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

Abstract

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.

1. 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 (MAPbI₃) perovskite is very sensitive to electron beam irradiation. Figure 1 shows the MAPbI₃ degradation process under electron beam irradiation. Tetragonal perovskite decomposes into hexagonal lead iodide (PbI₂) 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, MAPbI₃ decomposes into PbI₂ 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 MAPbI₃ perovskite. Meanwhile, distances and angles between crystal planes in the decomposition product PbI₂ are very similar to MAPbI₃, such that PbI₂ 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 2 shows typical HRTEM images with missing crystal planes. (1$\overline{1}$0), (002) planes are missing along the [110] zone axis (Figure 2A) [21], (020) plane is missing along the [101] zone axis (Figure 2B) [27], (11$\overline{2}$), (112) planes are missing along the [$\overline{2}$01] zone axis (Figure 2C) [13], and (002) plane is missing along the [$\overline{1}$20] zone axis (Figure 2D) [28]. However, missing crystal planes have been observed in low-dose imaging. Thus far, Song et al. have observed the missing (1$\overline{1}$0), (002) planes with 0.63 nm along [110] zone axis at total doses of 1.5 eÅ⁻² (Figure 3A) [32]. Similarly, Zhu et al. have observed the missing (1$\overline{1}$2), (112) planes with 0.44 nm along the [$\overline{2}$01] zone axis by Cryo-TEM at a low electron dose of about 3 e Å^-2 (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.

2. Results and Discussion

Here misidentifications in MAPbI₃ perovskite are summarized, classified and corrected. Figure 4 shows simulated ED patterns of MAPbI₃ and PbI₂ along different axis zones. Figure 4A is the ED pattern of MAPbI₃ along [110] zone axis. ($\overline{1}$10), (002) crystal planes are present in intrinsic tetragonal perovskite, but they are missing in a number of HRTEM characterizations [12,13,14,15]. Figure 4B shows the simulated ED patterns of PbI₂ along [4$\overline{4}$1] zone axis. (014), ($\overline{1}$04) crystal planes of PbI₂ have the similar distances and angles between crystal planes compared to ($\overline{2}2$0), (004) crystal planes of MAPbI₃. In the actual experiment, MAPBI₃ would decompose into PBI₂ and the missing crystal planes would be oftentimes ignored, leading to aforementioned mischaracterizations. Similarly, Figure 4C–H show ED patterns of MAPbI₃ along [101] and PbI₂ along [8 10 1], MAPbI₃ along [$\overline{2}$01] and PbI₂ along [8$\overline{8}$1], MAPbI₃ along [$\overline{1}$20] and PbI₂ along [$\overline{4}$11], respectively. The missing crystal planes in HRTEM characterizations are indicated by red circles in Figure 4.

Misidentifications exist in the following fields: phase identification and structure determination [16], grain, nanowire, and microwire orientation [9,14,17,18,19], morphology analysis and shape control of nanocrystals [20], growth direction of perovskite materials [21], degradation process and kinetics of perovskite [22], phase transition research in perovskite [26], and ion migration characterization in photoelectric devices [27]. Misidentifications not only occur in single component materials, but also in heterostructures, such as lattice matching and kinetic study in epitaxial growth of perovskite films on 2D materials [25], PbS quantum dots in perovskite materials [12,13], and lattice-anchoring stabilized perovskite research [14]. HRTEM is an auxiliary tool to identify the phase of perovskite materials and misinterpretations would not influence their device performances and conclusions. However, in some cases, such as the growth direction, orientation, and lattice matching of perovskite materials, misinterpretations in HRTEM would result in wrong conclusions. To make the comparisons and corrections clearer,

Table 1. Detailed crystallographic parameters of MAPbI₃ and PbI₂ along different zone axes corresponding to Figure 4, the missing low-order crystal planes of MAPbI₃ are marked in red. The similar crystal angles and interplanar spacings make them to be easily confused.

Material and Zone Axis	Characteristic Crystal Planes	Interplanar Spacing	Interplanar Angle	References
MAPbI3 [110]	($\overline{1}$10), ($\overline{2}$20) (002), (004)	d($\overline{1}$10) = 6.2Å. d($\overline{2}$20) = 3.1Å. d(002) = 6.3Å. d(004) = 3.2Å.	<($\overline{1}$10), (002)> =<($\overline{2}$20), (004)> = 90.0°	[9,15,16,17,18,19,20,21,22,23,24,25]
PbI2 [4$\overline{4}$1]	(014) ($\overline{1}$04)	d(014) = 3.2Å .d($\overline{1}$04) = 3.2Å.	<(014), ($\overline{1}$04)> = 87.5°	_______
MAPbI3 [101]	(020), (040) ($\overline{2}$02)	d(020) = 4.4Å. d(040) = 2.2Å. d($\overline{2}$02) = 3.6Å.	<(020), ($\overline{2}$02)> =<(040), ($\overline{2}$02)> =90.0°	[26,27]
PbI2 [8 10 1]	($\overline{1}$08) (1$\overline{1}$2)	d($\overline{1}$08) = 2.2Å. d(1$\overline{1}$2) = 3.7Å.	<($\overline{1}$08), (1$\overline{1}$2)> = 88.0°	_______
MAPbI3 [$\overline{2}$01]	(1$\overline{1}$2), (2$\overline{2}$4) (112), (224)	d(1$\overline{1}$2) = 4.4Å. d(2$\overline{2}$4) = 2.2Å. d(112) = 4.4Å. d(224) = 2.2Å.	<(1$\overline{1}$2), (112)> =<(2$\overline{2}$4), (224)> = 60.5°	[12,13,14,18]
PbI2 [8$\overline{8}$1]	(0$\overline{1}$ $\overline{8}$) (10$\overline{8}$)	d(0$\overline{1}$ $\overline{8}$) = 2.2Å. d(10$\overline{8}$) = 2.2Å.	<(0$\overline{1}\overline{8}$), (10$\overline{8}$)> = 57.2°	_______
MAPbI3 [$\overline{1}$20]	(002), (004) (211)	d(002) = 6.3Å. d(004) = 3.2Å. d(211)= 3.8Å.	<(002), (211)> =<(004), (211)> = 72.7°	[28,29,30,31]
PbI2 [$\overline{4}$11]	(104) (01$\overline{1}$)	d(104) = 3.2Å. d(01$\overline{1}$) = 3.9Å.	<(104), (01$\overline{1}$)> = 73.7°	_______

shows the detailed parameters of the intrinsic MAPbI₃ and confusable PbI₂ along different zone axis, where missing crystal planes are marked in red. Remarkably, the [110] and [001] zone axes are equivalent in tetragonal MAPbI₃ perovskite.

The MAPbI₃ 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⁻²s⁻¹ [43]. The intrinsic structure of MAPbI₃ 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 MAPbI₃ 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 MAPbI₃, otherwise substantial damage occurs during TEM characterizations.

3. Conclusions

The MAPbI₃ perovskite is very sensitive to electron beam irradiation and easily decomposes into PbI₂. 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 MAPbI₃. The summarized and analyzed misidentifications in HRTEM characterization of MAPbI₃ 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 measuring 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.

4. Methods

Corresponding crystal structures cif files were downloaded from Crystallography Open Database (COD) website. COD IDs of MAPbI₃ and PbI₂ are 4124388 and 9009141 respectively [45,46]. MAPbI₃ is I4/mcm space group with tetragonal structure, cell parameters: a = b = 8.839Å, c = 12.695Å; α = β = γ = 90°. PbI₂ is P-3m1 space group with hexagonal structure, cell parameters: a = b = 4.555Å, c = 20.937Å; α = β = 90°, γ = 120°. The electron diffraction (ED) simulations of MAPbI₃ and PbI₂ were obtained using CrystalMaker Software. The interplanar spacings and interplanar angles can be calculated from the cell parameters. During the process of phase identification, we also tried other polytypes of PbI₂, but the results did not match well.