Spectra stable blue perovskite light-emitting diodes

Device performance and in particular device stability for blue perovskite light-emitting diodes (PeLEDs) remain considerable challenges for the whole community. In this manuscript, we conceive an approach by tuning the ‘A-site’ cation composition of perovskites to develop blue-emitters. We herein report a Rubidium-Cesium alloyed, quasi-two-dimensional perovskite and demonstrate its great potential for pure-blue PeLED applications. Composition engineering and in-situ passivation are conducted to further improve the material’s emission property and stabilities. Consequently, we get a prominent film photoluminescence quantum yield of around 82% under low excitation density. Encouraged by these findings, we finally achieve a spectra-stable blue PeLED with the peak external quantum efficiency of 1.35% and a half-lifetime of 14.5 min, representing the most efficient and stable pure-blue PeLEDs reported so far. The strategy is also demonstrated to be able to generate efficient perovskite blue emitters and PeLEDs in the whole blue spectral region (from 454 to 492 nm).

can also be calculated that the d of 3D perovskite Rb0.6Cs0.4PbBr3 is around 5.6 Å from the diffraction peak of (1 0 0) and (2 0 0). As expected, the d of <n> = 2 is measured to be 22.4 Å, which fully fits the expected n = 2 quasi-2D layered perovskite structure perovskite reached 104.6 cd m -2 , which is about four-times higher than pristine <n>Rb0.6 = 3. However, luminance is proportional to the product of current density and EQE.
Even though luminance is higher, pristine <n>Rb0.6 = 5 displayed a fairly large current density indicating charge carrier leakage, which results in an overall low EQE of 0.11%.
In addition, the pristine <n>Rb0.6 = 5 devices display a brightness roll-off at around 6 V, indicating strong non-radiative recombination was taking place.

Supplementary Figure 11 | UV-vis absorption spectra of quasi-2D perovskite films.
UV-vis absorption spectra of <n>Rb0.6 = 3 perovskites with different content of excess CsBr. The corresponding CsBr-rich <n>Rb0.6 = 3 perovskites are also synthesized to explore the different passivation effect between CsBr and RbBr. Unfortunately, cation exchange (Rb + replace with Cs + ) is taking place when excessive amount of Cs + introduced due to the more reasonable tolerance factor of Cs + . It means PEA2(Rb0.2Cs0.8)2Pb3Br10 perovskite is formed when we introduced 100% excess of CsBr into pristine <n>Rb0.6 = 3 perovskite. Anyhow, excessive amount of CsBr essentially induce a cation exchange leading to a green emitter, which is not interested in this research. As shown, three dominant diffraction peaks could be observed, and the diffraction peak at 28° can be indexed to crystallographic plane (2 0 2) of quasi-2D perovskite 6 .

Supplementary
With the treatment of excess RbBr, the intensity of (2 0 2) continuous enhanced, indicating that the crystallinity along (2 0 2) plane of RbBr-rich <n>Rb0.6 = 3 is much higher than that of pristine analogue. The highly oriented structure reduced the disorder of the crystal and then suppressed the defect state formation 3 . Thus, the passivation effect of excess RbBr is pluralism, which can also passivate crystal defect by affecting the crystallization process as well as crystal orientation. However, while the content of excess RbBr exceeded 100%, the crystallinity of Rb-Cs alloyed quasi-2D perovskite was significantly reduced, demonstrating that too much RbBr did inhibit the crystallization.

Supplementary Table 3 | The corresponding formation time constant (τet) and
first-order decay (τ1) of the each GSBs shown in Fig. 2 and Supplementary Figure   9. The kinetics are fitted by a multiple-exponential function 9 , Supplementary Figure 7a shows the absorption spectra of the prepared quasi-2D perovskites. As for <n>Rb0.6 ≥ 2, the absorption spectra exhibit several absorption peaks  Figure 7g) were also used to investigate the defect density of <n>Rb0.6 = 3, 4 and 5 perovskites. As shown, the <n>Rb0.6 = 5 perovskite exhibits almost one-order of magnitude higher trap density than <n>Rb0.6 = 3 analogue, due to the larger n-values (n >> 5) perovskite domain which possess higher defect density.
Moreover, though the <n>Rb0.6 = 4 perovskite shows slightly lower trap density than <n>Rb0.6 = 3 analogue, they are still in the same order. Furthermore, the PL lifetime of the perovskites show the same evolutionary trend. The PL lifetime of <n>Rb0.6 = 5 is significantly shorter than <n>Rb0.6 = 3 and 4 perovskites, further demonstrating the serious non-radiative recombination. Thus, the modulation of the n distribution should be more precisely.
Strikingly, another broad GSB peak at approximately 475 to 500 nm, was gradually emerged with increasing delay time. The broad GSB peak refers to a group of higher-  Figures 7b, 9d). In addition, the PLQY obtained from pristine <n>Rb0.6 = 5 perovskite was inferior to <n>Rb0.6 = 3 analogue, deriving from that the larger-n values (n >> 5) perovskite domain possess higher defect density because of their decreased stability (Supplementary Figure 7e, f and g).
Therefore, there are two aspects needed to pay attention to obtain ideal blue emitter according to the above analysis. First, we should modulate the n distribution more precisely. That means we need to increase the population of narrower bandgap species to facilitate the energy transfer, but should avoid forming n >> 5 species. Secondly, surface passivation needs to be conducted to reduce the intrinsic defect density. In brief, the unapparent signal of n = 3 species GSB peak probably derived from two main reasons. Firstly, the closed packed GSB peaks are too close to be distinguished with the increasing ratio of Rb to Cs. In addition, strong ESA signals at around 440 nm might also significantly influence the signal of n = 3 GSB peak.

Supplementary Note 3 | Characteristics for GSB peaks of
Supplementary Note 4 | Full spectrum emission perovskites in the entire bluespectra region. Efficient and spectral stable pure-blue LEDs with λEL < 480 nm is highly desired for display application 26 , and also is as the key issue in our manuscript. As calculated, the valence band (EV) and conduction band (EC) of RbBr-rich <n>Rb0.6 = 3 perovskite film is 6.29 eV and 3.68 eV, respectively.