Undesirable role of remnant PbI₂ layer on low temperature processed planar perovskite solar cells

Abstract

Two-step deposition methods for preparing CH₃NH₃PbI₃ films are becoming increasingly competitive when considering fabricating perovskite solar cells (PSCs) under ambient conditions, along with possessing the ability to better control their morphology. The surplus PbI₂ phase can be detected within the formed CH₃NH₃PbI₃ films due to partial conversion of PbI₂ to CH₃NH₃PbI₃, which causes variation in the PbI₂ stoichiometry of the resultant films. In this work, we carefully study the influence of the remnant PbI₂ on the performance of planar PSCs including efficiency and thermal stability by varying the PbI₂ stoichiometry of the resultant CH₃NH₃PbI₃. Through a further comparative study, the undesirable role of the remnant PbI₂ layer on planar PSC performance is exposed, indicating that the remnant PbI₂ layer not only greatly impedes carrier extraction and transport, but also accelerates the degradation of the CH₃NH₃PbI₃ film. To further eliminate the interference of this phenomenon, low temperature processed planar PSCs approaching 13% efficiency are attained in ambient atmosphere by the modified two-step method with enhanced thermal stability, which has significant potential for future mass production of PSCs and provides insight into the correlation between device performance and the PbI₂ stoichiometry of CH₃NH₃PbI₃ films.

---

1. Introduction

Organic–inorganic hybrid perovskite solar cells (PSCs) have witnessed unprecedented efficiency progress in the past few years and have been considered as one of the most promising candidates in the community of solar cells due to their solution-processable technology.^1–5 To date, by the improvement of device construction, the optimization of the light capture layer, and the enhancement of the interface characteristics, the fabrication of cells with a high power conversion efficiency (PCE) in excess of 22.1% has been achieved.^6–8 In addition, tandem perovskite configurations exceeding 25% also can be obtained with silicon solar cells as top cells.⁹ Meanwhile, after continuous efforts,^10–13 long-term stability tests for PSCs stored in air at room temperature without encapsulation have shown that they can last over 1000 h under full sunlight. However, before PSCs reach the market, there is a serious problem required to be addressed, that the fabrication of efficient PSCs is often restricted to being in glove boxes filled with nitrogen as a consequence of the poor stability against water and oxygen.^10,14

Fortunately, a two-step deposition method has been developed to offer high quality perovskite films in ambient air with excellent morphology characteristics. And it is therefore justified that a high PCE of 15% has been obtained via a conventional two-step sequential deposition method (CSD) under relatively relaxed conditions, in which PbI₂ is first deposited on a mesoporous TiO₂ substrate by spin-coating a dimethylformamide (DMF) solution of PbI₂, followed by exposing it to an anhydrous isopropanol (IPA) solution of CH₃NH₃I.^1,15 However, in spite of the advantages, another problem concerning remnant PbI₂ has also caused notable attention in turn which can result in an uncertain control of the amount and exact location of remnant PbI₂.^16–22 And the resultant PbI₂ stoichiometry usually depends on the length of the dipping time. At the same time, surplus PbI₂ precursor solution can be cast away producing a potential crisis of environmental pollution.

Currently, although many researchers have pointed out that a slight excess of PbI₂ in the perovskite film is beneficial and the best performing devices have a few percent excess PbI₂ in them, not everyone finds the excess PbI₂ to be beneficial.^{17–19,23,24} That makes it an open question. For those who hold the positive view, excess PbI₂ can passivate the TiO₂/CH₃NH₃PbI₃ interface to decrease carrier recombination,^18–20 and passivate the grain boundary of the perovskite resulting in good termination with fewer intra-band gap states as possible recombination centers.²¹ Besides, they also propose that PbI₂ may be not beneficial in itself, but rather be correlated with other secondary advantageous effects, such as increasing perovskite grains or improving crystallization.^18,22 Oppositely, those who hold the negative view argue that if the remnant PbI₂ layer is too thick, it can result in electronic insulation and blocking of the carrier extraction due to the unfavourable energy level alignment.^23–25 Moreover, some reports state that it is relatively difficult to realize highly efficient planar PSCs via CSD compared to the same situation with mesoporous PSCs.^23,24 This may be due to the fact that PbI₂ layers from PbI₂ DMF solution are often prone to form dense and highly crystalline films impeding the short diffusion of CH₃NH₃I into them and affecting the complete conversion of PbI₂ to perovskite. Also the correlation between remnant PbI₂ and device stability is uncertain. Motivated by the previous results, we here denote the situation of remnant PbI₂ classified into two categories. The remnant PbI₂, formed as a consequence of perovskite degradation or the incomplete conversion of the precursors where there is a slight excess of PbI₂, is beneficial by playing a role in passivation (Fig. 1a).^20,22 And the remnant PbI₂ layer, formed as a unreacted precursor composition via CSD, will to a greater extent be found close to the back contact (Fig. 1b).^17,25 Remarkably, when we here say the remnant PbI₂ layer, we technically mean the situation illustrated in Fig. 1b, where the influence on the planar PSC performance is distinct from that in Fig. 1a.

Fig. 1 (a) A diagram illustrating the passivation role of remnant PbI₂ on the perovskite grain boundary; (b) a schematic explaining the remnant PbI₂ layer in the bottom of the formed CH₃NH₃PbI₃ via CSD; (c) the cross-section scanning electron microscopy image corresponding to the overall device structure.

In this article, all low temperature processed planar PSCs are fabricated containing ITO/compact TiO₂ (c-TiO₂)/CH₃NH₃PbI₃/spiro-OMeTAD/Au (Fig. 1c). By controlling the length of the dipping time in the CSD method, the CH₃NH₃PbI₃ films with various PbI₂ stoichiometry are prepared. And it is further observed that the device performance based on those films originally increases and then decreases with prolonging the dipping time, which are in good consistency with the result reported in ref. 19. Meanwhile, the fact that the remnant PbI₂ layer should be mainly responsible for the inferior CH₃NH₃PbI₃ thermal stability is experimentally demonstrated. Hence, the confluence of those two findings unveils the undesirable role of the remnant PbI₂ layer on planar PSCs. Furthermore, it is also indicated that CSD is not an effective method for the fabrication of planar PSCs. And a modified two-step deposition method in ambient atmosphere is proposed to fabricate PbI₂-free planar PSCs possessing superior device performance.

2. Experimental methods

2.1. Device fabrication

The 3 cm × 3 cm ITO glass substrates were firstly rinsed with a mixed solution of ethanol and acetone (1 : 1 volume ratio) via sonication for 15 min, and then sequentially rinsed with IPA solution for 15 min. Next, ITO glass substrates were treated in a UV-ozone clearer for 12 min to further remove attached residual organics on the surface. After that, a low temperature processed c-TiO₂ electron transport layer was deposited on the ITO substrate by a previously reported anodic oxidation method.²⁶ And the scanning electron microscopy (SEM) images of the formed dense c-TiO₂ are shown in Fig. S1† showing good uniformity and crystallization, apart from full coverage. For comparison, the substrates based on mesoporous TiO₂ (m-TiO₂) were also prepared according to ref. 27. With regard to the fabrication of CH₃NH₃PbI₃ films, we adopted CSD as a starting point, where the PbI₂ DMF solution (40 wt%) was primarily spun on c-TiO₂ and m-TiO₂ substrates at 5000 rpm for 40 s in ambient air (about 20% humidity) respectively, followed by exposing the formed PbI₂ film to an IPA solution of CH₃NH₃I (10 mg mL⁻¹). To observe the effect of the remnant PbI₂ layer on PSC performance, the length of the dipping time was controlled to vary the stoichiometry of the resultant films. In contrast, the modified two-step method was also proposed, in which PbI₂ powder (around 50 mg) was deposited on a c-TiO₂ substrate via thermal evaporation in a low vacuum (∼1 Pa) with 60 A evaporation current, and subsequently surplus CH₃NH₃I powder was evaporated on the formed PbI₂ film as well to ensure full conversion from PbI₂ to CH₃NH₃PbI₃. The distance between the c-TiO₂ substrate and evaporation source was kept at 15 cm. Then, the resultant CH₃NH₃PbI₃ film was further assembled by the interaction and inter-diffusion between the solid-state PbI₂ and CH₃NH₃I during the process of annealing at 100 °C. Finally, surplus CH₃NH₃I was removed by rinsing CH₃NH₃PbI₃ with IPA solution. A spiro-OMeTAD solution was prepared by dissolving 80 mg spiro-OMeTAD in 1 mL chlorobenzene, to which 18 μL of lithium bis(trifluoromethanesulfonyl)imide (LiTFSI) solution (280 mg mL⁻¹ acetonitrile) and 10 μL of 4-tert-butylpyridine (tBP) were added. The hole transport layer (HTL) was prepared by spin-coating spiro-OMeTAD solution at 3500 rpm for 40 s. Finally, a 150 nm thick gold counter electrode was deposited by thermal evaporation in a high vacuum (under 10⁻⁴ Pa) atmosphere and the dot area of the planar PSCs was 0.06 cm². The whole fabrication temperature could be confined to 120 °C.

2.2. Characterization

The surface morphology and cross-section characteristics were observed by scanning electron microscopy (SEM) (Hitachi SU8010). The crystallization and phase identification of the CH₃NH₃PbI₃ and PbI₂ films were performed using X-ray diffraction (XRD) (Philips PANalytical X’Pert Pro, Cu Kα). The CH₃NH₃PbI₃ absorption spectrum was recorded by a UV-vis spectrophotometer (Cary 5000). The steady-state photoluminescence (PL) measurement was recorded using a steady-state fluorescence spectrometer (Jobin Yvon FluoroLog-3). Electrochemical impedance spectroscopy (EIS) (METEK) was employed to characterize the carrier mechanism of the planar PSCs. Thermogravimetric analysis (TGA) was conducted to illustrate the thermal stability property of the CH₃NH₃PbI₃ films. The J–V characteristics were measured with Abet sun 2000 solar simulator under illumination of AM 1.5G (100 mW cm⁻²). The external and quantum efficiency (EQE) spectrum was measured to determine the photoelectric response of the planar PSCs.

3. Results and discussion

SEM was used to investigate the surface morphology evolution of the corresponding perovskite films. As shown in Fig. 2a and b, a dense and smooth PbI₂ layer was formed by spinning the precursor PbI₂ DMF solution which could directly block the rich contact of CH₃NH₃I from the bottom PbI₂ layer, although surface PbI₂ could well convert CH₃NH₃PbI₃ once interfaced with the CH₃NH₃I component. It was worth noting that the CH₃NH₃PbI₃ crystallization characteristic gradually became enhanced and the grain size also showed a growing trend with prolonging the dipping time, facilitating a lowering of the defect state density of the perovskite grain boundary and thereby decreasing the carrier recombination to some extent. Also, a long dipping time could better remove the remnant PbI₂ layer compared to a short dipping time, as supported by previous experimental evidence.^1,16,25 Unfortunately, for the length of dipping time slightly longer than that required for complete conversion (>9 min), apparent pinholes on the formed perovskite films were observed. That might result from either the slow dissolution of the formed CH₃NH₃PbI₃ into IPA solution or the reversible decomposition of CH₃NH₃PbI₃ into PbI₂. Theoretically, a full coverage perovskite film can reduce carrier recombination coming from a contact between TiO₂ and the HTL, which is essential for efficient planar PSCs.^18,24,27 Consequently, a long dipping time should be not suitable for fabricating planar PSCs, leading to a dilemma. To better support the hypothesis that the remnant PbI₂ layer was attributed to the situation mentioned in Fig. 1b, a cross-section SEM image was taken. As shown in Fig. S2,† the spun PbI₂ exhibited a uniform sectional characteristic. With prolonging the dipping time, the film volume turned to be expanding, agreeing with the reported transformation mechanism from PbI₂ to CH₃NH₃PbI₃. However, comparing the samples with various dipping times, the sectional morphology presented an obvious difference. A long dipping time (9 min) could form a close complete perovskite grain bulk. In comparison, the samples with a short dipping time possessed an obviously ambiguous layer structure, suggesting a remnant PbI₂ layer sandwiched between c-TiO₂ and CH₃NH₃PbI₃. At the same time, distinct from the situation of planar construction, the mesoporous construction could achieve complete conversion of PbI₂ to perovskite in a relatively short time with a capping layer (Fig. S3†). It is likely that the PbI₂ layer deposited on top of the mesoporous substrate has increased the roughness and contact area between PbI₂ and CH₃NH₃I which allows the conversion reaction to proceed faster.

Fig. 2 SEM images of PbI₂ film fabricated by spinning PbI₂ DMF precursor solution (40 wt%) (a and b) and the corresponding CH₃NH₃PbI₃ films fabricated by dipping the spun PbI₂ into IPA solution of CH₃NH₃I (10 mg mL⁻¹) with various dipping times: (c and d) 1 min; (e and f) 3 min; (g and h) 5 min; (i and j) 7 min; and (k and l) 9 min.

As further information to substantiate our hypothesis, X-ray diffraction (XRD) and UV-vis absorption spectra of CH₃NH₃PbI₃ were also recorded. As shown in Fig. 3a, the spun PbI₂ film on a planar substrate obviously presented intensity diffraction peaks near 12.6° indicating excellent crystallization. With delaying the length of the dipping time, the CH₃NH₃PbI₃ (near 14°) diffraction peaks started to compete with those of PbI₂. Notably, by comparing the XRD results, no systematic peak shifts were observed, which showed that the composition of the crystalline CH₃NH₃PbI₃ is unaffected by the overall stoichiometry and perovskite phase, which was essentially the same in the different samples. On the other hand, UV-vis absorption was measured on the same subset of samples (Fig. 3b). Also the samples with 30 s and 15 min dipping time were incorporated into them as extreme situations. As shown in Fig. 3b, the absorption of the CH₃NH₃PbI₃ films presented a large difference as a consequence of the variation of the PbI₂ stoichiometry. This may be because the reduced remnant PbI₂ firstly strengthened the ability of absorption but then vast pinholes in turn weakened it. Moreover, although all those samples showed typical CH₃NH₃PbI₃ absorption characteristics, there existed sharp absorption losses at a wavelength around 500 nm. It is known that the band gap of PbI₂ is around 2.3–2.5 eV, which corresponds to a threshold absorption wavelength around 496–539 nm.^20,28 Therefore, major sunlight absorption is lost due to the influence of the remnant PbI₂ layer. Furthermore, the pictures of the corresponding CH₃NH₃PbI₃ films are also shown in Fig. 3c and are in good consistency with the findings concluded from Fig. 3a and b.

Fig. 3 (a) XRD pattern of CH₃NH₃PbI₃ films by CSD with various dipping times; (b) the corresponding absorption spectrum; (c) the corresponding authentic CH₃NH₃PbI₃ pictures.

Furthermore, the planar PSCs were fabricated according to the structure provided in Fig. 1c via CSD. Fig. 4a showed the whole planar device performance and the corresponding parameters are summarized in Table 1. Also mesoporous PSCs were fabricated by CSD with a 5 min dipping time and the device performance was presented in Fig. S4† for comparison. It was observed that the planar devices exhibited an improved short circuit density (J_sc) with prolonging the length of the dipping time. However, when the dipping time reached as long as 15 min, the J_sc value reduced 16.20 mA cm⁻². As an extreme situation, the device with a full PbI₂ film replacing CH₃NH₃PbI₃ was also probed (Fig. S5†). It exhibited obvious ohmic contact characteristics and possessed an enormous resistance, indicating that pure PbI₂ has the potential to impede carrier extraction and transport. Correspondingly, the EQE response (Fig. 4b) presented nearly the same trend as that of the J_sc variation observed in Table 1. For the situation of mesoporous construction, it could exhibit a decent performance via CSD with just 5 min dipping time because of the prosperous conversion of PbI₂ to CH₃NH₃PbI₃ as well as a favourable morphology (Fig. S3 and S4†). In general, the dependence of V_oc is closely connected with the rising concentration of the photo-generated carriers and the associated separation of the electron and hole quasi-Fermi levels. It is therefore justified that the favourable carrier collection should facilitate the achievement of a high V_oc value. Surprisingly, although the planar device with a dipping time of 5 min had an inferior J_sc and V_oc, it possessed a distinctly superior fill factor (FF).

Fig. 4 (a) J–V curves showing performance of planar PSCs along with dipping time when fabricated via CSD without encapsulation; (b) corresponding EQE spectra. All the measurements for the J–V curves were carried out at regular time intervals (0.02 V s⁻¹ scanning rate) and in the forward scanning direction (J_sc to V_oc) in air atmosphere.

Table 1 Photovoltaic parameters derived from J–V measurements corresponding to Fig. 4a

Sample	Jsc (mA cm^−2)	Voc (V)	FF (%)	PCE (%)
30 s	0.32	0.52	54.08	0.09
1 min	9.06	0.61	32.38	1.79
3 min	15.49	0.69	38.64	4.13
5 min	16.76	0.80	60.71	8.14
7 min	17.24	0.82	53.34	7.54
9 min	17.91	0.84	47.39	7.13
15 min	16.20	0.66	50.70	5.42

---

To better illustrate our conclusion, the energy band alignment was further shown in Fig. 5. Theoretically, the PSC energy band alignment should be considered as Fig. 5a,²⁷ in which the photo-generated carriers could successfully be conveyed to the relevant electrode with a low potential barrier resistance facilitating the carrier collection. However, due to the presence of the remnant PbI₂ layer, the energy band alignment became the situation shown in Fig. 5b. The remnant PbI₂ layer would be considered as a tunnelling layer for its undesirable energy alignment, implying that the carriers were collected unless they went through the PbI₂ layer by the tunnelling effect with a high potential barrier. And the tunnelling current could be ascribed as eqn (1), where V_b represents the tunnelling voltage, ϕ represents the average work function, A represents the constant usually set as 1, and d represents the tunnelling layer thickness. Therefore, when the remnant PbI₂ layer was too thick, few carriers could be injected and collected. Meanwhile, steady-state photoluminescence (PL) and electrochemical impedance spectroscopy (EIS) were together displayed in Fig. S6 and S7† to confirm this description. In our case, the difference between all samples was only the variation of the dipping time resulting in different PbI₂ stoichiometry of the resultant CH₃NH₃PbI₃. With an approximately identical CH₃NH₃PbI₃ thickness (determined by cross-section SEM shown in Fig. S8†), all the CH₃NH₃PbI₃ films exhibited a significant PL peak around 780 nm, but the intensity was entirely different. The large difference in PL intensity should be mainly attributed to the number of photo-generated carriers resulting from the sunlight absorption difference. Additionally, owing to the influence of the remnant PbI₂, it was observed that when contacted with c-TiO₂, the PL quenching was also not as apparent as that we previously reported,²⁶ which implied an inferior carrier collection efficiency. Furthermore, the EIS results also demonstrated that the remnant PbI₂ layer seriously enhanced the carrier transport resistance by blocking carrier transport according to the feature at a low frequency response shown in the Nyquist plot.^29,30

Fig. 5 The relative energy band alignments. (a) Ideal PSC device without a PbI₂ film in the bottom of the perovskite film; (b) practical PSC device with the remnant PbI₂ layer existing in the bottom of the formed perovskite film.

With respect to the difference between the planar and mesoporous PSCs fabricated by the CSD method, several reports have shown positive effects of remnant PbI₂ in mesoporous PSCs.^20,21 Having studied and compared our results to the related studies reported in the literature we discuss here the comprehensive roles of remnant PbI₂ in PSCs to clarify this observation. Firstly, remnant PbI₂ indeed can reduce carrier recombination in the perovskite layer as well as at the interface of TiO₂/CH₃NH₃PbI₃ to some extent due to its increased barrier, preventing electrons from combining with grain boundary defects and oxygen vacancy defects.^17,20,21 Secondly, the remnant PbI₂ also can block the electron extraction and transport from the perovskite into TiO₂, especially when a thick and dense PbI₂ layer at TiO₂/CH₃NH₃PbI₃ is not thin enough to allow quantum tunnelling. Thirdly, remnant PbI₂ located in the perovskite grain boundary region alleviates the barrier of electron extraction from CH₃NH₃PbI₃ to the HTL by upward energy band bending, where electrons can be repelled thereby facilitating the improvement of hole extraction.²⁰ Moreover, PbI₂ has the ability to increase the crystallization growth of the CH₃NH₃PbI₃ film.²² Since it is a difficulty to control the amount and exact location of the remnant PbI₂ by the CSD method, competition concerning the above-mentioned effects together exists in typical PSCs.

A superior device performance is strongly dependent on photo-generated carrier extraction and transport, which avoid accumulation and decrease the corresponding recombination rate. According to a previous report, it is found that carrier extraction mainly arises from the interface of TiO₂/perovskite.³¹ For the case of mesoporous PSCs, it is better for carrier extraction owing to the increased TiO₂/perovskite interface area. However, abundant defects are located in the perovskite grain boundary and more oxygen vacancy defects are exposed to the interface of TiO₂/CH₃NH₃PbI₃, greatly weakening the carrier extraction efficiency in turn by capturing electrons into defects.^20,32 Therefore, a PbI₂ passivation of these defects, which reduces the density of defects as much as possible, can be a very important key technique. Also because there is a high chance that remnant PbI₂ does not make a continuous layer on the mesoporous TiO₂ scaffold, a sufficient TiO₂/CH₃NH₃PbI₃ interface area for carrier extraction and transport still exists despite slight excess remnant PbI₂ decreasing the interface area. Moreover, remnant PbI₂ passivation of defects may make up those losses. On the other hand, for planar PSCs, since the interface area is too little in comparison to the mesoporous structure, it is reasonable that there is not enough area for carrier extraction and transport if PbI₂ is there. As our results shows, the remnant PbI₂ is likely to be present at the interface of TiO₂/CH₃NH₃PbI₃ by means of a thick and dense layer structure in planar PSCs fabricated by CSD. In this case, it may completely prevent carrier extraction from perovskite to TiO₂ surpassing its other positive effects.

Besides, a high efficiency has been obtained in PSCs with a slight excess of PbI₂, formed as the incomplete conversion of the precursors where a slight excess of PbI₂ is intentionally used.²² This is because the remnant PbI₂ used in the precursor solution itself usually results in a uniform distribution in the bulk of CH₃NH₃PbI₃ grain boundaries, preferably not at the TiO₂/CH₃NH₃PbI₃ interface. Consequently, it can not only passivate defects, but possesses enough TiO₂/CH₃NH₃PbI₃ interface area to ensure carrier extraction and transport in this situation.

Moreover, the influence of the remnant PbI₂ layer on the thermal stability of the perovskite films was also observed. And thermogravimetric analysis (TGA) was conducted to characterize the thermal property of the CH₃NH₃PbI₃ film on a silicon substrate. Fig. 6a demonstrated that the weight loss was mainly concentrated at 200–300 °C and 400–500 °C. According to the previously reported results, the weight loss for the absorption peak (AP) at 200–300 °C was the result of CH₃NH₃I loss decomposed from CH₃NH₃PbI₃.³³ With regard to the weight loss at 400–500 °C, it should be as a consequence of the PbI₂ melting which was beyond the scope of our discussion. As shown in Fig. 6b, the CH₃NH₃PbI₃ film possessing various stoichiometries via CSD had a detectable difference in the derivative weight value. And the greater the PbI₂ stoichiometry in the perovskite is, the higher the derivative weight value is, suggesting that the remnant PbI₂ layer indeed should be responsible for accelerating the degradation rate of the CH₃NH₃PbI₃ film. Besides, the colour of the formed CH₃NH₃PbI₃ film changed from dark black to dark yellow at 100 °C annealing in just 120 min (Fig. 6c), consistent with the observed results.

Fig. 6 Thermogravimetric analysis (TGA) of the CH₃NH₃PbI₃ films fabricated via the CSD method on a silicon substrate. (a) Weight loss percentage; (b) derivative of weight loss as a function of temperature; (c) the pictures of formed CH₃NH₃PbI₃ showing the colour with 5 min dipping time as a function of annealing time at 100 °C.

According to some reported works, a porous PbI₂ film can more effectively promote the conversion of PbI₂ to CH₃NH₃PbI₃, consequently facilitating the formation of PbI₂-free perovskite films.^34,35 Thus, a modified two-step method was proposed to fabricate planar PSCs, in which PbI₂ and CH₃NH₃I were successively deposited by thermal evaporation in an approaching ambient atmosphere. Different from the situation of the PbI₂ film by spinning PbI₂ DMF solution, a porous snowflake shape PbI₂ film was obtained (Fig. 7a and b), which increased the contact area between PbI₂ and CH₃NH₃I thereby reducing their diffusion length to form CH₃NH₃PbI₃. Moreover, the resultant CH₃NH₃PbI₃ film not only exhibited full coverage, but the grain size was also improved to about 1 μm (Fig. 7c and d). It was therefore justified that the modified two-step method was more suitable for the fabrication of planar PSCs to address the dilemma situation faced by CSD.

Fig. 7 SEM images: (a and b) PbI₂ film fabricated by the thermal evaporation deposition method in close to ambient atmosphere; (c and d) CH₃NH₃PbI₃ film fabricated by the modified two-step deposition method.

Therefore, the planar devices were prepared with an identical structure except for the fabrication method of CH₃NH₃PbI₃ and the corresponding results are shown in Fig. 8a. Obviously, with the backward and forward direction scanning, the J–V curves are closely coincident, indicating a relatively credible efficiency and little hysteresis. Meanwhile, compared to the samples made via CSD, all the device parameters were enhanced and the whole EQE response was superior in the whole wavelength range, agreeing with the improvement of EIS measurement (Fig. S9†). As a result, the planar PSCs exhibited a V_oc value of 1.01 V, J_sc of 19.19 mA cm⁻², FF of 65.7%, and PCE value of 12.74%. On the other hand, as shown in Fig. 8b, two preferential and high-purity XRD diffraction peaks near 14.17° and 28.51° were detected, as evidence of the CH₃NH₃PbI₃ phases of (110) and (220), indicating an enhanced crystallization structure facilitating carrier transport without the remnant PbI₂ layer. Moreover, compared to CH₃NH₃PbI₃ via CSD, the CH₃NH₃PbI₃ films via the modified two-step method could suffer from long-term thermal annealing even for as long as 2 days at 100 °C in air atmosphere. And the device performance corresponding to thermal annealing reaching 2 days is also shown in Fig. S10,† implying that the efficiency value could still reach 34% of its initial PCE value. The degraded efficiency might arise from the influence of LiTFSI and tBP on the CH₃NH₃PbI₃ film. Anyway, the enhanced thermal stability due to complete conversion of PbI₂ to CH₃NH₃PbI₃ was observed.

Fig. 8 (a) Device performance for the planar PSCs fabricated by the modified two-step deposition method; (b) the XRD image of the CH₃NH₃PbI₃ film fabricated by the modified two-step deposition method, in which the variation of the CH₃NH₃PbI₃ colour with annealing time at 100 °C is shown in the inset.

4. Conclusions

In summary, we have demonstrated that the remnant PbI₂ layer has an undesirable effect on planar PSC performance including efficiency and thermal stability by varying the PbI₂ stoichiometry via CSD. Different from the situation in mesoporous PSCs, it is observed that there is a dilemma between complete conversion of PbI₂ to CH₃NH₃PbI₃ and favourable morphology characteristics in planar PSCs via CSD. That leads to either a high carrier transport barrier due to unsatisfied energy level alignment or serious leakage current loss due to the emergence of rich pinholes. Additionally, the undesirable effect of the remnant PbI₂ layer on the CH₃NH₃PbI₃ thermal stability is also experimentally displayed. And a modified two-step deposition method is proposed for fabricating PbI₂-free perovskite films. Furthermore, all low temperature processed planar PSCs approaching 13% PCE in close to ambient atmosphere can finally be achieved. These results represent an important step forward in developing low-cost and stable planar PSCs and provide insight to investigate connections between device performance, stoichiometry and thermal stability, which is valuable for further solar cell development.

Conflict of interest

The authors declare no competing financial interest.

Acknowledgements

We appreciate financial support from the National Natural Science Foundation of China (Grant No. 61504068).

References

1. J. Burschka, N. Pellet, S. Moon, R. Humphry-Baker, P. Gao, M. K. Nazeeruddin and M. Gratzel, Sequential Deposition as a Route to High-performance Perovskite-sensitized Solar Cells, Nature, 2013, 499, 316–319.
2. J. You, Z. Hong, Y. Yang, Q. Chen, M. Cai, T. Song, C. Chen, S. Lu, Y. Liu, H. Zhou and Y. Yang, Low-temperature Solution-Processed Perovskite Solar Cells with High Efficiency and Flexibility, ACS Nano, 2014, 8, 1674–1680.
3. W. S. Yang, J. H. Noh, N. J. Jeon, Y. C. Kim, S. Ryu, J. Seo and S. I. Seok, High-performance Photovoltaic Perovskite Layers Fabricated Through Intramolecular Exchange, Science, 2015, 348, 1234–1237.
4. M. Saliba, T. Matsui, J. Seo, K. Domanski, J. Correa-Baena, M. K. Nazeeruddin, S. M. Zakeeruddin, W. Tress, A. Abate, A. Hagfeldt and M. Gratzel, Cesium-containing Triple Cation Perovskite Solar Cells: Improved Stability, Reproducibility and High Efficiency, Energy Environ. Sci., 2016, 9, 1989–1997.
5. Y. Wu, J. Wang, X. Qiu, R. Yang, H. Lou, X. Bao and Y. Li, Highly Efficient Inverted Perovskite Solar Cells With Sulfonated Lignin Doped PEDOT as Hole Extract Layer, ACS Appl. Mater. Interfaces, 2016, 8, 12377–12383.
6. W. Chen, Y. Wu, Y. Yue, J. Liu, W. Zhang, X. Yang, H. Chen, H. Chen, E. Bi, I. Ashraful, M. Gratzel and L. Han, Efficient and Stable Large-area Perovskite Solar Cells with Inorganic Charge Extraction Layers, Science, 2015, 350, 944–948.
7. X. Li, D. Bi, C. Yi, J. Decoppet, J. Luo, S. M. Zakeeruddin, A. Hagfeldt and M. Gratzel, A Vacuum Flash-assisted Solution Process for High-efficiency Large-area Perovskite Solar Cells, Science, 2016, 8060.
8. NREL efficiency chart.
9. D. P. McMeekin, G. Sadoughi, W. Rehman, G. E. Eperon, M. Saliba, M. T. Horantner, A. Haghighirad, N. Sakai, L. Korte, B. Rech, M. B. Johnston, L. M. Herz and H. J. Snaith, A Mixed-Cation Lead Mixed-Halide Perovskite Absorber for Tandem Solar Cells, Science, 2016, 351, 151–155.
10. D. Bryant, N. Aristidou, S. Pont, I. Sanchez-Molina, T. Chotchunangatchaval, S. Wheeler, J. R. Durrant and S. A. Haque, Light and Oxygen Induced Degradation Limits the Operational Stability of Methylammonium Lead Triiodide Perovskite Solar Cells, Energy Environ. Sci., 2016, 9, 1655–1660.
11. B. Conings, J. Drijkoningen, N. Gauquelin, A. Babayigit, J. D’Haen, L. D’Olieslaeger, A. Ethirajan, J. Verbeeck, J. Manca, E. Mosconi, F. D. Angelis and H. Boyen, Intrinsic Thermal Instability of Methylammonium Lead Trihalide Perovskite, Adv. Energy Mater., 2015, 5, 1500477.
12. K. Domanski, J. Correa-Baena, N. Mine, M. K. Nazeeruddin, A. Abate, M. Saliba, W. Tress, A. Hagfeldt and M. Gratzel, Not All That Glitters Is Gold: Metal-Migration-Induced Degradation in Perovskite Solar Cells, ACS Nano, 2016, 10, 6306–6314.
13. T. A. Berhe, W. Su, C. Chen, C. Pan, J. Cheng, H. Chen, M. Tsai, L. Chen, A. A. Dubale and B. Hwang, Organometal Halide Perovskite Solar Cells: Degradation and Stability, Energy Environ. Sci., 2016, 9, 323–356.
14. A. M. Leguy, Y. Hu, M. Campoy-Quiles, M. I. Alonso, O. J. Weber, P. Azarhoosh, M. W. Schilfgaarde, M. T. Weller, T. Bein, J. Nelson, P. Docampo and P. R. F. Barnes, Reversible Hydration of CH₃NH₃PbI₃ in Films, Single Crystals, and Solar Cells, Chem. Mater., 2015, 27, 3397–3407.
15. H. Ko, J. Lee and N. Park, 15.76% Efficiency Perovskite Solar Cells Prepared under High Relative Humidity: Importance of PbI₂ Morphology in Two-step Deposition of CH₃NH₃PbI₃, J. Mater. Chem. A, 2015, 3, 8808–8815.
16. Q. Chen, H. Zhou, Z. Hong, S. Luo, H. S. Duan, H. H. Wang, Y. Liu, G. Li and Y. Yang, Planar Heterojunction Perovskite Solar Cells Via Vapor-assisted Solution Process, J. Am. Chem. Soc., 2014, 136, 622.
17. T. Jacobsson, J. Correa-Baena, E. H. Anaraki, B. Philippe, S. Stranks, M. Bouduban, W. Tress, K. Schenk, J. Teuscher, J. Moser, H. Rensmo and A. Hagfeldt, Unreacted PbI₂ as a Double-Edged Sword for Enhancing the Performance of Perovskite Solar Cells, J. Am. Chem. Soc., 2016, 138, 10331–10343.
18. D. Bi, W. Tress, M. I. Dar, P. Gao, J. Luo, C. Renevier, K. Schenk, A. Abate, F. Giordano, J.-P. Correa Baena, J.-D. Decoppet, S. M. Zakeeruddin, M. K. Nazeeruddin, M. Graetzel and A. Hagfeldt, Efficient Luminescent Solar Cells Based On Tailored Mixed-cation Perovskite, Sci. Adv., 2016, 2, e1501170.
19. D. H. Cao, C. C. Stoumpos, C. D. Malliakas, M. J. Katz, O. K. Farha, J. T. Hupp and M. G. Kanatzidis, Remnant PbI₂, an Unforeseen Necessity in High-Efficiency Hybrid Perovskite-Based Solar Cells, APL Mater., 2014, 2, 091101.
20. Q. Chen, H. Zhou, T. B. Song, S. Luo, Z. Hong, S. B. Duan, L. Dou, Y. Liu and Y. Yang, Controllable Self-induced Passivation of Hybrid Lead Iodide Perovskites Toward High Performance Solar Cells, Nano Lett., 2014, 14, 4158.
21. L. Wang, C. McCleese, A. Kovalsky, Y. Zhao and C. Burda, Femtosecond Time-resolved Transient Absorption Spectroscopy of CH₃NH₃PbI₃ Perovskite Films: Evidence for Passivation Effect of PbI₂, J. Am. Chem. Soc., 2014, 136, 12205.
22. C. Roldan-Carmona, P. Gratia, I. Zimmermann, G. Grancini, P. Gao, M. Graetzel and M. K. Nazeeruddin, High Efficiency Methylammonium Lead Triiodide Perovskite Solar Cells: the Relevance of Non-Stoichiometric Precursors, Energy Environ. Sci., 2015, 8, 3550–3556.
23. Y. Zhao and K. Zhu, Three-step Sequential Solution Deposition of PbI₂-free CH₃NH₃PbI₃ Perovskite, J. Mater. Chem. A, 2015, 3, 9086–9091.
24. H. Zhang, J. Mao, H. He, D. Zhang, H. L. Zhu, F. Xie, K. S. Wong, M. Gratzel and W. C. H. Choy, A Smooth CH₃NH₃PbI₃ Film via a New Approach for Forming the PbI₂ Nanostructure Together with Strategically High CH₃NH₃I Concentration for High Efficient Planar-Heterojunction Solar Cells, Adv. Energy Mater., 2015, 5, 1501354.
25. Y. H. Lee, J. Luo, R. Humphry-Baker, P. Gao, M. Gratzel and M. K. Nazeeruddin, Unraveling the Reasons for Efficiency Loss in Perovskite Solar Cells, Adv. Funct. Mater., 2015, 25(25), 3925–3933.
26. Y. Du, H. Cai, H. Hong, Y. Wu, L. Huang, J. Ni, J. Li and J. Zhang, Novel Combination of Efficient Perovskite Solar Cells with Low Temperature Processed Compact TiO₂ Layer via Anodic Oxidation, ACS Appl. Mater. Interfaces, 2016, 8, 12836–12842.
27. Y. Du, H. Cai, J. Ni, J. Li, H. Yu, X. Sun, Y. Wu, H. Wen and J. Zhang, Air-processed, Efficient CH₃NH₃PbI_3−xCl_x Perovskite Solar Cells with Organic Polymer PTB7 as a Hole-transport Layer, RSC Adv., 2015, 5, 66981–66987.
28. A. F. Silva, N. Veissid, C. Y. An, I. Pepe, N. B. D. Oliveira and A. V. B. Silva, Optical Determination of the Direct Bandgap Energy of Lead Iodide Crystals, Appl. Phys. Lett., 1996, 69, 1930–1932.
29. A. R. Pascoe, N. W. Duffy, A. D. Scully, F. Huang and Y. Cheng, Insights Into Planar CH₃NH₃PbI₃ Perovskite Solar Cells Using Impedance Spectroscopy, J. Phys. Chem. C, 2015, 119, 4444–4453.
30. Y. Ma, Y. H. Chung, L. Zheng, D. Zhang, X. Yu, L. Xiao, Z. Chen, S. Wang, B. Qu, Q. Gong and D. Zou, Improved Hole-Transporting Property via HAT-CN for Perovskite Solar Cells without Lithium Salts, ACS Appl. Mater. Interfaces, 2015, 7, 6406–6411.
31. Y. H. Lee, J. Luo, M. Son, P. Gao, K. T. Cho, J. Seo, S. M. Zakeeruddin, M. Gratzel and M. K. Nazeeruddin, Enhanced Charge Collection with Passivation Layer in Perovskite Solar Cells, Adv. Mater., 2016, 28, 3966–3972.
32. Y. Du, H. Cai, H. Hong, Y. Wu, Z. Li, J. Xu, L. Huang, J. Ni, J. Li and J. Zhang, Revealing the Unfavorable Role of Superfluous CH₃NH₃PbI₃ Grain Boundary Traps in Perovskite Solar Cells on Carrier Collection, RSC Adv., 2016, 6, 83264–83272.
33. N. Yantara, F. Yanan, C. Shi, H. A. Dewi, P. P. Boix, S. G. Mhaisalkar and N. Mathews, Unravelling the Effects of Cl Addition in Single Step CH₃NH₃PbI₃ Perovskite Solar Cells, Chem. Mater., 2015, 27, 2309–2314.
34. J. Cao, F. Wang, H. Yu, Y. Zhou, H. Lu, N. Zhao and C. Wong, Porous PbI₂ Films for the Fabrication of Efficient, Stable Perovskite Solar Cells via Sequential Deposition, J. Mater. Chem. A, 2016, 4, 10223–10230.
35. T. Zhang, M. Yang, Y. Zhao and K. Zhu, Controllable Sequential Deposition of Planar CH₃NH₃PbI₃ Perovskite Films via Adjustable Volume Expansion, Nano Lett., 2015, 15, 3959–3963.

---

Footnote

† Electronic supplementary information (ESI) available: The SEM images of low temperature processed compact TiO₂ film. The cross-section SEM images of CH₃NH₃PbI₃ films on c-TiO₂ via CSD with various dipping times. The cross-section SEM image of CH₃NH₃PbI₃ film on m-TiO₂ via CSD. The J–V and EIS performance of mesoporous PSCs via CSD. The steady-state photoluminescence (PL) measurement. The electrochemical impedance spectroscopy (EIS) measurement. The diagram of planar PSCs replacing CH₃NH₃PbI₃ with PbI₂ and corresponding J–V curve. See DOI: 10.1039/c6ra19265c

---