Nanoscale characterization of halide perovskite phase stability with scanning electron microscopy

Abstract

In light of the high potential of halide perovskites for solar energy applications, continued focus has been placed on long-term stability of these materials. Significant work has focused on the phase instabilities of both cesium- and formamidinium-based halide perovskites since the photoactive perovskite phases of both materials have intrinsic and extrinsic instabilities. Scanning electron microscopy (SEM) is routinely used to investigate the morphology and structure of halide perovskite materials. In this study, SEM techniques are extended to an investigation of material phase instability. This allows for high-resolution observations of the change from the perovskite to non-perovskite phase with insight into phase nucleation and growth. When combining secondary electron imaging with backscatter electron imaging, we show the facile growth of the non-perovskite phase across grain boundaries and a decrease in crystallinity after the phase transition. This work yields a closer look at this important phase transition and provides a roadmap for imaging materials with similar property differences between crystal phases.

---

Jeffrey A. Christians

Jeffrey A. Christians is an Associate Professor of Engineering at Hope College where he teaches, directs the Hope College Green Revolving Fund, and leads an undergraduate research team focused on halide perovskites. His research has focused on material stability questions for halide perovskites with a significant focus on phase changes and phase stability, among other topics. Prior to joining Hope, Jeff was an EERE postdoctoral fellow at the National Renewable Energy Laboratory (2015–2018) under Dr Joseph Luther and received his PhD in Chemical and Biomolecular Engineering from the University of Notre Dame in 2015 while studying under Dr Prashant Kamat.

Introduction

Characterization of materials at the nanoscale has revolutionized our understanding and control of materials. This is true of halide perovskites as well, where high spatial resolution characterization has been key for both material development and device work. In particular, scanning electron microscopy (SEM) is one of the main tools for high spatial resolution imaging of thin films, like the halide perovskites presented in this paper, because of its ease of sample preparation, resolution, and the variety of detectors/techniques which can provide contrast. SEM has been ubiquitous in investigations of halide perovskites, revealing film morphology,^1,2 crystallinity and grain structure,^3,4 defect chemistry and composition,^5,6 and more. This deeper understanding has led to improved material control allowing halide perovskite semiconductors to begin to make an impact in photovoltaics and many other optoelectronic applications.

In this work we show SEM imaging of halide perovskites undergoing phase transitions from their perovskite phase(s) (there are several different perovskite phases with slight distortions termed α, β, and γ) and into their non-perovskite phase(s) (which are typically called the δ-phase). This phase change has been widely discussed and studied in halide perovskites due to its impact on the material's optoelectronic properties.^7–15 While δ-CsPbI₃ is thermodynamically favored at room temperature,¹¹ it is possible, and often desirable, to kinetically trap materials into the γ-phase for long periods of time.¹² This phase transition can generally be understood in light of the Goldschmidt tolerance factor, an indicator of crystal structure stability that describes the ratio of the ionic size of the A-site cation relative to the “hole” formed by the corner-sharing BX₆²⁻ octahedra of the perovskite phase.¹⁶ During this phase transition, the ABX₃ crystal structure changes from a perovskite crystal structure, where adjacent BX₆²⁻ octahedra are corner-sharing, to a non-perovskite phase, where these BX₆²⁻ octahedra become edge-sharing (CsPbI₃, Fig. 1A) or face-sharing (as in formamidinium lead iodide).¹⁷ The nucleation of non-perovskite germs dominates the kinetics of the phase transition from the γ-phase into the δ-phase in CsPbI₃.^12,18

Fig. 1 Contrast between γ-CsPbI₃ and δ-CsPbI₃. (A) Crystal structures of γ-CsPbI₃ and δ-CsPbI₃. (B) Schematic showing general relative secondary electron yield of γ- and δ-CsPbI₃ with electron beam primary energy. Shaded areas show where positive (orange) and negative (purple) charging can occur. (C) Differences in charging between the γ- and δ-CsPbI₃ regions of the film. A δ-CsPbI₃ phase region is imaged in a γ-CsPbI₃ thin film at E_PE = 3 kV without (D) and with (E) significant positive charging effects, and at E_PE = 5 kV without (F) and with (G) significant negative charging effects.

Conditions that influence the speed of the phase transition have been studied, such as humidity,¹¹ crystal structure strain,^14,19 and surface chemistry and defects,^12,13,18 which have been summarized in recent review articles.^20–22 Together, this work has played a key role in helping researchers understand and control the crystal phases of these materials. Despite this, examples of high spatial resolution imaging of the phase transition are lacking, outside of nanocrystals.²³ In this work, we present methods for achieving quality imaging and good contrast of the γ to δ phase transition in a typical SEM, along with a discussion of how contrast arises between these two crystal phases under different electron beam conditions. We believe that this work will aid future investigations of halide perovskite phase transitions.

Results and discussion

To understand how contrast arises between the perovskite and δ phases CsPbI₃ was used as a model system. CsPbI₃ is a useful model because it is relatively insensitive to electron beam damage²⁴ and has a facile perovskite to δ-phase transition.¹¹ In an SEM a high energy electron beam is raster scanned across a sample surface and images are formed by using various detectors. The most common detector is a secondary electron detector, which, by convention, typically measures emitted electrons with energies below 50 eV.²⁵ Secondary electrons are emitted from a specimen when a collision with the high energy electron beam causes electrons to reach the sample's surface with an energy that is higher than the material's work function. The secondary electron yield is the ratio of emitted secondary electrons to incident electrons from the electron beam, and it is proportional to image brightness. Differences in the secondary electron yield can be caused by differences in atomic number/electron density, surface morphology, conduction electron concentration, and charge carrier mobility.^26–30 These last two are of particular importance for halide perovskite phase changes due to the band gap and optoelectronic property differences between the perovskite and δ phases.

In conductive materials such as metals and, to a lesser extent, γ-CsPbI₃, secondary electrons tend to lose energy as they travel due to scattering interactions with conduction electrons.²⁵ Thus, conductive materials have a low escape/detection probability. Conversely, insulators tend to have a higher secondary electron yield because there is a lower probability of emitted secondary electrons scattering off conduction electrons. Given this, we anticipate that the wide bandgap δ-CsPbI₃ (E_g = 2.8 eV) will appear brighter than the low bandgap γ-CsPbI₃ (E_g = 1.7 eV) perovskite phase,³¹ even though they have the same chemical composition. As shown schematically in Fig. 1B, secondary electron yield also varies with beam primary energy, E_PE. Under low voltage (E₁ < E_PE < E₂), the secondary electron yield is greater than one. In this case a material could experience positive charging if it is not sufficiently conductive for the imaging conditions used (Fig. 1C). On the other hand, under higher energy (E_PE > E₂) the secondary electron yield decreases below unity and a material could experience negative charging if it is not sufficiently conductive (note, this will also occur with E_PE < E₁, but SEM imaging is rarely done with voltages this low). This general behavior is seen across materials.^32,33

To observe these effects, a CsPbI₃ perovskite thin film on a fluorine-doped tin oxide (FTO) conductive glass substrate was allowed to partially change from the γ-phase to the δ-phase. The phase transition was arrested when the film was removed from ambient air and placed under vacuum. The film was then imaged in an SEM using the secondary electron detector, initially using an E_PE set to 3 kV (Fig. 1D and E). When the sample was not experiencing any charging effects, the newly converted δ-CsPbI₃ region of the film was brighter than the original γ-CsPbI₃ region of the film (Fig. 1D). We attribute this contrast to the higher secondary electron scattering of the γ-CsPbI₃, which decreases secondary electron yield. The contrast was confirmed by comparing color photos taken inside the SEM with the secondary electron micrographs (Fig. S1†). X-ray diffraction imaging taken of the films before and after the phase transition confirms the assignment of the structural phases (Fig. S2†).

Under higher current conditions the contrast will reverse, and the δ-CsPbI₃ region of the film will appear darker (Fig. 1E). This contrast reversal can be understood by taking two points into consideration. First, there is a higher degree of charging in δ-CsPbI₃ than in γ-CsPbI₃ because the δ-phase has lower carrier mobility.³⁴ Second, positive charging decreases the secondary electron yield/brightness since secondary electrons require additional kinetic energy to escape into the vacuum due to electrostatic attraction.³⁵ Similar contrast reversal has been previously observed in metal-insulator systems and in samples where charging can occur.^26,36

We confirmed our contrast reversal hypothesis in two ways. First, under the same electron beam conditions we can cause the contrast to flip when changing the raster scan of the electron beam from a long dwell time single scan, where the sample undergoes charging, to a short dwell time scan, where multiple frames are integrated to form the final image and charging does not occur (Fig. S3†). Second, when a thin layer of gold was deposited on top of the film, the contrast reversal is still possible but required much higher beam currents (Fig. S4†).

Above E_PE ∼5 kV we observed the same contrast as was seen at lower voltages and low beam current, where the δ-CsPbI₃ region of the film appears brighter than the perovskite region of the film. Fig. 1F shows an image acquired with E_PE set to 5 kV, but this general contrast holds at higher beam voltages as well (Fig. S5†). As the incident current is increased, negative charge deposited by the electron beam will repel newly generated secondary electrons out of the film, increasing secondary electron yield.^30,35 Negative charging effects will be more pronounced in insulating materials, like δ-CsPbI₃, due to low charge carrier mobility. The high energy electron beam can also shift/deflect because of repulsive electrostatic forces generated by negative charging, degrading image quality as seen in Fig. 1G, but not resulting in contrast reversal.

As demonstrated, this understanding of contrast development between δ-CsPbI₃ and γ-CsPbI₃ in SEM imaging can be used to obtain high resolution images of both phases and the crystal phase boundary. Beyond secondary electron images, we used energy dispersive X-ray spectroscopy (EDS) and backscatter electron detection to provide additional insight into crystal growth and nucleation. To explore these other detection methods, another CsPbI₃ thin film on FTO glass was observed during the initial stages of the phase change. Most δ-CsPbI₃ regions were compositionally homogeneous with the γ-CsPbI₃ regions, at least within instrumental limits, but other δ-CsPbI₃ regions have a clearly visible object or defect located near their center. For example, Fig. 2 presents secondary electron images, backscatter electron images, and carbon EDS spectra of two separate locations in the same film. In the first δ-CsPbI₃ region (Fig. 2A–C) we saw a high carbon concentration in the EDS map centered at an area of the film that has a different surface texture, indicating some type of foreign inclusion in the film. The other elements detected by EDS appeared homogenous (Fig. S6†). This area also shows up as a dark region when imaging using the backscatter electron detector because of its lower average atomic number (Fig. 2B). The second δ-CsPbI₃ region shown (Fig. 2D–F) does not have the same large defect or inclusion near its center and shows a homogenous EDS signal across the major elements (Fig. S7†). EDS shows no discernible compositional difference, outside of contamination/defects, which further proves that contrast arises in the secondary electron images as a result of the phase difference, and not the film composition.

Fig. 2 Compositional analysis and δ-CsPbI₃ nucleation. Secondary (A) and backscatter electron images (B) (E_PE = 3 kV) with (C) an EDS map (E_PE = 8 kV) of the carbon Kα peak at 0.277 keV. All three images are of the same location. The inset in (A) shows a higher magnification secondary electron image of the film location outlined in red. (D–F) Secondary, backscatter, and EDS images, respectively, were taken using the same conditions at a second location in the film.

To better understand the early nucleation dynamics in a γ-CsPbI₃ thin film, we allowed a freshly fabricated CsPbI₃ film to partially change phase in ambient conditions. After a short time in low humidity ambient conditions (RH < 20%), the film was loaded into a SEM. At this point, most of the film was still γ-CsPbI₃ with small (<20 μm) δ-CsPbI₃ areas throughout. Twelve of these δ-CsPbI₃ inclusions were randomly selected across different areas of the film and imaged using multiple detectors. The results of these images are presented in Fig. S6–S13.† In this survey of locations, we often do not detect any anomalies by EDS, as is shown in Fig. 2C, although these can be seen in at least 3 of the 12 cases; however, even when not detected by EDS, there is often contamination or an inclusion present in both the secondary and backscatter images. EDS has low surface sensitivity, so it is difficult to reach conclusions on the compositional makeup of these spots. In comparison, the dark spots in the backscatter electron images, indicative of lower atomic number, are correlated with small features in the secondary electron images. This suggests that there is contamination in the δ-phase regions. Of the 12 randomly selected δ-CsPbI₃ regions we can detect some sort of anomaly in at least 5 of them, typically most clearly seen in the electron backscatter images. This leads us to believe that contamination is a key early driver of δ-CsPbI₃ nucleation in these low humidity conditions, at least in our CsPbI₃ films.

The high-resolution imaging techniques outlined in this work also provide an unprecedented view of the growth of the δ-phase germs. Secondary electron imaging provides contrast between δ-CsPbI₃ and γ-CsPbI₃, which allows for trivial observation of phase growth, however, characteristics within these regions are difficult to observe. Backscatter electron imaging can provide additional film information through generation of contrast between crystal grains via electron channeling effects, where brightness is dependent on crystal orientation (Fig. 3B).²⁵ This is seen using low voltage and high current conditions. The generation of contrast through electron channeling is confirmed by the observation of contrast changes when the sample was tilted slightly under the electron beam (Fig. S14†). Backscatter electron imaging does not provide the detailed look at grain structure that electron backscatter diffraction imaging does,^4,37 regardless, it can give an improved look at grain structure in the film beyond secondary electron images. One such example of improved observation of grain structure by backscatter imaging is that δ-CsPbI₃ regions are more polycrystalline than the original γ-CsPbI₃ film. This provides evidence that the phase transition occurs via a crystalline–amorphous–crystalline mechanism, as Liu et al. recently outlined.¹⁵

Fig. 3 Crystallinity changes between γ- and δ-CsPbI₃. Expanded views of secondary (E_PE = 3 kV, probe current = 30) (A) and backscatter (E_PE = 3 kV, probe current = 75) (B) electron images shown in Fig. 2D and E, respectively. The dashed red line in (B) was obtained by tracing the phase boundary in (A).

From secondary electron imaging we see that the δ-CsPbI₃ grain growth is relatively isotropic, with somewhat elliptical growth being common. The phase boundary itself is brighter than the surrounding regions of the film, and the film grain boundaries do not appear to significantly slow phase growth. This can be seen in images of the same location at various stages of δ-CsPbI₃ growth (Fig. 4). During the phase change, cracks developed in the newly formed δ-CsPbI₃ regions of the film, which we attributed to strain that results from the slightly higher density of δ-CsPbI₃ relative to γ-CsPbI₃.³¹

Fig. 4 Phase growth of δ-CsPbI₃. Secondary (A and C) and backscatter (B and D) electron images of a single location in a CsPbI₃ film. Images (A) and (B) were taken initially, then the film was exposed to ambient air for a short time before images (C) and (D) were obtained.

Lastly, when the imaging techniques were applied to formamidinium lead iodide (FAPbI₃) thin films we observed similar results to CsPbI₃. As shown in Fig. 5, the δ-FAPbI₃ (non-perovskite phase) and α-FAPbI₃ (perovskite phase) crystal phases show the same contrast features as previously described in the δ-CsPbI₃/γ-CsPbI₃ material system. The regions of the film which are in the δ-phase have higher brightness in typical imaging conditions because of the lower secondary electron scattering probability associated with its wider band gap (2.43 eV for δ-FAPbI₃vs. 1.48 eV for α-FAPbI₃).²¹ While we have seen a contrast reversal in some films under high current imaging conditions (low voltage, high current) where the films become positively charged (Fig. S15†), this was much less apparent than in CsPbI₃. As such, the highest resolution images for FAPbI₃ phase changes were typically achieved under conditions where charging was minimized.

Fig. 5 δ-FAPbI₃ characterization. Secondary electron SEM images (E_PE = 2 kV) of a FAPbI₃ thin film where the brighter regions are δ-FAPbI₃ and the darker regions are α-FAPbI₃, as indicated.

Conclusions

The SEM imaging described provides a high-resolution view of phase changes in halide perovskite thin films, a process which has a large effect on optoelectronic properties and thus is important to halide perovskite applications.^20,21 The innate contrast between the perovskite and non-perovskite phases of CsPbI₃ and FAPbI₃ can be understood by the lower secondary electron scattering probability from the wider band gap non-perovskite phases. Contrast can reverse when the non-perovskite regions of the film experience preferential positive charging, which can be taken advantage of to achieve high-contrast images of these two phases. Furthermore, electron channeling effects in electron backscatter imaging provide a simple method to assess film grain structure.

Using these principles, we observe that unintentional contamination can often be seen at the center of the initial δ-phase inclusions, suggesting that nucleation of the δ-phase begins most readily at these contamination/defect centers. Furthermore, we demonstrate that the δ-phase growth proceeds readily across crystal grain boundaries and appears to proceed through a crystalline–amorphous–crystalline pathway with the newly formed δ-phases showing significantly higher polycrystallinity. Such information will aid researchers as they aim to improve the phase stability of halide perovskites. The methods discussed will allow for continued investigation of these intriguing materials and could be extended to image other material systems with similar property differences between crystal phases.

Experimental methods

Materials

Dimethyl sulfoxide (DMSO), dimethylformamide (DMF), toluene, and methyl acetate were all purchased from Sigma Aldrich. PbI₂ (99.9985%) and CsI (99.98%) were both purchased from Alfa Aesar. Fluorine-doped tin oxide (FTO) glass (TEC-7) and formamidinium iodide (FAI) were purchased from Greatcell Solar Materials. All materials were used as purchased.

CsPbI₃ films

FTO glass (25 mm × 25 mm) was washed by sonication in a soap solution for 10 min followed by an additional 10 min sonication in IPA. Once completed, the slides were dried with nitrogen gas and cleaned by a 10 min UV-ozone treatment. The CsPbI₃ perovskite solution was created by weighing out a 1 : 1.05 mmol ratio of CsI to PbI₂. For each 1 mmol of CsI, 1 mL of a 4 : 1 v/v ratio DMF : DMSO mixed solvent was added and the solution stirred vigorously until all solids dissolved. The cleaned slides and perovskite solution were then transferred into a nitrogen glovebox. A small amount of solution (∼40 μL) was spread on the FTO slides which were then spin cast at 1500 rpm for 45 seconds. With 10 s remaining in this spin procedure, 150–300 μL of methyl acetate antisolvent was quickly dripped onto the film. The shiny light brown films were then annealed on a hot plate for about 2.5–3.5 min at 340 °C until the entire slide was dark brown. The films were removed from the hot plate and placed on the metal floor of the glovebox to cool.

FAPbI₃ films

The FAPbI₃ perovskite solution was created by weighing out a 1.05 : 1 mmol ratio of FAI to PbI₂ and DMF was added to create a 0.7 M solution based on mmol PbI₂. The solution was spread on the FTO slides which were then spin cast at 300 rpm for 3 seconds, 3000 rpm for 10 seconds, and 5000 rpm for 30 seconds. At the beginning of the third step, 100 μL of toluene antisolvent was quickly dripped onto the film. The shiny light brown films were then annealed on a hot plate for 10 min at 170 °C until the entire slide was dark brown. The films were removed from the hot plate and placed on the metal floor of the glovebox to cool.

SEM imaging

Scanning electron micrographs were collected on a JEOL JSM-IT700HR FESEM. Typical images were collected in either secondary electron imaging mode or backscatter electron imaging mode with a 2.3 kV accelerating voltage commonly used. The probe current used for secondary electron imaging was typically lower than that used for backscatter electron imaging, although the absolute probe current was not able to be measured. Perovskite films were imaged on FTO glass substrates. For EDS characterization the primary beam energy was set to 8 kV.

Data availability

The data supporting this article have been included as part of the ESI.† Additional ESI data supporting this article,† as well as the presented data, including original SEM images, will be made available through the Zenodo data repository.

Author contributions

JAC conceived of the project. JAC, RC, IJM, and BT performed experiments and analyzed the data. JAC wrote the first draft of the manuscript. All authors participated in contributing to the text and content of the manuscript.

Conflicts of interest

The authors declare no conflict of interest.

Acknowledgements

Research reported in this publication was supported by the National Science Foundation (DMR-2128632 and under Major Research Instrumentation Award No. 2117655). RC acknowledges support by the Hope College Haight research fund. JC acknowledges support by the Towsley Research Scholars Program grant from the Towsley Foundation of Midland, Michigan.

References

1. N. J. Jeon, J. H. Noh, Y. C. Kim, W. S. Yang, S. Ryu and S. Il Seok, Solvent engineering for high-performance inorganic–organic hybrid perovskite solar cells, Nat. Mater., 2014, 13, 897–903.
2. M. Saliba, T. Matsui, J.-Y. Seo, K. Domanski, J.-P. Correa-Baena, M. K. Nazeeruddin, S. M. Zakeeruddin, W. Tress, A. Abate, A. Hagfeldt and M. Grätzel, Cesium-containing triple cation perovskite solar cells: improved stability, reproducibility and high efficiency, Energy Environ. Sci., 2016, 9, 1989–1997.
3. S. Jariwala, H. Sun, G. W. P. Adhyaksa, A. Lof, L. A. Muscarella, B. Ehrler, E. C. Garnett and D. S. Ginger, Local Crystal Misorientation Influences Non-radiative Recombination in Halide Perovskites, Joule, 2019, 3, 3048–3060.
4. H. Sun, G. W. P. Adhyaksa and E. C. Garnett, The Application of Electron Backscatter Diffraction on Halide Perovskite Materials, Adv. Energy Mater., 2020, 10, 2000364.
5. D. W. DeQuilettes, S. M. Vorpahl, S. D. Stranks, H. Nagaoka, G. E. Eperon, M. E. Ziffer, H. J. Snaith and D. S. Ginger, Impact of microstructure on local carrier lifetime in perovskite solar cells, Science, 2015, 348, 683–686.
6. C. G. Bischak, E. M. Sanehira, J. T. Precht, J. M. Luther and N. S. Ginsberg, Heterogeneous Charge Carrier Dynamics in Organic–Inorganic Hybrid Materials: Nanoscale Lateral and Depth-Dependent Variation of Recombination Rates in Methylammonium Lead Halide Perovskite Thin Films, Nano Lett., 2015, 15, 4799–4807.
7. J. A. Steele, H. Jin, I. Dovgaliuk, R. F. Berger, T. Braeckevelt, H. Yuan, C. Martin, E. Solano, K. Lejaeghere, S. M. J. Rogge, C. Notebaert, W. Vandezande, K. P. F. Janssen, B. Goderis, E. Debroye, Y. K. Wang, Y. Dong, D. Ma, M. Saidaminov, H. Tan, Z. Lu, V. Dyadkin, D. Chernyshov, V. Van Speybroeck, E. H. Sargent, J. Hofkens and M. B. J. Roeffaers, Thermal unequilibrium of strained black CsPbI3 thin films, Science, 2019, 365, 679–684.
8. Y. Wang, M. Ibrahim Dar, L. K. Ono, T. Zhang, M. Kan, Y. Li, L. Zhang, X. Wang, Y. Yang, X. Gao, Y. Qi, M. Grätzel and Y. Zhao, Thermodynamically stabilized b-CsPbI3–based perovskite solar cells with efficiencies >18%, Science, 2019, 365, 591–595.
9. S. Dastidar, D. A. Egger, L. Z. Tan, S. B. Cromer, A. D. Dillon, S. Liu, L. Kronik, A. M. Rappe and A. T. Fafarman, High Chloride Doping Levels Stabilize the Perovskite Phase of Cesium Lead Iodide, Nano Lett., 2016, 16, 3563–3570.
10. B. Wang, N. Novendra and A. Navrotsky, Energetics, Structures, and Phase Transitions of Cubic and Orthorhombic Cesium Lead Iodide (CsPbI3) Polymorphs, J. Am. Chem. Soc., 2019, 141, 14501–14504.
11. S. Dastidar, C. J. Hawley, A. D. Dillon, A. D. Gutierrez-Perez, J. E. Spanier and A. T. Fafarman, Quantitative Phase-Change Thermodynamics and Metastability of Perovskite-Phase Cesium Lead Iodide, J. Phys. Chem. Lett., 2017, 8, 1278–1282.
12. Z. Guo, J. Fu, G. Chen, F. Liu, C. Yu, R. Lu and W. J. Yin, First Domino in the Structural Collapse of Perovskite CsPbI₃ and Its Stabilizing Strategies, Adv. Funct. Mater., 2023, 34(14), 2308246.
13. J. Lin, M. Lai, L. Dou, C. S. Kley, H. Chen, F. Peng, J. Sun, D. Lu, S. A. Hawks, C. Xie, F. Cui, A. P. Alivisatos, D. T. Limmer and P. Yang, Thermochromic halide perovskite solar cells, Nat. Mater., 2018, 17, 261–267.
14. H. K. D. Le, C. K. Lin, J. Jin, Y. Zhang, Z. Lin, A. Vailionis, N. Tamura and P. Yang, Quantification of strain and its impact on the phase stabilization of all-inorganic cesium lead iodide perovskites, Matter, 2023, 6, 2368–2382.
15. J. Liu, X. Hao, M. A. van Huis and Z. Fan, Microscopic mechanism of water-assisted diffusional phase transitions in inorganic metal halide perovskites, J. Chem. Phys., 2024, 161, 174705.
16. W. Travis, E. N. K. Glover, H. Bronstein, D. O. Scanlon and R. G. Palgrave, On the application of the tolerance factor to inorganic and hybrid halide perovskites: A revised system, Chem. Sci., 2016, 7, 4548–4556.
17. C. C. Stoumpos, C. D. Malliakas and M. G. Kanatzidis, Semiconducting Tin and Lead Iodide Perovskites with Organic Cations: Phase Transitions, High Mobilities, and Near-Infrared Photoluminescent Properties, Inorg. Chem., 2013, 52, 9019–9038.
18. Z. R. Wylie, M. Al Katrib, R. Campagna, J. E. Outen, S. Smith, P. Ruffolo, B. Bérenguier, M. Bouttemy, P. Schulz and J. A. Christians, Surface Iodide Defects Control the Kinetics of the CsPbI ₃ Perovskite Phase Transformation, ACS Energy Lett., 2024, 4378–4385.
19. A. Swarnkar, A. R. Marshall, E. M. Sanehira, B. D. Chernomordik, D. T. Moore, J. A. Christians, T. Chakrabarti and J. M. Luther, Quantum dot–induced phase stabilization of α-CsPbI₃ perovskite for high-efficiency photovoltaics, Science, 2016, 354, 92–95.
20. S. Masi, A. F. Gualdrón-Reyes and I. Mora-Seró, Stabilization of Black Perovskite Phase in FAPbI ₃ and CsPbI ₃, ACS Energy Lett., 2020, 5, 1974–1985.
21. Y. An, J. Hidalgo, C. A. Ricardo Perini, A.-F. Castro-Mendez, J. N. Vagott, K. Bairley, S. Wang, X. Li and J.-P. Correa-Baena, Structural Stability of Formamidinium- and Cesium-Based Halide Perovskites, ACS Energy Lett., 2021, 6, 1942–1969.
22. X. Tan, S. Wang, Q. Zhang, H. Liu, W. Li, L. Zhu and H. Chen, Stabilizing CsPbI₃ perovskite for photovoltaic applications, Matter, 2023, 6(3), 691–727.
23. T. Udayabhaskararao, M. Kazes, L. Houben, H. Lin and D. Oron, Nucleation, Growth, and Structural Transformations of Perovskite Nanocrystals, Chem. Mater., 2017, 29, 1302–1308.
24. C. Xiao, Z. Li, H. Guthrey, J. Moseley, Y. Yang, S. Wozny, H. Moutinho, B. To, J. J. Berry, B. Gorman, Y. Yan, K. Zhu and M. Al-jassim, Mechanisms of Electron-Beam-Induced Damage in Perovskite Thin Films Revealed by Cathodoluminescence Spectroscopy, J. Phys. Chem. C, 2015, 119, 26904–26911.
25. L. Reimer, Scanning Electron Microscopy, Springer Berlin Heidelberg, Berlin, Heidelberg, 1998, vol. 45.
26. J. Cazaux, On some contrast reversals in SEM: Application to metal/insulator systems, Ultramicroscopy, 2008, 108, 1645–1652.
27. Y. Liao, Practical Electron Microscopy and Database, 2006.
28. M. Ohring, in Materials Science of Thin Films, Elsevier, 2002, pp. 559–640.
29. M. Chodorow, R. Symons, D. Abe, R. Abrams, B. Danly, H. Freund, K. Jensen, B. Levush, R. Myers, R. Parker, J. Shaw, A. Shih and J. Yater, in Reference Data for Engineers, Elsevier, 2002, pp. 16-1–16-59.
30. H.-B. Zhang, R.-J. Feng and K. Ura, Utilizing the Charging Effect in Scanning Electron Microscopy, Sci. Prog., 2004, 87, 249–268.
31. B. Wang, N. Novendra and A. Navrotsky, Energetics, Structures, and Phase Transitions of Cubic and Orthorhombic Cesium Lead Iodide (CsPbI3) Polymorphs, J. Am. Chem. Soc., 2019, 141, 14501–14504.
32. Y. Lin and D. C. Joy, Surf. Interface Anal., 2005, 37, 895–900.
33. J. J. Scholtz, D. Dijkkamp and R. W. A. Schmitz, Secondary electron emission properties, Philips J. Res., 1996, 50, 375–389.
34. L. K. Gao and Y. L. Tang, Theoretical Study on the Carrier Mobility and Optical Properties of CsPbI3 by DFT, ACS Omega, 2021, 6, 11545–11555.
35. M. T. Postek and A. E. Vladár, Does your SEM really tell the truth? How would you know? Part 4: Charging and its mitigation, Proc. SPIE 9636, Scanning Microscopies, 2015, p. 963605.
36. M. Zhao, B. Ming, J.-W. Kim, L. J. Gibbons, X. Gu, T. Nguyen, C. Park, P. T. Lillehei, J. S. Villarrubia, A. E. Vladár and J. Alexander Liddle, New insights into subsurface imaging of carbon nanotubes in polymer composites via scanning electron microscopy, Nanotechnology, 2015, 26, 169601.
37. S. Jariwala, H. Sun, G. W. P. Adhyaksa, A. Lof, L. A. Muscarella, B. Ehrler, E. C. Garnett and D. S. Ginger, Local Crystal Misorientation Influences Non-radiative Recombination in Halide Perovskites, Joule, 2019, 3, 3048–3060.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d5ta03723a

---