Molecular I₂ can be produced from iodide‐based lead perovskites under thermal stress; triiodide, I₃⁻, is formed from this I₂ and I⁻. Triiodide attacks protic cation MA⁺‐ or FA⁺‐based lead halide perovskites (MA⁺, methylammonium; FA⁺, formamidinium) as explicated through solution‐based nuclear magnetic resonance (NMR) studies: triiodide has strong hydrogen‐bonding affinity for MA⁺ or FA⁺, which leads to their deprotonation and perovskite decomposition. Triiodide is a catalyst for this decomposition that can be obviated through perovskite surface treatment with thiol reducing agents. In contrast to methods using thiol incorporation into perovskite precursor solutions, no penetration of the thiol into the bulk perovskite is observed, yet its surface application stabilizes the perovskite against triiodide‐mediated thermal stress. Thiol applied to the interface between FAPbI₃ and Spiro‐OMeTAD (“Spiro”) prevents oxidized iodine species penetration into Spiro and thus preserves its hole‐transport efficacy. Surface‐applied thiol affects the perovskite work function; it ameliorates hole injection into the Spiro overlayer, thus improving device performance. It helps to increase interfacial adhesion (“wetting”): fewer voids are observed at the Spiro/perovskite interface if thiols are applied. Perovskite solar cells (PSCs) incorporating interfacial thiol treatment maintain over 80% of their initial power conversion efficiency (PCE) after 300 h of 85 °C thermal stress.

Designing a high-quality control is crucial for reliable quantum computation. Among the existing approaches, closed-loop leaning control is an effective choice. Its efficiency depends on the learning algorithm employed, thus deserving algorithmic comparisons for its practical applications. Here we assess three representative learning algorithms, including GRadient Ascent Pulse Engineering (GRAPE), improved Nelder-Mead (NMplus), and Differential Evolution (DE), by searching for high-quality control pulses to prepare the Bell state. We first implement each algorithm experimentally in a nuclear magnetic resonance system and then conduct a numerical study considering the impact of some possible significant experimental uncertainties. The experiments report the successful preparation of the high-fidelity target state by the three algorithms, while NMplus converges fastest, and these results coincide with the numerical simulations when potential uncertainties are negligible. However, under certain significant uncertainties, these algorithms possess distinct performance with respect to their resulting precision and efficiency, and DE shows the best robustness. This study provides insight to aid in the practical application of different closed-loop learning algorithms in realistic physical scenarios.

The chiral nematic phase of poly-γ-benzyl-l-glutamate (PBLG) formed in a chloroform–DMSO co-solvent system can be used as a versatile alignment medium for the acquisition of high quality anisotropic NMR data for molecules of varying polarities.