Zhen Li and Guanjun Yang

School of Materials Science and Engineering, Xi'an Jiaotong University, Xi'an 710049, China

Abstract: Repressing the thermal decomposition during the process of heat treatment plays an indispensable part in the preparation of perovskite films. Here, a methylammonium iodide healing method was applied to prevent the volatilization of the organic component inside the perovskite structure during the heat treatment. High-quality CH3NH3PbI3 film with a much larger grain size over 800 nm was successfully fabricated via this healing method. Besides, the absorption and photoluminescence intensity were also both improved. Finally, the best power conversion efficiency of 18.89% with a fill factor over 80% was realized in an n-i-p configuration while possessing outstanding stability. This work suggests that methylammonium iodide healing method is a reliable way to promote crystal growth and improve the photovoltaic performance and humidity stability of the CH3NH3PbI3 solar cells.

Key words: perovskite; solar cells; heat treatment; film quality; photovoltaic performance; stability

1. Introduction

Metal halide perovskites, with formula ABX3, are emerging as a promising photoelectric material owing to their high absorption coefficient, slow carrier recombination and large carrier diffusion lengths[1-3]. Moreover, the low-priced processing technique via a solution method makes the perovskite material possess a bright future in commercialization[4,5]. Through long-term research, the power conversion efficiency (PCE) of the perovskite solar cells (PSCs) has increased significantly from 3.8%[6]to 25.6%[7], which is already close to that of silicon solar cells.

To realize high-efficiency PSCs, many aspects should be taken into account to reduce the amount of defects that might exist inside the solar cells[8]. To some extent, the perovskite grain boundary, where the light scattering and carrier recombination usually happens, is considered to be one kind of the defects that significantly affects the film quality[9]. Generally, a larger perovskite grain was supposed to be beneficial to the performance and stability of the PSCs[10,11].Therefore, promoting the crystal growth and reducing the grain boundary density are the pursuit of the preparation process. With regard to improving the film quality and increasing the grain size, much technological improvement has been developed by researchers, including the usage of chemical additives[12-15], the modification of interfacial energy[16,17], solvent engineering[18,19]and many other effective methods.

Since the thermal transmission during the stage of the heat treatment is a significant factor that influences the crystallinity and grain size[20,21], elevating the temperature or extending the duration of the heat treatment is supposed to promote the crystal growth of the perovskite films[22]. However,more does not always mean better; too high temperature or too long processing time of the heat treatment probably causes the decomposition of the perovskite material. For example, CH3NH3PbI3(MAPbI3), the most common perovskite composition, tends to decompose into lead iodide (PbI2) at a relatively high temperature, since the organic component is easy to volatilize during the conventional post-annealing treatment[23-25]. The PbI2emerging inside the film is usually located at the perovskite grain boundaries[26,27], which might put obstacles in the way of the perovskite grain boundary migration and crystal growth. Consequently, on the basis of enough thermal transmission, repressing the decomposition of the perovskite material is significant for the crystal growth.

Atmosphere healing treatment is also an optional method to improve perovskite film quality[28], among which the methylamine (MA) gas healing method could be applied to eliminate the dendrite-like crystals and obtain a dense and smooth perovskite film through a solid-liquid-solid phase transition[29]. However, the grain size usually tends to become smaller after the MA gas healing treatment[30]. Moreover, the usage of the explosive MA gas may increase the potential danger in the process of the gas healing treatment. So, it is still necessary to find a reliable atmosphere healing method to increase the perovskite grain size while keeping the films dense, smooth and pinhole-free.

Herein, we adopt a methylammonium iodide (MAI) healing method to replace the MA gas with the MAI atmosphere in an enclosed space[31]. High-quality MAPbI3films with a much larger average grain size over 800 nm were realized after the healing treatment. The absorption and photoluminescence intensity were also both improved. Moreover, a best PCE of 18.89% with an ultrahigh fill factor (FF) of over 80%was obtained in an n-i-p configuration while possessing an outstanding stability under ambient conditions. This work provides a new reliable fabrication method for high-quality perovskite film and high-efficiency PSCs with enhanced humidity stability based on MAPbI3absorbers via atmosphere healing treatment.

Fig. 1. (Color online) Schematic diagram of (a) the MAPbI3 perovskite fabrication treated by MAI healing method and (b) MAI healing process.

2. Material and methods

2.1. Materials

PbI2(99.99%), MAI (99.5%) and Spiro-MeOTAD (99.8%)were provided by Xi’an Polymer Light Technology Corp. Fluorine-doped tin oxide (FTO)-coated glasses (TEC7, 2.2 mm)were provided by Ying Kou You Xuan Trade Co., Ltd. N,N-dimethylformamide (DMF, anhydrous, 99.8%) was provided by Sigma-Aldrich.

2.2. Film and solar cell fabrication

The acetone, ethanol, and deionized water were used to clean the FTO glasses sequentially, each for 15 min respectively. Afterwards, the FTO glasses were dried by nitrogen and cleaned in UV ozone for 30 min. Then, MAPbI3precursor (1M)in DMF was spin-coated on the FTO substrate at 2000 rpm for 8 s. Immediately after the spin-coating process, the samples were transferred into a homemade gas-pump device[32], inside which the samples were treated at a pressure of 3000 Pa.Then, MAI precursor (0.2 M) in DMF was spin-coated on another FTO substrate at 4000 rpm for 20 s, which were also treated by the gas-pumping method at a higher pressure of 10 000 Pa.

Then, we used a MAI healing method to prevent MAPbI3material decomposing into PbI2due to the loss of organic components during the process of heat treatment. The elimination of the PbI2could remove the obstacle of the grain boundary migration and promote the crystal growth. As shown in Fig. 1(a), MAI and MAPbI3films were placed on a pair of faceto-face heating plates. Then, the gas-driven heating plates were moved to fit tightly and created an enclosed space.After we elevated the temperature of the heating plates,which were identical for the MAI and MAPbI3films, the atmosphere rich with MAI molecule could prevent the component loss inside the perovskite films due to the saturation states of the organic ammonium in this enclosed space, as shown in Fig. 1(b).

For PSCs fabrication, a tin dioxide (SnO2) modified titanium dioxide (TiO2) film was fabricated on the FTO substrate according to our previous report[33]. Then, the MAPbI3perovskite films and MAI films were both prepared via the gaspumping method as mentioned above. Afterwards, the films were treated with a heating temperature of 100 °C and a processing time of 15 min respectively for the post-annealing or MAI healing treatment. Spiro-MeOTAD was fabricated via a spin-coating method at 3000 rpm for 30 s onto the perovskite layer. Finally, the Au metal electrode was thermally evaporated, the thickness of which was controlled around 100 nm.

2.3. Material characterizations

Surface morphology was performed on a field-emission scanning electron microscope (SEM, MIRA3 TESCAN). Roughness of the film surface was tested on a high-resolution atomic force microscopy (AFM, INNOVA). X-ray diffraction (XRD) patterns were tested on an Ultima IV X-ray Diffractometer. Ultraviolet visible (UV-vis) absorption spectra were tested on a PerkinElmer Lambda 950 spectrophotometer. The steadystate photoluminescence (PL) spectra, time-resolution photoluminescence (TRPL) spectra (Excitation wavelength: 375 nm)were performed on a transient steady-state fluorescence spectrometer (Edinburgh FLS9). Photocurrent-voltage (J-V)curves of the PSCs (active area: 0.1 cm2) were studied by a source meter (2400, Keithley). The solar cells were tested with a sweep speed of 0.05 V/s under AM 1.5 G condition of 100 mW/cm2illumination by using a solar simulator (Peccell Technologies PEC-L01).

3. Results and discussion

3.1. Process parameter optimization

Surface morphology of the MAPbI3films with different heating temperature and processing time were studied to determine the optimal process parameters of the MAI healing method. The statistics of the grain size in Figs. 2 and 3 showed that increasing the heating temperature and processing time could promote the growth of the perovskite crystal. However, when the heating temperature increased to 150 °C, pinholes started to appear on a large scale and the grain boundary was difficult to distinguish. Moreover,the grain size no longer increased obviously when we extended the processing time from 15 to 18 min. Therefore, the heating temperature and processing time of the MAI healing method were set as 130 °C and 15 min respectively to realize the fabrication of perovskite films in subsequent research.

3.2. Film quality comparison

Fig. 2. (Color online) SEM images of MAPbI3 perovskite treated by MAI healing method with heating temperature of (a) 100 °C, (b) 120 °C,(c) 130 °C and (d) 150 °C. (e) Comparison of grain size of the MAPbI3 film with different heating temperature. (The processing time was set as 10 min.)

Fig. 3. (Color online) SEM images of MAPbI3 perovskite treated by MAI healing method with processing time of (a) 12 min (low magnification), (b)12 min (high magnification), (c) 15 min and (d) 18 min. (e) Comparison of grain size of the MAPbI3 film with different processing time. (The heating temperature was set as 100 °C.)

To investigate the effect on the surface morphology of the atmosphere healing, the MAPbI3films were fabricated with the same heating temperature and processing time,which were both treated at 130 °C for 15 min for conventional post-annealing treatment and the MAI healing method.Fig. 4 showed the surface morphology and corresponding statistics of grain size and grain boundary density of the MAPbI3films fabricated by different treatments. As shown in Figs.4(a) and 4(b), through the comparison of the SEM images,the MAI healing method is supposed to contribute to the crystal growth, while the film could still keep dense and smooth.Besides, as shown in Figs. 4(c)-4(f), the AFM images also demonstrated that larger grain was realized through the MAI healing method. As shown in Fig. 4(g), according to the results of the corresponding statistics, the average grain size treated by MAI healing method was 880 nm, nearly double that of the film treated by conventional post-annealing. The statistics of grain boundary density in Fig. 4(h) also decreased obviously after the healing process. The much larger grain size should be associated with the atmosphere rich with MAI in the enclosed space, which could repress the decomposition of MAPbI3and the production of PbI2around the grain boundaries. Thus, the elimination of the PbI2obstacle around the grains was supposed to contribute to the perovskite crystal growth.

Fig. 5 showed the XRD patterns and their partial enlarged views of the MAPbI3films prepared by different heat treatments. The dominant diffraction peaks of the MAPbI3perovskite were at 2θ= 14.11°, 28.43°, and 31.87°, corresponding to (110) (220) (310) crystal planes respectively, which were identical between these two different treatments.However, the diffraction peak at 2θ= 12.36° of the perovskite film, treated by conventional post-annealing, indicated that a small amount of PbI2had generated during the heat treatment, while there were no obvious diffraction peaks of PbI2in the XRD pattern of the perovskite treated by MAI healing method. These results demonstrated that the MAI healing method is an effective way to repress the production of PbI2during the heat treatment process of MAPbI3films.

Fig. 4. (Color online) SEM and AFM images of MAPbI3 perovskite films treated by (a, c, d) post-annealing, (b, e, f) MAI healing. (g) Statistics of the average grain size and (h) grain boundary density of MAPbI3 perovskite film with different treatment.

Fig. 5. (Color online) Phase structure of MAPbI3 films with different treatments. (a) XRD patterns. (b) Partial enlarged views.

To further study the absorption and photoluminescence properties, we tested the UV-vis, PL and TRPL spectra of the MAPbI3films prepared by different treatment. The absorbance of the MAPbI3films treated by MAI healing was higher than that of the film treated by post-annealing due to reduced grain boundaries, where the light scattering usually happens, as shown in Fig. 6(a). Besides, we exposed a sample treated by MAI healing under ambient condition (relative humidity: 30%-40%) for 10 days. The absorption intensity kept steady during this testing period, indicating the perovskite film possessed an excellent stability, as shown in Fig. 6(b).

Fig. 6. (Color online) Optical and electrical spectra of MAPbI3 films with different treatments. (a) UV-vis spectra of MAPbI3 film with different treatments. (b) UV-vis spectra of MAPbI3 film treated by MAI healing as-prepared and after storage in ambient air (relative humidity: 30%-40%) for 10 days. (c) PL and (d) TRPL spectra of MAPbI3 films with different treatment.

PL spectra were also tested to further investigate the photoluminescence properties. Fig. 6(c) showed that the emission peaks were both located at 770 nm, while the PL intensity of MAPbI3films treated by the MAI healing method was much higher than that of the MAPbI3films treated by conventional post-annealing. Moreover, TRPL measurements in Fig. 6(d) showed that the MAPbI3perovskite film treated by the MAI healing method had a significantly longer average lifetime of 93.5 ns, while that of the MAPbI3perovskite film treated by conventional post-annealing was only 40.8 ns.

3.3. Photovoltaic performance and stability

PSCs in a n-i-p configuration were fabricated to evaluate the photovoltaic performance and stability of the PSCs prepared by post-annealing and MAI healing treatment. The cross section SEM image in Fig. 7(a) demonstrated that a vertical columnar grain layer of perovskite was formed via the MAI healing treatment. Fig. 7(b) showed that the best device treated by the MAI healing method obtained a PCE of 18.89%with an open-circuit voltage (Voc) of 1.05 V, a short-circuit current density (Jsc) of 21.98 mA/cm2, and an FF of 81.6%. The value of FF was much higher than most of the MAPbI3-based PSCs without any additive or post-treatment, indicating that the MAI healing method was an effective way to increase the FF of the PSCs based on the MAPbI3absorber. For comparison, the best device treated by conventional postannealing obtained a PCE of 18.4% with aVocof 1.08 V, aJscof 21.28 mA/cm2, and an FF of 79.9%, respectively. The average statistical results in Table 1 indicated that the MAI healing method could improve the PCE by nearly 1% compared with the devices treated by conventional post-annealing. A lower current loss due to the reduced defects inside the PSCs was supposed to contribute to the evidently enhancedJscand FF.

Furthermore, we recorded the photovoltaic performance of the PSCs treated by post-annealing and the MAI healingmethod as-prepared and after storage under ambient conditions (relative humidity: 30%-40%) for 192 h continuously. As shown in Fig. 7(c), theVocfor both of the devices remained steady, while theJscand FF of the device treated by conventional post-annealing dropped significantly during this period, causing the PCE to decrease to less than 60% of the initial state. In contrast, the device treated by the MAI healing method still retained 90% of its initial PCE after the testing period, showing outstanding humidity stability.

Table 1. Statistic results of the photovoltaic performance of the PSCs prepared with different method.

4. Conclusion

In conclusion, we used a MAI healing method to prevent the decomposition of MAPbI3and the production of PbI2during the heat treatment. Through the healing treatment, a larger grain size over 800 nm was successfully realized.Moreover, the absorption and photoluminescence intensity were both improved obviously in comparison with the conventional post-annealing treatment. Finally, the best PCE of 18.89% with an ultrahigh FF over 80% was realized in an n-i-p configuration via the healing method, while retaining 90% of its initial state after storage in ambient conditions for 192 h continuously. This work indicates that adopting the MAI healing method to repress the volatilization process of the organic component during heat treatment contributes to the films quality, photovoltaic performance and humidity stability of the MAPbI3-based PSCs. The application of this healing method in other perovskite compositions should be studied in the future.

Fig. 7. (Color online) (a) Cross section SEM image of the PSCs treated by MAI healing. (b) J-V curves of the best performing device of MAPbI3 absorber with different treatment. (c) Photovoltaic performance of PSCs with different treatment as-prepared and after storage in ambient air (relative humidity: 30%-40%) for 192 h.

Acknowledgements

This work was fianancially supported by the National Key R&D Program of China (Grant No. 2019YFB1503200). We thank Dr. Ying Hao and Dr. Lu Bai for their kind assistance in UV-vis spectra tests.