Low-Cost Synthesis of Highly Luminescent Colloidal Lead Halide Perovskite Nanocrystals by Wet Ball Milling.

Lead halide perovskites of APbX3 type [A = Cs, formamidinium (FA), methylammonium; X = Br, I] in the form of ligand-capped colloidal nanocrystals (NCs) are widely studied as versatile photonic sources. FAPbBr3 and CsPbBr3 NCs have become promising as spectrally narrow green primary emitters in backlighting of liquid-crystal displays (peak at 520-530 nm, full width at half-maximum of 22-30 nm). Herein, we report that wet ball milling of bulk APbBr3 (A = Cs, FA) mixed with solvents and capping ligands yields green luminescent colloidal NCs with a high overall reaction yield and optoelectronic quality on par with that of NCs of the same composition obtained by hot-injection method. We emphasize the superiority of oleylammonium bromide as a capping ligand used for this procedure over the standard oleic acid and oleylamine. We also show a mechanically induced anion-exchange reaction for the formation of orange-emissive CsPb(Br/I)3 NCs.

Introduction

Bottom-up synthesis methods with surfactant (ligand)-capped colloidal nanocrystals (NCs), such as high-temperature hot injection or heating methods, have afforded an unprecedented variety of NC compositions and morphologies, with a high level of uniformity and narrow size distribution.[1−5] Colloidal synthesis in apolar solvents has enabled the development and recent commercial applications of colloidal semiconductor NCs (also known as quantum dots, QDs) composed of typical binary compound semiconductors: II–VI (CdS, CdSe, CdTe),[1] III–V (InP, InAs, InSb),[6−9] and IV–VI (PbS, PbSe, PbTe).[10−19] In addition, QDs with high structural complexity have been achieved, with the most notable examples being core–shell QDs (CdSe/ZnS[5,20] and PbX/CdX,[21−23] X = S, Se, Te), nanowires,[24] nanodisks,[25] nanoplatelets (NPLs),[26] rods, and tetrapots.[27]

Unsurprisingly, the most recent addition to the family of colloidal QDs, NCs of lead halide perovskites [LHPs; APbX3-type; A = Cs, formamidinium (FA), methylammonium (MA); X = Br, I] were initially approached with the same experimental mindset.[28] For example, the first synthesis of CsPbX3 NCs was accomplished by injecting a Cs precursor into a PbX2 solution at elevated temperatures.[29] Similarly, highly luminescent FA- and MA-based LHPs were synthesized by injecting the sources of MA and FA cations.[30,31] Colloidal Cs- and FA-based LHP NCs[29,31,32] exhibit broadly tunable photoluminescence (PL), spanning the entire visible spectral range (410–700 nm), small PL line widths (full width at half-maximum, fwhm, 12–40 nm for blue-to-red), and high PL quantum yields (QYs, 50%–90%), thus providing a broad color gamut.

LHPs are ionic and are characterized by low formation and lattice energies; therefore, they do not require thermal activation during formation to achieve high crystallinity. Subsequent studies have shown that the synthesis of colloidal LHP NCs involves a surfactant-controlled coprecipitation of ions that proceeds with fast kinetics even at room temperature (RT). Facile formability of LHP NCs has been successfully achieved using an alternative strategy, the reprecipitation method, wherein an ionic solution of the respective ions (A+, Pb2+, and X–) in a polar solvent is rapidly destabilized by mixing with a nonsolvent, inducing a burst of nucleation. This method was originally proposed for MAPbBr3 NCs[33] and has since been extended to Cs and FA systems. Noncuboidal shapes, such as NPLs, nanosheets, and nanowires, can also be obtained using both methods.[34−37]

The soft nature of LHPs and their facile crystallization suggest that top-down methods might also be applicable for LHP NC preparation. In this study, we report a simple mechanochemical synthesis of LHP NCs using a commercial ball mill. Such a synthesis essentially involves simple mechanical grinding of bulk APbX3 materials or AX + PbX2 mixtures in the presence of a solvent (mesitylene) and ligand (oleylammonium halide, OAmX, or mixture of oleic acid, OA, and oleylamine, OLA). These ligands are chosen based on their success in hot-injection synthesis methods.[29,31]

High-energy ball milling is a type of mechanical grinding of materials. This process can be conducted in a dry (without solvents) or wet (with solvents) fashion. Laboratory-scale ball milling (5–100 mL scale) is a batch process, whereas industrial ball mills can be operated in a continuous mode, thus offering a very high synthesis throughput. The milling occurs due to mechanical friction between the grinding medium, such as balls of the same or various sizes, and the ground material (Figure ). The container (bowl) and the grinding balls are typically made of the same, high-hardness material (zirconia, corundum, or stainless steel). Mechanical energy is provided to the system by the rotary motion of the bowl, as in the case of a planetary ball mill. The rotation of the bowl and the speed are optimized such that the maximum speed, the speed at which balls do not move within the bowl, is not exceeded. The milling can have diverse effects on the ground material: different extents of downsizing of the final powder (from microns to tens of nanometers), efficient mixing (e.g., production of slurries in battery manufacturing or preparation of pigments), solid-state chemical transformation (mechanochemical synthesis), or a combination of these effects.[38,39] Wet ball milling is also considered to be a green chemistry approach, as it does not require high temperatures (energy saving) and consumes minimal quantities of solvents.[40] Ball milling has been popular since the 1970s for producing oxide dispersions such as Al2O3, Y2O3, and ThO2. Grinding techniques are also widely used for alloying materials,[41] in the synthesis of metal oxides,[39,42,43] and for mechanical exfoliation of graphene.[44] However, in terms of the synthesis of semiconductor NCs, ball milling has achieved very limited success (examples include CdSe[45] and CdTe[46]) due to the lack of bright emission from the resulting QDs. This is because, on the one hand, ball milling generates numerous structural defects in rigid lattices of these materials, and on the other hand, it does not allow surface passivation. The combined effect of the trap states, which are abundant on the pristine (uncoated) NC surfaces, and other structural defects leads to very low PL QYs. Therefore, it is necessary to coat conventional QDs with wide-bandgap materials, such as in the canonical example of core–shell CdSe-CdZnS NCs, thereby decoupling the excitonic recombination from the detrimental surface states.

Figure 1

Photographs of (a) a typical planetary ball mill, (b) zirconia bowl filled with zirconia balls, and (c) crystals of bulk CsPbBr3 used as starting materials. (d) Schematic of the working principle of the planetary ball mill showing the horizontal cross-section of the bowl during the wet ball-milling experiments. The bowl undergoes two motions: orbital rotation of the entire ball and its simultaneous spinning. (e) Schematic of the processes occurring during the mechanochemical synthesis of LHP NCs. Photographs showing the colloid of CsPbBr3 NCs (f) directly after the ball-milling experiment and (g) after dilution with toluene and filtration (taken under visible light and under a UV lamp, λ = 365 nm). (h) TEM image of the resulting CsPbBr3 NCs.

On the contrary, LHP NCs are unique in that they do not require surface passivation with epitaxial wide-bandgap semiconductor layers for exhibiting bright emission in the green-red spectral region. This is one of the manifestations of the rare phenomenon of defect tolerance: structural defects are nearly fully benign with respect to the carrier dynamics. Theoretical calculations indicate that dominant point defects, primarily vacancies, in the bulk material,[47] at grain boundaries,[48] and on the NC surfaces,[49] are all shallow or intraband (in the valence or conduction bands). Other defects, for instance, those of the antisite or interstitial type, are not common in perovskites due to their crystal structure and ionic bonds. Defect tolerance makes LHPs vastly different from all known types of colloidal QDs and enables many synthesis pathways for LHP NCs. Besides the mechanochemical synthesis presented here and by others (see further discussion in the next section),[50,51] perovskite NCs with bright PL were obtained by sonication,[52] microwave irradiation,[53,54] or templating of crystallization using the nanoscale pores of mesoporous silica.[55]

Results and Discussion

The goal of this study was to produce highly luminescent CsPbBr3 and FAPbBr3 NCs in one step by ball milling. Both of these nanomaterials can be readily synthesized on microfluidic platforms[56] and by the hot-injection colloidal method,[29,31,57,58] ultrasonication,[52] the reprecipitation method at RT,[59,60] and microwave-assisted growth.[53,54,61] Mechanochemical syntheses were initially used only for MAPbI3,[51,62] wherein relatively large micro- and nanoparticles (>200 nm) are formed on Al2O3 carrier particles. In a recent report,[50] dry ball milling of CsX and PbX2 powders followed by the addition of OLA yielded luminescent colloidal NCs. In our own experiments, we were unable to produce stable and bright colloids with only OLA as the ligand. Also, dry-milling was ineffective in our experiments, irrespective of the ligands used afterward or whether bulk APbX3 or AX + PbX2 mixtures were used as starting precursors. Below, we detail our own study, wherein different ligand systems were used: (i) a mixture of OLA and OA and (ii) OAmX. These ligand systems are typically used in colloidal syntheses of perovskite NCs, and the general consensus is that OAm+ coordinates the surface anions, whereas Br– or oleate anions locate themselves close to surface cations, thereby maintaining the overall charge neutrality of the NC.[49,63,64]

The ball-milling method (Figure ) employs two rotational movements: one of the milling bowl and the other of the supporting disk (on which the bowl with NCs is mounted). The combined effect enables the efficient movement of the balls within the bowl, causing grinding of the material. If only circular motion of the bowl was employed, such as in centrifuges, all components of the mixture would be statically held by the centrifugal force. As precursor materials, either a bulk APbX3 compound or an equimolar mixture of AX and PbX2 is used. The combined effect of milling and the presence of the capping ligands and solvents allows for a simple one-step conversion of bulk precursors into colloids of APbX3 NCs. The optimal milling time at RT and 500 rpm is dependent on the material: 2 h for CsPbBr3 and 1 h for FAPbBr3.

In a typical experiment (see Methods section for the detailed protocols), 0.035 mmol of CsPbBr3 or 0.04 mmol of FAPbBr3 was loaded into a zirconia bowl with 25 zirconia balls (4 mm and 5 mm in diameter). OAmBr (0.03 mmol) was added as the ligand, and 0.4 mL of mesitylene was added as the solvent. The milling speed was set to 500 rpm and milling time to 2 h (for CsPbBr3) and 1 h (for FAPbBr3), as found to be optimal for obtaining NCs (Figure b). Powder X-ray diffraction (XRD) patterns confirmed the complete conversion of the precursors into nanogranular products with the expected crystal structure (orthorhombic for CsPbBr3 and cubic for FAPbBr3).

Figure 2

(a) Powder XRD patterns of bulk CsPbBr3 (orange pattern) used for the ball-milling experiment, and the resulting CsPbBr3 NCs (green pattern) showing an identical orthorhombic perovskite crystal structure (Pnma space group); (b) photograph showing the obtained colloidal CsPbBr3 NCs diluted in toluene; (c) photograph showing the initial bulk CsPbBr3 crystals; (d) comparison of the XRD patterns of bulk FAPbBr3 (orange pattern) used for the ball-milling experiment and the resulting FAPbBr3 NCs (green pattern), indicating the retention of the cubic lattice; (e–h) TEM images of CsPbBr3 NCs and FAPbBr3 NCs at various magnifications.

The ball-to-material weight ratio, wherein “material” is the mixture of precursors, ligands, and solvents, is an important parameter for obtaining uniform NCs: the higher the ratio, the shorter the time required for complete milling of the bulk materials. A broad range for this parameter is reported in the literature for various materials (from 1 to 220).[38,65,66] By keeping the number of balls constant and changing the amount of material, the ball-to-material ratio was varied from 21.7 to 167 (Figure S1). The optimal ball-to-material weight ratio was found to be 80 [∼10 balls (4 mm) and 15 balls (5 mm), 0.035 mmol CsPbBr3 (0.04 mmol for FAPbBr3), 0.1 g of OAmBr, 0.4 mL of mesitylene], yielding PL fwhm = 27 nm for CsPbBr3 NCs (PL peak at 510 nm) after 2 h of milling. Very long milling times often caused the emergence of mixed shapes, in particular, a large fraction of NPLs. NPLs were apparent in the blue-shifted bands in the PL spectra (Figure S2a, Figure S4). For example, for FAPbBr3, NCs with an emission peak at 537 nm could be readily obtained within 1 h (Figure a). A longer milling (ca. 2 h) time led to the appearance of an emission peak at 450 nm. This band was dominant in the PL spectrum after 3 h (Figure S2b). OAmBr was the most suitable ligand for both perovskite materials, giving stable colloidal dispersions and narrow emission line widths. When only OLA was used, no NCs were formed (no PL was observed), whereas using OA as the sole ligand yielded bright suspensions but unstable colloids with fast decay of the PL QY upon storage. A mixture of OA and OLA allowed the fabrication of stable colloids of FAPbBr3, but with broader PL line widths (Figure S3b, fwhm = 34 nm), most likely due to acid/base equilibrium reactions between the ligands; therefore, the existence of protonated and unprotonated species in the same time increased the probability to obtain NCs with a broader size distribution. The attempts to form CsPbBr3 NCs with OA and OLA resulted in a mixture of NCs and NPLs (Figure S3a). Other ligands, such as tetraoctylammonium bromide, induced the formation of bright FAPbBr3 NCs but without satisfactory colloidal stability (Figure S3). Mesitylene was identified as the best solvent for colloidal stability; other tested solvents included octadecene, toluene, diphyl, hexane, and chloroform. Dry milling, for example, without solvents and ligands, leads only to poorly luminescent microcrystalline powders.

Figure 3

(a) Absorbance and PL spectra for CsPbBr3 NCs (green) and FAPbBr3 NCs (black) obtained using the ball-milling method. (inset) Photograph of CsPbBr3 NCs under UV lamp (λ = 365 nm), (b) time-resolved PL of CsPbBr3 NCs (green) and FAPbBr3 NCs (black) measured in solutions, (c) absolute QY of CsPbBr3 NCs (green) and FAPbBr3 NCs (black) measured in solutions and films, and (d) decay time for CsPbBr3 NCs (green) and FAPbBr3 NCs (black) measured in solutions and films.

Time-resolved PL traces of both CsPbBr3 and FAPbBr3 samples were characterized by multiexponential decay behavior (Figure b). This could be explained by the broad size and shape distribution of the NCs after ball milling. This was also consistent with the broader emission band observed for NCs obtained by ball milling than for NCs obtained by hot injection.[29] For CsPbBr3 NCs, the fastest decay component in the biexponential fitting model was at least 2 times longer than nearly monoexponential decay parameter for NCs produced by hot injection (10.6 vs 5 ns). FAPbBr3 NCs, however, showed notably faster relaxation times than colloidally synthesized NCs of the same composition.[31] We associate this to the larger fraction of smaller NCs in the ball-milled FAPbBr3 product. In smaller NCs, higher quantum confinement accelerates the radiative rate. The acceleration due to surface states (the effect is rather common for II–VI QDs) can be ruled out by the observation of PL QYs. PL QYs were high and similar in solutions (>75% for CsPbBr3 NCs, >80% for FAPbBr3 NCs). Films of FAPbBr3 NCs nearly retained their high PL QYs (>70%), whereas CsPbBr3 NC films exhibited a significant decrease to 45% (Figure c). High initial QY and better stability in film for FAPbBr3 NCs can be attributed to the better tolerance to defects, lower density of defects, and also higher defect formation energy.[67] In films of FAPbBr3 NCs, the relaxation speed was slower than in the solution, as opposed to CsPbBr3 NC, where relaxation is slightly faster than in the corresponding solution (Figure d). Such different behavior of NC with organic and inorganic cations can be explained by the higher tendency of FAPbBr3 NCs to sintering at room temperature (bulk material has lower melting point) and higher density of surface defects in CsPbBr3 with lower defect formation energy.[68]

Attempts to compositionally tune the bandgap by methods such as milling bulk CsPbBr3 and oleylammonium iodide (OAmI) as starting materials showed limited success, as the resulting NCs exhibited low PL QYs. A better approach was to conduct an anion-exchange reaction[32] by adding OAmI to CsPbBr3 NCs (Figure ).

Figure 4

Example of an anion-exchange reaction performed by adding OAmI to the bowl immediately after the synthesis of CsPbBr3 NCs (520 nm). Within several minutes, the PL peak shifted from 520 to 580 nm, indicating the formation of CsPb(Br/I)3 NCs.

The results of the mechanochemical synthesis using CsBr and PbBr2 as precursors are illustrated in Figure and compared with those of the identical procedure using the CsPbBr3 precursor. Even after 14 h, the fwhm of the obtained NCs was 49 nm (centered at ca. 500 nm), indicating that it was very important to use a bulk ternary compound as a precursor rather than the mixture of two binary. Using two precursors, we can expect two simultaneous phenomena: the mechanical downsizing of the precursors and the chemical reaction between CsBr and PbBr; the nucleation of CsPbBr3 NCs could be initiated any time during those processes, and therefore different NCs will have different growth history, yielding polydisperse ensemble or remanence of some bulk material. If a shorter ball-milling reaction time is considered (1 h), the obtained NCs with emission at 522 and fwhm of 27 nm were not colloidally stable. This suspension contained most of the material in the nonluminescent bulk phase. After it was centrifuged and filtered, only a highly dilute colloid was obtained (≤1 mg/mL, <5% of theoretical reaction yield). Hence, the narrow PL band could be attributed to the strong reabsorption of PL by the bulk material. Unbalanced kinetics of the formation of new NCs and downsizing of the earlier formed NCs eventually led to a very broad PL peak.

Figure 5

Comparison between ball-milling synthesis of CsPbBr3 NCs employing bulk CsPbBr3 and that employing a mixture of CsBr and PbBr2 as precursors.

Finally, we would like to point out that this ball-milling synthesis method was essentially inapplicable to iodide systems (CsPbI3 and MAPbI3, Figure S5).

Conclusions

In summary, this study explored the utility of wet mechanical grinding for obtaining colloidal NCs of lead halide perovskite. The method yielded FAPbBr3 and CsPbBr3 NCs with optoelectronic quality on par with that required for application in the backlighting of liquid-crystal displays (as the green primary color). The utmost simplicity and speed of mechanochemical synthesis indicated its utility for future research. For instance, the fast downsizing achieved by this synthesis could be used as a general method for testing whether certain bulk materials, such as soft metal halides, can become bright emitters in the form of NCs.

Methods

Synthesis of Bulk CsPbBr3, Adapted from Ref (69)

Bulk CsPbBr3 crystals were obtained from dimethyl sulfoxide (DMSO, 99.8%, Fluka) solution at 110 °C. First, CsBr (Aldrich, 99.9%) and PbBr2 (99.999%, ABCR) were dissolved in DMSO with [Cs] and [Pb] of 0.5 and 1 M, respectively. This solution was filtered at RT and then slowly heated to 110 °C. Typically, numerous sub-millimeter crystals appeared just above 90 °C and continued to grow with increasing temperature. After ∼4 h of growth, the crystals were taken out of the solution, wiped with filter paper to remove the solvent, and dried overnight in a vacuum oven at 50 °C. Synthesis of bulk FAPbBr3. FA acetate (3.2 mmol, 0.33 g, Aldrich, 99%) was dissolved in 1 mL of HBr (48%, aqueous solution, Aldrich). PbBr2 (3.2 mmol, 1.174 g, 99.99%, Aldrich) was dissolved in 2 mL of HBr, and this solution was heated to 80 °C to fully dissolve the salts. To this warm solution, the FA solution was added, forming a red precipitate. The mixture was cooled to RT and centrifuged. The precipitate was rinsed with diethyl ether several times to remove the residual acid and dried on a filter paper.

Synthesis of Bulk MAPbI3, Adapted from Ref (70)

Pb(OAc)2·3H2O (11.33 mg, 0.03 mmol, ≥99.99%, Aldrich) was dissolved in 38.7 mL of hydroiodic acid (HI, 57 wt %, stabilized with 1.5% H3PO2, ABCR or Aldrich) and heated to 100 °C in an oil bath. Then, a mixture of 2.52 g of methylamine aqueous solution (40 wt %, Fluka) and 8.45 mL of HI was added. Small-grained black powder precipitated within a few minutes. Then, the solution was cooled to 75 °C and maintained at this temperature for 1 d. The obtained powder was washed with diethyl ether and dried in vacuum at 25 °C.

Preparation of Oleylammonium Halide (OAmX)[32]

Ethanol (100 mL, Aldrich, absolute, >99.8%) and OLA (12.5 mL, Acros Organics, 80%–90%) were combined in a 250 mL two-necked flask and vigorously stirred. The reaction mixture was cooled in an ice–water bath, and 8.56 mL of HBr (48% aqueous solution, Aldrich) for preparing OAmBr or 10 mL of HI (HI 57%, Aldrich, no stabilizer) for preparing OAmI was added. The reaction mixture was left to react overnight under N2 flow. Then, the solution was dried under vacuum, and the obtained product was purified by rinsing several times with diethyl ether. The product (white powder) was obtained after vacuum-drying at 80 °C overnight.

Mechanochemical Synthesis of CsPbBr3 NCs

Bulk CsPbBr3 (0.02 g, 0.035 mmol, prepared as described above) was loaded into a zirconia bowl with 25 zirconia balls (4 mm and 5 mm) and with OAmBr (0.01 g, 0.03 mmol, prepared as described above), and 0.4 mL of mesitylene (98%, Sigma-Aldrich). Another tested ligand system was a mixture of OA (0.05 mL, Sigma-Aldrich, 90%) and OLA (0.05 mL, Strem, 95%). Other tested solvents were octadecene (Sigma-Aldrich, 90%), Dowtherm A (diphyl, eutectic mixture of 26% diphenyl + 73.5% dipheniloxide), toluene (Sigma-Aldrich, 99.5%), hexane (>95%, Sigma-Aldrich), and chloroform (HPLC grade, Fisher Chemicals). The bowl was mounted on a planetary ball mill (Fritsch, Pulverisette 7, classic line), and the speed was set to 500 rpm. The time was varied from 30 min to 24 h. After the milling, a bright green suspension was obtained, which was diluted with toluene (2 mL) and used as-obtained or precipitated one time with acetonitrile (0.5 mL, 99.8%, Aldrich). The precipitate was redispersed in 2 mL of toluene.

Alternatively, CsBr (0.011 g, 0.05 mmol, Aldrich, 99.9%) and PbBr2 (0.018 g, 0.05 mmol) were loaded into a zirconia bowl with 25 zirconia balls (4 mm and 5 mm) and with 1.6 mL of mesitylene and 0.01 g of OAmBr (0.03 mmol). The subsequent milling lasted 2 h at 500 rpm. The bright green crude solution was diluted with 2 mL of toluene.

Mechanochemical Synthesis of FAPbBr3 NCs

Bulk FAPbBr3 (0.02 g, 0.04 mmol, prepared as described) was loaded into a zirconia bowl with 25 zirconia balls (4 mm and 5 mm) and with OAmBr (0.01 g, 0.03 mmol) and 0.4 mL of mesitylene as the solvent. Other tested ligand systems were a mixture of OA (0.1 mL) and OLA (0.05 mL) and tetraoctylammonium bromide (0.01 g, Aldrich, 98%). The bowl was mounted on a planetary ball mill, and the speed was set to 500 rpm. The milling time was varied from 30 min to 3 h. After the milling, a bright green suspension was obtained, which was diluted with toluene (2 mL) and used as-is or precipitated one time with acetonitrile (0.5 mL). The precipitate was redispersed in 2 mL of toluene.

Anion-Exchange Procedure

CsPbBr3 NCs were prepared as described above. After the ball-milling experiment was completed, an additional 0.02 g of OAmI (0.05 mmol) was added to the crude solution, and the ball-milling experiment was continued for 1 h at 500 rpm. A bright-orange suspension was obtained.

Film Preparation

Toluene solution of the perovskite NCs (10 mg/mL) was filtered using a 0.45 μm poly(tetrafluoroethylene) (PTFE) filter and drop-casted onto acetone/ethanol-cleaned glass slides.

Characterization

UV–Vis absorption spectra for the colloidal solutions were recorded using a Jasco V670 spectrometer in transmission mode. PL and absolute QY measurements were performed using a Fluorolog iHR 320 Horiba Jobin Yvon spectrofluorimeter equipped with a photomultiplier tube (PMT) detector, used to acquire steady-state PL spectra from solutions and films. QY values from NC dispersions were estimated according to the standard procedure using fluorescein as the reference.[71] Powder XRD patterns were recorded using a STOE STADI P powder diffractometer, operating in transmission mode. A germanium monochromator, Cu Kα1 irradiation, and a silicon strip detector, Dectris Mythen, were used. Transmission electron microscopy images were recorded using a Philips CM 12 microscope operating at 120 kV. Time-resolved PL measurements were performed using a time-correlated single-photon counting setup, equipped with an SPC-130-EM counting module (Becker & Hickl GmbH) and an IDQ-ID-100-20-ULN avalanche photodiode (Quantique) for recording the decay traces. The emission of the perovskite NCs was excited by a BDL-488-SMN laser (Becker & Hickl) with a pulse duration of 50 ps, wavelength of 488 nm, and continuous-wave (CW) power equivalent of ∼0.5 mW, externally triggered at a 1 MHz repetition rate. PL emission from the samples was passed through a long-pass optical filter with an edge at 500 nm to reject the excitation laser line. Average radiative lifetimes were determined as , where A and τ are corresponding amplitudes and exponential decay parameters in biexponential analysis. PL QY measurements of the films were conducted using a method similar to that reported by Semonin et al.[72] With an integrating sphere (IS200-4, Thorlabs) with a short-pass filter (FES450, Thorlabs), the absorbance was corrected to reflectance, and the scattering losses were estimated. A CW laser diode with a wavelength of 405 nm and a power of 0.2 W modulated at 30 Hz was used as the excitation source. The emitted light was measured using long-pass filters (FEL450, Thorlabs). The light intensity was measured by a broadband (0.1–20 μm) UM9B-BL-DA pyroelectric photodetector (Gentec-EO). The modulated signal from the detector was recovered by a lock-in amplifier (SR 850, Stanford Research). The ratio between the emitted and absorbed light gave the energy yield. The PL QY was obtained from the value of the energy yield, corrected to the ratio of photon energies of the laser beam and PL bands. The effect of emission reabsorption was taken into account in the final calculation.