Introduction

The 5d–4f transition in lanthanides (Ln) has been studied for decades on the luminescent mechanism and potential applications in various fields1,2,3,4,5,6,7,8. For the well-established luminescence of lanthanide ions, ff transition can be observed under ambient conditions, while the 5d–4f transition is usually absent due to thermally quenching by fast intersystem crossing from 4fn−15d1 to 4fn configuration. In divalent lanthanide systems, 5d–4f transition is much more prominent for its spin-allowed nature and the stabilization of the 5d orbitals6,7,8.

Among all Ln2+ ions, Eu2+ ions exhibit strong 5d–4f transition and great applications for two reasons: (1) the 5d level is near or below 6P7/2, decreasing the multiphoton relaxation6, (2) the reduction potential of Eu3+/Eu2+ is not too negative. The research on luminescent properties of Eu2+ ions can be roughly divided into two categories: Eu2+ dopants in inorganic matrix and molecular Eu2+ complexes. The first one has been extensively studied while the latter remains unexplored in many aspects. The physicochemical properties of Eu2+ complexes are mainly studied using cyclopentadienyls, hydrotris(pyrazolyl)borates, silylamides and their derivatives as ligands9,10,11. Recently, Allen et al. reported series of Eu2+-containing azacryptate complexes which have aroused growing interest for their attractive luminescent properties, photoredox catalytic performance and magnetic resonance imaging3,12,13,14,15.

The uniqueness of 5d–4f luminescent mechanism enables Eu2+ complexes to have great potential in high-performance organic light-emitting diodes (OLEDs), a technology has successfully been commercialized in cutting-edge displays and is under developing in solid-state lighting. To reach 100% theoretical exciton utilization efficiency (EUE), which is the key parameter to enhance the energy efficiency, phosphorescence16,17, thermally activated delayed fluorescence (TADF)18,19 and organic radical materials20 were discovered in succession and applied as emitters in OLEDs. Comparing with the traditional f–f transition and other currently used emitters, divalent europium compounds have the following significant advantages: (i) short decay lifetime: the f–f transition is spin-forbidden with long lifetimes up to milliseconds, strongly limiting their maximum luminance, while 5d–4f transition is spin-allowed with typical lifetimes in nanosecond scale, which significantly reduce the excited-state quenching to reach higher luminance and lower efficiency roll-off, (ii) high EUE: Eu2+ ion exhibits a unique transition for the open-shell electron from 4f65d1 to 4f7, which can harvest 100% exciton energy theoretically20,21, (iii) easily tunable emission by varying coordinate environment: the 5d orbitals are sensitive to the ligand field while the 4f orbitals, effectively shielded by 5s5p, are not sensitive to surroundings22,23,24,25. (iv) high abundance: europium has crustal abundance of 10−6 wt, much higher than the noble metals (Ir, Pt) used in commercial OLEDs. Thus, we believe that 5d–4f transition materials, represented by Eu2+ complexes, will be the next unexplored but promising field in OLED emitters.

Despite the advantages mentioned, Eu2+ complexes are strongly limited by their air stability according to the standard potential φ(Eu3+/Eu2+) = −0.38 V. To the best of our knowledge, there was only one report of OLED device based on Eu2+ complexes, with unsatisfied performance in external quantum efficiency (EQE) of 0.01% and maximum luminance of 10 cd m−2 considering the high photoluminescence quantum yield (PLQY) of the complex to be 85%26. Thus, more efforts must be worked on the rational design of Eu2+ complexes and the deep understanding of the electroluminescent process to boost efficiency and luminance. We propose that the steric effect of cryptate ligands and coordinate interaction could improve the stability of Eu2+ complexes. The steric effect prevents Eu2+ from O2 by a more rigid structure. Improving the coordination interaction between the ligand and Eu2+ can largely enhance the thermodynamic stability. Thus, two ligands, 1,4,7,10-tetraazacyclododecane (N4) and 1,4,7,10,13,16,21,24-octaazabicyclo[8.8.8]hexacosane (N8) are chosen for the design of four Eu2+-containing azacryptates named as EuX2–Nn (X = Br, I, n = 4, 8). Series of crystal analysis, spectral, stability, and theoretical studies were undertaken to reveal the photophysical nature of these Eu2+ complexes. Then EuX2–N8 complexes were chosen as emitters in OLEDs for their high efficiency and good thermal/air stability. As a breakthrough, the optimized device using EuI2–N8 exhibits excellent performance with a maximum EQE of 17.7% and a maximum luminance of 25470 cd m−2.

Results

Synthesis and structural analysis

The four Eu2+ complexes EuX2–Nn (X = Br, I, n = 4, 8) were synthesized in glovebox by mixing EuX2 and corresponding ligands in methanol3,27. The purified products were identified by elemental analysis. Then, single-crystal X-ray diffraction (SCXRD) was performed to investigate the coordinate geometry of these Eu2+-containing azacryptates (Fig. 1). EuBr2–N4 crystallizes in space group P21/n and one unit contains two azacryptate cations, four bromide ions in the outer sphere and four methanol. The N4 ligands have two possible conformations, 50% for each. Thus, the Eu2+ center, coordinated by eight nitrogen atoms from two ligands, adopts an unusual geometry with averagely half in square antiprism and half in distorted cube. Likely, the same coordinate geometry is found in EuI2–N4, which crystallizes in a higher-symmetry space group of Cmca without solvent. There are two sets of [Eu(N4)2]2+ with different orientations in one cell, locating in the edge center and body center, while the eight iodide ions intersperse therein (see Supplementary Fig. 1). The crystal structures of EuX2–N8 show that the center Eu2+ is coordinated by eight nitrogen atoms and one halide ion as a distorted “hula-hoop” geometry, with the other halide in the outer sphere as a counterion, which are consistent with the reported structures1,3. As shown in Table 1, the bond lengths of Eu–N in EuX2–N4 are relatively shorter than those in EuX2–N8, indicating the N4 complexes have stronger coordinate interaction between Eu2+ and ligands. Considering the charge separation in crystals, the EuX2–N4 compounds behave more like ionic crystals with relatively stronger electrostatic attraction between the counterion halides and the [Eu(N4)2]2+ ions.

Fig. 1: The crystal structure of four Eu2+ complexes.
figure 1

ORTEP drawings of the crystal structures of (a) EuBr2–N4, (b) EuI2–N4, (c) EuBr2–N8, and (d) EuI2–N8, respectively. e The coordination polyhedrons (from top to bottom): square antiprism in EuX2–N4, distorted cube in EuX2–N4, distorted “hula-hoop” in EuBr2–N8 and EuI2–N8. The nitrogen atoms are labeled from N1–N8 for the data in Table 1. The solvent methanol (in EuBr2–N4 and EuI2–N8) and all the hydrogens are omitted for clarification. Atom notation: Eu (cyan), C (gray), N (blue), Br (brown), I (purple).

Table 1 The bond lengths (distances) around Eu2+ center in EuX2–Nn.

Photophysical properties

To systematically study the photophysical properties of Eu2+ complexes, steady-state spectra and transient spectra were collected. Crystalline powder of EuX2–N4 shows orange-red emissions with maximum wavelength (λmax) of 605 nm (X = Br) and 613 nm (X = I), respectively (Fig. 2). Changing the azacryptates from N4 to N8, the EuX2–N8 complexes exhibit strong hypsochromic shift induced by the weaker crystal field of N8 ligands, with λmax of 510 nm (X = Br) and 515 nm (X = I). The lifetimes for these complexes were found to be hundreds of nanoseconds (Table 2, Supplementary Fig. 2), within the expected range for 5d–4f transition3,28. Full widths at half maximums (FWHMs) for these complexes in solid powder (40–45 nm) are relatively narrow comparing with luminescent materials featuring in charge-transfer (CT) mechanisms. The excitation bands of these complexes are broad and featureless, ranging from 230 to 500 nm (EuX2–N8) and 230 to 600 nm (EuX2-N4) as shown in Supplementary Fig. 3. Based on aforementioned photophysical studies and considering that the ligands in our system are saturated organic compounds with high-energy levels, it is reasonable to rule out the possibilities of ligand-metal charge transfer (LMCT). Hence, the excitation and emission processes can be regarded as the electronic transitions in Eu2+ ion, where the ground state is 4f7 [8S7/2] and the excitation state is 4f6[7F0]5d1, specifically.

Fig. 2: The photophysical properties of EuX2–Nn compounds in solid and solution.
figure 2

a The emission and (b) decay spectra of solid EuX2–Nn (X = Br, I, n = 4, 8). c The excitation (Ex), emission (Em), and (d) decay spectra of EuX2–N8 in methanol solution (1.5 mM). e The emission spectra of EuI2–N8 and (f) the decay spectra of EuI2–N8 of crystals (C), ground samples (G), and powder (P).

Table 2 The summary of photoluminescent properties of the four EuX2–Nn compounds.

Due to the insolubility of EuX2–N4 in common solvents, we only studied the photophysical properties of EuX2–N8 in methanol solution (1.5 mM) under N2 atmosphere. The EuX2–N8 solutions show bright yellow emission with λmax of 579 nm, and the emission spectra of the two complexes are almost identical. The emission is red-shifted by about 70 nm comparing with their solid samples, which is presumed to be caused by the differences in conformation of N8 ligand in solid and solution, outweighing the effect of different halogens. As shown in Fig. 2c, the excitation bands of two compounds are similarly located at 280 nm and 410 nm, attributed to the transition from 4fz3 to 5dz2 and from 4fz3 to 5dxy, respectively29,30. The UV–visible spectra (Supplementary Fig. 4) show that EuX2–N8 complexes have high-energy absorption around 250 nm (ε > 1000 L mol−1 cm−1) and low-energy absorption peak at 404 nm (ε = 644 L mol−1 cm−1, X = Br) and 405 nm (ε = 512 L mol−1 cm−1, X = I), respectively, which is consistent with their excitation bands. The large molar absorptivity is on par with the reported Eu2+ complexes due to the Laporte- and spin-allowed nature of fd transition10,11,31. The time-dependent density functional theory (TD-DFT) calculation was conducted for EuX2–N8 and EuBr2–N4. The calculation prediction of EuX2–N8 is very close to the experimental data. For the N4 complex, EuBr2–N4 has two possible conformations (high symmetry: cubic geometry and low symmetry: square antiprism), and calculation result suggests that the different conformations exhibit distinct absorption bands.

Interestingly, EuI2–N8 exhibits mechanochromic property, showing a fluorescence color change from green to yellow under moderate mechanical grinding in Fig. 2e. The emission spectra show that a new peak emerges in a longer wavelength region after grinding. Then the excitation and transient spectra of 515 nm and 580 nm in the ground sample were collected to probe possible explanation as shown in Supplementary Fig. 5 and Fig. 2f. The similar excitation characteristics and a slightly longer decay of the new peak at 580 nm infer that the longer-wavelength emission is still from Eu2+ center at a marginally different coordination environment. We tentatively attribute the mechanochromic property to the change in ligand conformation upon grinding, resulted from the relatively weak lattice energy of EuI2–N8. And the reversible process is essentially recrystallization in certain solvent atmosphere, like methanol32,33,34. Furthermore, the longer-wavelength emission can be pronouncedly enhanced by fast precipitation in antisolvent (tetrahydrofuran) to get amorphous powder of EuI2–N8. As shown in Fig. 2f, the emission shifts to a longer wavelength of 560 nm and the decay lifetime also increases from crystalline to amorphous state. However, a similar phenomenon was not observed in EuBr2–N8 (Supplementary Fig. 6), which indicates that EuBr2–N8 has a higher lattice energy, so it is harder to change the ligand conformation by such small mechanical stimulation.

Thermal and air stability

Thermal properties of these four compounds are studied by thermogravimetric analysis (TGA), which is of great significance for their further applications in OLEDs. The deposition temperature (Td, corresponding to 5% weight loss) are around 270 °C, 265 °C, 393 °C, and 436 °C for EuBr2–N4, EuI2–N4, EuBr2–N8, and EuI2–N8 in Fig. 3a, respectively. After 550 °C, the unchanged residue weight percentages of these compounds should be the mass percentage of metal halides, for that the decomposition process is tentatively attributed to the break of coordinate bonds followed by sublimation of organic ligands. Then the relative error (RE) of residue weight is calculated to verify that speculation. As shown in Supplementary Table 2, the REs of EuX2–N4 are reasonably low (~3%) while the REs of EuX2–N8 are too high (~6% and ~10% for X = Br, I, respectively). The element analysis is employed to exclude the possibility of impurities in EuX2–N8. Thus, we believe these EuX2–N8 compounds undergo both decomposition and sublimation around Td, which will result in a large deviation in the final weight percentage.

Fig. 3: The thermal properties and buried volume calculation of EuX2-Nn compounds.
figure 3

a Thermogravimetric analysis (TGA) of EuX2–Nn, where the 95%-weight line is shown to determine the Td. b The change in PLQY of EuX2–N8 complexes in air. c Buried volume %Vbur calculated of EuBr2–N8 (77.0%). d Buried volume %Vbur calculated of EuI2–N8 (75.1%).

Then the sublimation properties of these compounds were tested under high vacuum of 10−5 Pa and gradient heating. The EuBr2–N8 was found to be completely sublimable around 320 °C (tube temperature, which is different from the sample temperature) at a small scale of 50 mg. It is notable that there will be obvious decomposition at large-scale sublimation, probably due to the uneven heating in the sublimation boat. A similar property is found for EuI2–N8 with a higher sublimation temperature at 350 °C.

Considering the high thermal stability and near-unity PLQY, the N8 complexes are potential candidates used in OLEDs as emitters. However, Eu2+ ion is traditionally known to be easily oxidized to Eu3+ by O2, hence the air stability is a critical parameter in terms of further applications. To shed light on their air stability at room temperature, the PLQYs of EuX2–N8 were measured as the function of time. As shown in Fig. 3b where the PLQY value change reflects their respective air stability, the quantum yield of EuBr2–N8 does not change after exposure in air over 450 h and EuI2–N8 is metastable towards air. To explain the differences in stability for future design of Eu2+ complexes, the analysis of buried volume (%Vbur) was calculated to estimate the steric protection by N8 ligands as shown in Fig. 3c, d35,36. %Vbur is defined as the fraction of volume of ligand over the total volume of sphere centered on the metal. It determines the steric effect of a given ligand regard to the first coordination sphere (Supplementary Fig. 7). The two complexes all exhibit high %Vbur values (77% to 75.1% for X = Br, I, respectively). EuBr2–N8 has a slightly higher %Vbur than EuI2–Nn due to the closer distance between Eu2+ and Br-. Thus, the divergence in air stability between these two N8 complexes is related to the different lattice energy considering the similar %Vbur. The EuI2–N8 has smaller lattice energy than EuBr2–N8 because of weaker static interaction and the existing of solvent methanol in crystal, which is also applied to explain their different mechanochromic behaviors.

Electroluminescence performance

Based on the photophysical and stability studies, the EuX2–N8 complexes are better candidates used in OLEDs as emitters. Prior to device fabrication, the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) energy levels of the two complexes were deduced from their ultraviolet photoelectron spectroscopy (Supplementary Fig. 8) and ultraviolet absorption spectra data. Then, great efforts have been devoted to optimizing the device structure due to the lack of experiences of Eu2+ complexes used in OLEDs. The EuBr2–N8 was first chosen for device optimization, which includes screening host materials, finding the best combination of hole transporting layer (HTL) and electron transporting layer (ETL), adjusting the thickness of the emission layer in Supplementary sections 14. Then, we followed the optimized conditions and further adjusted the doping concentration and the thickness of the emission layer of the EuI2–N8 device in Supplementary sections 56. The details of materials used, device optimization, and performance are shown in Supplementary Figs. 917 and Supplementary Tables 38.

On the base of the aforementioned process, the optimized OLED structure is ITO/MoO3 (2 nm)/N,N′-bis(1-naphthalenyl)-N,N′-bis-phenyl-(1,1′-biphenyl)-4,4′-diamine (NPB, 50 nm)/cyclohexylidenebis[N,N′-bis(p-tolyl)aniline] (TAPC, 10 nm)/EuX2–N8:4,4′,4″-tris[phenyl(m-tolyl)amino]triphenylamine (m-MTDATA, 25 nm)/diphenyl[4-(triphenylsilyl)phenyl]phosphine oxide (TSPO1, 10 nm)/4,7-diphenyl-1, 10-phenanthroline (Bphen, 30 nm)/LiF (0.7 nm)/Al. The best EuBr2–N8 device gives pretty good performance with a turn-on voltage (Von) of 6.2 V, a maximum luminance (Lmax) of 10,200 cd m−2, a maximum current efficiency (CEmax) of 52.8 cd A−1 and a maximum EQE of 15.5%. While the champion device is obtained by using EuI2–N8 as the emitter, the Von, Lmax, CEmax, and EQEmax are 6.5 V, 25,470 cd m−2, 62.4 cd A−1, and 17.7%, respectively. These results shown in Fig. 4 far exceed the only previously reported Eu2+-based OLEDs, with EQEmax of 0.01%, Lmax of 10 cd m−2 and Von of 20 V26. In addition, the electroluminescence lifetimes of the champion device are determined at 2.5 mA cm−2 and 10 mA cm−2, which are little shorter than tris(2-phenylpyridine)iridium (Ir(ppy)3) based control device, a well-studied phosphorescence emitter in the same spectral region (Supplementary Fig. 18). It should be noted that both devices showed very short lifetime, since which is not only related to the emission material, but also host material, charge transport material, device fabrication, seal conditions, and so on37.

Fig. 4: The device structure and electroluminescence properties of EuX2-N8 compounds.
figure 4

a The optimized OLEDs device structure with frontier orbital energy levels of all organic materials and corresponding thickness. The HOMO and LUMO level of EuI2–N8 are noted as green dash lines in the EML. b Electroluminescence spectra of the champion device at varying voltage from 7 V to 13 V. c Current efficiency–luminance–external quantum efficiency (CE–L–EQE) curve of the champion device. d Current density–voltage–luminance (J–V–L) curve of the champion device.

It is notable that the EuX2–N8 devices have rather high Von values over 6 V considering the bandgap of host material is only 3 eV (~400 nm). To understand this phenomenon, the photophysical properties of films fabricated by doping 10 wt% EuX2–N8 in m-MTDATA onto quartz substrates in a vacuum chamber at high vacuum (10−5 Pa) were studied. The pure films of EuX2–N8 were also fabricated as reference. The emission of doping films is mainly located as two bands, where the 400–470 nm band (τ~1.2 ns) is attributed as fluorescence from host materials and the 500–650 nm band (τ~102 ns) is from EuX2–N8 in Supplementary Fig. 19. The excitation spectra show that in pure film, the two main excitation bands are located at around 320 nm and 390 nm, which is close to the studies showed in solution and solid state. The two doping films have almost identical excitation band at 350 nm, indicating the photoenergy first excites host materials and then transfers to doping Eu2+ complexes, without complete energy transfer in the photoluminescence process. Intriguingly, the electroluminescence spectra only exhibit emission from Eu2+ complexes with doping concentrations (7 wt%) lower than those in the photoluminescence study. Thus, we tentatively propose that the carrier recombination dominantly occurs in the doping materials instead of host materials, where the excitation of ligand results in a high Von.

Discussion

In summary, four Eu2+-containing azacryptates EuX2–Nn (X = Br, I, n = 4, 8) were synthesized, showing promising photoluminescent properties: high PLQY (~100% for N8 complexes), short excited-state lifetime (102 ns) and easily tunable emission by ligand field. Intriguingly, EuI2–N8 exhibits reversible mechanochromic property under grinding, which is attributed to the potential flexibility of N8 ligand and recrystallization. Furthermore, the EuX2–N8 complexes were chosen as the emissive materials in OLEDs due to their good air-/thermal-stability. After optimization of design, the best device showed excellent performance with a maximum EQE of 17.7% and luminance of 25,470 cd m−2. Our work deepens the understanding of photoluminescence and electroluminescence properties in Eu2+ complexes and proves their promising applications in OLEDs.

Methods

All chemical reagents used in the synthesis process were commercially available and were used as received unless otherwise mentioned. The N4 ligand was commercially available. 1H-NMR spectra were recorded on a Bruker-400 MHz NMR spectrometer. Tetramethylsilane (TMS) was used as an internal reference for the chemical shift correction, where δ(TMS) equals 0. Elemental analyses were performed on a VARIO EL analyzer (GmbH, Hanau, Germany). All the synthesis of Eu2+ complexes was conducted in glovebox. All spectral tests of solid Eu2+ complexes were carried out by paraffin encapsulation between two quartz plates and the solution was protected by capped cuvettes under N2 atmosphere. The commercially available paraffin was purified by oxidation using KMnO4 and column chromatography to remove fluorescent whitening agents.

Synthesis

1,4,7,10,13,16,21,24-octaazabicyclo[8.8.8]hexacosane (N8 ligand): The synthesis of N8 ligand is carried out by an improved version of a reported method38. Tris(2-aminoethyl)amine (4.9 g, 33.5 mmol), NEt3 (12 mL), and 2-propanol (250 mL) were added to a 2-neck 1-L round-bottom flask equipped with mechanical stirring and a drip funnel containing a dilute solution of glyoxal (7.5 g). The flask was cooled to −78 °C and the glyoxal solution was added slowly (1 drop s−1). After the completion of addition, the yellow solution was stirred at room temperature overnight. Then the solvent was removed under vacuum at 40 °C, yielding a yellow solid which was dispersed in 300 mL CHCl3 and stirred for 2 h with the generation of lots of yellow translucent gels. The gels were removed by filtration and the resulting CHCl3 was removed under vacuum at 40 °C. The crude intermediate was dissolved in 300 mL MeOH, cooled with ice water. Excess NaBH4 (14 g) was gradually added to the solution to prevent an intensive reaction. The cloudy solution was stirred for 4 h and the solvent was removed under vacuum yielding white solid, which was extracted by CH2Cl2 (200 mL×3). The removal of CH2Cl2 gave the crude product N8. Further purification was conducted by thermal gradient sublimation (160−80 °C) at low pressure (~5 Pa). 1H-NMR (400 MHz, D2O): δ 2.79 (s, 12H), 2.75 (m, 12H), 2.58 (m, 12H).

EuBr2-N4: EuBr2 (78 mg, 0.29 mmol) was dissolved in 6 mL MeOH in a clean glass bottle. N4 (85 mg, 0.50 mmol) was dissolved in 3.5 mL MeOH, which was slowly added to the EuBr2 solution without stirring. The colorless solution turned orange-red and red crystals suitable for SCXRD analysis formed as the evaporation of solvent (yield is 85% based on Eu). Elemental analysis for C16H40Br2EuN8, C, 29.28%, H, 6.14%, N, 17.07%. Found: C, 29.05%, H, 6.05%, N, 16.69%.

EuI2-N4: EuI2 (55 mg, 0.14 mmol) was dissolved in 6 mL MeOH in a clean glass bottle. N4 (50 mg, 0.29 mmol) was dissolved in 3.5 mL MeOH, which was slowly added to the EuI2 solution without stirring. The light-yellow solution turned orange-red and red crystals suitable for SCXRD analysis formed soon after the completion of mixing (yield is 70% based on Eu). Elemental analysis for C16H40EuI2N8, C, 25.61%, H, 5.37%, N, 14.93%. Found: C, 25.84%, H, 5.54%, N, 14.42%.

EuBr2-N8: EuBr2 (0.156 g, 0.500 mmol) was dissolved in 10 mL MeOH in a 50-mL round-bottom flask under magnetic stirring. N8 (0.185 g, 0.500 mmol) was dissolved in 10 mL MeOH, which was slowly added to the EuBr2 solution. The colorless solution turned orange. The solvent was removed under reduced pressure to get crude product. (Yield is 86% based on Eu). The complex was further purified by thermal gradient sublimation (320–250–60 °C) at low pressure (10−5 Pa). Green crystals suitable for SCXRD analysis were obtained by slow evaporation of MeOH as a solvent. Elemental analysis for C18H42Br2EuN8, C, 31.68%, H, 6.20%, N, 16.42%. Found: C, 31.75%, H, 6.18%, N, 16.39%.

EuI2-N8: EuI2 (0.260 g, 0.64 mmol) was dissolved in 10 mL MeOH in a 50-mL round-bottom flask under magnetic stirring. N8 (0.260 g, 0.70 mmol) was dissolved in 10 mL MeOH, which was slowly added to the EuI2 solution. The light-yellow solution turned orange and green crystals suitable for SCXRD analysis formed as the evaporation of solvent without stirring (yield is 72% based on Eu). Elemental analysis for C18H42I2EuN8, C, 27.85%, H, 5.45%, N, 14.43%. Found: C, 28.04%, H, 5.41%, N, 14.28%.