HTML
-
The complex Ce-1 was synthesized by stirring potassium hydrotris(3, 5-dimethylpyrazolyl)borate (KTpMe2)22 with Ce(CF3SO3)3 in tetrahydrofuran (THF), accompanied by hydrolysis due to a trace amount of water in the solvent (Fig. 1a). Though the reaction was found accidentally, it is repeatable and stable by adopting the conditions presented in the experimental section. Intriguingly, when KTpMe2 was replaced by potassium trispyrazolylborate (KTp) in this reaction, the expected complex Ce(Tp)3 was successfully obtained. The molecular structure and photoluminescence (PL) spectrum are exhibited in Supplementary Fig. S1. These results indicate that the methyl group on the pyrazole group has a crucial influence on the reaction. Most likely, the huge steric hindrance of the methyl group makes Ce(TpMe2)3 unstable and tend to react with trace amounts of water in the solvent to form hydrolysate Ce-1.
Fig. 1 Synthesis and structure of the investigated cerium(III) complex Ce-1.
a Synthetic route for the complex. b Single crystal structure of the complex shown as ellipsoids at the 50% probability level, where yellow represents Ce, pink represents B, blue represents N, red represents O, grey represents C, and the hydrogens are omitted for clarity. c Single crystal structure of the complex shown in a space-fill style, where hydrogens are shown in whiteThe thermal stability of Ce-1 is favourable, showing a decomposition temperature (Td, corresponding to a 1% weight loss) of 265 ℃ (Supplementary Fig. S2). Hence, the complex Ce-1 was purified by thermal gradient sublimation at 230 ℃ and 2 × 10−4 Pa, during which single crystals were collected. The sublimated Ce-1 was identified by high-resolution mass spectrometry, elemental analysis, and single crystal X-ray diffraction. No 1H NMR data were collected since Ce-1 is paramagnetic, as demonstrated by electron paramagnetic resonance (EPR) spectroscopy (Supplementary Fig. S3). There are two almost identical coordination environments for Ce3+ ions in the crystal. The central Ce3+ ions are coordinated by tridentate and pentadentate ligands with one and two negative charges, respectively. There are seven nitrogen atoms and one oxygen atom surrounding cerium in Ce-1, with Ce–N distance in the range of 2.573–2.680 Å and a Ce–O distance of ~2.399 Å (Supplementary Table S1). Figure 1b, c shows one of the structures in ellipsoid and space-fill styles. From the space-fill model, it is found that the ligands shield the Ce3+ ion completely, which is important for high PL efficiency, since coordinating solvent molecules are frequently efficient quenchers of the lanthanide luminescence23.
-
Figure 2a shows UV–Vis absorption spectra of Ce-1 and KTpMe2 in DCM solution. The three bands at 373, 322, and 280 nm with a molar extinction coefficient (ε) of ~102 L mol−1 cm−1 can be assigned to the 4f–5d transitions of the Ce3+ ion by comparing the absorption spectra of Ce-1 and KTpMe2. Another absorption band below 260 nm originates from the π–π* transition of the ligand. The excitation spectrum of Ce-1 in DCM (Supplementary Fig. S4) has a similar pattern to its absorption spectrum. The peak at ~240 nm is attributed to ligand absorption, exhibiting cascade-type energy transfer from the ligand to the central Ce3+ ion under high-energy excitation. Due to parity-allowed characteristics, the complex can also be easily excited by direct f–d absorption (the peak at ~375 nm), while f–f excitation is difficult in a dilute solution of a typical Eu(III) complex due to the parity-forbidden characteristics of f–f absorption.
Fig. 2 Photophysical properties of Ce-1 in solution and as crystalline powder.
a UV–Vis absorption spectra of Ce-1 and KTpMe2 in solution (DCM, 10−5 M), and PL spectra of Ce-1 in solution (DCM, 10−5 M) and as crystalline powder at room temperature (RT) and 77 K. The excitation wavelength is 370 nm. Inset, photographs of the complex in solution and as crystalline powder under 365 nm irradiation. b Transient PL decays of the complex in solution and as crystalline powder at room temperature and 77 K. The excitation wavelength is 375 nm. Inset, schematic diagram to illustrate the PL mechanism of the cerium(III) complex, where Abs, IC, and PL represent absorption, internal conversion, and photoluminescence, respectively. 2D3/2 is the lowest excited state of the Ce3+ ion. 2F5/2 and 2F7/2 are two ground levels of the Ce3+ ionThe dilute DCM solution and crystalline powder of Ce-1 both exhibit strong blue emission. Their PL spectra (Fig. 2a) show two broad bands with different relative amplitudes. The redshift can be ascribed to molecule packing in crystals, while the relative amplitude may be sensitive to the coordination environment24. At 77 K, the emission spectrum of the crystalline powder splits into two peaks at 436 and 477 nm. The energy difference between the two peaks is close to 2000 cm−1, in agreement with the energy splitting between 2F5/2 and 2F7/2, two ground levels of the Ce3+ ion20. Moreover, the decay lifetimes of the solution at room temperature, crystalline powder at room temperature and crystalline powder at 77 K were measured to be approximately 53, 42, and 42 ns (Fig. 2b), respectively, which are comparable with those of previously reported cerium(III) complexes12-14. In addition, the emissions of Ce-1 in solvents with different polarities showed almost identical spectra (Supplementary Fig. S5). This illustrates that unlike the famous radical emitter TTM-3NCz11, the electron transition in Ce-1 is highly localized.
The PLQYs of Ce-1 in dichloromethane (DCM, 10−5 M) and as crystalline powder are 48% and 82%, respectively. It should be noted that the PLQYs of Ce-1 in DCM can be promoted by increasing the concentration, reaching ~100% at 10−3 M (Supplementary Table S2). By combining the lifetime and PLQY values, the radiative and nonradiative decay rates (kr and knr) are calculated as 9.1 × 106 and 9.8 × 106 s−1 for the DCM solution at 10−5 M and as 2.0 × 107 and 4.3 × 106 s−1 for the crystalline powder, respectively. The higher PLQY in the solid state is attributed to both an increased radiative rate constant and a decreased nonradiative rate constant, indicating that rigid molecular packing is beneficial for promoting radiative transition and inhibiting nonradiative transition. In addition, these values are comparable to those of organic radical materials and an order of magnitude larger than those of phosphorescence and TADF materials.
To obtain more in-depth insights into the excited state of Ce-1, theoretical calculations were carried out (Supplementary Fig. S6). The donor and acceptor for the first symmetry allowed transition were recognized as the 4f and 5dz2 orbitals, respectively. This result further proves the transition mechanism of Ce-1. Therefore, the observed strong blue emission can be attributed to the Ce3+ ion, more specifically to the two electric-dipole 5d–4f transitions of the Ce3+ ion from the lowest excited state (2D3/2) to the ground state 2F5/2 and 2F7/2, as shown in the inset of Fig. 2b.
-
The Ce-1 complex shows high PLQY, excellent thermal stability, good stability in air (Supplementary Fig. S7), and short excited-state lifetime; it is worth investigating as the emitter in OLEDs with a vacuum thermal deposition method. The frontier molecular orbital energy levels of Ce-1 were obtained as −6.2 eV for the highest occupied molecular orbital (HOMO) and −3.1 eV for the lowest unoccupied molecular orbital (LUMO) by investigating the ultraviolet photoelectron spectrum (Supplementary Fig. S8) and UV–Vis absorption spectrum. To optimize the OLED structure, the theoretical equation for calculating the EQE of OLEDs should be considered
$$ \mathrm{EQE}=\mathrm{ŋ}_{\mathrm{e}-\mathrm{h}} \times \mathrm{ŋ}_{\mathrm{PL}} \times \mathrm{ŋ}_{\mathrm{out}} $$ (1) where ŋe–h is the recombination efficiency of injected holes and electrons, ŋPL is the intrinsic PL efficiency, i.e., the PLQY of the emission layer, ŋexciton is the EUE, i.e., the ratio of radiative excitons to total formed excitons, and ŋout is the light out-coupling efficiency. Based on the equation, the key is to screen host and carrier transport materials, hence achieving high PLQY in thin Ce-1 film, and to balance and confine hole and electron recombination within/near the emission layer.
First, bis[4-(N-carbazolyl)-phenyl]phenylphosphine oxide (BCPO) was chosen as the host due to its high triplet energy level and excellent bipolar carrier mobility. Device D1 was fabricated with a structure of indium tin oxide (ITO)/MoO3 (2 nm)/MoO3-doped 9-(4-tert-butylphenyl)-3, 6-bis(triphenylsilyl)-9H-carbazole (CzSi:MoO3, 20 wt%, 10 nm)/CzSi (30 nm)/BCPO:Ce-1 (10 wt%, 20 nm)/diphenyl-4-triphenylsilylphenyl-phosphine oxide (TSPO1, 10 nm)/2, 2′, 2′′-(1, 3, 5-benzinetriyl)-tris(1-phenyl-1-H-benzimidazole) (TPBi, 40 nm)/LiF (0.7 nm)/Al (100 nm). The frontier orbital energy levels of the utilized materials25-27 are shown in Fig. 3a. The device characteristics are plotted in Fig. 3b, c. The device emits deep-blue light similar to the Ce-1 solution with CIE coordinates of approximately (0.15, 0.08) (Supplementary Fig. S9). Considering that the doped BCPO:Ce-1 (10 wt%) film has a moderate PLQY of 26%, the device shows a relatively good maximum EQE of 5.1% and low efficiency roll-off (Fig. 3b). To estimate the EUE of D1 more accurately, we used the variable-angle spectroscopic ellipsometry (VASE) method to determine the molecular orientation distribution in the emission layer (Supplementary Fig. S10 and Table S3)28. The Ce-1 molecules show a horizontal dipole ratio of 68.1% in the BCPO film, which is close to the randomly oriented ratio (66.7%)29, 30, indicating ~20% ŋout, and hence, an ~100% EUE is demonstrated in D1. Thus, we expect that the performance of our deep-blue OLEDs could be further improved by screening the host material to obtain a higher PLQY.
Fig. 3 OLED demonstration.
a Frontier orbital energy levels for materials used in OLEDs; the energy levels of Ce-1 were deduced from its ultraviolet photoelectron spectrum and ultraviolet absorption spectrum. b Current density–voltage–luminance traces for the key OLEDs. c Power efficiency–luminance–EQE traces for the key OLEDs. d Maximum EQE histogram of 18 devices with the optimized device structure (average, ~12.4%; highest, 14.0%). e PL spectrum of the TSPO1:CzSi:Ce-1 (0.18:0.72:0.1 in weight ratio, excitation 280 nm) film, and EL spectrum of D2 at the luminance of 100 cd m−2. f Transient EL spectrum of D2Therefore, a series of Ce-1-doped films were fabricated in a vacuum chamber and subsequently measured for PLQYs and PL spectra (Supplementary Fig. S11 and Fig. 3e). Among these films, the TSPO1:Ce-1 (10 wt%), CzSi:Ce-1 (10 wt%), and TSPO1:CzSi:Ce-1 (0.18:0.72:0.1 in weight ratio) films show relatively high PLQYs up to 70%, 83%, and 93%, respectively. With substantial device optimization (see Supplementary Table S4 for details), a champion performance was obtained for D2 with the structure of ITO/MoO3 (2 nm)/CzSi:MoO3 (20 wt%, 30 nm)/CzSi (10 nm)/TSPO1:CzSi:Ce-1 (0.18:0.72:0.1 in weight ratio, 20 nm)/TSPO1 (10 nm)/bathophenanthroline (Bphen, 40 nm)/LiF (0.7 nm)/Al (100 nm). The device characteristics are plotted in Fig. 3b, c for comparison. The champion performance includes a turn-on voltage of 3.6 V, a high maximum EQE of 14.0%, and a typical emission from Ce-1 with CIE coordinates of (0.146, 0.078). This result is comparable to the best reported deep-blue OLEDs with a platinum complex, an iridium complex or TADF as the emitter (Supplementary Table S5)31-33. To estimate the EUE of D2, we measured the horizontal dipole ratio of Ce-1 in TSPO1:CzSi (0.18:0.72 in weight ratio) as only 56.0% (Supplementary Fig. S10 and Table S3), which is even lower than that of the random orientation (66.7%), leading to an ŋout lower than 20%. Considering the PLQY of 93% and EQE of 14.0%, the EUE of D2 is deduced to be higher than 70% (Table 1).
Device Host Vona [V] EQEmaxb [%] CEmaxc [cd A−1] Lmaxd [cd m−2] CIEe D1 BCPO (0.148, 0.083) Averagef 3.6 5.0 3.4 3457 Champion 3.6 5.1 3.5 3494 D2 TSPO1:CzSi (0.146, 0.078) Averageg 3.8 12.4 8.7 948 Champion 3.6 14.0 10.3 1008 aTurn-on voltage, taken as the reference point at which the luminance is 1 cd m−2
bMaximum EQE
cMaximum current efficiency
dMaximum luminance
eCoordinates at 100 cd m−2
fAverage of three devices
gAverage of 18 devicesTable 1. Summarized parameters of key OLEDs with Ce-1 as the emitter
The details of carrier recombination are critical for explaining the high EUE (100% for D1 and > 70% for D2) of these devices. As depicted in Fig. 3e, Supplementary Figs. S9 and S11, the steady-state PL spectra of emission layers contain deep-blue emissions that are substantially identical to the solution emission of Ce-1 and weak ultraviolet emissions from host materials, indicating inefficient energy transfer from hosts to Ce-1. In contrast, the EL spectra of these devices exhibit pure emissions from Ce-1, implying that carrier recombination dominantly occurs on Ce-1 rather than on host molecules. This inference can be further confirmed by the transient EL spectrum of D2 (Fig. 3f). The spike (in the red dashed square) is proven to be evidence of charge trapping on the guest34. Interestingly, the response time of the EL intensity to the applied electric field is slow, implying that there may be a slow process of charge migration and carrier trapping in D2. After the polarity of the electric field is reversed, the remaining carriers in the device de-trap and continue to recombine on the guest, resulting in a microsecond delay that is much longer than the excited-state lifetime of Ce-1. Moreover, the absorption edge of the ligand is ~250 nm, corresponding to a bandgap of ~5 eV. However, the turn-on voltages of D1 and D2 are in the range of 3.6–3.8 V, which are much lower than the bandgap of the ligand. This implies that carriers recombine on the Ce3+ ions with a narrower bandgap instead of on the ligands. The speculated EL mechanism is exhibited in Fig. 4. Thus, doublet exciton formation is dominant in these OLEDs. Since Ce(III) ions are more likely to lose only one electron from the 4f orbital to form Ce(IV), the bottom route would be the main route for exciton formation in the device.
Fig. 4 Speculated electroluminescence mechanism.
The electron or hole is captured by a Ce(III) ion to form Ce(II)* (top route) or Ce(IV) (bottom route); then, hole or electron injection brings the intermediate species to the excited Ce(III)* ion stateAlthough device D2 shows high efficiency, the maximum luminance of D2 is only 1008 cd m−2, and the efficiency roll-off is severe, which is contrary to the common knowledge that a short excited-state lifetime emitter is beneficial for reducing efficiency roll-off. In addition, the lifetime (LT50) of D2 was measured as only 147 s (Supplementary Fig. S12). To obtain more in-depth insights into the device efficiency roll-off and stability, we measured device D2 over three cycles and found that the device performance rapidly degrades (Fig. 5), indicating that the efficiency roll-off should be ascribed to the instability of the device. Furthermore, we introduced a florescence deep-blue emitter, 1-(10-(4-methoxyphenyl)anthracen-9-yl)-4-(10-(4-cyanophenyl)anthracen-9-yl)tetraphenylethene (TPEA)35, and a classic host material, 1, 3-bis(N-carbazolyl)benzene, as references to fabricate control devices (Supplementary Fig. S12). Detailed results indicate that the instability of the host material, especially TSPO136-38, is an important reason for device degradation. In addition, the Ce(III) complex is known to be a strong photo-reductant. This could be another reason for the device instability. Thus, the stability of Ce-1 in terms of EL remains to be explored.