84% efficiency improvement in all-inorganic perovskite light-emitting diodes assisted by a phosphorescent material

A novel mixed perovskite emitter layer is applied to design all-inorganic cesium lead halide perovskite light-emitting diodes (PeLEDs) with high electroluminescence (EL) performance, by combining CsPbBr3 with iridium(iii)bis[2-(4′,6′-difluorophenyl)pyridinato-N,C2′]-picolinate (FIrpic), where FIrpic is a phosphorescent material with very high internal quantum efficiency (IQE) approaching 100%. The CsPbBr3:FIrpic PeLEDs show a maximum luminance of 5486 cd m−2, and an external quantum efficiency of 0.47%, which are 1.84 and 1.76 times that of neat CsPbBr3 PeLEDs, respectively. It is found that FIrpic molecules as an assistant dopant can efficiently transmit energy from the excitons of FIrpic to the excited state of the CsPbBr3 emitter via a Förster energy transfer process, leading to enhanced EL efficiency in the CsPbBr3:FIrpic PeLEDs.

In this work, FIrpic is adopted as an assistant to form CsPbBr 3 :FIrpic composite lm. The PeLEDs based on this composite emissive layer can harvest both singlet and triplet excitons. And a maximum luminance of 5486 cd m À2 , a maximum current efficiency of 1.80 cd A À1 , and a maximum EQE of 0.47% are obtained.

PeLEDs fabrication
Patterned indium-tin oxide (ITO) glass substrates were cleaned successively using deionized water, ethanol, and acetone, and dried in an oven at 100 C for 10 min. Aer 5 min treatment with ultraviolet (UV)-ozone plasma, PEDOT:PSS was spin-coated onto ITO glass substrates at 4500 rpm for 40 s and annealed in air at 120 C for 20 min. Then, the samples were placed in a nitrogen-lled glovebox for 30 min to cool down. The perovskite lm was prepared by one-step spin-coating the perovskite precursor solution at 4000 rpm for 60 s and placed in a low vacuum environment of À1 bar for 20 min to remove residual DMSO solvent. Finally, TPBi (65 nm), Liq (2.5 nm), and Al (120 nm) were sequentially deposited by vacuum thermal depositing system, under a high vacuum (#5 Â 10 À4 Pa). Preparation of the perovskite lms, and encapsulation of PeLEDs were carried out in a nitrogen-lled glove box. The active area of the device was 6 mm 2 .

Measurements and characterizations
We used the UV-vis spectrophotometer (Shimadzu UV-2600) and Fluorolog-3 luminescence spectroscopy to measure the absorption and PL spectra, respectively. The XRD pattern and time-resolved PL spectra were acquired using a TD-3500 X-ray diffractometer and uorescence spectrometer (Fluorolog-3) severally. The PR670 spectrophotometer was used to collect the EL spectra. The current density-luminance-voltage (J-L-V) characterizations and stability of all PeLEDs were carried out by a light-emitting diode measurement system, including a Keith-ley2400 source meter, and a calibrated Si photodiode (Photoelectric Instrument Factory of Beijing Normal University, ST-86LA). All measurements were carried out at room temperature under ambient conditions.  Fig. 1(a) and (b). Device B with FIrpic exhibits better EL performance than Device A with higher current density, luminance, current efficiency and EQE at each applied voltage, and shows maximum luminance of 5486 cd m À2 , maximum current efficiency of 1.80 cd A À1 , and maximum EQE of 0.47%, which are 1.84, 1.76 and 1.76 times to that of the Device A (2974 cd m À2 , 1.02 cd A À1 , and 0.26%), respectively. The corresponding characteristics data about the repeatability of the devices are summarized in Table S1. † The Device B emits green light from only CsPbBr 3 despite of FIrpic (peaked at 470 nm) 22 with high color purity of FWHM $18 nm and a peak intensity located at 522 nm, as shown in Fig. 1(c). In the right inset of Fig. 1(c), it could be found that the shapes of EL spectra These results reveal that the energy transfer from FIrpic to CsPbBr 3 was efficient and complete. And the detailed mechanism to improve the EL performance in Device B are investigated by the following sections.

Characterization of perovskite lm
In order to investigate the physical mechanism of the enhanced EL performance in the PeLEDs, the XRD pattern, absorption and PL spectra, and the time-resolved PL spectra of the CsPbBr 3 lms FIrpic (0.5 mg ml À1 ) are exhibited in Fig. 2(a). In Fig. 2(a), by comparing the intensity changes of peaks at 15.7 and 31.1 assigned to the (101) and (202) planes of perovskite structure in the XRD pattern of both lms, it can conclude that peaks of CsPbBr 3 lm with FIrpic are sharper and more intense, which demonstrate that the presence of FIrpic can induce the crystal growth along the (101) and (202) planes and may do good to the EL performance. 8,10,13 And the XRD patterns match well with orthorhombic crystal structure. 13 The top-view and crosssectional SEM images of the neat CsPbBr 3 lm and the CsPbBr 3 :FIrpic lm (0.5 mg ml À1 ) are shown in Fig. S3 and S4, † respectively. It can be found that coverage is improved in the CsPbBr 3 :FIrpic lm. And the thickness of both the neat CsPbBr 3 lm and CsPbBr 3 :FIrpic lm (0.5 mg ml À1 ) are estimated to $30 nm. UV-vis absorption and PL emission spectra of CsPbBr 3 lm with 0 and 0.5 mg ml À1 FIrpic are plotted in Fig. 2(b). The two kinds of lms exhibit similar shapes of absorption spectra (centered at $518 nm) and PL spectra (peaked at $525 nm with a narrow FWHM of $18 nm). It is worthwhile mentioning that sufficient spectral overlap between the absorption spectra of the CsPbBr 3 emitter and the PL spectra of the assistant dopant FIrpic, which is benecial to energy transfer from FIrpic to CsPbBr 3 . And the shape of PL spectra of CsPbBr 3 :FIrpic lm is almost same to the neat CsPbBr 3 lm without the sub-peak from FIrpic, which may demonstrate that the energy transfer from FIrpic to the CsPbBr 3 are effective and complete. The FIrpic-doped CsPbBr 3 lm presents a much longer PL lifetime than that of the neat CsPbBr 3 lm, which is revealed by the time-resoled PL spectra (Fig. 2(c)). The average lifetime (s avg ) extracted from the PL decay curve for the composite perovskite lm is about 1.69 ns, while the neat CsPbBr 3 has a shorter lifetime of 0.82 ns, which imply that the FIrpic additive can effectively reduce non-radiation recombination and result in enhanced the EL performance of PeLEDs. The PL curves can be tted by the following eqn (1): The average lifetime (s avg ) of the entire decay process can be calculated by the following formula (2): s 1 , s 2 and s 3 are the lifetimes of the three decay components; and A 1 , A 2 , and A 3 are the fractions of the three decay components, respectively. According to the study by Zheng et al., 23 the s 3 (slow) decay component is related to radiative recombination inside the grains, while the s 1 (fast) and s 2 (middle) decay components are attributed to "two kinds of trap-assisted recombination at grain boundaries." The tting parameters are collected in Table 1. It can be found that the proportions of s 1 and s 2 are reduced, which suggests that traps in the FIrpic assisted CsPbBr 3 lm are reduced, due to the effective passivation brought in by the FIrpic assistant. Moreover, the lifetime of s 3 is extended and the its proportion is improved, which

Energy transfer process
The energy level diagram of each layer of Device B is shown in Fig. 3(a), and all energy level values are taken from the literature. 14,[24][25][26] There are two exciton generating interfaces, which are located at the PEDOT:PSS/CsPbBr 3 :FIrpic interface and CsPbBr 3 :FIrpic/TPBi interface, due to no injection barrier for electrons injected from TPBi to CsPbBr 3 :FIrpic lm and high injection barrier for holes at PEDOT:PSS/CsPbBr 3 :FIrpic interface (0.6 eV) and CsPbBr 3 :FIrpic/TPBi interface (0.45 eV). And excitons can also be generated on FIrpic and CsPbBr 3 . The energy transfer diagram is described schematically in Fig. 3(b). As shown in Fig. 3(b), excitons generated in FIrpic with 25% singlets (S 1 ) and 75% triplets (T 1 ). And in FIrpic, the energy of S 1 of FIrpic can be transfer to T 1 through inter system crossing process (ISC). And the energy of both S 1 and T 1 of FIrpic can be transferred to the excited state of CsPbBr 3 via Förster energy transfer process 21 and thus leading to potentially 100% IQE in CsPbBr 3 :FIrpic PeLEDs.

The stability of PeLEDs
In order to further reveal the effect of FIrpic, the half lifetimes of PeLEDs (Devices A and B) are performed and are shown in Fig. 4. The half lifetime is the time duration from the initial luminance of 100 cd m À2 decreasing to the half luminance of the initial luminance. Device B (with CsPbBr 3 :FIrpic lm) exhibit a much longer half lifetime of 147 s, which is 1.6 times than that of Device A (87 s). The better stability of Device B may be benet from reduced current leakage owing to the better coverage 9,12,13,15 and enhanced radiative recombination inside the grain due to the reduced traps, better passivation and higher exciton harvesting efficiency in the in CsPbBr 3 :FIrpic lm compared to that of Device A.

Conclusion
In summary, the enhancements of EL performance and stability can be attributed three factors in the CsPbBr 3 :FIrpic PeLEDs. Firstly, the orientation of the crystallization could be remarkably enhanced by FIrpic along with the reduced traps, better passivation in the CsPbBr 3 :FIrpic lm. Second, FIrpic improves the internal quantum efficiency in CsPbBr 3 :FIrpic, due to almost 100% IQE of FIrpic which can permit efficient transfer of all electrically excitons from the assistant dopant (FIrpic) to the perovskite emitter (CsPbBr 3 ) via Förster energy transfer process. And last, the higher coverage of the CsPbBr 3 :FIrpic benet to reduce current leakage. The CsPbBr 3 :FIrpic composite lm provide a facile method to achieve highly efficient PeLEDs and a new route for advanced light emission applications.

Conflicts of interest
The authors declare no conicts of interest.