed, as evidenced by EDS mapping at both micrometer and nanometer scales (fig. S23). As a complementary experiment, we also attempted to extend the two-component glasses to other molten metal halides (e.g. Sb- and Pb-based OIMHs). As shown in fig. S24 (A and B), the introduction of (ETP)₂SbCl₅ (ETP, ethyltriphenylphosphonium) enhanced the GFA of (Bzmim)₂SbCl₅ (Bzmim, 1-benzyl-3-methylimidazolium) and increased its T_g value to 34.0°C. In addition, the recrystallization behavior of (BPy)₂PbBr₄ was also suppressed by co-melting with (ETP)₂PbBr₄, leading to an increase in T_g to 37.9°C (fig. S24, C and D). Meanwhile, we have further extended this strategy to prepare the three-component Mn-based glass consisting of (TPG)₂MnBr₄, (DPG)₂MnBr₄, and (TMG)₂MnBr₄. In addition, another three-component glass containing different metal-halide units was also prepared by co-melting (TPG)₂MnBr₄, (ETP)₂SbCl₅, and (ETP)₂PbBr₄. Their single T_g values between the parent materials support the occurrence of homogeneous mixing (fig. S25).

On the basis of the above discussion on preparing two-component OIMH glasses, it can be envisaged that parent materials should meet the following requirements. Firstly, they need to have considerable thermal stability to ensure a stable melting process. Second, it is necessary to screen a high-GFA parent material that can regulate the GFA of another one by co-melting. Third, their rheological properties at the preparation temperature must be similar, which helps to achieve homogeneous melting (51). Furthermore, the scintillation performance of two-component OIMH glasses has been evaluated in fig. S26, where α_G(MTP₁TPG₁) and α_G(PTP₁TPG₁) have higher light yields (8810 and 10,240 photons MeV⁻¹) and lower LoD (203.9 and 164.0 nGy s⁻¹) compared with (TPG)₂MnBr₄ glass, while others exhibit the suboptimal scintillation performance (table S10). This discrepancy is due to the low T_g value (<0°C) of low-GFA A₂MnBr₄. The absence of large rigid aromatic cations reduces molecular structural rigidity and leads to severe nonradiation recombination at room temperature, resulting in inferior scintillation performance.

In situ crystallization of two-component OIMH glasses

In situ crystallization of the glassy matrix by annealing is an effective post-processing strategy for enhancing the scintillation performance. This method allows for the combination of high optical transparency of the glassy state and efficient luminescence of the crystalline state (52). However, the in situ crystallization in the single-component OIMH glass is challenging to control, which often results in diminished transparency and pronounced light scattering. As shown in fig. S27, both single-component (MTP)₂MnBr₄ and (PTP)₂MnBr₄ glasses underwent a glass-crystal transition process upon being heated at 70°C for 20 hours, while the α_G(TPG₁MTP₁) and α_G(TPG₁PTP₁) with better GFA remained amorphous even after being heated for 100 hours. The PXRD patterns before and after heat treatment support the above phenomenon. The two-component OIMH glass discussed above may serve as an appropriate choice. The GFA can be modulated by regulating the mass ratio of parent materials, thus achieving controlled in situ crystallization without notable reduction in light transmittance. Given a high light yield of 10,240 photons MeV⁻¹ and a suitable T_g of 59.7°C for the α_G(PTP₁TPG₁), a series of α_G(PTP_xTPG_y) with varying x:y ratios were annealed to prepare the glass-crystal composite, expressed as α_G–C(PTP_xTPG_y). As shown in Fig. 4A, we first measured the T_g values of α_G(PTP_xTPG_y) and found that there is a strong linear dependence between T_g and the mass ratio of parent materials, which proves the homogeneous mixing and allows us to modulate the T_g value within the range of 47.0° to 72.6°C. In general, the annealing temperature of glass for in situ crystallization is slightly higher than its T_g. At lower temperatures, the diffusion and migration of atoms in melts are impeded, while noticeable softening deformation or excessive crystallization occurs at higher temperatures. Accordingly, we tried to anneal the α_G(PTP_xTPG_y) samples with different x:y ratios (1:1 to 499:1) and (PTP)₂MnBr₄ glass at 55°C for 1 day (Fig. 4B).

Fig. 4. In situ crystallization and scintillation properties of α_G–C(PTP_xTPG_y).

(A) The T_g values of α_G(PTP_xTPG_y) versus the mass ratio of parent materials. (B) Schematic diagram of the in situ crystallization process. (C) Light transmittance spectra of annealed α_G–C(PTP_xTPG_y) with varying x:y ratios. (D) PXRD patterns of α_G–C(PTP_xTPG_y). (E) Light yields of α_G(PTP_xTPG_y) and α_G–C(PTP_xTPG_y). (F) Fitting signal-to-noise ratio (SNR) of α_G–C(PTP_xTPG_y) under low-dose-rate irradiation.

Photophysical and structural analyses were conducted to characterize these annealed samples. As shown in Fig. 4C, α_G–C(PTP_xTPG_y) samples with an x:y ratio of ≤99:1 exhibit high light transmittance achieving 80% in the visible light range. In contrast, other annealed samples with a higher content of (PTP)₂MnBr₄ exhibit severe devitrification and poor transparency (fig. S28). The presence of a crystalline phase has been verified through PXRD patterns. As shown in Fig. 4D, α_G–C(PTP_xTPG_y) samples with an x:y ratio of ≤19:1 remained amorphous because of the addition of increased amounts of high-GFA (TPG)₂MnBr₄, while those with an x:y ratio of ≥39:1 exhibited high-intensity, sharp crystalline peaks. However, the PXRD patterns of annealed α_G–C(PTP_xTPG_y) are different from the simulated pattern of (PTP)₂MnBr₄ crystals grown by the solution method. A reasonable explanation is that (PTP)₂MnBr₄ crystals undergo a phase transition before melting, which is supported by the DSC curve with an extra exothermal peak at 120.3°C (fig. S29A). The difference in PXRD patterns of (PTP)₂MnBr₄ crystals before and after heating at 130°C supports this explanation (fig. S29B). Thermogravimetric-infrared spectroscopy (TG-IR) results proved that the phase transition came from the removal of acetone molecules, where the IR spectrum of gas-phase acetone was observed at 120°C (fig. S30).

Since the poor transparency of α_G–C(PTP_xTPG_y) samples with an x:y ratio of ≥149:1 is detrimental to practical application in radiation detection and x-ray imaging, we focus on the scintillation performance of transparent α_G–C(PTP_xTPG_y) samples with an x:y ratio of ≤99:1. The presence of crystalline (PTP)₂MnBr₄ enhances the luminescence efficiency of annealed samples due to its higher PLQY of 86.8% than that of glassy state (44.0%) (fig. S31). As shown in Fig. 4E, the light yield of annealed samples with an x:y ratio ranging from 39:1 to 99:1 is 86 to 95% higher than that of amorphous α_G(PTP_xTPG_y). For instance, after being annealed at 55°C for 1 day, the light yield of α_G–C(PTP₉₉TPG₁) increased from 18,800 photons MeV⁻¹ to 35,140 photons MeV⁻¹, as well as an LoD value as low as 43.6 nGy s⁻¹, which is about 126 times below the medical x-ray diagnostic dose of 5.5 μGy s⁻¹ (Fig. 4F and table S11) (53). The high light transmittance of up to 80% also makes it promising for a variety of photonics applications such as scintillators, photoconductive fibers, light-emitting diodes, and laser crystals. The scintillation properties of annealed samples with an x:y ratio ≤19:1 were not improved because of the stronger resistance to crystallization, remaining in an amorphous state after annealing treatment. Overall, the combination of the preparation and in situ crystallization of two-component OIMH glass provides a robust approach for fabricating high-performance scintillation screens.

Application in the x-ray imaging and mechanical properties

To assess the feasibility of α_G–C(PTP_xTPG_y) scintillators for practical x-ray imaging, a self-built x-ray imaging system was constructed, comprising an x-ray tube, a shading film, a high-reflection mirror, and a digital camera (Fig. 5A). Firstly, these scintillators with varying x:y ratios were used to inspect the internal structure of a chip. As illustrated in Fig. 5B, there are discernible disparities in the image quality. The imaging brightness and contrast improve as the x:y ratio increases from 1:1 to 99:1. The annealed samples with a larger x:y ratio of 149:1, 299:1, 499:1, and even pure (PTP)₂MnBr₄, suffer an obvious loss in optical transparency and lead to a notable decline of x-ray imaging quality (fig. S32A). The imaging quality of annealed (TPG)₂MnBr₄ is close to the α_G–C(PTP₁TPG₁) due to the strong resistance to crystallization (fig. S32B). The extracted gray value contours along the lines of the red box highlight the advantages of α_G–C(PTP₉₉TPG₁), which provide the highest brightness and contrast because of its high light yield of 35140 photons MeV⁻¹ (Fig. 5C). At the same time, the spatial resolution of α_G–C(PTP₉₉TPG₁) was determined to be 20 lp∙mm⁻¹ using a standard line pair card, outperforming most of the reported OIMH scintillation, which is also proved by corresponding gray value contours (Fig. 5D, E). Moreover, we demonstrated the feasibility of the α_G–C(PTP₉₉TPG₁) in dynamic x-ray imaging. As shown in Fig. 5F and movie S1, the x-ray images of the rotating lead fan blade were monitored. The x-ray images at 0.1-s intervals show clear outlines of the lead blade without ghosting effects. Considering the inspiring scintillation performance, the α_G–C(PTP₉₉TPG₁) distinguishes itself as a potential candidate for dynamic x-ray imaging.

Fig. 5. Scintillation application of α_G–C(PTP_xTPG_y).

(A) Schematic representation of the home-built x-ray imaging system, the target subject was placed between the x-ray source and shading film. (B) X-ray images of a chip using α_G–C(PTP_xTPG_y) with varying x:y ratios (total dose of one image is 0.4 mGy). (C) Gray value contours along the lines of the red box extracted from Fig. 4B. (D) Spatial resolution of the α_G–C(PTP₉₉TPG₁) determined by a lead-made line pair card. (E) Gray value contours extracted from Fig. 4D. (F) Dynamic x-ray images of rotating lead fan blade using the α_G–C(PTP₉₉TPG₁) (total dose of one image is 22.6 μGy).

Nanoindentation technology has been used to investigate the mechanical properties of the scintillation screens, including a single-component (TPG)₂MnBr₄ glass, a two-component α_G(PTP₉₉TPG₁) glass, and an annealed α_G–C(PTP₉₉TPG₁) sample. As shown in Fig. 6A, the load-displacement curves obtained from five parallel tests are smooth with no breaks or turning points, indicating that these screens have relatively uniform and stable microstructures, and will not experience crack propagation or local collapse due to local stress concentration. The Young’s moduli (E) and hardness (H) were illustrated in Fig. 6B and fig. S33. Within the displacement range from 600 to 1000 nm, the E and H values tend to be stable and are less affected by the depth of the indentation. The three OIMH-based screens exhibit similar mechanical properties, with E values ranging from 5.52 to 5.79 GPa and H values ranging from 0.229 to 0.269 GPa. In addition, we compared the E and H values between OIMH-based screens and other reported or normal materials. As shown in Fig. 6C and table S12, these values of OIMH glasses are approximately located between those of organic polymers and inorganic or metallic glasses and are close to those reported organic-inorganic hybrid glasses, such as ZIF glasses (54–56).

Fig. 6. Mechanical properties.

(A) Load-displacement curves recorded in five parallel tests. (B) Young’s modulus (E) as a function of indentation depth for (TPG)₂MnBr₄ glass, α_G (PTP₉₉TPG₁), and annealed α_G–C(PTP₉₉TPG₁). (C) The summary of E and H values for polymers, OIMH glasses, ZIF glasses, inorganic glasses, and metallic glasses.

DISCUSSION

In conclusion, we have developed a universal strategy for preparing two-component OIMH glasses and transparent glass-crystal composites. The introduction of (TPG)₂MnBr₄ with high-GFA provides a platform to modulate the crystallization of A₂MnBr₄ melts by homogeneous melting. The idea of two-component OIMH melting allows those that are easy to crystallize to remain amorphous glassy state, which is supported by related thermodynamic and rheological analysis. The stable formation of multiple two-component OIMH glasses demonstrates the generality and scalability of the present strategy, which will greatly expand the scope of the glassy OIMH system. Moreover, the GFA modulation enables controlled in situ crystallization of two-component OIMH glasses, leading to the formation of transparent glass-crystal composites with higher luminescence efficiency. The facile implementation of high-performance glass-crystal composites will greatly stimulate the promising applications of OIMH materials in various photonics and electronics fields.

MATERIALS AND METHODS

Materials

The chemicals used are as follows: TPG (TCI, 97%), MTP·HBr (Aladdin, 99%), PTP·HBr (Aladdin, 99%), TMG (Bidepharm, 97%), DPG·HBr (Bidepharm, 98%), EPy·HBr (Bidepharm, 99%), BPy·HBr (Bidepharm, 95%), EmIm·HBr (Aladdin, 99%), BmIm·HBr (Aladdin, 98%), TBP·HBr (Aladdin, 98%), TBA·HBr (Aladdin, 99%), MnBr₂·4H₂O (Aladdin, 98%), hydrobromic acid (Aladdin, 48 weight% in water,), anhydrous methanol (Guangzhou Chemical Reagent, ≥99.5%), anhydrous ethanol (Guangzhou Chemical Reagent, ≥99.5%), isobutanol (Macklin, ≥99.0%), dichloromethane (Guangzhou Chemical Reagent, ≥99.9%), acetone (Guangzhou Chemical Reagent, ≥99.9%), acetonitrile (Guangzhou Chemical Reagent, ≥99%), and ethyl acetate (General-reagent, ≥99.5%). All the chemicals were used without further purification.

Synthesis of TPG·HBr single crystals

TPG was dissolved in a mixed solution of methanol and hydrobromic acid. The precursor solution was stirred and then slowly evaporated at 80°C for several hours. The crystals were precipitated and washed with ethyl acetate.

Synthesis of (TPG)₂MnBr₄ single crystals

TPG·HBr (2 mmol) and MnBr₂·4H₂O (1 mmol) were added to ethanol and then stirred at 80°C to form a clear solution. The precursor solution was slowly evaporated at 80°C for 2 days. Transparent crystals were precipitated and washed several times with ethyl acetate.

Synthesis of A₂MnBr₄ single crystals (A, organic cations)

The A₂MnBr₄ single crystals were synthesized through solvent evaporation or cooling-induced crystallization methods. Specifically, the organic salt and MnBr₂·4H₂O at a molar ratio of 2:1 were dissolved in the organic solvent under magnetic stirring. Then, the clear precursor solution was evaporated at 40° to 80°C or was slowly cooled to room temperature to obtain transparent crystals.

Preparation of A₂MnBr₄ and (TPG)₂MnBr₄ glasses

All OIMH glasses mentioned in the manuscript were prepared using the melt-quenching method. First, the hybrid manganese bromide crystals were heated in a muffle furnace for 10 to 20 min to form a clear melt. Then, the bubbles in the melt were removed using a vacuum-drying oven. The transparent glass was formed upon cooling to room temperature naturally.

Preparation of two-component α_G(A_xTPG_y) and α_G–C(A_xTPG_y)

The amorphous α_G(A_xTPG_y) sample was prepared by the melt-quenching process. The (TPG)₂MnBr₄ glass was mixed with A₂MnBr₄ crystals in a special mass ratio. The mixture was then heated in a vacuum muffle furnace at 220°C to form a homogenous melt. Last, the melt was cooled to room temperature to form a transparent α_G(PTP_xTPG_y). The α_G–C(PTP_xTPG_y) samples were obtained by annealing the α_G(PTP_xTPG_y) at 55°C for 1 day.

Crystallography measurements

SCXRD was performed on a Bruker D8 VENTURE diffractometer with Mo-Kα radiation (λ = 0.71073 Å). The data were solved by the intrinsic phasing method (SHELXT) and refined by the SHELXL package included in the Olex2 software (57, 58). Deposited CCDC numbers are 2,378,331 [for (TPG)₂MnBr₄], 2,380,213 [for (PTP)₂MnBr₄], 2,382,136 [for (TMG)₂MnBr₄], 2,408,182 [for (BmIm)₂MnBr₄], and 2,408,056 [for TPG·HBr]. PXRD patterns were collected using Cu-Kα radiation (λ = 1.54184 Å) on a Rigaku Mini-flex600 diffractometer.

Thermodynamics and rheological measurements

TGA and DSC data were obtained using NETZSCH TG 209F1 Libra system and NETZSCH DSC 214, respectively. Synchronous TG-IR spectra were carried out on a NETZSCH STA449F3/Nicolet 6700 system. All measurements were performed in a nitrogen atmosphere with a heating rate of 10 K min⁻¹. Temperature-dependent viscosity data were collected on a HAAKE MARS40 rotary rheometer. The samples were preheated to form melt and then cooled down at a rate of 10 K min⁻¹, and the viscosity values were recorded. A shear rate of 10 s⁻¹ was used for the measurements.

Spectroscopic and mechanical measurements

Relevant photoluminescence spectra were obtained using an Edinburgh FLS980 spectrometer equipped with the Oxford cryostat. PLQY values were measured by using the Hamamatsu C9920 instrument. FTIR spectra were performed on the PerkinElmer Frontier spectrophotometer. The light transmittance spectra were recorded by a Shimadzu UV-3600 spectrophotometer. The XAFS spectra on (TPG)₂MnBr₄ crystals and glass were obtained using the RapidXAFS (Anhui Absorption Spectroscopy Analysis Instrument Co. Ltd.) by transmission mode at 10 kV and 20 mA, and a Si (440) spherically bent crystal analyzer was used for Mn K-edge. The XAFS experiments for α_G(PTP₁TPG₁) were carried out on the sample BL11 second hatch x-ray absorption beamline of the SAGA Light Source. The energy resolution of the station ranges from 10⁻⁴ to 10⁻³ ΔE/E, with a photon flux ranging from 2 × 109 to 1 × 108 photons s⁻¹, corresponding to x-ray energies in the range of 2.1 to 14 keV. Nanoindentation experiments were performed using the KLA iMicro equipped with a Berkovich diamond indenter tip. All tests were conducted under the continuous stiffness measurement. The EDS mapping was carried out on the GeminiSEM 500.

PDF analysis

The high-resolution synchrotron x-ray total scattering measurements were performed at beamline ID31 of the European Synchrotron Radiation Facility. The sample powders were loaded into cylindrical slots (approx. 1-mm thickness) held between Kapton windows in a high-throughput sample holder. Each sample was measured in transmission with an incident x-ray energy of 75.051 keV (λ = 0.16520 Å). Measured intensities were collected using a Pilatus CdTe 2M detector (1679 × 1475 pixels, 172 μm by 172 μm each) positioned with the incident beam in the corner of the detector. The sample-to-detector distance was approximately 0.3 m for the total scattering measurement. Background measurements for the empty windows were measured and subtracted. National Institute of Standards and Technology (NIST) SRM 660b (LaB₆) was used for geometry calibration. The integration of the 2D XRD pattern was performed with the software pyFAI, followed by image integration, including flat-field, geometry, solid-angle, and polarization corrections. The PDF data obtained from the experiment were processed using PDFgetX3 software (59–61).

Scintillation and x-ray imaging

The x-ray attenuation efficiency (AE) is calculated using the following equation

$$\mathit{AE}=(1-e^{-\mathrm{t\rho d}})\times 100\%$$

where t is the total attenuation coefficient obtained from the XCOM database of the NIST. ρ and d denote the density(gram per cubic centimeter) and thickness (centimeter) of scintillators. The commercial LuAG: Ce single crystal was used as a reference to estimate the light yield of these samples. An Amptek Mini-X2 x-ray tube (target: Ag) and an Ocean Optics portable spectrometer (QEpro) equipped with an integrating sphere were used to construct the measurement system. The size of these samples is consistent with the LuAG: Ce (diameter = 13 mm, thickness = 1 mm). The measured photon counts are normalized via the equation below

$$\begin{array}{c}{\mathit{LY}}_{\text{sample}}={\mathit{LY}}_{\text{LuAG}:\text{Ce}}\times \frac{P_{\text{measured}}\left(\mathit{\text{Sample}}\right)/\mathit{AE}\left(d\right)}{P_{\text{measured}}(\mathit{\text{LuAG}}:\mathit{Ce})/\mathit{AE}\left(d\right)}\hfill \end{array}$$

where P_measured denotes the absolute number of photons and AE(d) is the attenuation efficiency (%) of the sample at a certain thickness. The LoD is defined as the dose rate when the signal-to-noise ratio equals 3. An H10721-210 photomultiplier tube was used to collect the photons produced by the sample under different dose rates and convert them to signal currents. The noise data were obtained in the absence of the x-ray radiation. The dose rate was calibrated by the Radical x-ray dosimeter.

Supplementary Materials

The PDF file includes:

Figs. S1 to S33

Tables S1 to S12

Legend for movie S1

Other Supplementary Material for this manuscript includes the following:

Movie S1