Hybrid organic–inorganic perovskite materials have attracted tremendous attention due to their excellent properties, such as large charge carrier mobility,^[1] long diffusion length,^[2] high absorption coefficient,^[3,4] tunable band gap,^[5–8] and unique tolerance to structural defects.^[9,10] The outstanding optoelectronic characteristics combined with multiple manufacturing methods^[11–19] have presented such materials with extensive applications in light emitting devices (LEDs),^[20,21] thin film transistors (TFTs),^[22] photodetectors,^[23,24] solar cells,^[25–28] and others. During preparation, a distinct advantage of vacuum deposition over solution processing is the ability to prepare layered uniform, compact, and multi-stack thin films,^[15] which enables multi-layered structures. However, the efficiencies are lower and successful cases of efficiencies over 15% are few for vacuum deposited solar cells^[15,29] compared to solar cells prepared with solution processing,^[30] which mainly results from the small density and high vapor pressure of organic methyl ammonium iodide (MAI) diffusing inside the vacuum chamber, making it very difficult to control the ratio of organic MAI to inorganic lead salt.^[31–34] To improve the controllability during vacuum deposition, inorganic lead halides and organic MAI have been prepared in separated processes to synthesize perovskite films, as shown in the scheme in Fig. 1(a). Vapor-assisted solution processing,^[16,35,36] sequential vacuum deposition,^[37,38] evaporation–solution processing,^[39–45] hybrid chemical vapor deposition,^[46] layer-by-layer growth,^[47] and alternating precursor layer deposition^[48] have been developed and can enhance the experimental repeatability, as these preparation routes can provide strict control over the stoichiometry of organic MAI vapor. However, as common methods of control, monitoring the organic vapor rate with quartz monitor crystals^[36,48] or evaporating organic halide at a fixed temperature^[32] are invalid for controlling the penetration ability of the randomly diffused organic halide vapor. In addition, the relationship between the thickness of the inorganic lead halide layer and the concentration of MAI/isopropyl alcohol (IPA) for optimized device performance is unclear, which is vital for evaporation–solution methods widely used in semitransparent and tandem devices.^{[40–42,49,50]}

Fig. 1.

Figure Option ViewDownloadNew Window

Fig. 1. (color online) Scheme illustration of the intercalation reaction to form the perovskite film and schematic structure of the PSC. (a) First, the inorganic lead halide was sublimed, and then, the as-deposited inorganic seeding layer was reacted with organic molecules through a penetration process. (b) Schematic structure of the PSC with the perovskite film prepared by sequential vacuum deposition and evaporation–solution methods.

In fact, an intercalation process is required to manufacture a perovskite film with a step-by-step deposition process.^[51–53] During the organic molecule penetration process into the inorganic skeleton, the thickness of the inorganic seeding layer and the penetration ability and chemical quantity of the organic molecules are core factors. Heating the substrate has been shown as a successful method to enhance the penetration ability of organic MAI molecules.^[37] In our work, first, we prepared a high-quality perovskite film by controlling the penetration ability and chemical quantity of the MAI vapor through monitoring the evaporation pressure and time. The efficiency of the perovskite solar cell (PSC) reached over 15% with the all-vacuum prepared perovskite film. Moreover, the thickness of the inorganic lead halide and the concentration of the organic molecule solution show a linear relationship with the optimized perovskite film, and an empirical formula was proposed for the evaporation–solution method.

The perovskite film was manufactured with sequential vacuum deposition,^[37,54] and the detailed process has been described in the experimental section. A 100 nm inorganic PbCl₂ seeding layer was first sublimed onto the substrate with a controlled rate. In addition to the concentration and temperature, the pressure of the organic MAI vapor was manipulated to realize an effective diffusion reaction. The grain size at a low organic vapor pressure (Fig. 2(a), 5 × 10⁻⁵ Torr) was smaller than that at a high vapor pressure (Figs. 2(b) and 2(c), ). The three XRD peaks of CH₃NH₃PbI₃ (MAPbI₃) corresponding to the (110), (220), and (330)^[15,55] lattice planes were stronger and shaper at an organic vapor pressure of , as shown in Fig. 2(d), indicating the formation of larger perovskite grains. The increased crystalline grain size of the perovskite at high pressure was attributed to the improved reaction rate of PbCl₂ and MAI,^[36] referred to as Ostwald ripening.^[56,57] At high MAI vapor pressure, more and higher-reactive MAI molecules will react with PbCl₂, which favors secondary nucleation at the primary sites. The nucleation sites of the large crystallites grow faster because of the larger contact area. As a result, the larger crystalline grains grow even larger and diminish the small perovskite nucleation sites. Therefore, a suitable high pressure of organic MAI is beneficial for the formation of high-quality perovskite films.

Fig. 2.

Figure Option ViewDownloadNew Window

Fig. 2. (color online) Surface morphology, phase, and electrical properties of the perovskite film prepared by sequential vacuum deposition before annealing. SEM top-view images of the perovskite films prepared with various organic MAI sublimation pressures and duration time: (a) , 10 h; (b) , 5 h; (c) , 10 h. (d) XRD patterns of the perovskite films prepared with various sublimation conditions of organic MAI. (e) TRPL spectroscopy of the FTO/TiO2/perovskite film prepared with a MAI vapor pressure of or . The inorganic seeding layer is 100 nm PbCl2. The sublimation duration time of organic MAI reflects its chemical quality and is sufficient when the time is 10 h but insufficient at 5 h.

Transient TRPL decay measurements were conducted to investigate the electrical properties of the perovskite film (Fig. 2(e)). A single-exponential function was used to fit the PL decay time. The decay time of the sample with an organic vapor pressure of is ns, while the decay time of is ns. The rapid decay for the sample prepared at an organic vapor pressure of 10 Torr results from the quick recombination in the perovskite bulk caused by more grain boundaries.^[58]

The chemical quantity of MAI is also important in the formation of MAPbI₃. As shown in Fig. 2(d), the peaks at and 15.56° corresponding to PbI₂ (001) and CH₃NH₃PbCl₃ (MAPbCl₃) (100)^[15,59] appear at an MAI evaporation time of 5 h, indicating the formation of non-pure MAPbI₃. However, pure MAPbI₃ can be obtained when prolonging the evaporation time of MAI. Based on previous studies^[59–61] and the XRD patterns, the reaction steps are listed as follows: 1 2 3 4 First, reactions (1) and (2) occur. Then, MAI reacts with PbI₂ and MAPbCl₃ through reactions (3) and (4), respectively. With an insufficient amount of MAI, PbI₂ and MAPbCl₃ cannot be consumed, leaving in MAPbI₃. With a sufficient amount of MAI, pure MAPbI₃ is formed.

The efficiency of a PSC (structural schematic shown in Fig. 1(b)) is optimal only when the perovskite film is formed with a high organic MAI vapor pressure and sufficient MAI content. The J–V characteristics and corresponding EQE curves are shown in Fig. 3, and the device performance parameters are summarized in Table 1. The current density ( from the J–V curves was larger than integral from EQE, resulting from lateral collection.^[62] The PSC with the absorber formed at an MAI vapor pressure of Torr demonstrated higher , open-circuit voltage ( ), and fill factor ( ), which are attributed to the larger perovskite grain size, reducing recombination in the perovskite bulk. The PSC fabricated at an MAI vapor pressure of Torr but with an MAI evaporation time of 5 h displayed lower , FF and higher than the cell with an MAI evaporation time of 10 h, which may result from PbI₂ and MAPbCl₃ reducing the absorption^[63] but enhancing the band gap.^[64]

Fig. 3.

	Figure Option ViewDownloadNew Window
	Fig. 3. (color online) Performance of PSCs prepared using sequential vacuum deposition. (a) J–V characteristics with an effective area of 0.1 cm2 (FB-SC) under 1 sun illumination and (b) the corresponding EQE curves for PSCs with various thermal evaporation deposition conditions.

Table 1.

Table 1.

Table 1. PSC parameters with different organic vapor pressures and duration time. . Conditions /mA cm−2 /mV Eff./% / 5.0×10−5 Torr, 10 h 10.5 850 0.310 2.790 606.9 1.5×10−4 Torr, 10 h 22.6 1050 0.686 16.28 4.6 1.5×10−4 Torr, 5 h 17.6 1090 0.711 13.63 10.0

Table 1. PSC parameters with different organic vapor pressures and duration time. .

For the intercalation reaction of the forming perovskite film, the thickness of the inorganic seeding layer and the concentration of the organic molecules are vital.^[18] However, the quantitative relationship between the thickness of the inorganic seeding layer and the concentration of the organic molecules for optimal device performance is unclear. We investigated this relationship via an evaporation–solution method^[43] because thermally evaporated PbI₂ is non-destructive to substrates and can be prepared with controlled morphology and thickness, and the concentration of the MAI/IPA solution can be easily changed. PbI₂ layers with thicknesses of 85 nm, 150 nm, 200 nm, 270 nm were prepared, and MAI/IPA solutions with four concentrations were spin coated onto each PbI₂ thickness. The performance parameters of PSCs with different PbI₂ thicknesses and MAI/IPA concentrations are shown in Fig. 4. When PbI₂ possessing each thickness reacts with a particular MAI/IPA solution concentration, the PSC can acquire an optimal performance. Maximum photoelectric conversion efficiencies (PCEs) of 11.88%, 13.99%, 14.91%, and 13.33% were obtained with 85 nm, 150 nm, 200 nm, 270 nm PbI₂ reacting with 16 mg/mL, 27.5 mg/mL, 33 mg/mL, 46.5 mg/mL MAI/IPA, respectively.

Fig. 4.

	Figure Option ViewDownloadNew Window
	Fig. 4. (color online) Performance parameters of the PSCs prepared using the evaporation–solution method. Performances of the PSCs with PbI2 thicknesses of (a) 85 nm, (b) 150 nm, (c) 200 nm, and (d) 270 nm reacting with various concentrations of the MAI/IPA solution.

The optimum matching of the MAI/IPA concentration with the PbI₂ thickness is depicted in Fig. 5. Obviously, the optimum MAI/IPA concentration to acquire the maximum PSC efficiency linearly increases with the PbI₂ thickness, and the linear formula fit to the MAI/IPA concentration versus PbI₂ thickness data is 5 where y is the MAI/IPA concentration, and x is the PbI₂ thickness. To examine the feasibility of the above formula, we prepared 175 nm PbI₂ and reacted the deposited film with 25 mg/mL, 28 mg/mL, 31 mg/mL, and 35 mg/mL MAI/IPA. According to the above formula, the optimum match is 175 nm PbI₂ reacting with 30.5 mg/mL MAI/IPA. Within the margin of error, the optimum efficiency of 14.69% was achieved with the PSC with the film formed using 31 mg/mL MAI/IPA, and the performance curves showed negligible hysteresis, as shown in Fig. 6.

Fig. 5.

	Figure Option ViewDownloadNew Window
	Fig. 5. (color online) Relationship between the PbI2 thickness and MAI/IPA concentration to achieve the highest PSC efficiency. The thickness of the PbI2 film has a linear correlation with the MAI/IPA concentration to achieve the optimum perovskite film.

Fig. 6.

	Figure Option ViewDownloadNew Window
	Fig. 6. (color online) Performance of the PSC with 175 nm PbI2 and 30.5 mg/mL MAI/IPA. (a) J–V curve with an effective area of 0.1 cm2 and (b) EQE curve of the cell under 1 sun illumination.

As seen from Fig. 5, the PSC efficiency increases with an increasing PbI₂ thickness up to 200 nm and decreases with 270 nm PbI₂. Figures 7(a)–7(d) show the cross-sectional SEM images of the perovskite films with 85 nm, 150 nm, 200 nm, and 270 nm PbI₂, respectively. These perovskite films are approximately twice as thick as the PbI₂ films, agreeing with the previous reports.^[65,66] The thickness of the perovskite film increases with an increasing PbI₂ thickness, leading to obviously enhanced ultraviolet-visible absorption between 550 nm to 800 nm, as shown in Fig. 7(e). When the thickness of the perovskite film is less than 400 nm, as shown in Figs. 7(a)–7(c), the, the crystalline grains are perpendicular to the substrates and larger than the film thickness, which are beneficial to transporting photogenerated carriers. For a thicker perovskite film (Fig. 7(d)), the film has a reduced quality with more grain boundaries, resulting in more loss of electricity. Balancing the optical absorption and carrier transport, the optimum thickness of the perovskite film is 400 nm manufactured with 200 nm PbI₂ and 33 mg/mL MAI/IPA. The J–V curves and performance parameters of the perovskite films of different thicknesses are shown in Fig. 7(f) and Table 2.

Fig. 7.

Figure Option ViewDownloadNew Window

Fig. 7. (color online) SEM images and absorption properties of the perovskite films and performances of the PSCs. SEM cross-sectional images of perovskite films with thicknesses of (a) 160 nm, (b) 300 nm, (c) 400 nm, and (d) 540 nm manufactured by reacting 85 nm, 150 nm, 200 nm, and 270 nm PbI2 with 16 mg/mL, 27.5 mg/mL, 33 mg/mL, and 46.5 mg/mL MAI/IPA, respectively. (e) Absorption spectra of the perovskite films, and (f) J–V curves (FB-SC) of the PSCs under 1 sun illumination with different PbI2 thicknesses matched with the optimal concentrations of MAI/IPA.

Table 2.

Table 2.

Table 2. Performance parameters of the optimal PSCs using different PbI2 thicknesses reacting with the optimal concentrations of the MAI/IPA solution. . PbI2 thickness/nm 85 150 200 270 MAI concentration/mg mL−1 16 27.5 33 46.5 Perovskite thickness/nm 160 300 400 540 /mA cm−2 17.4 19.3 20.8 19.7 /mV 1060 1090 1060 1020 /% 0.642 0.664 0.677 0.663 Maximum Eff./% 11.88 13.99 14.91 13.33

Table 2. Performance parameters of the optimal PSCs using different PbI2 thicknesses reacting with the optimal concentrations of the MAI/IPA solution. .

We obtained large-grain, pure CH₃NH₃PbI₃ using a sequential vacuum deposition, and an efficiency of over 15% was achieved with the PSC fabricated using a sufficient and high pressure of organic MAI vapor. In addition, we investigated the relationship between the inorganic seeding layer thickness and organic molecule concentration to achieve optimum devices using the evaporation–solution method. For the optimum PSC at each PbI₂ thickness, the thickness of the inorganic lead halide and concentration of organic molecule solution showed a linear relationship, and an empirical formula was determined.

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