X-ray scintillators, capable of converting high-energy X-ray into visible low-energy light, are widely used in the fields of industrial non-destructive testing,^[1-3] medical imaging,^[4-8] etc. The most researched scintillators are inorganic compounds, such as CsI:Tl, bismuth germanate (BGO), and CsPbBr₃, which demonstrate strong X-ray absorption capability and excellent X-ray scintillation properties, but generally suffer from the disadvantages of ultrahigh preparation temperature, high toxicity, and inferior environmental stability.^[9-10] Pure organic scintillators, on the other hand, are of significant advantages such as low preparation temperature, ease of modification, and the ability to fabricate large-area flexible scintillation screens.^[11-16] Despite these benefits, pure organic scintillators, typically composed of elements with small atomic numbers like carbon and hydrogen, have weak X-ray absorption, leading to low radioluminescence (RL) efficiency and posing a significant challenge to their widespread application. Combined the advantages of inorganic and organic scintillators, organic-inorganic hybrid scintillators simultaneously exhibit accomplished X-ray scintillation performance and simple preparation process, possessing more practical application prospect.^[17-19]

Recently, organic-inorganic hybrid Cu(I) clusters,^[20-27] exhibiting the superior RL efficiencies in the solid state, have been a research focus in the scintillator field. Cu(I) clusters contain a rigid [CuX]_n (X=Cl, Br, I) core (Figure 1a), which plays an important role for high photoluminescence quantum yield (PLQY) due to the inhibited structural relaxation and strong X-ray absorption owing to high atomic number copper and halogen atoms. Under X-ray excitation, the RL of Cu(I) clusters are mainly derived from metal/halogen-to-ligand [(M＋X)LCT] excited state.^[20,28] Due to the heavy atom effect, both ¹(M＋X)LCT and ³(M＋X)LCT excitons can emit RL, enabling 100% exciton utilization. Benefited from these inherent advantages, the single crystals of Cu(I) clusters show the comparable scintillation performance with the commercial inorganic counterparts. However, it is extremely challenging for them to cultivate large area single crystals for X-ray imaging due to the weak intermolecular interactions and complex molecular structures. Hence, they are usually doped into polymer matrix to prepare scintillator films, whereas the polymer matrix has low X-ray absorption ability and is hardly radioluminescent, resulting in deteriorated scintillation efficiency. In addition, the uneven dispersion of scintillators within the polymer matrix can also lead to poor X-ray imaging resolution. Hence, there is an urgent need for innovative film fabrication techniques for Cu(I) cluster scintillators.

Vacuum evaporation deposition that can prepare high-quality films over large area has been extensively applied in organic light-emitting diodes,^[29-31] perovskite photovoltaics,^[32-34] and organic field-effect transistors.^[35-37] We assumed that the pure Cu(I) cluster films prepared by vacuum evaporation deposition would be homogeneous without sacrificing the scintillation performance (Figure 1b). To proof our hypothesis, two dinuclear Cu(I) clusters [CuX(dppb)]₂ (X=Br, I) with admirable thermal stability were prepared (Figure 1c). The decomposition temperatures of the two clusters are higher than 440 ℃, making them suitable for vacuum evaporation deposition. Due to the strong X-ray absorption and high exciton utilization, the clusters display excellent X-ray scintillation performance with the high light yields of up to 19356.7 photons/MeV. The scintillation screen based on the Cu(I) clusters prepared by vacuum evaporation deposition is smooth and uniform with a small surface roughness value of 1.04 nm (Figure 1d), which is exploitable in X-ray imaging for various objects. The current work suggests great potential of vacuum evaporation deposition in fabricating scintillation screens for X-ray radiography.

The dinuclear Cu(I)-halide clusters, [CuBr(dppb)]₂ and [CuI(dppb)]₂, could be synthesized in a large scale through a simple room-temperature solution reaction of 1,2-bis(diphenylphosphino)benzene (dppb) with CuX (X=Br or I) and purified by vacuum sublimation. The chemical structures of the products were confirmed by proton (¹H) and carbon (¹³C) nuclear magnetic resonance (NMR) and crystallographic analyses. The thermal stability of [CuX- (dppb)]₂ was subsequently determined by thermogravimetric analysis (TGA). [CuX(dppb)]₂ clusters exhibit a high decomposition temperature of up to 440 ℃, indicating that the material has good thermal stability which is conducive to fabricate evaporation-processed scintillator film.

The basic photophysical properties of the two clusters were investigated through ultraviolet-visible (UV-Vis) absorption spectroscopy and photoluminescence (PL) spectroscopy. Both the pure [CuBr(dppb)]₂ and [CuI- (dppb)]₂ films exhibit two absorption peaks at ca. 290 and ca. 330 nm (Figure 2a). The ca. 290 nm absorption is attributed to π-π* transition of dppb ligand, while ca. 330 nm absorption should be assigned to CuX-involved electronic transition. The steady-state PL (SSPL) spectra of [CuBr-(dppb)]₂ and [CuI(dppb)]₂ crystals are similar with the bluish green emission bands centered at 505 nm, which shows the lifetimes of ca. 4 µs at room temperature (Figure 2b). Notably, the SSPL spectrum of [CuBr(dppb)]₂ crystal is red-shifted than that of [CuI(dppb)]₂ crystals. The reason is that Br^－ has stronger ligand field strength than I^－, which would increase the energy separation of d-orbitals, resulting in a smaller metal-to-ligand charge transfer (MLCT) energy. In addition, both [CuBr(dppb)]₂ and [CuI(dppb)]₂ crystals exhibit the satisfactorily high photoluminescence quantum yields (PLQYs) of 49.5% and 56.9% respectively (Figure 2b), indicating that non-radia- tive transition is inhibited, which might be due to the tight molecular packing in crystalline state. With the temperature increasing from 78 to 288 K, the lifetimes of the [CuI(dppb)]₂ crystals constantly decrease (Figure 2c), illustrating that the bluish green emission is mainly derived from phosphorescence. The phosphorescence-dominant emission is further confirmed by temperature-dependent SSPL spectra, which show slightly decreased emission intensity with increasing temperature from 80 to 180 K (Figure 2d). The emission has virtually no change above 180 K, which also manifests the rigid structure of [CuI- (dppb)]₂ crystals.

Theoretical calculations were performed to get an insight into the photophysical properties of [CuX(dppb)]₂ clusters. The frontier molecular orbits of the two clusters are similar, with the highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) located on Cu₂X₂ core and dppb ligand, respectively, indicating their charge transfer characteristic. The separated HOMO and LUMO distributions render the small single-triplet energy gaps (ΔE_ST), which are 0.15 and 0.13 eV for [CuBr(dppb)]₂ and [CuI(dppb)]₂, respectively. The small ΔE_ST indicate that the clusters may have thermally activated delayed fluorescence (TADF) properties.^[38] The natural transition orbital (NTO) calculations were also performed to analyze the electron transition components of [CuX(dppb)]₂ at triplet (T₁) states. The highest occupied NTOs (HONTOs) are primarily distributed on Cu₂X₂ core and dppb ligand (Figure 2e), while the lowest unoccupied NTO (LUNTOs) are dominated by dppb ligand; the overlaps between HONTOs and LUNTOs for [CuBr(dppb)]₂ and [CuI(dppb)]₂ are 38.68% and 35.14%, respectively, guaranteeing the high PLQYs.^[39] Additionally, the molecular cores of the two copper clusters show minimal conformational variation between the singlet (S₁) and ground (S₀) states with a low root mean square displacement (RMSD) values of approximately 0.12 Å (Figure 2f). This indicates that they have a rigid core structure, which can suppress the non-radiative transition and improve the luminous efficiency.

The excellent crystal luminescence and high atomic number of halogen atoms (Br, I) forebode the promising potential of [CuX(dppb)]₂ clusters as X-ray scintillators. Therefore, a systematic study of the scintillation performance of [CuX(dppb)]₂ under X-ray excitation was conducted. Since the X-ray absorption coefficient (α) is proportional to the fourth power of the atomic number,^[40] [CuI(dppb)]₂ (Z_max=53, K_α=33.17 keV) containing iodine atoms exhibits the stronger X-ray absorption in comparison with [CuBr(dppb)]₂ (Z_max=35, K_α=13.47 keV) (Figure 3a). Subsequently, the attenuation of [CuX- (dppb)]₂ as a function of thickness was calculated at an X-ray photon energy of 68 keV. It was found that a 0.4-mm-thick [CuI(dppb)]₂ can attenuate 99% of X-ray photon energy, while [CuBr(dppb)]₂ requires a thickness of 0.68 nm (Figure 3b). Figure 3c shows the RL spectra of 50 mg [CuX(dppb)]₂ crystals compared to commercial scintillators CsI:Tl, Ce:YAP, and Ce:YAG powders under the same X-ray dose rate irradiation. [CuX(dppb)]₂ crystals manifest the excellent X-ray scintillation with RL intensity comparable to these commercial scintillators (Figure 3d). Although the PLQY of [CuI(dppb)]₂ crystal is lower than [CuBr(dppb)]₂ crystal, it has higher RL intensity probably due to its stronger X-ray absorption of I atom. Benefited from the strong X-ray absorption and high exciton utilization, [CuI(dppb)]₂ crystal exhibits a high light yield of up to 19356.7 photons/MeV. The RL spectra of [CuX(dppb)]₂ at different X-ray dose rates were recorded. As shown in Figure 3e and Figure S14 (see Supporting Information), the RL intensity of [CuX(dppb)]₂ increases monotonically and exhibits a good linear response in the range of 0.688 μGy/s to 278 μGy/s. The limit of detection (LOD) values at a signal-to-noise ratio of 3 were evaluated to be 186 and 158 nGy/s for [CuBr- (dppb)]₂ and [CuI(dppb)]₂, respectively (Figure 3f), which are significantly lower than the limit for medical X-ray diagnostics (5.5 μGy/s). Moreover, the [CuI- (dppb)]₂ crystals reveal excellent radiation stability with RL intensity keeping nearly unchanged for 130 on-off excitation circles (Figure 3g). The radiation intensities maintain over 99% even after continuous irradiation at a dose rate of 278 μGy/s for 1800 s.

The RL and SSPL (Figure 2a) spectra of [CuX(dppb)]₂ crystals closely align with the emission peak at ca. 500 nm, indicating that the emissions might arise from the same radiative channels under UV and X-ray irradiation. Thus, a rational mechanism for the RL of [CuX(dppb)]₂ crystals can be proposed. Upon X-ray excitation, the heavy atoms (Br and I atoms) in [CuX(dppb)]₂ interact with X-ray photons through the photoelectric effect and Compton scattering,^[41] releasing high-energy electrons. These hot electrons further interact with atoms, producing a large number of secondary electrons. After losing excess energy, the electrons are eventually captured by the ligands, forming hole-electron pairs. The recombination of hole-electron pairs produces 25% singlet excitons and 75% triplet excitons according to the spin rules (Figure 3h). Due to the small ΔE_ST between the ¹(M＋X)LCT and ³(M＋X)LCT states, the S₁ and T₁ excitons of [CuX(dppb)]₂ are interconvertible through intersystem crossing (ISC) and reverse intersystem crossing (RISC). The RL emission mainly comes from phosphorescence generated by ³(M＋X)LCT excitons, and partly comes from fluorescence derived from ¹(M＋X)LCT excitons.

Benefiting from the excellent X-ray scintillation performance and thermal stability, [CuI(dppb)]₂ was used to fabricate scintillation screen through vacuum evaporation deposition for X-ray imaging (Figure 4a). Atomic force microscopy (AFM) images indicate that the vacuum- evaporated [CuI(dppb)]₂ film has a smooth and uniform surface with a small surface roughness value of 1.04 nm (Figure 4b). Line pair card imaging further demonstrates its high spatial resolution, where the line pairs corresponding to >8 lp/mm is identifiable (Figure 4c). Then the film was further applied to X-ray imaging for various objects (Figures 4c and 4d). Under X-ray irradiation, the paper clip inside a capsule and the internal structure of an electronic chip could be clearly revealed, highlighting the potential applications of such scintillator films in non-destructive testing and medical imaging.

In summary, we have proposed an effective strategy to prepare X-ray imaging screen through vacuum evaporation deposition. Two coper clusters [CuBr(dppb)]₂ and [CuI- (dppb)]₂ appropriate for evaporation process were synthesized, which exhibit meritorious luminous efficiencies due to the rigid molecular structures. The X-ray scintillation performance of the coper clusters is comparable to those commercial inorganic scintillators with a high light yield of up to 19356.7 photons/MeV and a detection limit as low as 158 nGy/s. The thin film of [CuI(dppb)]₂ fabricated by vacuum evaporation deposition is smooth and uniform, which has been applied to X-ray imaging for various objects. Though the imaging spatial resolution of the thin film needs to be improved, the current work highlights the po-tential of vacuum evaporation deposition method in the development of high-quality imaging screen for advanced X-ray radiography.