Ethylammonium Lead Iodide Formation in MAPbI 3 Precursor Solutions by DMF Decomposition and Organic Cation Exchange Reaction

: Extra peaks have constantly been observed in the X ‐ ray diffraction measurement for the CH 3 NH 3 PbI 3 film. Such mysteries have now been uncovered in this paper, in which powder X ‐ ray diffraction, in situ X ‐ ray diffraction, and scanning electron microscopy measurements were conducted, and these peaks were attributed to the ethylammonium lead iodide (CH 3 CH 2 NH 3 PbI 3 /EAPbI 3 ). It was found that the formation of EAPbI 3 was triggered by the breakdown of N, N ‐ dimethylformamide (DMF), which was adopted as the solvent in the preparation of the precursor solutions. EAPbI 3 was generated by the organic cation exchange reaction in the subsequent annealing process. A simple solution for this problem is proposed in this paper as well, which would hopefully help the community to eradicate this impurity.


Introduction
The power conversion efficiency of CH3NH3PbI3(MAPbI3) and HC(NH2)2PbI3 (FAPbI3) perovskite solar cells has soared from 3.8% to 25.2% in the past decade [1,2], which now rivals that of other technologies such as silicon or cadmium telluride (CdTe) solar cells [3,4]. While such record performance seems to be inspiring for renewable energy development, the device lifetime, especially under moisture environment, remains at a low level, which prevents its commercialization. The devices usually degrade within a day unless an extra encapsulation layer is adopted [5][6][7], and even with the encapsulation layer the longest lifetime reported is still less than 4000 hours, which is far from enough for commercialization [8][9][10].
Although various strategies have been proposed, such as doping at A, B, and X sites of the ABX3 perovskite structure [11][12][13], grain boundary passivation [7,14,15], or replacing hydrophilic hole transport layer [16], none has resolved the stability issue successfully. Therefore, uncovering the scientific mechanism of the decomposition is imperative to address this issue. Despite many mechanisms having been reported before [17][18][19][20][21][22], they were often inconsistent and self-contradictory. Hopefully, some of the conflicts can be neutralized by our recently published work, where we revealed the importance of hydrogen bonding for the perovskite structural stability and discovered the true cause of the irreversible degradation [10]. However, there remains some other substance detected during the synthesis of the CH3NH3PbI3 thin film, which has yet to be identified by the perovskite solar cell community, despite the fact that a similar phenomenon was also reported before elsewhere [23,24]. As the impurity could influence the device performance, and thus it calls for a systematic investigation.
It is, therefore, the purpose of the current article to systematically investigate this substance, discovered in the perovskite thin film during the synthesis process, using powder X-ray diffraction (PXRD), in situ X-ray diffraction, and scanning electron microscopy (SEM). We prepared various PbI2 and CH3NH3I precursor solutions to investigate the influence of heating temperature and time on the CH3NH3PbI3 thin film formation. It was discovered that N, N-dimethylformamide (DMF) could decompose at a higher temperature, which enabled the organic cation exchange reaction and led to the formation of CH3CH2NH3I (EAI). Subsequently, EAI could react with PbI2 and form EAPbI3, which, as confirmed by the PXRD results, was identified to be the impurity presented in the CH3NH3PbI3 perovskite film. This discovery is expected to help the community, to eliminate the EAPbI3 impurities and to enhance the CH3NH3PbI3 perovskite film morphology, thereby leading to a better device performance.
XRD, instead of chemical characterization methods such as FTIR or NMR, was adopted for material characterization, as XRD is more powerful for material structure and impurity analysis. In situ XRD measurement was conducted by a Bruker D8 ADVANCE diffractometer in an enclosed chamber. The thin film sample was placed in the sample stage of the chamber. Desiccant was also adopted, which has a composition of 98% CaSO4 and 2% CoCl2 (purchased from McMaster chemical store, Hamilton, ON, Canada). The measurement range was 24°-26.5° due to the long time required to scan the whole two theta range. This measurement lasted for 3 days. Powder XRD was conducted by Bruker D8 DISCOVER diffractometer (Bruker Corporation, Billerica, Massachusetts, USA) with a range of 10° to 76°. A diagram of the XRD chamber is shown in Figure 1

Results & Discussion
To identify the influence of the dissolving temperature and the stirring time on the CH3NH3PbI3 perovskite formation, both powder and in situ XRD measurements were adopted in a methanol and water environment. As circled in Figure 2a, CH3CH2NH3PbI3 (EAPbI3) was detected when the precursor solutions were heated at 120 °C for 2 h; while such structure disappeared when the heating temperature and time were reduced to 90 °C and 45 min. These extra peaks exactly matched the peak of EAPbI3 from the database. The peaks for EAPbI3 were also observed in the in situ measurement, where a more controlled observation was conducted. As shown in Figure 2b, the peaks at 24.8° and 25.8° were attributed to EAPbI3. This proved that EAPbI3 was generated gradually as time passing by, and we speculated that the solvent (DMF) could be the cause of the EAPbI3 formation, as only DMF contains the structure of ethylammonium. In the subsequent analysis, we found the supporting evidence from the literature, where the DMF was proven to be unstable at a temperature higher than 100 °C [25,26]. Therefore, DMF could be decomposed at higher heating temperatures, starting from the breakdown of the bond between the nitrogen atom and the carboxyl group, generating ethylammonium ion intermediate, which reacted with CH3NH3I by an organic cation exchange reaction. Eventually, EAI was generated. In the following annealing process, EAI reacted with PbI2 to form EAPbI3. Only weak peaks were observed for the first few hours in the XRD diagram, indicating a slow decomposition rate of DMF under heating. A schematic illustration of this process is shown in Figure 3. It is thus clear now, that the substance detected in the CH3NH3PbI3 film is indeed EAPbI3, which not only reduces the film purity, but also affects the film morphology, as can be confirmed in Figure  4a, b. When there was no EAPbI3 present in the film, large CH3NH3PbI3 grains could be observed with a size of several microns, as shown in Figure 4; while such morphology changed greatly when EAPbI3 was generated. In Figure 4b, the large grain was damaged, and voids were also generated, which proved that EAPbI3 is detrimental for the CH3NH3PbI3 film growth and device performance. To summarize the results, the formation of EAPbI3 in the CH3NH3PbI3 film is triggered by the degradation of DMF solvent, which is unstable at a higher temperature and decomposes into EA + intermediate. This is followed by an organic cation exchange reaction, where EA + switches position with MA + , resulting in the formation of EAI, which subsequently reacts with PbI2 in the following annealing process. Although other methods could be adopted to avoid the formation of impurities, including chemical vapor deposition or physical vapor deposition [27,28], a much easier method has been proposed here by reducing the heating time and temperature during the preparation of the precursor solution.

Conclusion
To conclude the manuscript, the extra peaks detected by many others in the CH3NH3PbI3 film have been successfully identified, and attributed to EAPbI3, a perovskite that has a similar structure to CH3NH3PbI3, which is why it was so difficult to identify. The formation of EAPbI3 in the CH3NH3PbI3 film was found to be detrimental to the grain growth, which could lower the power conversion efficiency. The EAPbI3 could easily be eradicated by reducing the heating time and temperature when preparing the precursor solution. Hopefully, our discovery could be of immediate help to the community in the near future.
Author Contributions: R.T.W. conducted the experiments and wrote the manuscript. E.E.L. and L.W.Y. completed the diagram plot and proofreading. A.F.X. assisted with the SEM and XRD measurements. J.Y.Z. assisted with the materials synthesis. G.X. provided the funding and overall supervision. All authors have read and agreed to the published version of the manuscript.
Funding: This work was financially supported by the Natural Sciences and Engineering Research Council of Canada, grant number 10546964.