Despite numerous technical advances over the last century, the field of academic organic chemistry still relies heavily on manual labor and non-systematic experiment design. Discovery
and optimization of chemical reactions are subject to arbitrarily large reaction spaces, leading to
an extreme amount of possible combinations of reagents and reaction conditions. The practical
aspect often takes up the largest portion of available working time, reducing opportunities
for conceptualization, planning and analysis. Hence, chemists often find themselves performing non-value adding activities most of the time. To eliminate this bottleneck, two strategies
were explored in this work: increasing synthetic throughput using machines, and increasing
information density per experiment using statistical methods. The use of a recommissioned
parallel synthesizer in synergy with Design of Experiments allowed for the high-throghput generation of reproducible, meaningful data. Using interdisciplinary tools such as programming,
microcontroller prototyping and 3D printing, the capabilities of the automated equipment
were enhanced even more, for example through the development of automated, quantitative
thin-layer chromatography. The new workflows were demonstrated in three use cases based
on the molecule ferrocene. The use cases were centered around parameter investigation in
complex reactions, library synthesis and the exploration of unknown reaction spaces.
|