A modified two-step sequential deposition method for preparing perovskite CH₃NH₃PbI₃ solar cells

Abstract

Two-step sequential deposition of CH₃NH₃PbI₃ (MAPI) films is widely used in the preparation of perovskite solar cells, but still suffers from some drawbacks such as residual PbI₂ and uncontrollable surface morphology of MAPI films. Here a modified two-step sequential deposition is reported to prepare MAPI films with pure phase and smooth surface. The process involves the deposition of a low-crystallinity PbI₂ film firstly and subsequent spraying of CH₃NH₃I solution. A well-converted and extremely surface-smooth MAPI film was successfully prepared via this two-step method. The weak crystallinity of raw PbI₂ film leads to the pure MAPI film. The following spraying of CH₃NH₃I solution in a spin-coating process results in MAPI films with much smoother surface, better crystallization and higher absorption compared with the traditional method of immersing in CH₃NH₃I solution. Consequently, the corresponding mesostructured perovskite solar cells based on these MAPI films achieve a high efficiency of 14.3% and good reproducibility, which demonstrates the advantages and promising applications of this modified two-step sequential deposition method.

---

Introduction

Due to their advantages of direct band gap,¹ large absorption coefficient,² ambipolar diffusion,^1,3 weakly bound excitons⁴ and long carrier diffusion length,^5–8 organic–inorganic lead halide perovskites have become a promising class of light absorbers for low-cost, high-efficiency solar cells over the past few years.^9–11 A certified power conversion efficiency (PCE) of 20.1% has been achieved by Seok's group,¹² and PCE higher than 15% was also achieved by a large-area device recently,¹³ implying their rapid development towards large-scale manufacture and practical application. CH₃NH₃PbI₃ (MAPI) is the most common member of the organic–inorganic lead halide perovskites in the field of photovoltaic application. Despite the inherent virtue of MAPI material and great progress of perovskite solar cells (PSCs), it is urgent to find a facile and controllable preparation technology to deposit MAPI films with good quality and reproducibility. Traditional deposition methods of MAPI films include the sequential deposition method¹⁴ and one-step deposition method,¹⁵ both of which have their own advantages, especially the former method, which was found to exhibit better photovoltaic performance than the one-step deposition method due to better morphology and interfaces.¹⁶ But, there are some problems for the sequential deposition method, such as incomplete conversion of PbI₂ to MAPI and the uncontrolled surface morphology of MAPI films, which seriously limit its large-scale application for high-efficiency PSCs with good reproducibility.

Lots of works have been developed to overcome the problems mentioned above.^16–20 For example, Han et al. prepared amorphous PbI₂ film firstly by dissolving PbI₂ in DMSO solvent, and then completely transformed MAPI absorber was obtained by soaking the amorphous PbI₂ film in CH₃NH₃I (MAI) solution.¹⁷ But the surface topography of the absorber obtained in this way was still coarse. In Seigo Ito's work, a perovskite absorption layer with high smoothness of surface morphology was obtained by the drop-casting method, leading to good conditions for application of new hole transport materials (HTM).¹⁸ Unfortunately, residual PbI₂ can still be observed with this method. Therefore, it is necessary to develop a better deposition method to solve the two big problems at the same time.

Consequently, we have developed a modified two-step sequential deposition method and successfully fabricated compact, extremely smooth and pinhole-free MAPI films. At first, PbI₂ film with low crystallinity is deposited by the solvent–solvent extraction (SSE) method,²¹ which is helpful for complete transformation of PbI₂ to MAPI in subsequent reaction with MAI. In order to achieve flat and smooth MAPI absorption layer, a MAI solution spraying technique, namely spraying a MAI isopropanol solution during spin-coating process of PbI₂ films, is used in the second step, which shows better effects than the traditional immersing method. The MAPI films deposited by this method possess pure MAPI phase and flat surface with high smoothness, which overcome the difficulties in the traditional sequential deposition method. The MAPI film-based mesostructured perovskite solar cells achieved a PCE of 14.3% and good reproducibility.

Experimental

Transparent conducting substrate and mesoporous TiO₂ thin film

Fluorine-doped SnO₂-coated transparent conducting glass substrate (FTO) was washed by ultrasonic cleaning with detergent firstly, then with deionized water, acetone and ethanol, and finally purged with dry nitrogen air. A 50 nm-thick TiO₂ compact layer was then deposited on the substrates by sputtering a TiO₂ target at 0.6 Pa with 120 W for 30 min. The mesoporous TiO₂ film was prepared by spin-coating a 20 nm-sized TiO₂ nanoparticulate paste (diluted in ethanol with a ratio of 1 : 5 by weight, OPV-Tech, DHS-TPP3) 4 times at 2500 rpm for 30 s, and dried at 125 °C for 10 min, then heated at 500 °C for 30 min.

Deposition of perovskite thin films

CH₃NH₃I (MAI) was synthesized by reacting 19.5 mL methylamine (40 wt% aqueous solution) and 32.3 mL hydroiodic acid (55–58 wt% aqueous solution) in an ice bath for 2 h with stirring. The product was obtained by rotary evaporation at 45 °C and cleaning with diethyl ether followed by drying in a vacuum oven at 60 °C overnight.

First step: deposition of PbI₂ films. There are two kinds of PbI₂ films in this work: solvent–solvent extracted PbI₂ films (SSE-PbI₂) and spin-coated PbI₂ films (SC-PbI₂). For the SSE-PbI₂ films, the precursor solution was prepared by dissolving 1 mmol of PbI₂ (97%, Sigma-Aldrich) into 1 mL mixed solvents (2 : 8 v/v) of N,N-dimethylformamide (DMF) and 1-methyl-2-pyrrolidinone (NMP) at 60 °C for 6 h. The precursor was spin-coated on the mesoporous TiO₂ films at 4000 rpm for 15 seconds, followed by quickly dipping in a ∼50 mL anhydrous diethyl ether bath for 2 min. Then the substrate was heated on a hotplate at 100 °C for 10 minutes, resulting in low-crystallinity PbI₂ film. For the SC-PbI₂ films, the precursor solution was prepared by dissolving 1 mmol of PbI₂ (97%, Sigma-Aldrich) into 1 mL DMF at 6 °C for 6 h. Then it was spin-coated on the mesoporous TiO₂ films at 4000 rpm for 15 seconds and finally heated on a hotplate at 100 °C for 10 min. The entire PbI₂ film fabrication process was performed in atmosphere with ∼30% humidity.

Second step: fabrication of the CH₃NH₃PbI₃ (MAPI) films. The high-quality MAPI perovskite films without PbI₂ residue were prepared with the low-crystallinity SSE-PbI₂ by a spin-spray method, denoted as SSE-SS-MAPI films. Briefly, a 2-propanol solution of MAI (10 mg mL⁻¹) was sprayed on the SSE-PbI₂ films using an airbrush pen with 2 mm nozzle. During spraying, the substrates were spun at 2000 rpm at room temperature, films were sprayed about 10 s at a nozzle height of 10 cm and an air carrier pressure of 15 psi, followed by annealing on a hotplate at 100 °C for 10 min in air. The low-crystallinity SSE-PbI₂ films and high-crystallinity SC-PbI₂ films were also treated with the traditional two-step method²² for the purpose of comparison. The detailed process was as follows. The PbI₂ films (low- or high-crystallinity) were infiltrated in 2-propanol and blow-dried first, followed by immersing in 10 mg mL⁻¹ MAI 2-propanol solution for appropriate time, respectively (25 minutes for SC-PbI₂ films and 5 minutes for SSE-PbI₂ films). After being rinsed with 2-propanol, the perovskite films were annealed at 100 °C for 10 minutes. These two samples were named as SC-immerse-MAPI films and SSE-immerse-MAPI films, respectively.

Device fabrication

The hole transport material (HTM) solution was prepared as follow: 52.8 mg 2,29,7,79-tetrakis-(N,N-di-p-methoxyphenylamine)-9,99-spirobifluorene (spiro-MeOTAD) was dissolved in 640 μL chlorobenzene and mixed with 10 μL of 500 mg mL⁻¹ bis(trifluoromethane) sulfonimide lithium salt (LiTFSI) solution in acetonitrile and 14.4 μL 4-tert-butylpyridine (TBP). The HTM solution was spin-coated on the perovskite-covered TiO₂ electrodes at 2500 rpm for 30 seconds. For the counter electrode, a 100 nm-thick Au layer was deposited on the top of the HTM layer by thermal evaporation and the active area was 7 mm².

Characterization

The crystallization and phase identification of the thin films were performed by X-ray diffraction (XRD, Bruker D8 Focus) with a monochromatized source of Cu Kα1 radiation (λ = 0.15405 nm) at 1.6 kW (40 kV, 40 mA). UV-visible absorbance spectra were recorded using a Hitachi U-3010 spectrophotometer with a scanning velocity of 300 nm min⁻¹. Top-view field emission scanning electron microscopy (FESEM) images were taken with a ZEISS SUPRA 55 microscope. Photocurrent density–voltage characteristics were measured using a Keithley model 2440 source meter under AM 1.5 illumination. A 1000 W Oriel solar simulator was used as a light source and the power of the light was calibrated to one sun light intensity by using a NREL-calibrated Si cell (Oriel 91150). The external quantum efficiency (EQE) was measured by a Newport QE system equipped with a 300 mW xenon light and a lock-in amplifier.

Results and discussion

The crystallizations of PbI₂ films deposited by solvent–solvent extraction method (SSE-PbI₂) and spin-coating method (SC-PbI₂) were characterized by XRD patterns, as shown in Fig. 1a. According to the peak intensity and the half-peak width of the characteristic peak of PbI₂ at 12.7°, it is obvious that the crystallization of PbI₂ in SSE-PbI₂ film is much weaker than that in SC-PbI₂ film. For the obtained SSE-PbI₂ film with low crystallinity it would be easy for the transition of PbI₂ to MAPI in the second reaction step. Fig. 1b shows the UV-visible absorption spectra of SSE-PbI₂ film and SC-PbI₂ film. Due to the lower crystallinity of SSE-PbI₂ film, it exhibits lower light absorption in the range below 500 nm. Digital photos of these two films are also shown in the inset in Fig. 1b. It is easily to distinguish the films from each other by the naked eye. The surface morphologies of these two films were observed by FESEM. As shown in Fig. 1c and d, not only is the surface roughness of SSE-PbI₂ film higher than SC-PbI₂ film, but also some pinholes could be found in SSE-PbI₂ film. It was demonstrated that it is the features of low crystallinity and loose structure that facilitated the rapid and complete conversion of PbI₂ to MAPI.

Fig. 1 XRD patterns (a) and UV-visible absorption spectra (b) of SSE-PbI₂ (solvent–solvent extracted PbI₂) films and SC-PbI₂ (spin-coated PbI₂) films. Insert picture in (b) digital photos of the two films, left: SSE-PbI₂, right: SC-PbI₂. Top-view FESEM images of (c) SSE-PbI₂ films and (d) SC-PbI₂ films.

Compared with the spin-coating method, ether extraction can be used to obtain PbI₂ films with low crystallinity because extraction film forming is a transient film-forming process. The state of a solid PbI₂ film obtained from extraction is closer to its state in organic solution than from spin-coating, namely a long-range disordered state. However, spin-coating PbI₂ film forming involves a gradual process of solution evaporation and crystal growth. The difference in the transient and gradual film forming leads to a great difference in the crystallinity of films, as well as their surface morphology. As a result, the crystallinity of MAPI films obtained by the two processes will also be much different.

The XRD patterns of the MAPI films transformed from SSE-PbI₂ film and SC-PbI₂ film with different process methods are shown in Fig. 2a. Apparently, the MAPI films transformed from SSE-PbI₂ have no residual PbI₂, no matter what transformation method is adopted. In contrast to that, the XRD pattern of MAPI films transformed from the SC-PbI₂ films has an apparent peak at 12.7°, indicating the existence of residual PbI₂. In addition, after comparing their crystallinity from the analysis of their XRD patterns, we can draw the conclusion that: the MAPI films prepared from the SSE-PbI₂ films via spin-spray coating method (SSE-SS-MAPI films) have almost the same crystallinity as the MAPI films from the SC-PbI₂ films via traditional two-step immersion method (SC-immerse-MAPI films). The MAPI films prepared from the SSE-PbI₂ films through traditional two-step immersion method (SSE-immerse-MAPI films) show the lowest crystallinity among the three samples because of the low crystallinity of the SSE-PbI₂ films. However, it seems abnormal that SSE-SS-MAPI films have the same high crystallinity as the SC-immerse-MAPI films.

Fig. 2 XRD patterns (a) and UV-visible absorption spectra (b) of three kinds of MAPI films. Namely, MAPI films transformed from: SSE-PbI₂ by spin-spray method (SSE-SS-MAPI), SC-PbI₂ by immersing method (SC-immerse-MAPI), SSE-PbI₂ by immersing method (SSE-immerse-MAPI).

It can be concluded from the above results that the SSE-PbI₂ films are porous and of low crystallinity. During high-speed rotation in the spinning-spray process, if isopropanol solution from MAI is sprayed, the MAI molecules carried by the solution may quickly and comfortably enter the lattice of PbI₂ because of the dynamic interaction between the solution and the solid film, thus producing a homogeneously mixed MAPI precursor. As the high-speed rotation continues, the isopropanol solution volatilizes, and the perovskite crystals recrystallize quickly and become more ordered. Meanwhile, thanks to the effect of centrifugal force sustained by the film in the process of crystallization, the resultant film possesses a highly smooth and dense surface, which is unachievable by the traditional two-step process.

Besides, the light absorption of these MAPI films was further measured. As shown in Fig. 2b, the absorption abilities of these three MAPI films increase with an increase of the film crystallinity. This result reveals that good crystallinity of the MAPI films is important for the light absorption of the perovskite solar cell.

Fig. 3a–c show the micro surface morphologies of these three different MAPI films, i.e., SSE-SS-MAPI films, SC-immerse-MAPI films and SSE-immerse-MAPI films, respectively. From those images, we can visually get the information that the SSE-SS-MAPI films have better quality than the other two MAPI films in the aspects of smoothness and compactness, which are the most important quality parameters of perovskite films. Besides, SC-SS-MAPI film also exhibits a relatively smooth surface morphology (ESI Fig. S2†). However, restricted by its low conversion ratio of PbI₂ to MAPI, it showed a poor performance in the complete device. Fig. 3d shows the cross-sectional structure of the device made from SSE-SS-MAPI film. It clearly exhibits a layer structure from bottom to top with FTO, TiO₂ blocking layer, TiO₂ mesoporous layer infiltrated with MAPI, MAPI top layer, HTM and Au back electrode.

Fig. 3 Top-view FESEM images of MAPI films transformed from (a) SSE-PbI₂ by spin-spray method (SSE-SS-MAPI), (b) SC-PbI₂ by immersing method (SC-immerse-MAPI), (c) SSE-PbI₂ by immersing method (SSE-immerse-MAPI), and (d) cross-sectional FESEM image of complete solar cell device fabricated with the SSE-SS-MAPI film.

Furthermore, the AFM characterizations confirm that SSE-SS-MAPI films have the smallest surface roughness root mean square (RMS), as shown in Fig. S2 (see ESI†).

In current–voltage (J–V) curve measurements, the curve shape can be affected by scan rate significantly.²² In order to obtain a conventional J–V curve and avoid long time radiation, a scan rate of 57.5 mV s⁻¹ was applied in the J–V curve measurement.

To evaluate the practical effect of these three kinds of MAPI films, we have fabricated complete devices based on these MAPI films and carried out a photocurrent density–voltage characterization. Fig. 4a shows the typical current–voltage (J–V) curves of the three devices made from three different MAPI films, i.e., SSE-SS-MAPI films, SC-immerse-MAPI films and SSE-immerse-MAPI films. We chose the best performing three devices of each group in the same batch and did this characterization with the same conditions. Fig. 4b shows the incident-photon-to-current conversion efficiency (IPCE) spectra of these three devices, well consistent with the J–V curve (ESI Fig. S3†). And the device performance parameters are summarized in Table 1.

Fig. 4 (a) Current–voltage (J–V) curves of the three different MAPI film-based solar cells under a standard AM 1.5 solar illumination at an intensity of 100 mW cm⁻². (b) IPCE spectra of the three different MAPI film-based solar cells without any applied bias.

Table 1 Summary of three different kinds of MAPI film-based solar cell performance parameters

Sample	Voc (V)	Jsc (mA cm^−2)	FF (%)	η (%)
SSE-SS-MAPI	1.032	19.85	69.9	14.3
SC-immerse-MAPI	0.972	18.52	63.3	11.4
SSE-immerse-MAPI	0.918	11.36	48.6	5.07

---

As can be seen from Fig. 4, the photoelectric conversion efficiencies of the three devices, based on SSE-SS-MAPI films, SC-immerse-MAPI films and SSE-immerse-MAPI films respectively, show a decline in sequence. The device prepared by directly soaking the low-crystallinity PbI₂ films is inferior to the devices prepared by the traditional two-step method in terms of the photoelectric conversion efficiency. This is mainly because the crystallinity of SSE-immerse-MAPI films is too low and many pinholes exist, which have great negative effects on light absorption (Fig. 3c) and carrier transport of the device. Therefore, despite no PbI₂ residue, its device efficiency is the lowest one. In comparison, SC-immerse-MAPI films are superior in both density and crystallinity. Although PbI₂ residue exists, the energy conversion efficiency can still reach 11.4%. SSE-SS-MAPI films possess both of the advantages, namely no PbI₂ residue and high film quality, which results in better light absorption, carrier transportation and separation,²³ and consequently a higher energy conversion efficiency. Additionally, the open-circuit voltage V_oc, short-circuit current J_sc and filling factor FF are all improved. We can also evaluate their performance from the IPCE spectra, as shown in Fig. 4b. It is very clear that the device with high-quality SSE-SS-MAPI films shows better IPCE property than the other two devices for almost the whole spectrum, especially in the range between 600 and 750 nm. Although the SC-immerse-MAPI solar cell delivers good performance in the range between 300 and 500 nm, its IPCE decreases rapidly after 550 nm, which should result from the film's poor crystallinity.²⁴

In addition, we obtained statistics for the four performance parameters of the three solar cells, as displayed in Fig. S4.† The low error bars in all four parameters of the SSE-SS-MAPI solar cell indicate the good repeatability of our device fabrication procedures. Fig. S5† shows the J–V curves of the best performing solar cell based on SSE-SS-MAPI films with different scan direction at a scan rate of 57.5 mV s⁻¹. It shows a hysteresis with a difference of 16–17% in the efficiency between the forward scan and reverse scan.

Conclusions

In summary, a modified two-step sequential deposition method is developed by us to fabricate perovskite CH₃NH₃PbI₃ (MAPI) solar cells with a high efficiency exceeding 14%. A low-crystallinity PbI₂ film is firstly prepared through solvent–solvent extraction process and subsequently dense, smooth, and completely converted MAPI films are fabricated on mesoporous TiO₂ via the spin-spray method. Compared with the previously reported two-step fabrication process, our process solves for the first time two problems at the same time, i.e., the high roughness of the perovskite light absorption layer and incomplete PbI₂ conversion to MAPI, and succeeds in producing quality MAPI films with excellent crystallinity and light absorption performance. Furthermore, the spin-spray technique, considering its simple and controllable features and the economy and suitability of its raw material for large-area operation, offers a highly practical applicability prospect. This perovskite film fabrication process when used in combination with good element packaging and waterproof technology is expected to allow practical application of large-area high-performance perovskite solar cells.

Acknowledgements

This work was financially supported by National Science Foundation of China (grants 61376056, 51402335 and 51402341), Science and Technology Commission of Shanghai (grants 13JC1405700 and 14520722000), the Key Research Program of Chinese Academy of Sciences (grant no. KGZD-EW-T06) and the Science and Technology Projects of State Grid (SGRIDGKJ [2015]452).

References

1. C. C. Stoumpos, C. D. Malliakas and M. G. Kanatzidis, Inorg. Chem., 2013, 52, 9019–9038.
2. S. De Wolf, J. Holovsky, S. J. Moon, P. Loper, B. Niesen, M. Ledinsky, F. J. Haug, J. H. Yum and C. Ballif, J. Phys. Chem. Lett., 2014, 5, 1035–1039.
3. C. S. Ponseca, T. J. Savenije, M. Abdellah, K. B. Zheng, A. Yartsev, T. Pascher, T. Harlang, P. Chabera, T. Pullerits, A. Stepanov, J. P. Wolf and V. Sundstrom, J. Am. Chem. Soc., 2014, 136, 5189–5192.
4. T. J. Savenije, C. S. Ponseca, L. Kunneman, M. Abdellah, K. B. Zheng, Y. X. Tian, Q. S. Zhu, S. E. Canton, I. G. Scheblykin, T. Pullerits, A. Yartsev and V. Sundstrom, J. Phys. Chem. Lett., 2014, 5, 2189–2194.
5. A. Kojima, M. Ikegami, K. Teshima and T. Miyasaka, Chem. Lett., 2012, 41, 397–399.
6. C. R. Kagan, D. B. Mitzi and C. D. Dimitrakopoulos, Science, 1999, 286, 945–947.
7. 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, 2013, 342, 341–344.
8. G. C. Xing, N. Mathews, S. Y. Sun, S. S. Lim, Y. M. Lam, M. Gratzel, S. Mhaisalkar and T. C. Sum, Science, 2013, 342, 344–347.
9. M. Gratzel, Nat. Mater., 2014, 13, 838–842.
10. P. P. Boix, K. Nonomura, N. Mathews and S. G. Mhaisalkar, Mater. Today, 2014, 17, 16–23.
11. T. C. Sum and N. Mathews, Energy Environ. Sci., 2014, 7, 2518–2534.
12. W. S. Yang, J. H. Noh, N. J. Jeon, Y. C. Kim, S. Ryu, J. Seo and S. I. Seok, Science, 2015, 348, 1234–1237.
13. W. Chen, Y. Z. Wu, Y. F. Yue, J. Liu, W. J. Zhang, X. D. Yang, H. Chen, E. B. Bi, I. Ashraful, M. Gratzel and L. Y. Han, Science, 2015, 350, 944–948.
14. J. Burschka, N. Pellet, S. J. Moon, R. Humphry-Baker, P. Gao, M. K. Nazeeruddin and M. Gratzel, Nature, 2013, 499, 316–319.
15. H.-S. Kim, C.-R. Lee, J.-H. Im, K.-B. Lee, T. Moehl, A. Marchioro, S.-J. Moon, R. Humphry-Baker, J.-H. Yum, J. E. Moser, M. Grätzel and N.-G. Park, Sci. Rep., 2012, 2, 591.
16. J. H. Im, H. S. Kim and N. G. Park, APL Mater., 2014, 2, 081510.
17. Y. Wu, A. Islam, X. Yang, C. Qin, J. Liu, K. Zhang, W. Peng and L. Han, Energy Environ. Sci., 2014, 7, 2934.
18. S. Ito, S. Tanaka and H. Nishino, J. Phys. Chem. Lett., 2015, 6, 881–886.
19. Z. G. Xiao, C. Bi, Y. C. Shao, Q. F. Dong, Q. Wang, Y. B. Yuan, C. G. Wang, Y. L. Gao and J. S. Huang, Energy Environ. Sci., 2014, 7, 2619–2623.
20. B. Yang, O. Dyck, J. Poplawsky, J. Keum, S. Das, A. Puretzky, T. Aytug, P. C. Joshi, C. M. Rouleau, G. Duscher, D. B. Geohegan and K. Xiao, Angew. Chem., 2015, 54, 14862–14865.
21. Y. Y. Zhou, M. J. Yang, W. W. Wu, A. L. Vasiliev, K. Zhu and N. P. Padture, J. Mater. Chem. A, 2015, 3, 8178–8184.
22. Y. Xie, F. Shao, Y. Wang, T. Xu, D. Wang and F. Huang, ACS Appl. Mater. Interfaces, 2015, 7, 12937–12942.
23. G. Zhu, T. Lin, X. Lü, W. Zhao, C. Yang, Z. Wang, H. Yin, Z. Liu, F. Huang and J. Lin, J. Mater. Chem. A, 2013, 1, 9650.
24. F. Hao, C. C. Stoumpos, Z. Liu, R. P. H. Chang and M. G. Kanatzidis, J. Am. Chem. Soc., 2014, 136, 16411–16419.

---

Footnote

† Electronic supplementary information (ESI) available: AFM images of MAPI films, parameters of perovskite solar cells. See DOI: 10.1039/c6ra05718g

---