The use of perovskite materials (particularly methylammonium lead iodide) in solar cells has become very attractive due to the fast increase in reported power conversion efficiencies over the last few years, leading to values above 20%. While this value is competitive with established photovoltaic technologies, the stability of perovskite-based solar cells is still insufficient for commercial applications. In particular, it is very well known that some components, including the perovskite layer and the hole transporter, can degrade when exposed to a combination of heat and moisture. In situ TEM is an ideal tool for investigating such degradation and understanding the phenomena underpinning it.

In this work [1,2] we prepare methylammonium lead iodide cells using different approaches from the literature (with the perovskite conversion carried out in single- and double-step in glovebox, in air or in vacuum); we prepare TEM cross-sectional samples using focused ion beam milling.

For each cell, we carry out scanning TEM imaging and EDX elemental mapping (shown in Figure 1) as they are heated in situ in the TEM. This is a procedure that requires careful control over the temperature and the electron dose. To that aim we exploit recent advances in TEM-related technology, such as Silicon Drift Detectors (SDD) for EDX, which collect energy-dispersed X-ray spectra with a good yield, and stable MEMS heaters, enabling the temperature to be cycled quickly and reproducibly. Moreover, we employ multivariate analysis (principal component analysis, PCA) to increase the signal-to-noise ratio of the spectral maps.

Cross-sectional views acquired after heating are reported in Figure 2. We do not observe changes in the morphology or the elemental distribution in the perovskite layer for heating up to 150°C for short times (employing a heating ramp with 30’ steps every 25°C). Since the ex-situ heating of the same samples above 90°C causes a significant decay in cell performance, we attribute such decay to the degradation of the charge transport properties of the hole transporter (spiro-OMeTAD in this case). Increasing the temperature further, different decomposition patterns emerge for the perovskite layer. In samples that had not been exposed to air, elemental migration of lead and iodine results in the formation of aggregates, which EDX suggests might be PbI₂, clustering on the FTO electrode. In the sample exposed to air, a different phenomenon occurs – instead of forming aggregates, the elemental species diffuse from the perovskite into the hole transporter. This is visible both as an increased contrast in the high-angle annular dark field images (HAADF) and as features in the EDX spectra; we hypothesise that the trapped moisture within the cell might be hindering the formation of PbI₂ and make elemental diffusion more favourable.

[1] Divitini, G. et al. – Nature Energy 201512 (2016)

[2] Matteocci, F. et al. – ACS Applied Materials & Interfaces 7, 26176 (2015)

Figures:

Figure 1: Cross-sectional view and EDX maps of a representative perovskite solar cell (adapted from [1]).

Figure 2: HAADF-STEM images of cross-sections of different solar cells (A – double step in vacuum, B – double step in glovebox, C – double step, air conversion, D – single step) after heating at the temperatures indicated in the labels (adapted from [1]).

« Back to The 16th European Microscopy Congress 2016

EMC Abstracts - https://emc-proceedings.com/abstract/in-situ-observation-of-heat-induced-degradation-of-perovskite-solar-cells/