Performance enhancement of perovskite solar cells using trimesic acid additive in the two-step solution method

Performance enhancement of perovskite solar cells using trimesic acid additive in the two-step solution method

Journal of Power Sources 426 (2019) 11–15 Contents lists available at ScienceDirect Journal of Power Sources journal homepage: www.elsevier.com/loca...

2MB Sizes 2 Downloads 72 Views

Journal of Power Sources 426 (2019) 11–15

Contents lists available at ScienceDirect

Journal of Power Sources journal homepage: www.elsevier.com/locate/jpowsour

Performance enhancement of perovskite solar cells using trimesic acid additive in the two-step solution method

T

Lijun Sua, Yaoming Xiaoa,b,∗, Gaoyi Hana, Liping Lua, Honggang Lic, Miaoli Zhua,∗∗ a Institute of Molecular Science, Key Laboratory of Chemical Biology and Molecular Engineering of Education Ministry, Innovation Center of Chemistry and Molecular Science, Key Laboratory of Materials for Energy Conversion and Storage of Shanxi Province, Shanxi University, Taiyuan, 030006, PR China b College of Chemical Engineering and Materials Science, Quanzhou Normal University, Quanzhou, 362000, PR China c School of Chemistry and Chemical Engineering, Liaocheng University, Liaocheng, 252059, PR China

H I GH L IG H T S

G R A P H I C A L A B S T R A C T

acid (TMA) was first used as • Trimesic an additive in the lead precursor solution.

TMA additive could control • Moderate the morphology improvement of perovskite film.

TMA additive could significantly im• prove the stability of perovskite solar cell.

solar cell with 5% TMA • Perovskite additive gained the highest efficiency of 17.21%.

A R T I C LE I N FO

A B S T R A C T

Keywords: Perovskite solar cell Additive Two-step solution method Trimesic acid

The biggest challenge for the pursuit of potentially available organic-inorganic perovskite solar cell is to achieve both high power conversion efficiency and good long-term stability. Here, the trimesic acid is first used as an additive in the lead precursor solution to prepare perovskite film by a two-step solution method to improve the performances of perovskite solar cell. After adding trimesic acid into the lead precursor, the device efficiency is increased from 14.27% to 17.21% and the device stability is improved. This can be attributed to the improvement of crystal structure and quality of perovskite film. The results indicate that the use of trimesic acid additive by a two-step solution is a great way to design high-efficiency and stable perovskite solar cells.

1. Introduction Photovoltaic cells based on two-dimensional and three-dimensional perovskites have become a hot research field and resulted in a certified power conversion efficiency (PCE) of 23.7% [1–8]. The quality of the perovskite film is important for the photovoltaic capacity of the corresponding cell [9–11]. However, the perovskite films prepared by solution method often cause many grain boundaries and defects, which

can be used as nonradiative recombination centers, thereby reducing the performance of solar cells [12,13]. In addition, the organic components can easily migrate under conditions of high temperature and/ or humidity, which can accelerate the decomposition of perovskite films [14,15]. Therefore, the high-quality perovskite film with good crystallinity, low defect, excellent surface morphology, and low migration rate of organic components could improve the photovoltaic performance and stability of the solar cell.



Corresponding author. Institute of Molecular Science, Key Laboratory of Chemical Biology and Molecular Engineering of Education Ministry, Innovation Center of Chemistry and Molecular Science, Key Laboratory of Materials for Energy Conversion and Storage of Shanxi Province, Shanxi University, Taiyuan, 030006, PR China. ∗∗ Corresponding author. E-mail addresses: [email protected] (Y. Xiao), [email protected] (M. Zhu). https://doi.org/10.1016/j.jpowsour.2019.04.024 Received 29 January 2019; Received in revised form 1 April 2019; Accepted 6 April 2019 0378-7753/ © 2019 Elsevier B.V. All rights reserved.

Journal of Power Sources 426 (2019) 11–15

L. Su, et al.

Much efforts have been expanded to modify the perovskite films, such as preparation methods [16–21], additive engineering [22–27] and antisolvents [28], etc. Currently, the two-step solution method is considered to be an efficient method to prepare perovskite films which have controllable morphology, crystallization and better reproducibility than the single-step method [16,17]. Compared to the deposition method, the effect of additive on the perovskite film is also crucial. The materials with hydroxyl group (-OH) have been introduced into the perovskite precursor solution, which would interact with ions of perovskite to influence their growth mechanism and form high-quality perovskite films [27]. Specifically, the interaction between hydrogen bond of hydroxyl group and iodide could passivate the surface of perovskite films [14,15]. Terephthalic acid was added into the perovskite precursor solution by one-step method to obtain more stable and highly efficient PSC than original PSC [14]. Jiang et al. used hydroxylamine hydrochloride as an additive into perovskite solution to improve the performance of PSCs [15]. However, there were few reports about using an additive with hydroxyl group by two-step method to enhance the PCE and stability of PSCs. In this paper, trimesic acid (TMA) as an additive was first incorporated into the lead precursor solution by a two-step method. On the one hand, the rigidity and the π-π bond effect of benzene ring make the PSC more stable. On the other hand, the interaction of the hydrogen bond between the hydrogen atom in the hydroxyl group and the iodine atom in the perovskite could suppress the ion migration of perovskite. At the same time, the two-step solution method significantly improved the quality of perovskite films and reproducibility of the experiments. Based on the high quality of perovskite films, PSC with TMA additive showed a PCE of 17.21% with great humidity and thermal stability, which was superior to that of original PSC without TMA.

Fig. 1. FTIR spectra of the pure TMA and the MAPbI3 film with TMA.

Device characterization: The infrared spectrometric analyzer (BRUKER TENSOR 27) was used to record the fourier transform infrared (FTIR) spectra. The steady-state photoluminescence (PL) curves were performed using an Edinburgh Instruments FLS980. Field emission scanning electron microscope (FESEM) measurements were conducted using a JEOL-JSM-6701F operated at 10 kV. The powder X-ray diffraction (XRD) patterns were recorded using a BRUKER D8ADVANCE. The ultraviolet to visible (UV–Vis) absorption spectra of films were measured using the Agilent 8453 UV–Vis diode array spectrophotometer. The photocurrent density-voltage (J-V) characteristic of the PSC was carried out using a computer-controlled CHI660D in ambient atmosphere. The incident light intensity was set under 100 mW cm−2 (AM 1.5 G). The active cell area was controlled to 0.09 cm2 by applying a black mask.

2. Experimental Reagents and Materials: Tetrabutyl titanate, methylamine alcohol solution, N,N-dimethyl sulfoxide (DMSO), trimesic acid (TMA), lead iodide (PbI2, 99.9%), acetonitrile, N,N-dimethylformamide (DMF), methanol, isopropyl alcohol, n-butanol, ether, hydroiodic acid, 4-tertbutyl-pyridine, and lithium bis(trifluoromethylsulfonyl) imide (Li-TFSI) were obtained from Aldrich. The tetrabutyl titanate and acetylacetone were purchased from the Sinopharm Chemical Reagent Co. Ltd. SpiroOMeTAD was gained from Xi'an Bolet Optoelectronics Technology Co., Ltd. The above agents were used for experiment directly. F-doped tin oxide (FTO) substrates were gained from NSG, Japan. CH3NH3I (MAI) and the TiO2 paste were prepared according to previous papers [29,30]. Device fabrication: A TiO2 precursor solution was spin-coated on a cleaned FTO substrate at a spin rate of 500 rpm for 9 s and 2500 rpm for 30 s, and then annealed at 450 °C for 30 min to obtain the TiO2 compact layer. The TiO2 precursor solution was a mixed solution of 25 mL nbutanol, 5 mL tetrabutyl titanate, 5 mL isopropyl alcohol, 5 mL acetylacetone, and 0.6 mL emulsifier of Triton X-100. On the surface of compact TiO2, TiO2 paste was deposited (at 500 rpm for 9 s and 2500 rpm for 30 s) and followed by annealing at 450 °C for 30 min to obtain TiO2 electronic transport layer (ETL) [29]. 1.0 mol L−1 PbI2 precursor solution without or with a certain amount of TMA (mass ratio of TMA vs PbI2) in a mixed solvent (VDMF: VDMSO = 7:3) was spincoated on the TiO2 ETL at 500 rpm for 9 s and 2500 rpm for 30 s, then baked at 100 °C for 10 min to prepare the PbI2 films without or with TMA. After drying, MAI solution (100 μL) was dropped on the PbI2 films, then the samples were allowed to rest for 1 min and rinsed with isopropyl alcohol (500 μL). Subsequently, the substrate was dried at 100 °C to prepare the MAPbI3 films without or with TMA. The hole transfer layer (HTL) with a concentration of 65 mM Spiro-OMeTAD, 30 mM Li-TFSI and 200 mM 4-tert-butyl-pyridine was spin-coated on MAPbI3 film at 500 rpm for 6 s and 2000 rpm for 20 s. After 8 h in air under room temperature and 0.10% of humidity, the Ag electrode was deposited by thermal evaporation.

3. Results and discussion The interaction between TMA and perovskite was collected by FTIR test (Fig. 1). The stretching vibration of C = O in TMA appears at 1718 cm−1, which shifts to 1631 cm−1 in the MAPbI3 film with TMA. The reduction of C = O stretching vibration frequency is due to the partial electron cloud migration from C = O to O·Pb (Fig. 2), which indicates that TMA strongly interacts with PbI2 [31]. FESEM was conducted to assess the impact of TMA additive on the perovskite films. The film quality with few grain boundaries and pin holes is regarded as a key factor to obtain high efficient PSCs, which could improve light harvesting and provide high-speed charge-transport channels [14,15,31]. As shown in Fig. 3, the grain size of the perovskite film without TMA is small (Fig. 3 a). When introducing TMA additive into the PbI2 solution, perovskite films of Fig. 3 b and c exhibit larger grains than that of Fig. 3 a. However, a poor-quality perovskite film with many small grains was formed after increasing the TMA amount to 10% (Fig. 3 d). It implies that moderate TMA additive could control the morphology improvement of MAPbI3 film. This is due to the controlled reaction kinetics of DMF-DMSO-PbI2 and MAI at the appropriate concentration of TMA, which is similar to the growth mechanism of terephthalic acid [32]. In detail, the retarding crystallization is

Fig. 2. Molecular structure of TMA and schematic of the proposed reaction between TMA and Pb2+. 12

Journal of Power Sources 426 (2019) 11–15

L. Su, et al.

Fig. 3. Top-view FESEM images of MAPbI3 films without (a) and with (b) 2.5, (c) 5, and (d) 10% TMA additive, respectively. Fig. 4. XRD patterns of the FTO, PbI2, MAPbI3 films without (a) and with (b) 2.5, (c) 5, and (d) 10% TMA additive, respectively.

without and with different amounts of TMA, respectively. The peaks at 2θ of 14.08, 19.89, 23.31, 24.57, 28.16, 31.72, 40.49, and 42.98° for MAPbI3 films can be indexed to the (1 1 0), (1 1 2), (2 1 1), (2 0 2), (2 2 0), (3 1 0), (2 2 4), and (3 1 4) diffraction of perovskite grains, respectively [29]. And other peaks observed at 26.59, 37.95, 51.78 and 65.8° are assigned to the FTO substrate. Compared to pristine MAPbI3, the absolute intensities of the (1 1 0) peak of MAPbI3 with TMA became stronger, which implied that TMA additive could improve the perovskite crystallinity. It is well known that the crystallinity of the perovskite layer determines the performance of the solar cell because defects in the crystals will result in severe trapping sites and shorting for charge recombination. Higher crystallinity of the perovskite will also affect charge dissociation and diffusion length [34]. The peak at 12.54° in MAPbI3 films without or with different amounts of TMA demonstrates the presence of some PbI2 in the perovskite films, which attributes to the incomplete reaction between PbI2 and MAI. Moreover, the peak intensity at 12.54° of PbI2 decreased when the TMA amount increased, but when the TMA amount increased to 10%, the peak intensity at 12.54° again increased. This phenomenon indicates the TMA additive could influence the reaction extent of the PbI2 with MAI by controlling the crystal growth rate. These changes in surface morphology and crystallinity of the perovskite film also mean that photovoltaic performance can be adjusted with TMA additive, which are validated in subsequent UV–Vis absorption, steady-state PL, and J-V tests. The UV–Vis absorption spectra (Fig. 5 A) of MAPbI3 samples were measured to evaluate the characteristics of light absorption. The light absorption intensities of samples were gradually enhanced with TMA amount increasing owing to the combined effects of crystallinity and larger grains of perovskite. However, when the amount of the additive is 10%, the light absorption intensity is lowered and the reason for the decrease might due to the poor perovskite film with many small perovskite grains. In order to further research the effect of additive on the quality of perovskite films, we tested the optoelectronic behavior of the perovskite films by means of steady-state PL. As shown in Fig. 5 B, the samples display emission peaks at about 783 nm, closing to the previous paper [35]. The perovskite film with TMA additive shows stronger peak intensity than that of the film without TMA. It demonstrates that the enhanced crystallinity and reduced surface defect could suppressed the non-radiative decay [36]. For the perovskite film with 10% TMA additive, it leads to the poor morphology with much more pinhole and small perovskite grains, which results in carrier shunting and

Scheme 1. Representation of the interaction between perovskite and TMA.

generally recognized as an effective approach to control the course of crystallization to obtain perovskite films with high quality [33]. After adding appropriate amount of TMA additive into the PbI2 solution, the TMA additive with a benzene ring and three carboxyl groups has a strong interaction with the metal cation by a chelation effect (Scheme 1), which could coordinate with Pb2+ to form an intermediate phase. When the PbI2 film reacting with MAI, the intermediate phase changed into the perovskite phase gradually with the removal of solvents during the heating process, which could lengthen the crystallization time of the perovskite film. Moreover, the intermediate phase could also regulate crystallization kinetics by reducing the configuration entropy of perovskite film formation and resulting in a smaller Gibbs free energy, which could increase the crystallization rate of the perovskite film. Since the dynamic balance of the perovskite film is affected during heating process, the perovskite film has better crystallinity and surface coverage. Furthermore, TMA can regulate the internal grain of perovskite during perovskite deposition. The carboxylic acid group could form hydrogen bond with the halide anion of the perovskite (Scheme 1) due to that the hydrogen atom could covalently bond to an atom X with a large electronegativity (X = F, Cl, I, O), a similar mechanism has been proven in the previous paper [27]. When the three terminals of the carboxyl groups in the TMA combined with the surface of the perovskite grains, the TMA could crosslink adjacent perovskite grains, which promoted the cohesion of the perovskite grains and improved inter-connectivity of the perovskite film. Therefore the perovskite film with large grain size was obtained. However, after adding excess TMA additive into the PbI2 solution, the excess intermediate phase might inversely hinder the proceeding of perovskite film formation, resulting in a poor-quality perovskite film with many small grains. Fig. 4 reveals XRD patterns of the FTO, PbI2, and MAPbI3 films 13

Journal of Power Sources 426 (2019) 11–15

L. Su, et al.

Fig. 5. (A) UV–Vis absorption spectra, (B) steady-state PL spectra of MAPbI3 films without (a) and with (b) 2.5, (c) 5, and (d) 10% TMA additive, respectively.

of PSC with TMA addition can still maintain 49% under the same test conditions. The photographs of the perovskite layers with annealing at 100 °C for 10 h in air are shown in Fig. 7 B. The bare perovskite film easily decomposed to PbI2 with yellow color after baking 10 h at 100 °C. In contrast, the color of the perovskite layer with TMA additive remains dark brown after 10 h. The above phenomenon indicates that the TMA additive can effectively prevent the decomposition of MAPbI3 into PbI2 and MAI to improve the thermal stability of perovskite films. The air stability is another critical issue in assessing the performance of perovskite solar cells. Fig. 8 A shows the time change of PCE derived from the J-V curves for PSCs without encapsulation. The PSC with TMA additive kept 71% of the original PCE, whereas bare PSC without TMA only retained 46% of the original PCE value with a relative humidity (RH) of 30% after 20 days. The 12.7° peak of the PbI2 phase was formed during the perovskite degradation process (Fig. 8 B). The stronger PbI2 peak intensity reveals that the bare MAPbI3 film without TMA was seriously destroyed and exposed more PbI2 because of humidity immersion. From the insets in Fig. 8 A, it can be clearly seen that after 20 days, the perovskite film without TMA additive turns yellow, while the perovskite film with TMA additive is dark brown. The PSC with TMA additive maintains a good stability, which is determined by the structure of the TMA additive itself (TMA with a benzene ring and three carboxyl groups). The strong hydrogen bond between hydroxyl group and iodide could play an important role of anchoring effect to suppress the loss of iodide ion, therefore preventing the perovskite from decomposing. Moreover, the benzene ring with rigidity and the π-π bond effect, as the hydrophobic alkyl chains could further protect the perovskite from reacting with water [15]. At the same time, we can see that the silver electrode is no longer bright after 20 days regardless of whether or not additives are added, indicating that the silver electrode was also damaged by the oxygen in air and iodide from the decomposition of MAPbI3. Therefore, replacing the Ag electrode by gold or carbon-based electrode and packaging the device to avoid being oxidized by air could further enhance the long-term stability of the PSC device.

Fig. 6. Photocurrent density-voltage characteristics of PSCs based on MAPbI3 films without (a) and with (b) 2.5, (c) 5, and (d) 10% TMA additive, respectively. Table 1 The photovoltaic performances of the PSCs under sunlight illumination of 100 mW cm−2 (AM 1.5 G). In which the results are based on more than 20 cells for each type, and the average values are summarized. PSC

JSC (mA cm−2)

VOC (V)

a b c d

21.75 22.98 23.54 21.02

1.02 1.03 1.03 1.02

± ± ± ±

0.15 0.15 0.11 0.14

± ± ± ±

FF 0.03 0.02 0.01 0.03

0.67 0.69 0.71 0.66

PCE (%) ± ± ± ±

0.02 0.04 0.02 0.03

14.86 16.33 17.21 14.15

± ± ± ±

0.06 0.13 0.07 0.10

recombination. Fig. 6 shows the J-V curves of PSCs without and with TMA and the corresponding photovoltaic parameters are summarized in Table 1. The bare MAPbI3 without TMA showed short circuit current density (JSC) of 21.21 mA cm−2, open-circuit voltage (Voc) of 1.02 V, fill factor (FF) of 0.66, and PCE of 14.27%. When the TMA doped into the PbI2, the performances of PSCs were improved. The PSC with 5% TMA additive gained the highest PCE of 17.21% with a JSC of 23.54 mA cm−2, VOC of 1.03 V, and FF of 0.71. However, when the doping amount to 10%, the PCE values dropped rapidly which might due to the formation of the coarse perovskite film. The J-V results are consistent with those of results above. The thermal stability is an important characteristic for the PSCs. Fig. 7 A shows the PCE of PSC without TMA provided only 20% of its initial value after annealing at 100 °C for 10 h in air. Whereas, the PCE

4. Conclusions In summary, we demonstrated a significant enhancement PCE from 14% to over 17% in the PSCs via introduced TMA additive into the lead precursor solution by a two-step solution method. The TMA additive plays a key role in morphology and crystal structure of perovskite films, which is crucial for the photovoltaic capacity of the corresponding cell. More importantly, the TMA additive can significantly improve thermal stability and air stability of PSCs. The improvement of stability and PCE can be explained from the following aspects: (1) the rigidity and the π-π bond effect of benzene ring. (2) the interaction of hydrogen bond between the hydroxyl group in TMA and iodide in the perovskite could 14

Journal of Power Sources 426 (2019) 11–15

L. Su, et al.

Fig. 7. (A) Normalized PCE values of PSCs without and with TMA additive in air for various durations with annealed time at 100 °C, (B) The photographs of MAPbI3 without and with TMA additive annealing for 10 h at 100 °C in the air.

Fig. 8. (A) Normalized PCE values of PSCs with aging time, (B) XRD patterns of perovskite films, and insets of Fig. 8 A are the real photographs of PSCs without encapsulation at room temperature in a 30% relative humidity for 20 days.

suppress the ion migration of perovskite. (3) the two-step solution method not only provided high-quality perovskite films, but also significantly improved the reproducibility of the experiments. This work provides an effective and simple route to prepare the high performance perovskite solar cells.

[11] Y. Li, L. Ji, R. Liu, C. Zhang, C. Mak, X. Zou, H. Shen, S. Leuf, H. Hsu, J. Mater. Chem. 6 (2018) 12842. [12] Y. Wu, F. Xie, H. Chen, X. Yang, L. Han, Adv. Mater. 29 (2017) 1701073. [13] H. Zhang, J. Shi, L. Zhu, Y. Luo, D. Li, H. Wu, Nano Energy 43 (2017) 383. [14] X. Hou, S. Huang, W. Yang, L. Pan, Z. Sun, X. Chen, ACS Appl. Mater. Interfaces 9 (2017) 35200. [15] H. Jiang, Z. Yan, H. Zhao, S. Yuan, Z. Yang, J. Li, ACS Appl. Energy Mater. 1 (2018) 900. [16] F. Zhang, J. Cong, Y. Li, J. Bergstrand, H. Liu, B. Cai, A. Hajian, Z. Yao, L. Wang, Y. Hao, X. Yang, J. Gardner, H. Ågren, J. Widengren, L. Kloo, L. Sun, Nano Energy 53 (2018) 405. [17] J. Heo, S. Im, J. Noh, T. Mandal, C. Lim, J. Chang, Y. Lee, H. Kim, A. Sarkar, Md Nazeeruddin, M. Grätzel, S. Seok, Nat. Photon. 7 (2013) 486. [18] X. Huang, Z. Hu, J. Xu, P. Wang, J. Zhang, Y. Zhu, Electrochim. Acta 231 (2017) 77. [19] Y. Xiao, G. Han, Y. Li, M. Li, Y. Chang, J. Wu, J. Mater. Chem. 2 (2014) 16531. [20] Y. Xiao, G. Han, Y. Li, M. Li, J. Wu, J. Mater. Chem. 2 (2014) 16856. [21] X. Hou, L. Pan, S. Huang, W. Ou-Yang, X. Chen, Electrochim. Acta 236 (2017) 351. [22] Z. Huang, X. Hu, C. Liu, L. Tan, Y. Chen, Adv. Funct. Mater. 27 (2017) 1703061. [23] S. Wu, Q. Liu, Y. Zheng, R. Li, T. Peng, J. Power Sources 359 (2017) 303. [24] Y. Xiao, L. Yang, G. Han, Y. Li, M. Li, H. Li, Org. Electron. 65 (2019) 201. [25] R. Pathipati, N. Shah, Sol. Energy 162 (2018) 8. [26] C. Wang, D. Zhao, Y. Yu, N. Shrestha, R. Grice, W. Liao, A. Cimaroli, J. Chen, R. Ellingson, X. Zhao, Y. Yan, Nano Energy 35 (2017) 223. [27] X. Li, M. Dar, C. Yi, J. Luo, M. Tschumi, S. Zakeeruddin, M. Nazeeruddin, H. Han, M. Grätzel, Nat. Chem. 7 (2015) 703. [28] N. Jeon, J. Noh, Y. Kim, W. Yang, S. Ryu, S. Seok, Nat. Mater. 13 (2014) 897. [29] Y. Xiao, G. Han, Y. Chang, Y. Zhang, Y. Li, M. Li, J. Power Sources 286 (2015) 118. [30] Q. Chen, H. Zhou, T. Song, S. Luo, Z. Hong, H. Duan, Nano Lett. 14 (2014) 4158. [31] L. Han, S. Cong, H. Yang, Y. Lou, H. Wang, J. Huang, Sol. RRL (2018) 1800054. [32] C. Zhang, Q. Luo, X. Deng, J. Zheng, W. Yang, X. Chen, S. Huang, Electrochim. Acta 258 (2017) 1262. [33] M. Li, B. Li, G. Cao, J. Tian, J. Mater. Chem. 5 (2017) 21313. [34] W. Nie, H. Tsai, R. Asadpour, J. Blancon, A. Neukirch, G. Gupta, J. Crochet, M. Chhowalla, S. Tretiak, M. Alam, Science 347 (2015) 522. [35] Y. Tu, J. Wu, X. He, P. Guo, T. Wu, H. Luo, Q. Liu, Q. Wu, J. Lin, M. Huang, Z. Lan, S. Li, J. Mater. Chem. 5 (2017) 21161. [36] D. Yuan, A. Gorka, M. Xu, Z. Wang, L. Liao, Phys. Chem. Chem. Phys. 17 (2015) 19745.

Acknowledgments This work was financially supported by the National Natural Science Foundation of China (61504076, 61804091, 21571118, 21671124, 21574076, and U1510121) and the Fund for Shanxi “1331 Project” Key Innovative Research Team. And we are also grateful for the test platform provided by Shanxi University of Scientific Instrument Center. References [1] M. Lee, J. Teuscher, T. Miyasaka, T. Murakami, H. Snaith, Science 338 (2012) 643. [2] D. Zhao, Y. Yu, C. Wang, W. Liao, N. Shrestha, C. Grice, A. Cimaroli, L. Guan, R. Ellingson, K. Zhu, X. Zhao, R. Xiong, Y. Yan, Nat. Energy 2 (2017) 17018. [3] Y. Zhao, H. Tan, H. Yuan, Z. Yang, J. Fan, J. Kim, O. Voznyy, X. Gong, L. Quan, C. Tan, J. Hofkens, D. Yu, Q. Zhao, E. Sargent, Nat. Commun. 9 (2018) 1607. [4] M. Saliba, T. Matsui, K. Domanski, J. Seo, A. Ummadisingu, S. Zakeeruddin, J. Correa-Baena, W. Tress, A. Abate, A. Hagfeldt, M. Grätzel, Science 354 (2016) 206. [5] NREL best research-cell efficiencies chart, https://www.nrel.gov/pv/assets/ images/efficiency-chart.png. [6] X. Zeng, T. Zhou, C. Leng, Z. Zang, M. Wang, W. Hu, X. Tang, S. Lu, L. Fang, M. Zhou, J. Mater. Chem. 5 (2017) 17499. [7] M. Wang, Z. Zang, B. Yang, X. Hu, K. Sun, L. Sun, Sol. Energy Mater. Sol. Cells 185 (2018) 117. [8] T. Zhou, M. Wang, Z. Zang, X. Tang, L. Fang, Sol. Energy Mater. Sol. Cells 191 (2019) 33. [9] Y. Xiao, G. Han, J. Wu, J. Lin, J. Power Sources 306 (2016) 171. [10] Q. Jiang, Z. Chu, P. Wang, X. Yang, H. Liu, Y. Wang, Z. Yin, J. Wu, X. Zhang, J. You, Adv. Mater. 29 (2017) 1703852.

15