Rapid Online Analysis of Reactions (ROAR)

数据科学和数字技术被誉为21世纪的新工业革命。它已经对许多科学学科产生了变革性的影响，例如。物理学中的“大数据”、机器人中的人工智能、量子计算和数学生物学。然而，它对分子科学的潜在影响在很大程度上仍未被探索。目前，合成的速度和效率（包括规模化）仍然是医疗保健、农化、分子电子、智能材料等新兴领域发展的瓶颈。化学反应是一个包含许多相互依赖的变量的高度复杂的系统。完全自主反应器（“合成机”）的发展将取决于我们准确捕获和解释数据的能力，因此我们不仅可以产生新的知识，而且还可以设计出有效的方法来预测给定条件（前体，试剂，催化剂，溶剂，温度等）的反应结果。为了克服这一挑战，化学家需要熟悉并能够执行数据丰富的实验。这将需要在设备和专业知识方面的大量资本投资，而这目前超出了学术实验室的能力。ROAR的核心精神是通过促进高通量，自动化实验来促进数据丰富的化学合成方法，包括应用数学和统计工具（实验设计，多变量分析）以及更好的反应和过程理解（动力学和热剖面），从而改变分子科学的景观和实践。从长远来看，这将提高合成方案的可靠性，改进化学信息学工具，从而更好地预测反应结果和选择最有效的合成路线。

Data science and digital technologies have been hailed as the new industrial revolution for the 21st Century. It is already having a transformative effect on many scientific disciplines, e.g. 'Big data' in physics, artificial intelligence (AI) in robotics, quantum computing, and mathematical biology. However, its potential impact on the molecular sciences has remained largely unexplored. Currently, the speed and efficiency of synthesis (including scale-up) remain a bottleneck in the development of healthcare, agrochemicals, molecular electronics, smart materials and other emerging fields. A chemical reaction is highly complex system containing many interdependent variables. Development of fully autonomous reactor ('synthesis machine') will depend on our ability to capture and interpret data accurately, so that we can not only generate new knowledge, but also to devise effective methods to predict reaction outcomes from a given set of conditions (precursors, reagents, catalyst, solvent, temperature, etc). To overcome the challenge, chemists needs to be able to be familiar with and able to execute data-rich experiments. This will require substantial capital investment in equipment and expertise that are currently beyond the means of academic laboratories.The central ethos of ROAR is to transform the landscape and practice of molecular science, by promoting high-throughput, automated experimentation to promote a data-rich approach to chemical synthesis, including the application of mathematical and statistical tools (Design-of-Experiments, multivariate analysis) and greater reaction and process understanding (kinetic and thermal profiling). In the long run, this will enable greater reliability of synthetic protocols and refinement of chemoinformatics tools, leading to better prediction of reaction outcomes and selection of most effective synthetic routes.