Dimensional Control over Metal Halide Perovskite Crystallization Guided by Active Learning

Abstract

Metal halide perovskite (MHP) derivatives, a promising class of optoelectronic materials, have been synthesized with a range of dimensionalities that govern their optoelectronic properties and determine their applications. We demonstrate a data-driven approach combining active learning and high-throughput experimentation to discover, control, and understand the formation of phases with different dimensionalities in the morpholinium (morph) lead iodide system. Using a robot-assisted workflow, we synthesized and characterized two novel MHP derivatives that have distinct optical properties: a one-dimensional (1D) morphPbI3 phase ([C4H10NO][PbI3]) and a 2D (morph)2PbI4 phase ([C4H10NO]2[PbI4]). To efficiently acquire the data needed to construct a machine learning (ML) model of the reaction conditions where the 1D and 2D phases are formed, data acquisition was guided by a diverse-mini-batch-sampling active learning algorithm, using prediction confidence as a stopping criterion. Querying the ML model uncovered the reaction parameters that have the most significant effects on dimensionality control. Based on these insights, we propose a reaction scheme that rationalizes the formation of different dimensional MHP derivatives in the morph-Pb-I system. The data-driven approach presented here, including the use of additives to manipulate dimensionality, will be valuable for controlling the crystallization of a range of materials over large reaction-composition spaces.

Spatiotemporal Route to Understanding Metal Halide Perovskitoid Crystallization

“…23 Although their system could image reaction compositions over time, imaging the reaction system from above only allows for observing whether a given mother liquor and antisolvent composition gives rise to crystal formation. Similarly, Li et al described a liquid handling robot-based system for performing highthroughput ASVC experiments to explore mother liquor and antisolvent composition variables; likewise, only the final composition is reported, 24 obscuring the critical parameter required to induce crystallization. In contrast, here, we report a unique spatiotemporal approach that replaces a parallel set of experiments in space (i.e., conducted in separate spatial vials and with separate materials) with a smaller set of experiments conducted over time.…”

Principled Exploration of Bipyridine and Terpyridine Additives to Promote Methylammonium Lead Iodide Perovskite Crystallization

“…An enclosed Hamilton Microlab NIMBUS4 was used at Haverford College and a Hamilton Microlab NIMBUS4 liquid handling robot was used at Lawrence Berkeley National Laboratory in this study for the high-throughput antisolvent vapor-assisted crystallization (HT-ASVC) synthesis of MAPbI 3 -additive crystals. 14 ASVC is a slow process that relies upon variations of the perovskite solubility in different solvents. In other words, supersaturation can be attained by exposing a solution of the product to another solvent in which the product is sparingly soluble (otherwise called antisolvent).…”

“…However, considering the vast number of possible bipyridine- and terpyridine-based additive molecules with various functional groups, a systematic exploration of additive molecules coupled with a method to better understand the results is still needed. Previous studies on single-crystal LHPs using high-throughput experiments solely focused on exploring different A-site cations and reaction conditions; , only recently have additives such as water , been explored in the same manner.…”

Serendipity based recommender system for perovskites material discovery: balancing exploration and exploitation across multiple models

“…This can result in missing potentially new phases-for example, we recently reported a morpholinium lead iodide system where small concentration changes result in phases with distinct structural and optical properties. (20) The serendipity-based recommendation system can be applied to any model to increase the recommendation diversity while keeping the probability of success high. Laboratory comparisons indicate that serendipity-directed recommendation improves the diversity of recommendations, which in turn improves the robustness of the recommendation against initialization conditions, without substantially degrading recommendation success.…”