The mixed cation colloidal Cs1−XFAXPbI3 perovskite quantum dots (PQDs) obtained by cation exchange between CsPbI3 and FAPbI3 PQDs have been reported to exhibit enhanced photovoltaic performance. However, the cation exchange mechanism requires further in-depth investigation in terms of both material properties and device application. In this work, the impact of PQD weight ratio, PQD concentration, and host solvent polarity during cation exchange is comprehensively investigated for the first time. In addition, the whole exchange process under varying conditions is monitored by photoluminescence spectroscopy. As a result, we observe extremely fast cation exchange (∼20 min) under a condition at a CsPbI3/FAPbI3 PQD weight ratio of 1:1, a concentration of 70 mg/ml, and a host solvent using toluene. Moreover, we directly fabricate a PQD solar cell device using these obtained mixed cation Cs0.5FA0.5PbI3 PQDs and achieved an enhanced power conversion efficiency of 14.58%. We believe that these results would provide more insights into the cation exchange in emerging PQDs toward efficient photovoltaic fabrication and application.

Organic–inorganic lead halide perovskite materials with a general structure of APbX3 (where A is Cs or an organic cation and X is halide) prepared by the solution process have achieved a power conversion efficiency (PCE) of 25.6% in solar cell application during the past decade.1–5 Many previous studies have revealed excellent photoelectric properties of these emerging materials, such as high extinction coefficient, low exciton binding energy, and sufficient carrier diffusion length.6–9 However, the current perovskite solar cells (PSCs) still face challenges such as large open circuit voltage (VOC) deficit and phase segregation under thermal or light stresses.10,11 Concurrently, nanometer-sized colloidal perovskite quantum dots (PQDs) or more broadly, nanocrystals (NCs), with reduced in both materials size and dimensionality were reported as an effective approach to overcome these issues.12 

In comparison with bulk thin-film perovskite, these PQDs have shown tremendous advantages such as a tunable bandgap, multi-exciton generation, and near-unity photoluminescence quantum yield (PLQY).13–17 The advantages were further amplified when all-inorganic CsPbI3 PQDs were first used in solar cells in 2016, where a PCE of 10.77% was achieved together with a high VOC value of 1.23 V and improved phase stability relative to thin-film perovskite.18 The performance of CsPbI3 PQD-based solar cells has been steadily improved with the growing study on surface passivation methodology and device architecture engineering, and the PCE of CsPbI3 PQD solar cells has been rapidly improved to over 16%.19–22 In addition, CsPbI3 PQDs provide many benefits in device technology due to the advantages of film forming and crystallization stability.23–27 Unfortunately, CsPbI3 PQDs, with a relatively wide bandgap of 1.75 eV, are not ideal for single-junction solar cells. Incorporation of formamidinium (FA) cation can reduce the bandgap to about 1.55 eV as well as better phase stability than CsPbI3.26 However, the current FAPbI3 PQDs exhibit problems such as unfavorable size distribution and excess surface ligands.27 Under these circumstances, the mixed A-site cation doping strategy can integrate the advantages of different cations. This particular strategy opened up more opportunities to go beyond simple CsPbI3 or FAPbI3 quantum dot (QD) solar cell devices, and the certified record efficiency of 16.6% for QD solar cells was achieved using Cs0.5FA0.5PbI3 composition.30 

Cs1−XFAXPbI3 PQDs have been proven to show enhanced photovoltaic properties; however, it is difficult to synthesize high-quality PQDs directly through traditional hot-injection methods due to the large synthetic difference between CsPbI3 (160 °C) and FAPbI3 (80 °C) PQDs.31–34 If the temperature is too low, CsPbI3 would tend to form an unconducive δ-phase, and if the temperature is too high, FA containing organic components would decompose.18,35 In addition to temperature, the purification process is also complicated. Therefore, cation exchange through directly mixing the as-prepared CsPbI3 with FAPbI3 PQDs in the solution phase is proposed, and Hazarika et al. successfully prepared high-quality Cs1−XFAXPbI3 PQDs in 2018 for photovoltaic applications.33 The exchange activation energy of Cs+ and FA+ cations is about 0.65 eV, and the mixed cation PQDs show a continuously tuning range of the bandgap with a wide emission changing in 680–780 nm wavelength region, demonstrating that the cation exchange method is feasible. In 2020, Hao et al. reported that the surface oleic acid (OA) ligand-rich conditions can protonate A-site cations, thereby lowering the cation exchange barrier and greatly increasing the exchange rate.30 

However, the as-prepared Cs1−XFAXPbI3 PQDs need to be further purified for solar cells fabrication, which is undesired toward further development. In addition, except for the surface ligands, the impact of other deciding factors like the host solvent during the cation exchange process needs further investigation for a better understanding of the mechanism. Herein, we systematically studied the effect of the mixing weight ratio of CsPbI3 and FAPbI3 PQDs, the concentration of PQDs, and the solvent environment on the rate of cation exchange (as shown in Fig. 1). As a result, we observe an extremely fast cation exchange process and the obtained mixed cation Cs1−XFAXPbI3 PQDs could be directly adopted for solar cell fabrication without any further purification delivering a champion efficiency of 14.58%.

FIG. 1.

Schematic illustration of cation exchange between the colloidal CsPbI3 and FAPbI3 PQDs.

FIG. 1.

Schematic illustration of cation exchange between the colloidal CsPbI3 and FAPbI3 PQDs.

Close modal

Cesium carbonate (Cs2CO3, 99.9%, Sigma), formamidinium acetate (FAAc, 99%, Sigma), lead iodide (PbI2, 99.9, Sigma), 1-octandecene (1-ODE, 90%, J&K), oleic acid (OA, 90%, Alfa), oleylamine (OAm, 90%, Alfa), n-hexane (Hex, 97.5%, J&K), n-octane (Oct, anhydrous, ≥98%, Alfa), toluene (Tol, 99.7%, Enox), chloroform (CF, 99.7%, Enox), methyl acetate (MeOAc, anhydrous, 99.5%, Sigma), and Poly[bis(4-phenyl)(2,4,6-tri-methy-lphenyl)amine] (PTAA) are purchased from Xi’an Polymer Light Technology Corp (China). All the materials were used directly without further purification. Glass/FTO was purchased from Advanced Election Technology Co. Ltd. (China).

CsPbI3 and FAPbI3 PQDs are prepared by following previously reported methods. For CsPbI3, 2 g of Cs2CO3, 8 ml of OA, and 100 ml of ODE were added into 250 ml three-neck flask and were degassed under vacuum at 90 °C for 1 h. Then, the flask was heated up to 120 °C under N2 until Cs2CO3 and OA completely reacted to form the Cs-oleate (Cs-OA) solution. Cs-OA was kept at 80 °C for PQD synthesis. Then, 1 g PbI2 and 50 ml 1-ODE were added into a 250 ml three-neck flask and were degassed under vacuum at 90 °C for 1 h. Then, 5 ml of OA and 5 ml of OAm were injected into the flask under N2. After totally dissolving PbI2, the mixture was heated up to 160 °C under N2. Then, 4 ml of the Cs-OA solution was injected into the flask for 5 s, and the reaction was quenched by an ice bath. The crude solution was divided into three centrifuge tubes to separate the CsPbI3 PQDs from the reaction liquor. 60 ml of MeOAc was added to the centrifuge tubes each at room temperature and centrifuged at 8000 rpm for 5 min. The precipitate was dispersed in 10 ml of hexane, followed by adding 10 ml MeOAc, and centrifuged at 8000 rpm for 3 min. Finally, the precipitate was re-dispersed in 10 ml hexane each and centrifuged at 4000 rpm for 5 min. Then, the CsPbI3 PQD solution was kept overnight in a refrigerator. For FAPbI3, see a similar synthesis process with CsPbI3 in the supplementary material.

First, hexane in the CsPbI3 and FAPbI3 PQD solution is spun dry. Colloidal solutions of CsPbI3 and FAPbI3 PQDs dispersed in octane with the same concentration of 20 mg/ml were mixed in different weight ratios (CsPbI3/FAPbI3 = 3:1, 1:1, and 1:3). Different concentrations (20, 50, and 70 mg/ml) of CsPbI3 and FAPbI3 PQD octane solutions were mixed to produce alloyed PQD solution with the same CsPbI3/FAPbI3 weight ratio, which is 1:1. Colloidal solutions of CsPbI3 and FAPbI3 PQDs were dispersed in a different solvent (octane, toluene, and chloroform) with the same concentration of 70 mg/ml and the same CsPbI3/FAPbI3 weight ratio, which is 1:1. Before mixing, the absorption spectra of the individual samples were measured, and the concentration was adjusted so that each solution had a similar optical density near the band edge.

The FTO/glass substrate was cleaned by industrial acetone, deionized water, isopropanol, and acetone for 15 min, respectively. The 40 nm TiO2 electron transfer layer was deposited on the FTO/glass substrate by means of chemical bath deposition as reported previously. Before depositing PQDs, the TiO2 film was treated by UV–ozone for 10 min. Then, 15 μl of as-prepared toluene solution of PQDs was spin-coated on the FTO/TiO2 substrate at 1000 rpm for 15 s and 2000 rpm for 20 s. Then, 150 μl of MeOAc was dropped on the QD film for 5 s to remove excess long-chain ligands and then spun at 2000 rpm for 20 s. This process was repeated five times to obtain thick enough QDs film. 15 mg/ml of Tris(pentafluorophenyl) borane-doped PTAA toluene solution was spin-coated on the as-prepared film as the hole transporting layer (HTL) at 3000 rpm for 40 s. Finally, 8 nm of MoO3 and 100 nm of the Ag electrode were deposited by thermal evaporation with a shadow mask of 7.25 mm2. The current density–voltage (JV) characteristics of the solar cells were measured using a Keithley 2400 Digital Source Meter under N2 glovebox and simulated AM 1.5 G spectrum at 100 mW/cm2 with a solar simulator (Class AAA, 94023A-U, and Newport). Before the test, the light intensity of the xenon lamp was calibrated with a standard silicon solar cell (91150V, Newport Oriel). JV scans were measured by the forward direction (−1.25 to 1.25 V) with the speed of 0.016 V per point and a scan rate of 100 mV/s.

To conduct cation exchange for as-prepared Cs1−XFAXPbI3 PQDs, the colloidal halide perovskite CsPbI3 and FAPbI3 PQDs in this work were first prepared according to the previous report through hot injection methods (details shown in Sec. II B). Transmission electron microscopic (TEM) (Fig. S1, supplementary material) images demonstrate that we have successfully obtained CsPbI3 and FAPbI3 PQDs with the average size of ∼10 and ∼12 nm, respectively. To precisely determine the proportion of A-site cations in the Cs1−XFAXPbI3 PQDs, we then measured the absorption spectra of the pristine PQD solutions before performing cation exchange, with the result shown in Fig. S2. In these ternary PQDs, the conduction band and valence band of perovskite materials are mainly affected by the s and p orbitals of the Pb and X-site atoms; hence, CsPbI3 and FAPbI3 PQD solutions have similar concentrations under the condition of similar optical density near the band edge. The obtained colloidal CsPbI3 PQD solution shows a wide bandgap with an absorption peak at 680 nm; in contrast, FAPbI3 PQD exhibits red-shifted absorption with a peak at 780 nm.

To prepare the mixed cation Cs1−XFAXPbI3 PQDs, the proportion of Cs/FA cations is mainly determined by the mixed weight ratio of two PQDs. Along this line, we first investigate the impact of the weight ratio of CsPbI3 and FAPbI3 PQDs on the cation exchange process. The concentration of both CsPbI3 and FAPbI3 PQDs is controlled at 20 mg/ml, Octane (Oct) is used as the solvent, and varying CsPbI3:FAPbI3 PQD weight ratios of 3:1, 1:1, and 1:3 were selected. As shown in Fig. S3, after mixing colloidal PQD solutions at different weight ratios, continuous modulation of the spectrum from 680 to 780 nm can be obtained. The PL emission peak position of different weight ratios varies linearly with the predicted compositions [seen in Fig. S3(b)]. It should be noted that in order to realize the facile and effective preparation of alloyed PQDs, the post-cation exchange reaction was performed at room temperature (20 °C) under the ambient condition (relative humidity: 20%–30%) without any stirring.

In order to understand the weight ratio on the rate of the cation exchange process, as shown in Fig. 2, the time-dependent PL measurements were performed to monitor the A-site cation exchange as a function of the time scale. The PL measurements show that the two initial emission peaks merge into a single peak over time. As shown in Figs. 2(a)2(c), the two PL peaks do not change distinctly from the original positions of CsPbI3 and FAPbI3 at least up to ∼120 min in all three systems. Since then, the high-energy and the low-energy peaks merge rapidly at a higher FAPbI3 PQD content system [Fig. 2(c)]. In specific, the time consumption decreases from ∼360 min for Cs0.75FA0.25PbI3 to 300 min for Cs0.5FA0.5PbI3, and then further down to 270 min for Cs0.25FA0.75PbI3, respectively. It is clearly that the FAPbI3 PQD rich condition slightly accelerates the exchange rate, which may be attributed to the higher ligand density on the surface of FAPbI3 compared to CsPbI3 PQDs. The excess surface ligands can protonate A-site cations, decreasing the barrier of cation exchange.11 However, it remains experimentally difficult to further accelerate the exchange process in these systems; the average exchange time is around 5 h, which is not desired for a practical application such as solar cell fabrication.

FIG. 2.

PL evolution at different PQD weight ratios: (a) Cs:FA = 3:1; (b) Cs:FA = 1:1; and (c) Cs:FA = 1:3 over time. (The concentration of the PQD solution is 20 mg/ml in the Oct solvent.)

FIG. 2.

PL evolution at different PQD weight ratios: (a) Cs:FA = 3:1; (b) Cs:FA = 1:1; and (c) Cs:FA = 1:3 over time. (The concentration of the PQD solution is 20 mg/ml in the Oct solvent.)

Close modal

In order to realize a more efficient cation exchange process, we further studied the influence of the PQD concentration. The weight ratio of CsPbI3 and FAPbI3 PQDs is controlled at 1:1 and PQD solutions at concentrations of 20, 50, and 70 mg/ml in the Oct solvent were selected. Figure 3 displays the evolution of PL emission peaks with time at different PQD solution concentrations. As can be seen in Fig. 3(a), consistent with the result of weight ratio regulation, the peak position does not move in the first 120 min. In contrast, as the PQD concentration increases, we observe the shift of PL peaks as shown in Figs. 3(b) and 3(c); the two PL peaks start to move substantially from the original positions around 60 min at concentrations of 50 mg/ml and starts even faster within 30 min in the case of70 mg/ml. Such an increase in the cation exchange rate as a function of the PQD concentration may be attributed to the increased chance of collisions between neighboring PQDs and further accelerates the exchange of two cations. Expectedly, as the PQD concentration increases, the time consumption to complete the cation exchange gradually decreases from ∼300 min for 20 mg/ml to ∼270 min for 50 mg/ml and then down to 240 min for 70 mg/ml conditions, respectively. All these results demonstrated that the PQD solution at the sufficient concentration would be an effective approaching to balancing the cation exchange time and further PQD device fabrication because the PQD concentration is ideal to obtain the thick QD film.

FIG. 3.

PL evolution at different PQD concentrations: (a) 20 mg/ml; (b) 50 mg/ml; and (c) 70 mg/ml over time (with the PQD weight ratios CsPbI3/FAPbI3 = 1:1 and the solvent is Oct).

FIG. 3.

PL evolution at different PQD concentrations: (a) 20 mg/ml; (b) 50 mg/ml; and (c) 70 mg/ml over time (with the PQD weight ratios CsPbI3/FAPbI3 = 1:1 and the solvent is Oct).

Close modal

In order to achieve a more efficient PQD solar cell for practical application, fast and controllable cation exchange is necessary. Therefore, we continue to investigate the impact of the host solvent polarity on the cation exchange. In comparison with Oct, the polar solvents, such as toluene (Tol) and chloroform (CF), were selected, which is widely adopted as the processing solvent for solar cell fabrication. The CsPbI3 and FAPbI3 PQDs were dispersed in Oct, Tol, and CF at the same concentration of 70 mg/ml, respectively, and the PQD weight ratio is controlled at 1:1. As can be seen from Figs. 4(a)4(c), the two PL peaks in the Oct solvent system starts to move substantially from the original positions at about 30 min. On the contrary, the peaks start to move at the very beginning under the conditions of Tol and CF, which is around 4 min in Tol and even earlier in the case of CF. Eventually, when conducting the cation exchange in the polar solvent, such as Tol and CF, the time consumption to complete the exchange greatly decreases from ∼240 min for Oct to ∼20 min for Tol and then to ∼10 min for CF. These results suggest that the cation exchange process is greatly decided by the polarity of the host solvent. Under the polar solvent environment, the enhanced solvent molecule polarity would enhance cation movement, which could reduce the activation energy of the cation exchange reaction and accelerate the exchange rate.36 However, dispersing PQD in these polar solvents usually leads to aggregation and unstable PQD solution, especially for less stable CsPbI3 PQDs. As shown in Fig. S4, the PL peak position of CsPbI3 PQDs dispersed in the CF solvent starts to red shift after 20 min, which may be the aggregation induced by the damage to the surface of the PQDs caused by the polar solvent. By contrast, the PQDs dispersed in Tol demonstrate good stability, and the PL peak position does not shift significantly as shown in Fig. S5.

FIG. 4.

PL evolution with different solvents: (a) octane, (b) toluene, and (c) chloroform over time (with the PQD weight ratio CsPbI3/FAPbI3 = 1:1 and the concentration is 70 mg/ml).

FIG. 4.

PL evolution with different solvents: (a) octane, (b) toluene, and (c) chloroform over time (with the PQD weight ratio CsPbI3/FAPbI3 = 1:1 and the concentration is 70 mg/ml).

Close modal

As shown in Fig. 5(a), according to the aforementioned experimental results, we further summarize the rate of the cation exchange rate as a function of three different factors, including PQD weight ratios, PQD concentrations, and host solvent polarity. It is obvious that higher polarity of the solvent greatly improves the efficiency of cation exchange. After considering the cation exchange time, the stability of PQD solution, and further PQD device fabrication, we demonstrated the optimal condition for preparing Cs1−XFAXPbI3 PQDs:CsPbI3:FAPbI3 weight ratio of 1:1, the PQD concentration of 70 mg/ml, and toluene as the solvent. Under such an optimized condition, Cs1−XFAXPbI3 PQD post-synthesis through cation exchange can be complete in ∼20 min. As characterized by TEM images [Fig. 5(b)] together with the statistical averages of the QD size (Fig. S7), the Cs0.5FA0.5PbI3 PQDs exhibit similar size and distribution relative to that of pristine CsPbI3 and FAPbI3 PQDs, evidently confirming that that such cation exchange does not significantly alter the crystal structure and morphology of PQDs. To verify the application of such fast cation exchange, we further fabricated PQD solar cells adopting these obtained Cs1−XFAXPbI3 PQDs (without any further purification process). As shown in Fig. 5(c), we fabricated PQD solar cells with a planar structure of glass/fluorine-doped tin oxide (FTO)/TiO2/PQDs/PTAA/MoO3/Ag, and each layer of the device is clearly identified in the cross-sectional scanning electron microscopy (SEM) image (Fig. S6). Figure 5(d) shows the optimal current density–voltage (JV) characteristics of the CsPbI3, FAPbI3, and Cs0.5FA0.5PbI3 PQD solar cells under reverse scan at AM 1.5 G, 100 mW/cm2 simulated solar illumination, with device parameters shown in Table S1. The CsPbI3 PQD device shows a PCE of 13.44% with a VOC value of 1.251 V, a short-circuit current density (JSC) of 14.88 mA/cm2, and a fill factor (FF) of 0.722, and the FAPbI3 PQD device shows a PCE of 11.93% with VOC of 1.181 V, JSC of 15.63 mA/cm2, and FF of 0.646. In contrast, the best performing Cs0.5FA0.5PbI3 cell output with an enhanced PCE reached 14.58% with VOC of 1.191 V, JSC of 16.82 mA/cm2, and FF of 0.724. The top-view SEM images (Fig. S8) show the surface morphology of the PQD film. In comparison with the FAPbI3 PQD film, the mixed cation PQDs exhibit dense surface morphology with less pinholes.

FIG. 5.

(a) Comparison of exchange velocity under three different conditions. (b) TEM images of CsPbI3, Cs0.5FA0.5PbI3, and FAPbI3 PQDs. (c) The device structure of PQD solar cells. (d) JV curves of CsPbI3, FAPbI3, and Cs0.5FA0.5PbI3 PQD solar cells measured under the influence of the AM 1.5 G solar spectrum at 100 mW/cm2.

FIG. 5.

(a) Comparison of exchange velocity under three different conditions. (b) TEM images of CsPbI3, Cs0.5FA0.5PbI3, and FAPbI3 PQDs. (c) The device structure of PQD solar cells. (d) JV curves of CsPbI3, FAPbI3, and Cs0.5FA0.5PbI3 PQD solar cells measured under the influence of the AM 1.5 G solar spectrum at 100 mW/cm2.

Close modal

In conclusion, through adjusting the weight ratio of CsPbI3 and FAPbI3 PQDs, the PQD concentration, and the polarity of the host solvent, we in-depth investigated the key factors affecting the cation exchange of Cs1−XFAXPbI3 PQDs. A fast cation exchange (∼20 min) process was observed under a condition at a CsPbI3:FAPbI3 PQD weight ratio of 1:1, a PQD concentration of 70 mg/ml, and the host solvent being toluene. It should be noted that the fast cation exchange process here with any thermal treatment has not been observed in the previous report. More importantly, such a cation exchange allows us to directly adopt the obtained Cs0.5FA0.5PbI3 PQD for solar cell fabrication, delivering an enhanced PCE of 14.58%, outperformed either pristine CsPbI3 or FAPbI3 PQDs. These findings in this work would provide an efficient approach to the design and synthesis of mixed cation PQDs and enhance the fabrication and performance of PQD solar cells.

See the supplementary material for UV–visible absorption data, TEM, and SEM images of PQDs.

This work was financially supported by the National Key Research and Development Program of China (Grant No. 2019YFE0108600), the National Natural Science Foundation of China (Grant Nos. 52073198 and 22161142003), the Natural Science Foundation of Jiangsu Province (Grant No. BK20211598), the Science and Technology Program of Suzhou (Grant No. SYG202037), “111” project, the Young Elite Scientist Sponsorship Program by CAST, the Collaborative Innovation Center of Suzhou Nano Science and Technology, and Soochow University.

The authors have no conflicts to disclose.

The data that support the findings of this study are available from the corresponding author upon reasonable request.

1.
J.
Jeong
,
M.
Kim
,
J.
Seo
,
H.
Lu
,
P.
Ahlawat
,
A.
Mishra
,
Y.
Yang
 et al,
Nature
592
,
381
(
2021
).
2.
A.
Kojima
,
K.
Teshima
,
Y.
Shirai
, and
T.
Miyasaka
,
J. Am. Chem. Soc.
131
,
6050
(
2009
).
3.
M. M.
Lee
,
J.
Teuscher
,
T.
Miyasaka
,
T. N.
Murakami
, and
H. J.
Snaith
,
Science
338
,
643
(
2012
).
4.
M.
Liu
,
M. B.
Johnston
, and
H. J.
Snaith
,
Nature
501
,
395
(
2013
).
5.
Y.
Kanemitsu
,
J. Chem. Phys.
151
,
170902
(
2019
).
6.
T. C.
Sum
,
M.
Righetto
, and
S. S.
Lim
,
J. Chem. Phys.
152
,
130901
(
2020
).
7.
J.
Burschka
,
N.
Pellet
,
S.-J.
Moon
,
R.
Humphry-Baker
,
P.
Gao
,
M. K.
Nazeeruddin
, and
M.
Grätzel
,
Nature
499
,
316
(
2013
).
8.
S. D.
Stranks
,
G. E.
Eperon
,
G.
Grancini
,
C.
Menelaou
,
M. J. P.
Alcocer
,
T.
Leijtens
,
L. M.
Herz
,
A.
Petrozza
, and
H. J.
Snaith
,
Science
342
,
341
(
2013
).
9.
J.-W.
Lee
,
D.-H.
Kim
,
H.-S.
Kim
,
S.-W.
Seo
,
S. M.
Cho
, and
N.-G.
Park
,
Adv. Energy Mater.
5
,
1501310
(
2015
).
10.
S.-H.
Turren-Cruz
,
A.
Hagfeldt
,
M.
Saliba
,
S. H.
Turren-Cruz
,
A.
Hagfeldt
, and
M.
Saliba
,
Science
362
,
449
(
2018
).
11.
M.
Saliba
,
T.
Matsui
,
K.
Domanski
,
J.-Y.
Seo
,
A.
Ummadisingu
,
S. M.
Zakeeruddin
,
J.-P.
Correa-Baena
,
W. R.
Tress
,
A.
Abate
,
A.
Hagfeldt
, and
M.
Grätzel
,
Science
354
,
206
(
2016
).
12.
J.
Yuan
,
N.
Rujisamphan
,
W.
Ma
,
J.
Yuan
,
Y.
Li
, and
S.-T.
Lee
,
J. Energy Chem.
62
,
505
(
2021
).
13.
L.
Protesescu
,
S.
Yakunin
,
M. I.
Bodnarchuk
,
F.
Krieg
,
R.
Caputo
,
C. H.
Hendon
,
R. X.
Yang
,
A.
Walsh
, and
M. V.
Kovalenko
,
Nano Lett.
15
,
3692
(
2015
).
14.
X.
Li
,
Y.
Wu
,
S.
Zhang
,
B.
Cai
,
Y.
Gu
,
J.
Song
, and
H.
Zeng
,
Adv. Funct. Mater.
26
,
2435
(
2016
).
15.
M.
Li
,
R.
Begum
,
J.
Fu
,
Q.
Xu
,
T. M.
Koh
,
S. A.
Veldhuis
,
M.
Gratzel
,
N.
Mathews
,
S.
Mhaisalkar
, and
T. C.
Sum
,
Nat. Commun.
9
,
4197
(
2018
).
16.
J.
Pan
,
Y.
Shang
,
J.
Yin
,
M.
De Bastiani
,
W.
Peng
,
I.
Dursun
,
L.
Sinatra
,
A. M.
El-Zohry
,
M. N.
Hedhili
,
A.-H.
Emwas
,
O. F.
Mohammed
,
Z.
Ning
, and
O. M.
Bakr
,
J. Am. Chem. Soc.
140
,
562
(
2018
).
17.
D.
Rossi
 et al,
J. Chem. Phys.
153
,
184703
(
2020
).
18.
A.
Swarnkar
,
A. R.
Marshall
,
E. M.
Sanehira
,
B. D.
Chernomordik
,
D. T.
Moore
,
J. A.
Christians
,
T.
Chakrabarti
, and
J. M.
Luther
,
Science
354
,
92
(
2016
).
19.
J.
Chen
,
D.
Jia
,
E. M. J.
Johansson
,
A.
Hagfeldt
, and
X.
Zhang
,
Energy Environ. Sci.
14
,
224
261
(
2021
).
20.
E. M.
Sanehira
,
A. R.
Marshall
,
J. A.
Christians
,
S. P.
Harvey
,
P. N.
Ciesielski
,
L. M.
Wheeler
,
P.
Schulz
,
L. Y.
Lin
,
M. C.
Beard
, and
J. M.
Luther
,
Sci. Adv.
3
,
eaao4204
(
2017
).
21.
X.
Ling
,
S.
Zhou
,
J.
Yuan
,
J.
Shi
,
Y.
Qian
,
B. W.
Larson
,
Q.
Zhao
,
C.
Qin
,
F.
Li
,
G.
Shi
,
C.
Stewart
,
J.
Hu
,
X.
Zhang
,
J. M.
Luther
,
S.
Duhm
, and
W.
Ma
,
Adv. Energy Mater.
9
,
1900721
(
2019
).
22.
F.
Li
,
S.
Zhou
,
J.
Yuan
,
C.
Qin
,
Y.
Yang
,
J.
Shi
,
X.
Ling
,
Y.
Li
, and
W.
Ma
,
ACS Energy Lett.
4
,
2571
(
2019
).
23.
J.
Shi
,
F.
Li
,
J.
Yuan
,
X.
Ling
,
S.
Zhou
,
Y.
Qian
, and
W.
Ma
,
J. Mater. Chem. A
7
,
20936
(
2019
).
24.
Y.
Wang
,
J.
Yuan
,
X.
Zhang
,
X.
Ling
,
B. W.
Larson
,
Q.
Zhao
,
Y.
Yang
,
Y.
Shi
,
J. M.
Luther
, and
W.
Ma
,
Adv. Mater.
32
,
2070243
(
2020
).
25.
X.
Zhang
,
H.
Huang
,
X.
Ling
,
J.
Sun
,
X.
Jiang
,
Y.
Wang
,
D.
Xue
,
L.
Huang
,
L.
Chi
,
J.
Yuan
, and
W.
Ma
,
Adv. Mater.
34
,
2105977
(
2022
).
26.
K.
Ji
,
J.
Yuan
,
F.
Li
,
Y.
Shi
,
X.
Ling
,
X.
Zhang
,
Y.
Zhang
,
H.
Lu
,
J.
Yuan
, and
W.
Ma
,
J. Mater. Chem. A
8
,
8104
(
2020
).
27.
J.
Yuan
,
X.
Ling
,
D.
Yang
,
F.
Li
,
S.
Zhou
,
J.
Shi
,
Y.
Qian
,
J.
Hu
,
Y.
Sun
,
Y.
Yang
,
X.
Gao
,
S.
Duhm
,
Q.
Zhang
, and
W.
Ma
,
Joule
2
,
2450
(
2018
).
28.
J.
Xue
,
J.-W.
Lee
,
Z.
Dai
,
R.
Wang
,
S.
Nuryyeva
,
M. E.
Liao
,
S.-Y.
Chang
,
L.
Meng
,
D.
Meng
,
P.
Sun
,
O.
Lin
,
M. S.
Goorsky
, and
Y.
Yang
,
Joule
2
,
1866
(
2018
).
29.
M.
Que
,
Z.
Dai
,
H.
Yang
,
H.
Zhu
,
Y.
Zong
,
W.
Que
,
N. P.
Padture
,
Y.
Zhou
, and
O.
Chen
,
ACS Energy Lett.
4
,
1970
(
2019
).
30.
M.
Hao
,
Y.
Bai
,
S.
Zeiske
,
L.
Ren
,
J.
Liu
,
Y.
Yuan
,
N.
Zarrabi
,
N.
Cheng
,
M.
Ghasemi
,
P.
Chen
,
M.
Lyu
,
D.
He
,
J.-H.
Yun
,
Y.
Du
,
Y.
Wang
,
S.
Ding
,
A.
Armin
,
P.
Meredith
,
G.
Liu
,
H.-M.
Cheng
, and
L.
Wang
,
Nat. Energy
5
,
79
(
2020
).
31.
L.
Protesescu
,
S.
Yakunin
,
S.
Kumar
,
J.
Bär
,
F.
Bertolotti
,
N.
Masciocchi
,
A.
Guagliardi
,
M.
Grotevent
,
I.
Shorubalko
,
M. I.
Bodnarchuk
,
C.-J.
Shih
, and
M. V.
Kovalenko
,
ACS Nano
11
,
3119
(
2017
).
32.
I.
Lignos
,
V.
Morad
,
Y.
Shynkarenko
,
C.
Bernasconi
,
R. M.
Maceiczyk
,
L.
Protesescu
,
F.
Bertolotti
,
S.
Kumar
,
S. T.
Ochsenbein
,
N.
Masciocchi
,
A.
Guagliardi
,
C.-J.
Shih
,
M. I.
Bodnarchuk
,
A. J.
deMello
, and
M. V.
Kovalenko
,
ACS Nano
12
,
5504
(
2018
).
33.
A.
Hazarika
,
Q.
Zhao
,
E. A.
Gaulding
,
J. A.
Christians
,
B.
Dou
,
A. R.
Marshall
,
T.
Moot
,
J. J.
Berry
,
J. C.
Johnson
, and
J. M.
Luther
,
ACS Nano
12
,
10327
(
2018
).
34.
M.
Suri
,
A.
Hazarika
,
B. W.
Larson
,
Q.
Zhao
,
M.
Vallés-Pelarda
,
T. D.
Siegler
,
M. K.
Abney
,
A. J.
Ferguson
,
B. A.
Korgel
, and
J. M.
Luther
,
ACS Energy Lett.
4
,
1954
(
2019
).
35.
B. w.
Park
and
S. I.
Seok
,
Adv. Mater.
31
,
1805337
(
2019
).
36.
Y.
Sun
,
H.
Zhang
,
K.
Zhu
,
W.
Ye
,
L.
She
,
X.
Gao
,
W.
Ji
, and
Q.
Zeng
,
RSC Adv.
11
,
27333
(
2021
).

Supplementary Material