Lead-free perovskite Cs3Cu2I5 NCs were synthesized via a previously reported hot-injection method with some modifications using cesium iodide (CsI) and copper iodide (CuI) as precursors 32. We found that the synthesis conditions—ligand concentration and reaction temperature—were optimal for obtaining high-quality Cs3Cu2I5 NCs. CuI oleate was prepared by mixing CuI (6 mmol), oleic acid (OA, 6 mL), and oleylamine (OAM, 6 mL) with octadecene (ODE, 100 mL) in a three-neck flask under vacuum at 100°C. Subsequently, Cs oleate was quickly injected into the CuI oleate solution at 70°C (Fig. 1a). After 10 s, the reaction mixture was rapidly cooled in an ice bath. The as-synthesized Cs3Cu2I5 NCs were isolated via centrifugation at 10,000 rpm for 10 min, and then, the collected Cs3Cu2I5 NCs were redispersed in toluene, which was used as a nonpolar solvent. Cs3Cu2I5 NCs are very sensitive to polar antisolvents during the reprecipitation process. For instance, when ethyl acetate was used as an antisolvent in the reprecipitation of the NC colloid to remove impurities, the reprecipitated Cs3Cu2I5 NCs showed unfavorable crystal growth and a random crystal shape (Figure S1, supplementary information). Thus, we developed a GPC purification process for Cs3Cu2I5 NCs to remove impurities and form a high-quality film. The major benefit of GPC purification is that only toluene is used as a developing solvent in this process, which can produce stable Cs3Cu2I5 NCs. The packing medium—polystyrene beads—is a porous and spherical ball, which enables the removal of small molecular hydrocarbon impurities such as ODE or unbound ligands, as shown in Fig. 1b. The Cs3Cu2I5 NC toluene colloid (3 mL) was injected into a polystyrene bead-packed column tube (see the Experimental section for details). Purified Cs3Cu2I5 NCs (1 mL) were obtained at a concentration of 10 mg mL-1 (Fig. 1c). The Cs3Cu2I5 NCs had high colloidal stability, considering that the NC solution remained without causing precipitate at least a month. The quantified zeta potential both before and after GPC purification had positive valves: 18.2 and 15.6 mV, respectively. The formation of egg-like Cs3Cu2I5 NCs was confirmed by transmission electron microscopy (TEM) and histograms, as shown in Fig. 1d and Figure S2. The average sizes of the NCs before and after GPC purification were 20.7 and 22.7 nm, respectively. The GPC-purified NCs had an interplanar distance of 5.61 Å, which corresponded to the (020) crystal phase.
1H-NMR analysis was performed to determine residual impurities, as shown in Fig. 2a. The Cs3Cu2I5 NCs without GPC purification clearly showed alkene resonance related to the ODE at 5.7 to 5.9 ppm, which indicated that the purification was not completed; the OA and OAM related to the range at 5.2 to 5.4 ppm. However, after GPC purification, we successfully collected Cs3Cu2I5 NCs without ODE and kept the surrounding ligands. The surface ligand composition of the Cs3Cu2I5 NC film was confirmed by Fourier transform infrared spectroscopy (FTIR) (Fig. 2b). The peak resonance of the purified Cs3Cu2I5 NCs was weakly related, implying the loss of OAM on the NC surface. The OA might tend to absorb onto the NCs, whereas the OAM was likely to desorb and leave the surface 38. This physical phenomenon was magnified during the GPC process, leading to a lack of ligands. It explains that a relative low zeta-potential after GPC purification but still keep the high colloid stability. In addition, X-ray photoelectron spectroscopy (XPS) was used to elucidate the chemical composition of the Cs3Cu2I5 NCs (Fig. 2c and Figure S3). The estimated average atomic percentages of Cs, Cu, and I were 26.99%, 17.53%, and 55.47%, respectively, indicating a slightly halide-rich chemical composition. After the GPC process, the chemical composition of the purified NCs remained the same as that before the GPC process (Table S1, supplementary information).
Cs3Cu2I5 single crystals were prepared from a saturated CsI and CuI precursor solution in dimethyl sulfoxide (DMSO), and then, single-crystal XRD (SXRD) analysis was performed. The obtained crystal data (Pnma, a = 10.1824(8) Å, b = 11.6655(11) Å, and c = 14.3687(12) Å) are consistent with those in previous reports (Table S2) 15,39. On the other hand, the thin-film XRD analysis of the GPC-purified Cs3Cu2I5 NC film showed a simple crystal pattern with peaks at (020), (040), and (060) (Fig. 3a and Table S3). This diffraction pattern matched the calculated diffraction pattern obtained from the single-crystal structure determination. Interestingly, the spin-coated Cs3Cu2I5 NC film exhibited a highly oriented single-crystal arrangement in which the b-axis was vertical to the substrate, whereas the drop-coated film was randomly arranged. DFT calculation results showed that the surface energy of the (020) phase (0.117 eV Å–2) was lower than that of the (040) phase (0.145 eV Å–2), indicating that it was easier for the (020) phase than for the (040) phase to form on the surface (Fig. 3b and Figure S4) [40]. In the crystal model, (020) showed an iodide-rich surface, which was consistent with the XPS results. Furthermore, the SEM image showed a close-packed thin film, which had egg-like shapes that tended to lie flat on the substrate (Fig. 3c).
The UV-vis absorption and PL spectra of the GPC-purified Cs3Cu2I5 NC films are shown in Fig. 3d. The optical band edge of the Cs3Cu2I5 NC film was at 316 nm, which corresponded to a bandgap of 3.92 eV. The PL spectra of the Cs3Cu2I5 NC film exhibited blue emission at 443 nm with an FWHM of 74.3 nm. The PLQY and PL decay time of the Cs3Cu2I5 NCs were 79.3% and 1.284 µs in the film and nearly 100% and 1.236 µs in the colloidal solution, respectively (Figure S5). It is worth mentioning that the PLQY and PL decay time of the colloid before and after GPC purification showed no obvious changes owing to the suppression of optical quenching. Moreover, we dried the colloid using a vacuum pump and redispersed it back into toluene. We found that the PLQY reduced by only approximately 2% without any change in the emission wavelength, which is unlikely to happen in conventional lead halide perovskite NCs (Figure S6). The excellent redispersibility indicates less intense aggregation, compared with a lead-based perovskite, owing to their iodide-rich surface and low surface energy. Therefore, for the first time, we succeeded in investigating the GPC purification of Cs3Cu2I5 NCs and their optical properties in the solid film state. The detailed optical properties of the Cs3Cu2I5 NCs are listed in Table S4.
The improvement of environmental stability properties such as thermal stability, photostability, and water resistance is a key requirement for various optoelectronic applications. However, Cs3Cu2I5 NC films exhibit poor water resistance. We demonstrated for the first time that by blending Cs3Cu2I5 NCs with a PMMA polymer composite, a free-standing film could be obtained without a substrate and this film could be used for down-conversion filtering applications. The use of chemically stable polymers (PMMA and polystyrene) creates a framework for embedding conventional lead halide perovskite NCs; therefore, the film has improved stability against heat, light, and water owing to the hydrophobic protective property of polymers 40–42. To prepare the free-standing Cs3Cu2I5 NCs blended with the PMMA film, a purified Cs3Cu2I5 NC colloidal solution (10 mg mL-1 in toluene) was mixed with a PMMA solution (25 wt% in chloroform) in a 1:3 volume ratio, and then, the mixture was kneaded for 2 min. After spin coating the blended film, compression treatment was carried out to produce flat films (Fig. 4a). The PL spectra of the blended film exhibited an emission wavelength of 443 nm with an FWHM of 74.9 nm and was identical to the PL spectra of the Cs3Cu2I5 NC neat film (Fig. 4b). The thermal stability of the blended film was tested in a temperature cycle between 25 and 100°C, as shown in Fig. 4c. The PL intensity decreased as the temperature increased owing to the thermal quenching effect [41]. As the temperature returned to room temperature, the PL intensity also recovered to its initial value, indicating that there was no thermal degradation of the blended film. The blended film exhibited excellent water resistance and maintained the PL intensity even after 10 days at room temperature, as shown in Fig. 4d.