Autonomous Chemical Experiments: Challenges and Perspectives on Establishing a Self-Driving Lab

Accounts of Chemical Research, Год журнала: 2022, Номер 55(17), С. 2454 - 2466

Опубликована: Авг. 10, 2022

We must accelerate the pace at which we make technological advancements to address climate change and disease risks worldwide. This swifter of discovery requires faster research development cycles enabled by better integration between hypothesis generation, design, experimentation, data analysis. Typical take months years. However, data-driven automated laboratories, or self-driving can significantly molecular materials discovery. Recently, substantial have been made in areas machine learning optimization algorithms that allowed researchers extract valuable knowledge from multidimensional sets. Machine models be trained on large sets literature databases, but their performance often hampered a lack negative results metadata. In contrast, generated laboratories information-rich, containing precise details experimental conditions Consequently, much larger amounts high-quality are gathered laboratories. When placed open repositories, this used community reproduce experiments, for more in-depth analysis, as basis further investigation. Accordingly, will increase accessibility reproducibility science, is sorely needed.In Account, describe our efforts build lab new class materials: organic semiconductor lasers (OSLs). Since they only recently demonstrated, little known about material design rules thin-film, electrically-pumped OSL devices compared other technologies such light-emitting diodes photovoltaics. To realize high-performing materials, developing flexible system synthesis via iterative Suzuki-Miyaura cross-coupling reactions. platform directly coupled analysis purification capabilities. Subsequently, molecules interest transferred an optical characterization setup. currently limited measurements solution. properties ultimately most important solid state (e.g., thin-film device). end different scientific goal, inorganic focused oxygen evolution reaction.While future very promising, numerous challenges still need overcome. These split into cognition motor function. Generally, cognitive related with constraints unexpected outcomes general algorithmic solutions yet developed. A practical challenge could resolved near software control because few instrument manufacturers products mind. Challenges function largely handling heterogeneous systems, dispensing solids performing extractions. As result, it critical understand adapting procedures were designed human experimenters not simple transferring those same actions system, there may efficient ways achieve goal fashion. carefully rethink translation manual protocols.

Язык: Английский

---

Closed-loop optimization of general reaction conditions for heteroaryl Suzuki-Miyaura coupling

Science, Год журнала: 2022, Номер 378(6618), С. 399 - 405

Опубликована: Окт. 27, 2022

General conditions for organic reactions are important but rare, and efforts to identify them usually consider only narrow regions of chemical space. Discovering more general reaction requires considering vast space derived from a large matrix substrates crossed with high-dimensional conditions, rendering exhaustive experimentation impractical. Here, we report simple closed-loop workflow that leverages data-guided down-selection, uncertainty-minimizing machine learning, robotic discover conditions. Application the challenging consequential problem heteroaryl Suzuki-Miyaura cross-coupling identified double average yield relative widely used benchmark was previously developed using traditional approaches. This study provides practical road map solving multidimensional optimization problems search spaces.

Язык: Английский

---

Predictive chemistry: machine learning for reaction deployment, reaction development, and reaction discovery

Chemical Science, Год журнала: 2022, Номер 14(2), С. 226 - 244

Опубликована: Ноя. 28, 2022

This review outlines several organic chemistry tasks for which predictive machine learning models have been and can be applied.

Язык: Английский

---

On the use of real-world datasets for reaction yield prediction

Chemical Science, Год журнала: 2023, Номер 14(19), С. 4997 - 5005

Опубликована: Янв. 1, 2023

The lack of publicly available, large, and unbiased datasets is a key bottleneck for the application machine learning (ML) methods in synthetic chemistry. Data from electronic laboratory notebooks (ELNs) could provide less biased, large datasets, but no such have been made available. first real-world dataset ELNs pharmaceutical company disclosed its relationship to high-throughput experimentation (HTE) described. For chemical yield predictions, task synthesis, an attributed graph neural network (AGNN) performs as well or better than best previous models on two HTE Suzuki-Miyaura Buchwald-Hartwig reactions. However, training AGNN ELN does not lead predictive model. implications using data ML-based are discussed context predictions.

Язык: Английский

---

Recent applications of machine learning in alloy design: A review

Materials Science and Engineering R Reports, Год журнала: 2023, Номер 155, С. 100746 - 100746

Опубликована: Июнь 17, 2023

Язык: Английский

---

Machine Learning‐Assisted Property Prediction of Solid‐State Electrolyte

Advanced Energy Materials, Год журнала: 2024, Номер 14(20)

Опубликована: Фев. 14, 2024

Abstract Machine learning (ML) exhibits substantial potential for predicting the properties of solid‐state electrolytes (SSEs). By integrating experimental or/and simulation data within ML frameworks, discovery and development advanced SSEs can be accelerated, ultimately facilitating their application in high‐end energy storage systems. This review commences with an introduction to background SSEs, including explicit definition, comprehensive classification, intrinsic physical/chemical properties, underlying mechanisms governing conductivity, challenges, future developments. An in‐depth explanation methodology is also elucidated. Subsequently, key factors that influence performance are summarized, thermal expansion, modulus, diffusivity, ionic reaction energy, migration barrier, band gap, activation energy. Finally, it offered perspectives on design prerequisites upcoming generations focusing real‐time property prediction, multi‐property optimization, multiscale modeling, transfer learning, automation high‐throughput experimentation, synergistic optimization full battery, all which crucial accelerating progress SSEs. aims guide novel SSE materials practical realization efficient reliable technologies.

Язык: Английский

---

Self-Driving Laboratories for Chemistry and Materials Science

Chemical Reviews, Год журнала: 2024, Номер 124(16), С. 9633 - 9732

Опубликована: Авг. 13, 2024

Self-driving laboratories (SDLs) promise an accelerated application of the scientific method. Through automation experimental workflows, along with autonomous planning, SDLs hold potential to greatly accelerate research in chemistry and materials discovery. This review provides in-depth analysis state-of-the-art SDL technology, its applications across various disciplines, implications for industry. additionally overview enabling technologies SDLs, including their hardware, software, integration laboratory infrastructure. Most importantly, this explores diverse range domains where have made significant contributions, from drug discovery science genomics chemistry. We provide a comprehensive existing real-world examples different levels automation, challenges limitations associated each domain.

Язык: Английский

---

Palladium-catalyzed decarboxylative (4 + 3) cycloadditions of bicyclobutanes with 2-alkylidenetrimethylene carbonates for the synthesis of 2-oxabicyclo[4.1.1]octanes

Chemical Science, Год журнала: 2024, Номер 15(34), С. 13942 - 13948

Опубликована: Янв. 1, 2024

While cycloaddition reactions of bicyclobutanes (BCBs) have emerged as a potent method for synthesizing (hetero-)bicyclo[

Язык: Английский

---

Probing the chemical ‘reactome’ with high-throughput experimentation data

Nature Chemistry, Год журнала: 2024, Номер 16(4), С. 633 - 643

Опубликована: Янв. 2, 2024

High-throughput experimentation (HTE) has the potential to improve our understanding of organic chemistry by systematically interrogating reactivity across diverse chemical spaces. Notable bottlenecks include few publicly available large-scale datasets and need for facile interpretation these data's hidden insights. Here we report development a high-throughput analyser, robust statistically rigorous framework, which is applicable any HTE dataset regardless size, scope or target reaction outcome, yields interpretable correlations between starting material(s), reagents outcomes. We data landscape with disclosure 39,000+ previously proprietary reactions that cover breadth chemistry, including cross-coupling chiral salt resolutions. The analyser was validated on hydrogenation datasets, showcasing elucidation significant relationships components outcomes, as well highlighting areas bias specific spaces necessitate further investigation.

Язык: Английский

---

Automatic feature engineering for catalyst design using small data without prior knowledge of target catalysis

Communications Chemistry, Год журнала: 2024, Номер 7(1)

Опубликована: Янв. 12, 2024

The empirical aspect of descriptor design in catalyst informatics, particularly when confronted with limited data, necessitates adequate prior knowledge for delving into unknown territories, thus presenting a logical contradiction. This study introduces technique automatic feature engineering (AFE) that works on small datasets, without reliance specific assumptions or pre-existing about the target catalysis designing descriptors and building machine-learning models. generates numerous features through mathematical operations general physicochemical catalytic components extracts relevant desired catalysis, essentially screening hypotheses machine. AFE yields reasonable regression results three types heterogeneous catalysis: oxidative coupling methane (OCM), conversion ethanol to butadiene, three-way where only training set is swapped. Moreover, application active learning combines high-throughput experimentation OCM, we successfully visualize machine's process acquiring precise recognition design. Thus, versatile data-driven research key step towards fully automated discoveries.

Язык: Английский

---