A Physical Data Fusion Approach to Optimize Compositional Stability of Halide Perovskites

Abstract

<p>Compositional search within multinary perovskites employing brute force synthesis are prohibitively expensive in large chemical spaces. To identify the most stable multi-cation lead iodide perovskites containing Cs, formamidinium (FA) and methylammonium (MA), we fuse results from density functional theory (DFT) calculations and <i>in situ</i> thin-film degradation test within an end-to-end machine learning (ML) algorithm to inform the compositional optimization of Cs<sub>x</sub>MA<sub>y</sub>FA<sub>1-x-y</sub>PbI<sub>3</sub>. We integrate phase thermodynamics modelling as a <i>probabilistic constraint</i> in a Bayesian optimization (BO) loop, which effectively guides the experimental search while considering both structural and environmental stability. After three optimization rounds and only sampling 1.8% of the compositional space, we identify thin-film compositions centred at Cs<sub>0.17</sub>MA<sub>0.03</sub>FA<sub>0.80</sub>PbI<sub>3</sub> that achieve a 3x delay in macroscopic degradation onset under elevated temperature, humidity, and light compared with the more complex state-of-the-art Cs<sub>0.05</sub>(MA<sub>0.17</sub>FA<sub>0.83</sub>)<sub>0.95</sub>Pb(I<sub>0.83</sub>Br<sub>0.17</sub>)<sub>3</sub>. We find up to 8% of MA can be incorporated into the perovskite structure before stability is significantly compromised. Cs is beneficial at low concentrations, however, beyond 17% is found to contribute to reduced stability<b>.</b> Synchrotron-based grazing-incidence wide-angle X-ray scattering (GIWAXS) further validates that the interplay of chemical decomposition and phase separation governs the non-linear instability landscape of this compositional space. We reveal the detrimental role of the ẟ-CsPbI<sub>3</sub> minority phase in accelerating degradation and it can be kinetically suppressed by co-optimising Cs and MA content, providing insights into simplifying perovskite compositions for further environmental stability enhancement. Our approach realizes the effectiveness of ML-enabled data fusion in achieving a holistic, efficient, and physics-informed experimentation for multinary systems, potentially generalisable to materials search in the vast structural and alloyed spaces beyond halide perovskites.</p><br>

Tailoring capping-layer composition for improved stability of mixed-halide perovskites

“…This image data is processed, where the red, green, blue (RGB) values for each image is extracted using the protocol that has been published previously. 3,9 The large dataset is subsequently analyzed using dissimilarity matrix to extract the most stable capping layer material for each absorber composition. Using this method, we can also compare the samples variance within the same composition and across 5 different compositions.…”

Becoming Good at AI for Good

“…In the industry, continued data collection and user-supplied labels allow models to be improved over time, the so-called "data flywheel" effect [16,66]. In a similar manner, many scientific fields have come up with labeled datasets and models [13,64,71] that are continuously updated as labeling techniques (such as physical simulations or data acquisition methods) are improved. A common industry practice for improving model performance is to iterate on improving datasets instead of iterating on improving models [31].…”