Common Phase and Structure Misidentifications in High-Resolution TEM Characterization of Perovskite Materials

High-resolution TEM (HRTEM) is a powerful tool for structure characterization. However, methylammonium lead iodide (MAPbI3) perovskite is highly sensitive to electron beams and easily decomposes into lead iodide (PbI2). Misidentifications, such as PbI2 being incorrectly labeled as perovskite, are widely present in HRTEM characterization and would negatively affect the development of perovskite research field. Here misidentifications in MAPbI3 perovskite are summarized, classified, and corrected based on low-dose imaging and electron diffraction (ED) simulations. Corresponding crystallographic parameters of intrinsic tetragonal MAPbI3 and the confusable hexagonal PbI2 are presented unambiguously. Finally, the method of proper phase identification and some strategies to control the radiation damage in HRTEM are provided. This warning paves the way to avoid future misinterpretations in HRTEM characterization of perovskite and other electron beam-sensitive materials.


Introduction
High-resolution transmission electron microscopy (HRTEM) is a powerful characterization tool and has been extensively and successfully used for analyzing crystal structures on an atomic scale [1][2][3][4]. Recently, halide perovskites have achieved substantial success in various optoelectronic devices owing to their solution-based growth and their remarkable physical properties [5][6][7][8][9]. However, the extensively studied methylammonium lead iodide (MAPbI 3 ) perovskite is very sensitive to electron beam irradiation. Figure 1 shows the MAPbI 3 degradation process under electron beam irradiation. Tetragonal perovskite decomposes into hexagonal lead iodide (PbI 2 ) by the dissociation of methylamine and hydrogen iodide molecules.
Comparing distances and angles between crystal planes to identify the material phase, it is easy to ignore the missing crystal planes, which leads to false identification results, such as labelling lead iodide as perovskite. In a proper phase identification, HRTEM images alone cannot be used for phase identification, but always have to be supported by other measurement results such as diffractograms, simulated ED, nanodiffractions, or XRD specimen data [1]. As a rough estimation, MAPbI 3 decomposes into PbI 2 at a total dose irradiation > 150 eÅ −2 [10,11]. However, the value of the electron dose in normal HRTEM is around 800-2000 eÅ −2 s −1 , so the experimental total dose would have been much higher than the critical dose of MAPbI 3 perovskite. Meanwhile, distances and angles between crystal planes in the decomposition product PbI 2 are very similar to MAPbI 3 , such that PbI 2 can easily be misidentified as perovskite. A typical feature of such misidentification is missing crystal planes in the results. Therefore, the absence of crystal planes serves as an indicator that the material is not in the perovskite, but another structural phase. Such misidentifications were widely ignored in literature studies [9,[12][13][14][15][16][17][18][19][20][21][22][23][24][25][26][27][28][29][30][31].  Figure 3B) [33]. Moreover, the missing crystal planes have also been observed in other characterization techniques based on the Bragg's law, such as selected area electron diffraction (SEAD) ( Figure 3C) [34] and X-ray diffraction (XRD) ( Figure 3D) [10,[35][36][37]. Finally, the low-order crystal plane (d(112) = 4.4 Å) has longer interplanar spacing than the high-order crystal plane (d(224) = 2.2 Å). Thus, the missing of low-order crystal planes should not result from the poor accuracy in the data interpretation. Although the mistakes are taken seriously [10,11,35], an urgently needed, complete summary and correction of that issue has not been published yet.

Results and Discussion
Here misidentifications in MAPbI 3 perovskite are summarized, classified and corrected. Figure 4 shows simulated ED patterns of MAPbI 3 and PbI 2 along different axis zones. Figure 4A is the ED pattern of MAPbI 3 along [110] zone axis. (110), (002) crystal planes are present in intrinsic tetragonal perovskite, but they are missing in a number of HRTEM characterizations [12][13][14][15]. Figure 4B shows the simulated ED patterns of PbI 2 along [441] zone axis. (014), (104) crystal planes of PbI 2 have the similar distances and angles between crystal planes compared to (220), (004) crystal planes of MAPbI 3 . In the actual experiment, MAPBI 3 would decompose into PBI 2 and the missing crystal planes would be oftentimes ignored, leading to aforementioned mischaracterizations. Similarly, Figure Figure 4, the missing low-order crystal planes of MAPbI 3 are marked in red. The similar crystal angles and interplanar spacings make them to be easily confused. The MAPbI 3 perovskite is extremely beam-sensitive with a critical dose around a few hundred electrons per square angstrom. For most measurements, and in particular HRTEM, keeping the dose below the critical value is extremely challenging [38][39][40]. However, there are some ways to reduce the radiation damage in HRTEM, which might be helpful in obtaining the intrinsic structure of perovskite materials. Low-dose imaging techniques are effective approaches for electron beam-sensitive materials [41]. Zhang et al. developed a program to achieve a direct, one-step alignment of the zone axis and obtained HRTEM images of electron beam-sensitive materials with a dose of 6-12 e Å −2 [42]. Carlino reported the in-line holography in TEM for the study of radiation-sensitive materials providing high-contrast holograms within 1-2 eÅ −2 s −1 [43]. The intrinsic structure of MAPbI 3 has been captured by the Direct-detection electron-counting (DDEC) cameras at doses below 3 e Å −2 [33,44]. Cooling the specimen also poses a feasible way to reduce damage. Using Cryo-TEM, HRTEM images of MAPbI 3 has been captured [11,33]. Lastly, encapsulation of specimen with a thin, stable, and continuous layer reduces radiation damage [40]. In any case, the total electron dose should be below the critical dose of MAPbI 3 , otherwise substantial damage occurs during TEM characterizations.

Conclusions
The MAPbI 3 perovskite is very sensitive to electron beam irradiation and easily decomposes into PbI 2 . HRTEM characterizations performed above the critical dose can easily lead to misinterpretations, falsely identifying decomposition products as the actual perovskite. An indicator of such confusable decomposition products is missing crystal planes when comparing the expected theoretical diffractograms to the actual experimental data of MAPbI 3 . The summarized and analyzed misidentifications in HRTEM characterization of MAPbI 3 may be helpful for researchers to avoid such misidentifications in perovskite research. Our findings show that it is unreliable to identify material phases only by measur-ing distances and angles between crystal planes. We need to take diffractograms, simulated ED, nanodiffractions, or XRD data into consideration in order to ensure complete crystal planes. We also provide a method of proper phase identification and some strategies to reduce the radiation damage in HRTEM.
Funding: This research received no external funding.
Data Availability Statement: All data are available from the corresponding author upon reasonable request.