Highly Soluble CsPbBr 3 Perovskite Quantum Dots for Solution-Processed Light-Emission Devices

: We report on the synthesis of CsPbBr 3 perovskite quantum dots (PeQDs) with a high solubility of 75 g/L in toluene and a good ﬁ lm-forming property, as enabled by a dense layer of didodecyldimethylammonium bromide and octanoic acid surface ligands. The crystalline and monodisperse PeQDs feature a cubic-like shape, with an edge length of 10.1 nm, and a high photoluminescence quantum yield of greater than 90% in toluene solution and 36% as a thin ﬁ lm. We ﬁ nd that the PeQDs are n-type doped following the synthesis but also that they can be p-type and additionally n-type doped by in situ electrochemistry. These combined properties render the PeQDs interesting for the emitter in solution-processed light-emitting electrochemical cells (LECs), and we report a PeQD-LEC with air-stabile electrodes that emits with a narrow emission spectrum ( λ peak = 514 nm, full width at half-maximum = 24 nm) and a luminance of 250 cd/m 2 at 4 V and a luminance of 1090 cd/m 2 at 6.8 V. To reach this performance, it was critical to include a thin solution-processed layer comprising p-type poly(vinyl carbazole) and a tetrahexylammonium tetra ﬂ uoroborate ionic liquid between the PeQD emission layer and the anode in order to compensate for the as-synthesized n-type doping of the PeQDs.


INTRODUCTION
−5 PeQDs, with the generic chemical formula of ABX 3 (A = monovalent cation, B = divalent cation, X = monovalent anion), are further highly tolerant to surface defects, which distinguishes them from the incumbent CdS, CdTe, and ZnInS quantum dots that depend on a complex core−shell structure for efficient emission. 2,3−9 The Zeng group was the first to report on the application of PeQDs as an electroluminescent material, when they introduced PeQDs as the emitter in a light-emitting diode (LED) in 2015. 2 The progress in the field has thereafter been rapid, and it was recently reported that a PeQD-LED can feature an impressive external quantum efficiency of 22%. 10 However, a drawback with the current generation of high- This is an open access article published under a Creative Commons Attribution (CC-BY) License, which permits unrestricted use, distribution and reproduction in any medium, provided the author and source are cited.
−13 An easier-to-fabricate and low-cost alternative to the LED is the light-emitting electrochemical cell (LEC). 14A characteristic feature of the LEC is the presence of mobile ions in the active material.During the initial operation (after a voltage is applied between the two electrodes), these mobile ions redistribute to form electric double layers (EDL) at the two electrode interfaces, which facilitate an efficient and balanced injection of electrons and holes. 15The first injected electrons and holes are compensated by a further redistribution of the mobile ions (that are not locked up in the EDLs), so that the active material is p-type doped at the positive anode and n-type doped at the negative cathode.After a turn-on time, a p−n junction doping structure has formed in the active material, which can allow for efficient operation also for a simple LEC device structure comprising a single active-material layer sandwiched between two air-stabile electrodes. 16,17−36 Specifically, Horvath et al. reported on a single-crystal perovskite LEC, which delivered a high peak luminance of 1800 cd/m 2 . 32Slinker and coworkers developed a composite perovskite LEC, comprising a CsPbBr 3 perovskite blended with a poly(ethylene oxide)/ lithium salt electrolyte, that emitted with a very high luminance of ∼15.000 cd/m 2 31 and also delivered a high power efficacy of 9.1 lm/W. 36owever, the performance of PeQD LECs is currently lagging behind that of the corresponding single-crystal and composite perovskite LEC, and it is thus one of the goals of this study to remedy this shortcoming.The first PeQD-LEC, published by Costa and co-workers, comprised a spray-coated active material containing organometallic PeQDs blended with a LiCF 3 SO 3 salt and a trimethylolpropane ethoxylate (TMPE) ion transporter, and it delivered a low but detectable luminance of 1−2 cd/m 2 . 34The same group later reported a PeQD-LEC, comprising all-inorganic mixed-halide PeQDs blended with a KCF 3 SO 3 salt and the TMPE ion transporter, which delivered an improved, but still modest, luminance of 8 cd/m 2 . 35−40 In this context, we mention that Bakr and co-workers report on a bidentate ligand passivation approach for CsPbI 3 PeQDs with improved PLQY and stability, 41 that Zhang introduced an an oleyl-/octyl-cocapping ligand system that reduces surface defects and improves the PeQD stability, 42 and that Zeng and co-workers employed benzenesulfonic acid as a ligand with notably high binding stability to PeQDs. 43We also wish to call attention to the fact that Luther, 44 Liu, 45 and Ma 46 reported on the synthesis of a CsPbI 3 PeQD dispersion, with a high concentration of 70−75 g/L, which was utilized for the spin-coating fabrication of a uniform and dense film for the active material in a solar cell device.
Moreover, it remains unclear whether ionic moieties can be released from the PeQD structure during LEC operation and, as such, contribute to the formation of EDLs at the electrode interfaces and the p−n junction doping structure in the bulk of the active material. 29,31−49 To address these issues, we synthesized CsPbBr 3 PeQDs with a high solubility of 75 g/L in toluene and good filmforming property.This was accomplished by anchoring a dense binary-ligand system, comprising branched and long-chain ACS Applied Nano Materials didodecyldimethylammonium bromide (DDAB) and intermediate chain-length octanoic acid (OTAc), onto the surface of the PeQDs. 50,51The monodisperse and crystalline PeQDs exhibited a high PLQY of greater than 90% in concentrated solution and 36% as a thin film, and they were found to be ntype doped in their pristine state.We successfully fabricated several different types of devices, which all comprised a thin film of solution-deposited PeQDs as the emissive layer, and unequivocally established that ions are liberated from the PeQD structure during operation.These liberated ions were shown to be mobile and to contribute to the electrochemical doping, which in turn verified that the devices comprising a PeQD layer void of an external electrolyte are indeed to be considered as LEC devices.The best-performing PeQD-LEC comprised a solution-processed layer of p-type poly(vinyl carbazole) (PVK) and a tetrahexylammonium tetrafluoroborate (THABF 4 ) ionic liquid between the PeQD emission layer and the anode in order to compensate for the as-synthesized ntype doping of the PeQDs, and it delivered a peak luminance of 250 cd/m 2 during constant-voltage operation at 4 V and a luminance of 1090 cd/m 2 at 6.8 V during a voltage-ramp experiment.

RESULTS AND DISCUSSION
2.1.PeQD Synthesis.One goal of the study was to develop a highly luminescent PeQD ink that enables a solution-based fabrication of uniform and pinhole-free thin films for the emissive active material in light-emission devices.For this end we modified a reported functional synthesis method, 50 so that the PeQD solute concentration could be increased from 20 to 75 g/L in toluene.Figure 1a presents the key steps in the synthesis procedure, and more details can be found in the Methods Section.In brief, three master solutions were prepared: a "Cs 2 CO 3 -in-OTAc solution" comprising 50 mmol/L of Cs 2 CO 3 in OTAc (octanoic acid), a "PbBr 2 -and-TOAB solution" comprising 10 mmol/L of PbBr 2 and 20 mmol/L tetraoctylammonium bromide (TOAB) in toluene, and a "DDAB solution" comprising DDAB dissolved in toluene in a concentration of 50 mmol/L.The chemical structures of the constituents TOAB, DDAB, and OTAc are presented in Figure 1b.
One milliliter of the Cs 2 CO 3 -in-OTAc solution was blended with 5 mL of the PbBr 2 -and-TOAB solution under vigorous stirring, whereafter 1.3 mL of the DDAB solution was added.The three-component blended solution was mixed with 11 mL of ethyl acetate and thereafter centrifuged at 10 000 rpm for 5 min.The collected precipitates were redispersed in toluene.The centrifugation and redispersing procedures were repeated a second time, after which the PeQDs-in-toluene liquid (i.e., the "PeQD ink") displayed in Figure 1c was obtained.We call attention to three key properties of the yellow-colored PeQD ink: (i) It is a stabile colloidal solution, as evidenced by the fact that it remained optically clear under daylight illumination following 30 d of storage without any indication of precipitate settlement and without obvious PLQY drop (see Figure S1 in the Supporting Information).(ii) It features a notably high PeQD concentration of 75 g/L, which is a critical criterion for the formation of pinhole-free and uniform solid thin films from ACS Applied Nano Materials solution processing.(iii) It is highly fluorescent, as indicated by the vibrant green UV-excited emission in the right photograph in Figure 1c and the near-unity PLQY (97.9%).
2.2.PeQD Characterization.The synthesized PeQD ink was deposited by different solution-casting methods onto appropriate substrates for the establishment of the QD morphology and crystalline structure.Figure 2a is a transmission electron microscopy (TEM) image of a dry PeQD film, as deposited by drop-casting onto a lacey carbon grid.We observe that the two-dimensional projection of a typical PeQD is square-shaped and therefore, by a symmetry reasoning, draw the conclusion that the three-dimensional PeQDs exhibit a cubic-like shape.Moreover, the appearance of distinct and continuous bright regions between the dots suggests that the surface ligands form a dense coverage on the surface of the PeQDs and, as desired, are capable of preventing direct aggregation.
Figure 2b presents a corresponding size analysis of a representative ensemble of 50 PeQDs.The majority of the cubic-shaped PeQDs features an edge length in a narrow range between 8 and 12 nm, with the calculated average edge length being 10.1 ± 1.5 nm. Figure 2c is a high-resolution TEM (HRTEM) image of the same PeQD film, which focuses on the internal structure of one typical QD.This closer inspection visualizes a well-ordered crystalline lattice throughout the entire square-shaped perovskite area, with the indicated interplanar distance being 0.58 nm.
Further information on the PeQD crystalline nature is provided by X-ray diffraction (XRD).Figure 2d displays measured XRD data recorded on a dry film of PeQDs deposited by drop-casting the PeQD ink on a glass substrate (red line) and a corresponding fit of the measured data using the monoclinic structure (black line); the fitted data were taken from the database PDF No. 18-0364.The good agreement between the measured and fitted data discloses that the PeQDs crystallize into the monoclinic structure, which is in agreement with earlier reports. 50Specifically, the fitting establishes that the measured diffraction peaks at 15.35°, 21.67°, 30.72°, 34.40°, 37.85°, and 44.02°in Figure 2d are due to diffraction from the (100), ( 110), ( 200), ( 201), (211), and (202) monoclinic crystal planes, respectively.We note a slight deviation between the experimentally measured diffraction peaks and those delivered by a single-crystal monoclinic lattice, and we speculate that some vacancies and/or defects exist on the PeQD surface, which in turn can induce strain-induced lattice distortions.
We also utilized Bragg's law to calculate the corresponding interplanar distances within the internal PeQD structure (see Table S1 for details), and we, for example, find that the interplanar distance for the (100) plane is 0.58 nm, which is in excellent agreement with the observation in Figure 2c.
X-ray photoelectron spectroscopy (XPS) and energydispersive X-ray spectroscopy (EDS) were employed to establish the elemental composition of the PeQDs.Figure 3a displays a broad-range XPS survey of a dry PeQD powder, with the assigned key XPS peaks marked in the graph; the corresponding high-resolution XPS spectra in the Cs 3d, Pb 4f, and Br 3d regions are presented in Figure S2.An integratedpeak-ratio analysis of two independent sets of XPS data yields a Cs/Pb/Br elemental number ratio of 1:1.16:3.29.As XPS is a surface-probing technique, this finding suggests that the surface of the PeQDs film is richer in Br and Pb than that implied by the nominal CsPbBr 3 perovskite composition.Figure 3b presents the EDS spectrum of the PeQD film, with the chemical origin of the different peaks indicated in the graph.An analysis of the integrated ratio of the different EDS peaks from five independent EDS measurements resulted in a Cs/ Pb/Br number ratio of 1:1.00 (±0.07):3.38 (±0.09).As EDS has a much larger penetration depth than XPS, this suggests that the bulk of the PeQDs are rich in Br.−55 The next object of investigation is the ligands attached to the surface of the PeQDs.Figure 3c is a high-resolution XPS spectrum of a dry PeQD powder in the N 1s region between 410 and 395 eV.A single peak is observed at 402.2 eV, which is attributed to the center nitrogen of the DDAB ligand due to the relatively low electron density around N. 8 We also call attention to the absence of an XPS peak at ∼401 eV, which reveals that the TOAB molecule included during the synthesis (see Figure 1a) is absent from the surface of the PeQDs. 11,50,56−58 Figure 3d displays the Fourier transform infrared (FTIR) spectrum of the PeQD powder, with the most prominent IR absorptions marked.The strong absorption bands centered at 2923 and 2852 cm −1 correspond to the asymmetric and symmetric stretching vibrations of CH 2 , respectively, the intermediate bands at 1467 and 2958 cm −1 are assigned to the CH 2 bending vibration and the C−H stretching vibration of the CH 3 group, respectively, 56 and the smaller-intensity band at 721 cm −1 corresponds to the rocking mode of a methylene (−CH 2 −) n chain. 59Importantly, the characteristic CO stretching vibration is located at 1710 cm −1 , 50 and its existence in the FTIR spectrum in Figure 3d demonstrates that the OTAc ligand is attached to the surface of the PeQDs (see Figure 1b for a comparison of the chemical structures of the different compounds).Moreover, the IR bands at 979 and 926 cm −1 represent C−N stretching vibrations, 59 which could originate from both DDAB and TOAB.However, since the existence of TOAB in the PeQD sample was excluded by the XPS result in Figure 3c, we can draw the conclusion that it is solely the OTAc and the DDAB molecules, and not the TOAB molecule, that are attached to the PeQDs as surface ligands.
Accordingly, these combined data show that it is the shortchain OTAc acid and the branched DDAB molecule that are attached to the PeQDs as surface ligands, which thereby endow the PeQDs with the observed high solubility and good stability in nonpolar solvents such as toluene. 60,61We anticipate that DDAB, in particular, is important for this task.−58 We further anticipate that the roles of the nonattaching TOAB molecule are to act as a solubilizing agent for the PbBr 2 salt and to possibly function as a Br source during the synthesis of the CsPbBr 3 PeQDs. 50,51e now shift our attention to the optical properties of the PeQDs.Figure 4a presents the normalized absorption and photoluminescence (PL) spectra of a PeQD colloidal solution in toluene with a concentration of 10 g/L.The absorption spectrum (solid black squares) exhibits an onset at ∼519 nm, which corresponds to an optical band gap of the PeQDs of 2.39 eV, as derived from the Tauc plot in Figure 4b

ACS Applied Nano Materials
emission peak located at 514 nm and the full width at halfmaximum (FWHM) being a mere 19 nm.The PLQY of the PeQD colloidal solution was measured on three different batches, which all feature a high PLQY in excess of 85%, with the highest value being 97.9%.The stability of the PeQD colloidal solution was investigated by storage under ambient air in dark conditions for 30 days.Figure S1 shows that both the overall appearance and the PLQY were left effectively invariant by the long-term storage, and we therefore draw the conclusion that the ink stability is adequate for many applications.
The photograph in the inset of Figure 4a is recorded on a dry spin-coated PeQD thin film under UV excitation.It exhibits bright and uniform green PL over its entire area, and the PL peak of the solid thin film is positioned still at 514 nm (Figure S3), while its PLQY is 36%.The observations of an invariant PL spectrum, and a relatively retained high PLQY, when going from dilute solution to highly concentrated solid state indicate that aggregation effects are minor in both solution and in a solid state.This conclusion is further supported by the TEM data in Figure 2a and in line with the binary-ligand (OTAc + DDAB) system being effective in coating the PeQD surface.In this context we mention that both we and others 11 find that CsPbBr 3 PeQDs capped with oleic acid and oleylamine instead of OTAc + DDAB feature an ∼50% lower PLQY in both concentrated solution and thin film but also that a recent study demonstrated that it is possible to obtain near-unity PLQY in the solid state for CsPbBr 3 PeQDs capped with a bipolar shell. 10ltraviolet photoelectron spectroscopy (UPS) was employed to determine the energy structure of the spin-coated PeQD thin film, positioned on an indium−tin oxide (ITO) coated glass substrate.Figure 4c presents the UPS data in the secondary electron cutoff (SECO) region, from which we derive that the SECO energy (E seco ) is 17.25 eV.The Fermi level can be calculated with the equation E FL = hv − E seco , with the value of hv being 21.22 eV (i.e., the standard He I line).Accordingly, we find that the Fermi level of the PeQDs thin film is positioned at 3.97 eV below the vacuum level.Figure 4d presents the UPS data in the low-energy region, which shows that the valence band maximum (VBM) is positioned at 1.86 eV below the Fermi level.Thus, the VBM is positioned at 5.83 eV below the vacuum level.Figure 4b showed that the optical bandgap of the PeQD thin film is 2.39 eV, and it is therefore straightforward to conclude that the conduction band minimum (CBM) is positioned at 3.44 eV below the vacuum level.
Accordingly, we find that the energy difference between the CBM and the Fermi level is small at 0.53 eV, whereas the corresponding energy difference between the VBM and the Fermi level is much larger at 1.86 eV.The observation of a Fermi level positioned much closer to the CBM than the VBM can be attributed either to surface defect states that pin the Fermi level at this energy or to the PeQDs being n-type doped following the synthesis.In this context, we remember the EDS and XPS data suggested that the PeQDs are relatively rich in Br, which has been reported to be concomitant with n-type doping. 62This suggests that the PeQDs are primarily n-type doped following the synthesis, although a contribution from surface defect state cannot be completely ruled out.
Figure 5a presents cyclic voltammetry (CV) traces measured on a reference electrolyte solution void of PeQDs (solid black line) and on the same electrolyte solution but with PeQDs added in a 2 g/L concentration (open red squares).A comparison between the reference trace and the PeQD trace reveals that the significant redox peaks are due to the existence of the PeQDs, which are electrochemically active in both reduction (as evidenced by the negative current peak at the negative potentials) and oxidation (the positive current peak at the positive potentials).More specifically, the onset of reduction of the PeQDs is identified to be located at −1.00 V versus Fc/Fc + , whereas the onset of oxidation is located at +1.03 V versus Fc/Fc + .With this information at hand, we can calculate the electrochemical energy structure with the equation  3 shows that the optical and electrochemical values for E VBM coincide perfectly, whereas the corresponding values for E CBM deviate by 0.36 eV.Accordingly, the derived electrochemical bandgap of 2.03 eV is 0.36 eV lower than the optical bandgap.We also note that the reduction reaction in Figure 5a is reversible in that the corresponding oxidation return reaction exhibits a similar integrated area (i.e., the same amount of electronic charge is added and removed from the PeQDs), while the oxidation reaction is more irreversible. 63 common challenge with the solution-based deposition of QDs (and small molecules) is that the resulting dry thin film suffers from nonuniform coverage and pinholes, which is a severe problem from a device stability and performance perspective.Although these problems can be addressed by, for example, the addition of a high-viscosity polymer to the solution ink or by a repetition of the deposition cycle, [29][30][31]34,64 it is preferable to perform the deposition in a single step without the employment of potentially performance-limiting additives.
Figure 5b is a top-view scanning electron microscopy (SEM) image of a 40 nm thick PeQD film, which was spin-coated from the 75 g/L PeQD ink on top of a PVK polymer on an ITO-coated glass substrate.The PeQD film appears highly uniform and pinhole-free over the entire 0.5 × 0.5 μm 2 probed area.Figure 5c is an atomic force microscopy (AFM) tapping image of the same PeQD film over a larger area of 6 × 6 μm 2 .The root-mean-square (RMS) height variation over the entire AFM-probed film surface is a mere 1.63 nm, and the film is again uniform and completely free from noncoated hole areas and significant spikes.For comparison, we also investigated the film-forming ability of a PeQD ink prepared by the alternative hot-injection method and with the PeQDs being capped with oleic acid and oleylamine surface ligands.This solution featured a lower maximum solute concentration of 20 g/L in toluene, and most importantly it failed to realize a uniform film following spin-coating, as visualized in Figure S4c.The merit of the PeQD ink developed herein for solution processing is further manifested in uniform and pinhole-free thin films being successfully fabricated on a number of different substrate surfaces, including PVK, ITO, and poly(3,4-ethylenedioxythiophene) polystyrene sulfonate (PEDOT:PSS) (see also Figure S4).In this context, we mention that the PeQD film deposited on the polymeric PVK and PEDOT:PSS surfaces was flatter than that deposited on ITO.
2.3.PeQD Light-Emitting Device.We first fabricated and characterized a device with an ITO/PEDOT:PSS/PeQD/Al architecture; its energy structure is depicted in Figure 6a.The active material, sandwiched between the PEDOT:PSS anode and the Al cathode, comprised solely of the PeQDs.The best device performance was obtained with a 40 nm thick PeQD active material, which was also the thickest uniform film that could be fabricated by spin-coating the 75 g/L PeQD-intoluene ink.The motivation for the addition of a transparent PEDOT:PSS layer on top of the more conductive and transparent ITO was that the dry PeQD film was found to be more flat and uniform when deposited on PEDOT:PSS than on ITO (see Figure S4) and that problems with shortcircuits through the thin PeQD active material thereby were suppressed.In this context, we wish to call attention to the average edge length of the cubic-shaped PeQD being 10.1 nm (see Figure 2a−c), which implies that the 40 nm thin PeQD film, on average, only comprises four layers of PeQD "cubes" stacked on top of each other.
Figure 6c presents the temporal evolution of the current density (solid black squares) and the luminance (open red circles) for the device when driven by a voltage of 4 V.During the first 20−40 s of constant-voltage operation, the current density is increasing by a factor of 2, while the luminance is increasing by more than an order of magnitude.The same qualitative behavior was consistently observed also at other drive conditions (see Figure S5).−68 The latter process results in the formation of a light-emitting p−n junction, when

ACS Applied Nano Materials
then p-type and n-type doping regions meet.Thus, it appears clear that mobile ions indeed are migrating within the PeQD active material when a voltage is applied and that these ions can perform electrochemical doping of the PeQDs.The latter conclusion is supported by the CV measurement in Figure 5a, which demonstrated that the PeQDs can be electrochemically p-type and n-type doped.Accordingly, the device in Figure 6a functions as a LEC, and we term it a "single-layer PeQD-LEC".
The source of the mobile ions within this device is not firmly established, but the most likely origin is the PeQDs themselves.We remember that the XPS and EDS data in Figure 3b implied that the synthesized PeQDs are relatively rich in Br − , and we thus speculate that the observed existence of mobile ions could be due to the liberation of Br − from the PeQDs during the device operation.In this context, we note two recent studies that report on a similar liberation of mobile ions from perovskite in an external electric field. 69,70mportantly, we find that the performance of the single-layer PeQD-LEC is quite promising, particularly in comparison to earlier studies in the field.Figure 6c shows that the pristine single-layer PeQD-LEC reached a respectable peak luminance of 70 cd/m 2 at 4 V after 40 s of operation.We further find that luminance can be detected already at a low voltage of 2.7 V (Figure S5a), which is close to the optical band gap potential of PeQDs at 2.39 eV.Driving with a constant current density of 154 mA/cm 2 resulted in a higher peak luminance of 140 cd/ m 2 , whereas driving with a lower constant current density of 23 mA/cm 2 yielded a peak current efficacy of 0.24 cd/A and a peak power efficacy of 0.22 lm/W (Figure S5f−h).
At this point, we remember that the analysis of the UPS data in Figure 4c,d revealed that the as-synthesized PeQDs are ntype doped.We thus anticipate that the initial electrochemical reaction at the positive anode in the single-layer PeQD-LEC is dedoping of the n-type doped PeQDs and that it is only when this dedoping is complete that the PeQDs can begin to be ptype doped.In contrast, the initiation of the electrochemical ntype doping at the negative cathode should be very fast, as the PeQDs are already (chemically) n-type doped at the onset of device operation.The consequence should be that the meeting point of the p-and n-type doping fronts, that is, the position of the light-emitting p−n junction, is very close to the PEDOT:PSS anode.The positioning of the light-emission region close to a conductive electrode is a well-known problem in thin-film devices, such as LECs and organic light-emitting devices (OLEDs), since the excitons (the bound electron−hole pairs that can decay under the emission of photons) that form in the emission zone are strongly quenched by a nearby conductive material. 71,72r hypothesis was then that this problem could be resolved by the introduction of a "spacer layer" between the PEDOT:PSS anode and the PeQD emission layer, and for this end we included a 70 nm thin PVK spacer layer in the "bilayer PeQD-LEC", as schematically depicted in Figure 6b.Interestingly, Figure 6d reveals that the peak luminance for this bilayer PeQD-LEC is 160 cd/m 2 at 4 V, that is, twice that of the single-layer PeQD-LEC (see Figure 6c), and also that the turn-on time to peak luminance has dropped by a factor of 2. Figure S6c shows a similar improvement at a higher drive voltage of 5 V, with the peak luminance being 200 cd/m 2 , the peak current efficacy being 0.31 cd/A, and the peak power efficacy being 0.19 lm/W.
An optical simulation verified that our Gedankenexperiment was correct, and Figure S7c,d and Table S2 reveal that also the simulated luminance was substantially improved by the introduction of the PVK spacer layer.Specifically, we find that the simulation suggests a luminance improvement of 28%, whereas the measured improvement in Figure 6c,d is 26%.We further find that the simulation predicts that the preferred thickness for the PVK spacer layer is in the range between 70 and 100 nm, which is in agreement with the 70 nm thickness derived from systematic experimentation.For details on the simulation procedure, we refer to the Methods Section and refs 73 and 74.
It has been firmly established that the concentration of the mobile ions in the active material has a profound influence on the performance of conventional LECs based on conjugated polymers and small molecules as the emitting species. 47,48,75e therefore investigated whether the addition of an external electrolyte to the PeQD active material, in the form of either THABF 4 or tetrabutylammonium hexafluorophosphate (TBAPF 6 ), 76−79 could result in a further improved device performance.Figure S8 shows that the device with an optimized concentration of 1.1 mass% of THABF 4 or TBAPF 6 in the PeQD active material actually performed worse than both the single-layer and bilayer PeQD-LECs.However, we also find that the emission from the devices with an external electrolyte included into the PeQD active material was nonuniform, which implies that the poorer device performance can be attributed to a significant phase separation between the electrolyte and PeQDs in the active material.
We therefore investigated the effects of adding the THABF 4 electrolyte to the PVK layer.Figure 7a shows that this "ionbilayer PeQD-LEC", with an optimized PVK:THABF 4 mass ratio of 10:1, exhibited the best performance of all the investigated devices.Specifically, the peak luminance is 252 cd/m 2 at 4 V, which is a factor of 1.7 higher than the best of the external-electrolyte-free devices.Moreover, the turn-on time to this peak luminance is the fastest (at 15 s), and the values for the peak current efficacy of 0.46 cd/A and the peak power efficacy of 0.40 lm/W are also the best of the PeQD-LECs.In addition, Figure S9 shows that this ion-bilayer PeQD-LEC delivers a notably high peak luminance of 1090 cd/m 2 at 6.8 V during a voltage-ramp operation.This implies that the addition of an external electrolyte to the device is beneficial for its operation, presumably because the highly mobile ions from the external electrolyte allows for a fast formation of efficient EDLs and electrochemical doping.
As the inclusion of the PVK spacer layer was found to be critical for the attainment of a good device performance, it is relevant to establish whether PVK also contributes directly to the actual emission or whether it originates solely from the PeQDs.Figure 7b provides a verification that it is the PeQDs and not the PVK that is contributing to the emission from the PeQD-LECs.The upper panel presents the broad PL spectrum of PVK (red line) and the characteristic narrow PL spectrum of the PeQDs (black line), and the comparison with the EL spectra measured on the three best-performing PeQD-LECs presented in the lower panels provides unambiguous evidence for the EL indeed originating from the PeQDs.We also monitored the temporal evolution of the EL spectrum of this ion-bilayer PeQD-LEC, during driving with a constant voltage and constant current density, and we find that the EL spectrum remains invariant during its entire operational lifetime (Figure S9c,d).
Finally, we performed a preliminary study aimed at establishing whether the high-concentration PeQD ink also can be utilized for the fabrication with more scalable solutionbased deposition techniques.We find that functional LEC devices indeed can be fabricated by spray-coating the PeQD ink into much thicker active material layers.In consideration of the spray-coating deposition being performed under a relatively open atmosphere, we selected to change the ink solvent from toluene to more benign octane.Figure S10 shows that a bilayer PeQD-LEC, featuring a 500 nm thick PeQD active material spray-coated from the octane ink, can deliver a luminance of 26 cd/m 2 at a drive voltage of 17 V.Further optimization of the fabrication protocol for the attainment of more efficient, larger-emission area devices is ongoing.

CONCLUSIONS
We report on the synthesis of CsPbBr 3 PeQDs with a high solubility of 75 g/L in nonpolar solvents, such as toluene and octane, and a good film-forming property during solution casting.A key to this accomplishment was the anchoring of a binary-ligand system onto the PeQD surface, which comprised branched and long-chain DDAB and intermediate chain-length OTAc.The monodisperse and crystalline PeQDs exhibited a high PLQY of greater than 90% in concentrated solution and 36% as a thin film, and they were found to be n-type doped in their pristine state.We successfully fabricated several different types of devices, which all comprised a thin layer of solutiondeposited PeQDs as the emissive layer, and unequivocally established that ions are liberated from the PeQD structure during device operation.These liberated ions are mobile and perform electrochemical doping, which in turn verifies that the devices comprising a PeQD layer void of an external electrolyte function as LEC devices.The best-performing PeQD-LEC comprised a solution-processed layer of p-type PVK and a THABF 4 ionic liquid between the PeQD emission layer and the anode in order to compensate for the as-synthesized n-type doping of the PeQDs, and it delivered a peak luminance of 250 cd/m 2 during constant-voltage operation at 4 V and a luminance of 1090 cd/m 2 at 6.8 V during a voltage-ramp experiment.
One milliliter of the Cs 2 CO 3 -in-OTAc solution was swiftly injected into 5 mL of the PbBr 2 -in-TOAB solution.The blend solution was vigorously stirred for 20 s, whereafter 1.3 mL of the DDAB solution was added.After 2 min, 11 mL of ethyl acetate (anhydrous, 99.8%, Sigma-Aldrich) was included.The final blend solution was centrifuged at 10 000 rpm for 5 min, and the collected precipitates were dispersed in toluene.The final process, encompassing centrifugation and redispersion in toluene, was repeated one more time, and the final PeQD precipitates were now either dispersed in toluene or dried under air atmosphere for further characterization.A calibrated commercial CsPbBr 3 PeQD colloidal solution in toluene (Quantum Solutions), with a PeQD concentration of 10 g/L, was used for the determination of the concentration of the synthesized PeQD colloidal solution by means of absorption spectroscopy and the Beer−Lambert law.
For comparison, PeQDs capped with oleic acid and oleylamine were synthesized via the hot injection method following the procedure outlined in ref 6.The crude solution was washed with ethyl acetate twice, and the resulting PeQDs were dispersed in toluene at 20 g/L concentration.PeQD thin films were fabricated by spin coating the PeQD solution at 1000 rpm for 60 s onto ITO-coated glass substrates.
4.2.PeQD Characterization.The absorption spectra were recorded with a UV−vis spectrophotometer (Lambda 35, PerkinElmer), and the PL spectra were measured with a fluorescence spectrometer (LS45, PerkinElmer).The PLQY was measured with a spectrometer (C9920-02G, Hamamatsu Photonics) equipped with an integrating sphere, using 360 nm as the excitation wavelength.The XRD data were recorded with an X-ray diffractometer (X'Pert3 Powder, PANalytical).The TEM images were captured with a TEM (FEI Talos L120C, Thermo Scientific).The XPS spectra were collected with a spectrometer (Axis Ultra DLD, Kratos) and a monochromatic Al Kα source (hν = 1486.6eV).The low background UPS measurements were performed in a high-precision UPS apparatus with an electron energy analyzer (MBS A-1) and monochromatized He Iα (hν = 21.218eV).The Fermi level was extrapolated from the SECO region (linear scale), and the VBM was extrapolated from the valence region (logarithmic scale).The FTIR spectra were recorded with a spectrometer (IFS/v66 FTIR, Bruker).The SEM images and the EDS data were acquired using a fieldemission scanning electron microscope (Carl Zeiss Merlin) with an energy-dispersive X-ray spectrometer (X-max 80 mm 2 , Oxford Instruments).The AFM images and the film thickness were measured using an AFM microscope (MultiMode SPM microscope equipped with a Nanoscope IV Controller, Veeco Metrology) operating under ambient conditions.
The CV was performed with a three-electrode setup connected to an electrochemical workstation (Autolab PGSTAT302).The working electrode was a Pt disk electrode, the counter electrode was a Pt wire, and the pseudoreference electrode was a freshly activated Ag wire.
The Ag/Ag + pseudoreference potential was calibrated versus the ferrocene/ferrocenium (Fc/Fc + ) redox couple at the end of each measurement.The supporting electrolyte was a 0.1 M solution of THABF 4 (Sigma-Aldrich) in anhydrous dichloromethane (Sigma-Aldrich).Twenty milligrams of PeQD powder was dissolved in 10 mL of the supporting electrolyte solution.The oxidation potential and the reduction potential were defined as first peaks on oxidation scan and reduction scan, respectively.The CV measurements were performed in a N 2 -filled glovebox ([O 2 ], [H 2 O] < 1 ppm).
4.3.Device Fabrication and Characterization.The ITO coated glass substrates (20 Ω sq −1 , Thin Film Devices) were carefully cleaned in sequence with diluted detergent, deionized water, acetone, and 2-propanol before use.A PEDOT:PSS (Clevios P VP AI 4083, Heraeus) layer was spin-coated on top of the ITO at 4000 rpm for 60 s and then dried at 120 °C for 30 min.The thickness of the dry PEDOT:PSS layer was 40 nm, as measured with a stylus profilometer (Dektak) and AFM.For the single-layer PeQD-LEC, the 75 g/L PeQD colloidal toluene solution was spin-coated on top of the PEDOT:PSS layer at 1000 rpm for 60 s and dried at 60 °C for 30 min.For the bilayer PeQD-LEC, the 20 g/L PVK (M w = 1 100 000, Sigma-Aldrich) in chlorobenzene (anhydrous, 99.8%, Sigma-Aldrich) was spin-coated on top of PEDOT:PSS at 2000 rpm for 60 s and dried 70 °C for 30 min; thereafter, the PeQD ink was spin-coated on top of PVK.The ion-bilayer PeQD-LECs were fabricated using the same procedures, except the 20 g/L PVK-chlorobenzene solution contained 2 g/L THABF 4 .The thicknesses of the PVK and PeQD layer in the optimized architecture are 70 and 40 nm, respectively.
A set of four Al cathodes was deposited on top of the PeQD layer by thermal evaporation at p < 5 × 10 −4 Pa.The light-emission area, as defined by the overlap of one Al cathode and the ITO anode, is 0.85 × 0.15 cm 2 .The LECs were characterized with a computer-controlled source-measure unit (Agilent U2722A) and a calibrated photodiode, equipped with an eye-response filter (Hamamatsu Photonics), connected to a data acquisition card (National Instruments USB-6009) via a current-to-voltage amplifier.The EL spectra were recorded with a calibrated fiber-optic spectrometer (USB2000 + , Ocean Optics).All of the above procedures, with the exception for the cleaning of the substrates and the coating of PEDOT:PSS, were performed in two interconnected N 2 -filled glove boxes ([O 2 ] < 1 ppm, [H 2 O] < 0.5 ppm).All of the chemicals and solvents were used as received without any purification.
The preferred thickness of the PVK film in LEC devices was determined for three different concentrations of PVK in chlorobenzene: 8, 16, and 20 g/L.The spin speed was 2000 rpm for 60 s, and the annealing was done at 70 °C for 30 min.The resulting dry PVK thicknesses, as established with a stylus profilometer (Dektak), were 40, 80, and 100 nm, respectively.However, the deposition of the PeQD-in-toluene ink on top of the dry PVK layer partially removed the PVK, and a separate experiment executed by spin-coating a toluene solution on top of the PeQD layer yielded the result that the effective thickness of the PVK layer in the LEC devices is 10, 50, and 70 nm, respectively.4.4.Optical Simulation.The optical simulation was executed with the software Setfos 4.6.11from Fluxim (www.fluxim.com).−84 The excitons were modeled as Hertzian dipoles positioned in a transparent environment.The exciton distribution was simulated as a δ-function positioned at four distinct locations in the PeQD layer, 0.125, 0.375, 0.625, and 0.875, with 0 being the PeQD/PEDOT:PSS interface (for the single-layer PeQD-LEC) or the PeQD/PVK interface (for the bilayer PeQD-LEC) and 1 being the PeQD/Al interface.The real part of the refractive index of the PeQDs was derived from a measurement on a 800 nm thick CsPbBr 3 film, 85 while the imaginary part was extracted from the absorption measurement in Figure 4a.A constant current density of 50 mA/cm 2 was driving the devices in the simulation.More details can be found in refs 73 and 74.

■ ASSOCIATED CONTENT
* sı Supporting Information PeQD solution PL spectra and photographs under 30-day storage; XRD data summarized in table; XPS spectra of the PeQDs; normalized PL spectra of PeQD solution and film; AFM height images of the PeQDs on different substrates; cross-sectional SEM image of spin-coated PeQD film; SEM image of a spin-coated film from hot-injection synthesized PeQDs capped with oleic acid and oleylamine; the temporal evolution of the current density/voltage and the luminance for single-layer PeQD-LEC; Optimization of PVK layer in bilayer PeQD-LEC; optical simulation data summarized table at three specific wavelengths; The simulated device structure and bilayer PeQD LEC performance as a function of PVK layer; the temporal evolution of the optoelectronic performance of bilayer PeQD-LEC; the performance of the optimized ionbilayer PeQD-LEC device; the performance of a bilayer-PeQD LEC, with the PeQD layer fabricated by spray-coating.The Supporting Information is available free of charge at https:// pubs.acs.org/doi/10.1021/acsanm.0c02797. (PDF)

Figure 1 .
Figure 1.Synthesis procedures of the PeQDs.(a) Schematic illustration of the key steps in the PeQD synthesis.(b) The chemical structures of the TOAB, DDAB, and OTAc compounds.(c) Photographs of the synthesized 75 g/L PeQD in toluene solution under illumination by (left) ambient daylight and (right) an ultraviolet lamp (λ peak = 365 nm).

Figure 2 .
Figure 2. PeQD morphology and crystal characterization.(a) A TEM image, (b) the corresponding size distribution, and (c) HRTEM image of a dry film of PeQDs deposited by drop-casting a PeQD-in-toluene solution onto a lacey carbon grid.(d) The measured XRD pattern (red line) of a dry film of PeQDs deposited by drop-casting a PeQD-in-toluene solution onto a glass substrate and the corresponding fit (black line) using the monoclinic crystalline structure.
. The PL spectrum (open red circles) is narrow and symmetric, with the

Figure 4 .
Figure 4. PeQD characterization for optical properties and energy structure.(a) The normalized absorption and PL spectra of a 10 g/L PeQD-intoluene colloidal solution.(inset) Photograph of the green PL from a spin-cast 40 nm thin PeQD film on a quartz substrate, as captured under UV illumination (excitation-peak = 365 nm).(b) The Tauc plot derived from the absorption spectrum in (a).The UPS spectra of a PeQD thin film on an ITO-coated glass substrate in (c) the SECO region for the establishment of the Fermi level, and (d) the low-energy region for the determination of the valence band maximum.
find that E VBM EC = −5.83eV and E CBM EC = −3.80eV relative to the vacuum level.A comparison with the optically derived values in Figure

Figure 5 .
Figure 5. PeQD characterization for redox property and film-forming ability.(a) CV traces measured on a reference solution void of PeQDs (solid black line) and the same solution with included PeQDs (open red squares).The electrolyte was 0.1 mol/L THABF 4 , the PeQD concentration was 2 g/L in dichloromethane, and the scan speed was 0.1 V/s.(b) Top-view SEM image and (c) the AFM height image recorded on a PeQD film spin-coated on top of PVK on an ITO-coated glass substrate.

Figure 6 .
Figure 6.LEC energy structures and their performances.The energy structures of (a) the "single-layer PeQD-LEC" and (b) the "bilayer PeQD-LEC".The temporal evolution of the current density (solid black squares) and the luminance (open red circles) for (c) the single-layer PeQD-LEC and (d) the bilayer PeQD-LEC, during driving by a constant voltage of 4 V.

Figure 7 .
Figure 7. LEC performance and emission spectra.(a) The temporal evolution of the current density (solid black squares) and luminance (open red circles) for an ITO/PEDOT:PSS/PVK+THABF 4 /PeQD/Al ion-bilayer PeQD-LEC during constant-voltage operation at 4 V.(b) Comparison of PL spectra of PVK and PeQD thin films (top panel) and EL spectra of LEC devices as identified in the insets (lower three panels).