2.1. Synthesis of Cs2NaInCl6 QDs
Cs
2NaInCl
6 double perovskite QDs were synthesized by a modified method developed by Wang et al. [
13]. Cs(OAc) (0.71 mmol), Na(OAc) (0.5 mmol), In(OAc)
3 (0.495 mmol), and Sb(OAc)
3 (0.055 mmol) in octadecene(ODE) (9 mL) were placed in a three-neck flask. Then, OA and OAm were placed in a three-necked flask with different ratios, i.e., [OA]/[OAm] is 4, 2, 1, 0.5 and 0.25, respectively, where [OA]+[OAm]=3.5 mL. The mixture was heated to 110 ℃ and stirred under vacuum for 50 minutes, and then heated to 170 ℃ under nitrogen atmosphere, and a GeCl
4 precursor solution containing 77 μL of GeCl
4 per 1 mL of ODE was swiftly injected. The solution was then heated to 180 ℃. After 5 min, the reaction mixture was rapidly cooled in ice water to terminate the reaction. The reaction mixture was centrifuged at 9500 rpm for 5 min and the precipitate was collected. This precipitate was mixed with 10 mL of chlorobenzene, centrifuged at 9500 rpm for 5 min, and the precipitate was collected. The precipitate was thoroughly dried, dispersed in 4 mL of hexane, centrifuged at 4000 rpm for 5 min, and the supernatant containing the target quantum dots were collected.
2.2. Properties of Cs2NaInCl6 double perovskite QDs
The morphology and crystal structure of the synthesized QDs under the condition of [OA]/[OAm] = 4 were characterized by TEM images and XRD pattern (
Figure 1). The QDs showed a cubic shape. The average size of the QDs obtained from the TEM images was 11.2 nm, and the size estimated from the XRD diffraction pattern was 11.6 nm using
Scherrer equation. The XRD pattern is consistent with the that of Cs
2NaInCl
6 [
14], indicating Cs
2NaInCl
6 QDs have been synthesized successfully.
Figure 2 shows typical PL, optical absorption and PL excitation (PLE) spectra of the synthesized QDs under the condition of [OA]/[OAm] = 4, respectively. From the PL spectrum, a blue emission with a peak wavelength of 450 nm can be observed clearly. The PLQY of this luminescence was about 85%. In
Figure 2(b), besides the optical absorption spectrum (top), its second derivative spectrum (bottom) is also shown. The second-order differentiation of the absorption spectrum can be very useful for analyzing the spectrum of which the peaks are difficult to see, because it is known that the minima of the second derivative spectrum correspond to the peaks in the optical absorption spectrum [
24]. Therefore, two absorption peaks at 320 nm and 335 nm can be observed clearly. These peaks are obtained by doping antimony (Sb), and when combined with the PLE (
Figure 2 (c)) peaks, the emission is strong at these absorption peaks. This broad blue luminescence has been reported to originate from the STE in the Cs
2NaInCl
6 QDs [
14].
2.3. OA and OAm states in Cs2NaInCl6 double perovskite QDs
OA and OAm are used in the synthesis of the QDs, which are expected to bind to the QDs produced. To confirm this expectation, FTIR and NMR measurements were carried out.
Figure 3 shows the FTIR spectra of four kinds samples: Cs
2NaInCl
6 QDs, OAm, OA and mixture of OAm and OA. The characteristic peak of OA was obtained at 1710 cm
-1. This peak is also observed in the QD solution. The broad peaks at 1640 cm
-1 and 1555 cm
-1 are thought to be the peaks of deprotonated OA and protonated OAm, which are also partly contained in the QD solution. Finally, the peak at 1585 cm
-1, which is only observed in the QD solution, is considered to be a shifted peak of ionized OA or ionized OAm. These results indicate that the QD solution contains OA, deprotonated OA, and protonated OAm.
However, it is impossible to determine whether the two kinds of ligands are bound to the QDs or not only by the FTIR results. Therefore, NMR measurements were performed.
Figure 4(a) shows the results of
1HNMR. In the QDs solution, the 3 and 6 protons common to OA and OAm, as well as the 1 and 2 peaks characteristic of OA, were observed. The sharp peak seen around 2.17 ppm was acetone used to clean the sample tube [
25]. The amine peak of OAm could not be observed because the amount of OAm in the solution was small or because it was close to the surface [
26]. Therefore, we dropped 10 μL of OA and OAm into the QD solution and measured them (
Figure 4(b)). In the sample with OA, the OA peak at 2.33 ppm was enhanced, but the peak did not shift, and no broadening of the peak was observed. This indicates that there is no interaction between OA ligand and the QDs, which suggest that OA is not bound to the QDs. On the other hand, the β peak at 2.66 ppm was shifted in the sample with the addition of OAm. This is because the OAm binding to the QDs restricts the longitudinal relaxation, resulting in a broadening of the peak due to the predominance of transverse relaxation [
27,
28]. These results indicate that there is an interaction between OAm ligand and the QDs.
Nuclear Overhauser Spectroscopy (NOESY) was further performed to provide evidence that OA is not bound to the QDs but OAm is bound (
Figure 5). First, nothing could be measured when only the QD solution was used. This is thought to be due to the same reason as for
1HNMR. Next, we measured the QD solution with OA or OAm added, as in
1HNMR measurements. Red contour lines were observed in the sample with OA addition. This is a positive NOE signal, which is evidence that the QDs are not bound with OA. Next, a negative NOE signal was observed in the sample to which OAm was added. This is the evidence that the OAm is bound to the QDs. This is thought to be because the unbound ligand behaves as a small molecule, while the bound ligand behaves as a large molecule [
29].
2.4. Roles of OA and OAm on Cs2NaInCl6 QDs
Based on the above results, it was shown that only OAm is bound to the QDs. Next, it is important and necessary to elucidate the respective roles of OA and OAm on the photophysical properties such PLQY and the stability. Therefore, we prepared five kinds of QDs by changing the ratio of [OA] and [OAm] in the synthesis process, i.e., [OA]/[OAm] is 4, 2, 1, 0.5, and 0.25, respectively.
The TEM images and the XRD pattern of the five kinds of QDs shown in
Figure 6. In a previous study, it was reported that the shape of the QDs changed when synthesized by changing the ratio of OA and OAm [
30]. However, the synthesized QDs here generally show cubic shapes (
Figure 6(a)-(e)). In addition, XRD results showed that the crystal structure of the QDs are cubic for all kinds of samples (
Figure 6(f)), and the average sizes of the QDs obtained from the TEM images were 11.2, 9.2, 9.9, 10.3, and 13.4 nm when [OA]/[OAm] was 4, 2, 1, 0.5, and 0.25, respectively.
Figure 7(a) and (b) show the PL spectra and PLQY of the five kinds of QDs, respectively. As the [OA]/[OAm] changed, the PL peak wavelength was almost the same and the shift of a few nm is maybe due to a little concentration change of the QD solution. On the other hand, the PLQY increased as [OA]/[OAm] was decreased, i.e., as the ratio of OAm increased as shown in
Figure 7(b). In addition, the PLQY could be achieved close to 100% when [OA]/[OAm] was 0.5 and 0.25, respectively. The high PLQY is thought to be due to the passivation of surface defects of the QDs when the OAm ratio was increased in the synthesis. From the previous result that only OAm was bound to the QDs, it is clear that it is OAm ligand has the role to passivate the surface defects of the Cs
2NaInCl
6 QDs.
FTIR measurement was performed to investigate the OA and OAm in the QD solution for different [OA]/[OAm] (
Figure 8). As a result, the C=O peak at 1710 cm
-1, which is a characteristic of OA, became smaller as [OA]/[OAm] decreased, and no peak was observed when [OA]/[OAm] was 0.5 and 0.25. This indicates that there is little residual OA in the QD solution.
Next, to investigate the change in stability when the [OA]/[OAm] changes, we measured and compared the PL spectra of the five kinds of samples after one month of storage at room temperature and in air. For the samples with [OA]/[OAm] of 4, 2, and 1, the PL spectra and the PL intensity were not changed after one month, indicating that the QDs were not degraded. On the other hand, for [OA]/[OAm] of 0.5 and 0.25, the PL intensity decreased and the PL spectrum also changed after one month, indicating that the stability of the QDs became worse when the OA ratio was decreased. In summary, the samples with high OA content showed high stability. This is thought to be because OA can ionize OAm, the latter are then reattached to the QDs after they leaved the QDs. In samples where the QD solution did not contain enough OA, it is thought that the OA could not reionize the released OAm.
Figure 9.
PL spectra of Cs2NaInCl6 QDs synthesized at (a) [OA]/[OAm]=4. (b) [OA]/[OAm]=2, (c) [OA]/[OAm]=1, (d) [OA]/[OAm]=0.5 and (e) [OA]/[OAm]=0.25, respectively, which were measured soon after preparation and stored at room temperature and in air for 1 month.
Figure 9.
PL spectra of Cs2NaInCl6 QDs synthesized at (a) [OA]/[OAm]=4. (b) [OA]/[OAm]=2, (c) [OA]/[OAm]=1, (d) [OA]/[OAm]=0.5 and (e) [OA]/[OAm]=0.25, respectively, which were measured soon after preparation and stored at room temperature and in air for 1 month.
To confirm that OA ionizes OAm and binds it back to the QDs, we compared the stability of a sample of QD solution with [OA]/[OAm] of 0.25 and a sample added with OAm when stored at room temperature and in air for 72 hours (
Figure 10). The QDs precipitated in the pure QD and OAm-added samples. On the other hand, no precipitation was observed in the sample to which oleic acid was added. This indicates that OA ionizes OAm and binds it to the QDs, as expected above.