ABSTRACT
We describe a system, ChemSCAD, for the creation of digital reactors based on the chemical operations, physical parameters, and synthetic sequence to produce a given target compound, to show that the system can translate the gram-scale batch synthesis of the antiviral compound Ribavirin (yield 43% over three steps), the narcolepsy drug Modafinil (yield 60% over three steps), and both batch and flow instances of the synthesis of the anticancer agent Lomustine (batch yield 65% over two steps) in purities greater than or equal to 96%. The syntheses of compounds developed using the ChemSCAD system, including reactor designs and analytical data, can be stored in a database repository, with the information necessary to critically evaluate and improve upon reactionware syntheses being easily shared and versioned.
ABSTRACT
The control and prediction of complex chemical systems is a difficult problem due to the nature of the interactions, transformations and processes occurring. From self-assembly to catalysis and self-organization, complex chemical systems are often heterogeneous mixtures that at the most extreme exhibit system-level functions, such as those that could be observed in a living cell. In this paper, we outline an approach to understand and explore complex chemical systems using an automated droplet maker to control the composition, size and position of the droplets in a predefined chemical environment. By investigating the spatio-temporal dynamics of the droplets, the aim is to understand how to control system-level emergence of complex chemical behaviour and even view the system-level behaviour as a programmable entity capable of information processing. Herein, we explore how our automated droplet-maker platform could be viewed as a prototype chemical heterotic computer with some initial data and example problems that may be viewed as potential chemically embodied computations.
ABSTRACT
Evolution, once the preserve of biology, has been widely emulated in software, while physically embodied systems that can evolve have been limited to electronic and robotic devices and have never been artificially implemented in populations of physically interacting chemical entities. Herein we present a liquid-handling robot built with the aim of investigating the properties of oil droplets as a function of composition via an automated evolutionary process. The robot makes the droplets by mixing four different compounds in different ratios and placing them in a Petri dish after which they are recorded using a camera and the behaviour of the droplets analysed using image recognition software to give a fitness value. In separate experiments, the fitness function discriminates based on movement, division and vibration over 21 cycles, giving successive fitness increases. Analysis and theoretical modelling of the data yields fitness landscapes analogous to the genotype-phenotype correlations found in biological evolution.
