Correlation of Mobile Ion Concentration to the Stability and Performance of Perovskite Solar Cells

Ion migration is one of the important factors affecting the stability and operational lifetimes of perovskite solar cells (PSCs). Hence, it is very important to understand the correlation of ion migration to stability and its impact on PSC power conversion efficiencies (PCE). But ion migration analysis is typically qualitative by either showing the presence of migrating ions in PSC or identifying the migrating species without any knowledge of the densities or mobilities of the ions. Our previous work has shown that ion migration in PSCs can be quantified in terms of mobile ion concentration (N_o) and ion mobility. N_o of PSC is obtained by determining the ionic charge by integrating the drift current (from the transient dark current response of PSC after the applied voltage of 0.8V for 10ms in forward bias) with time and using the empirical relation between N_o and ionic charge. N_o exhibits a wider range of magnitude over a range of 7 orders from 10¹¹ – 10¹⁸ cm^-3 based on the type of top electrode used for PSC.
In this work, we expand and deepen our understanding of N_o by correlating it with chemical species in a PSC and as a quantitative stability metric for devices. Our hypothesis was that mobile metal ions that diffuse from the top electrode through the fullerene-based ETL with low barrier properties can influence N_o and result in degradation through metal-halide reactions. In this study, we selected p-i-n devices that incorporated dense barrier layers of tin oxide (SnO₂) or ozone-nucleated SnO₂ (O₃-SnO₂) to stop the diffusion of metal ions along with different top-electrode chemistries compared with fullerene-only ETLs. We show the impact of top-electrode chemistry on N_o where the use of silver (Ag) electrodes results in significantly higher N_o values > 10¹⁴ cm^-3 when compared to a gold (Au) with N_o values of ~ 10¹³ cm^-3 or carbon top electrode with N_o values < 10¹³ cm^-3. PSCs that contain a SnO₂ barrier with an Ag electrode also have significantly lower N_o values < 10¹³ cm^-3. We confirm the involvement of Ag ion diffusion through the ETL and into the active layer while quantifying the increase of the Ag ions into the underlying device layers with N_o and Rutherford backscattering spectrometry (RBS), except in the case of PSCs with O₃-SnO₂ barrier layer. In situ measurements of N_o in PSCs with increasing temperatures showed a threshold N_o value (~2.8 * 10¹⁶ ± 8.5 * 10¹⁵ cm^-3) that correlates with device failure, which occurs at lower temperatures for non-barrier devices (~370 K) and higher temperatures for barrier devices (~450 K) and that O₃-SnO₂ stops Ag diffusion and improves thermal stability.
We also report on the correlation of N_o and performance of PSCs demonstrated by drift-diffusion simulations using a realistic inverted p-i-n device configuration with nickel oxide (NiO_x) as HTL and C₆₀ as ETL. In the model, we varied the N_o from 10¹¹ cm^-3 (extremely low) to 10¹⁸ cm^-3 (extremely high), keeping either the cations mobile, anions mobile, or both mobile. We found that there was essentially no change in PCE with a higher N_o which was unexpected. However, it is overly simplified because the simulations did not include any non-radiative recombination parameters. We also decided to use mobile cations with fixed anions to correspond to the positively charged halide vacancies that cause ion migration. While operating the PSC at a realistic surface recombination velocity (~10³ cm/s) simulations show a negligible impact on PCE with a N_o < 10¹⁵ cm^-3 but show a drastic impact on PCE with a N_o > 10¹⁵ cm^-3. Simulations also showed the need to reduce the recombination at the HTL as higher recombination velocities at HTL had a greater impact on the PCE of PSCs when compared to higher recombination at ETL. Further ongoing efforts involve modeling the transient dark current response of a PSC and extracting N_o using SETFOS to compare the ion dynamics with experimental measurements to develop a realistic parameter set for simulation.