Efficient Discovery of Visible Light-Activated Azoarene Photoswitches with Long Half-Lives Using Active Search

Abstract

Photoswitches are molecules that undergo a reversible, structural isomerization after exposure to different wavelengths of light. The dynamic control offered by molecular photoswitches is favorable for applications in materials chemistry, photopharmacology, and catalysis. Ideal photoswitches absorb visible light and have long-lived metastable isomers. We used high throughput virtual screening to predict the absorption maxima (λ max ) of the E-isomer and half-lives (t 1/2 ) of the Z-isomer. However, computing the photophysical and kinetic properties of each entry of a virtual molecular library containing 10 3 -10 6 entries with density functional theory is prohibitively time-consuming. We applied active search, a machine learning technique to intelligently search a chemical search space of 255 991 photoswitches based on 29 known azoarenes and their derivatives. We iteratively trained the active search algorithm based on whether a candidate absorbed visible light (λ max > 450 nm). Active search was found to triple the discovery 1 rate compared to random search. Further, we projected 1 962 photoswitches to 2D using the Uniform Manifold Approximation and Projection (umap) algorithm and found that λ max depends on the core, which is tunable with substituents. We then incorporated a second stage of screening with to predict the stabilities of the Z-isomers for the top 1% of candidates. We identified four ideal photoswitches that concurrently satisfy λ max > 450 nm and t 1/2 > 2 hours; the range of λ max and t 1/2 range from 465 to 531 nm and hours to days, respectively.

Thermal Half-Lives of Azobenzene Derivatives: Virtual Screening Based on Intersystem Crossing Using a Machine Learning Potential

“…Further, the program is quite fast, replicating the results of spin-flip, time-dependent density functional theory (SF-TDDFT) in milliseconds through use of a transferable ML potential. This program adds to the growing collection of computational models for predicting photoswitch properties. ,− Our software and pretrained models are freely available at .…”

Thermal half-lives of azobenzene derivatives: virtual screening based on intersystem crossing using a machine learning potential

“…Observing and explaining general trends will enable more focused candidate screening in the future. Previous papers have also explored these relationships computationally [11,12,14,15,49]; we build on their conclusions and extend them to other substitution patterns and groups.…”

“…Several works have predicted the thermal lifetimes of azobenzene derivatives with computational methods [10][11][12][13][14][15]. However, there are two main issues with previous calculations.…”

Data-driven discovery of molecular photoswitches with multioutput Gaussian processes

“…83 A further point of interest would be an investigation into how synthetic chemists may use model uncertainty estimates in the decision process to screen molecules e.g. via active learning 84 and Bayesian optimisation. The condence-error curves in the ESI † show initial promise in this direction and indeed understanding how best to tailor calibrated Bayesian models to molecular representations 65,85 is an avenue worthy of pursuit.…”

“…Recently work by Lopez and co-workers 55 employed machine learning to accelerate a quantum chemistry screening workflow for photoswitches. The screening library in this case is generated from 29 known azoarenes and their derivatives yielding a virtual library of 255 991 azoarenes in total.…”