Freeing hands and automating processes at a chemistry laboratory

The Challenge

There is a lot of equipment around and many tasks to be carried out: noting results, discussing them with the team, instruments maintenance, cleaning. Some manual tasks that require less “brainpower”, let's put it that way, may include moving the vials from one point to another, opening and closing them, putting them in a machine, etc. From that point of view, it would make sense to have a non-human assistant around, don't you think? Conventional, labor-intensive workflows for finding solubility values require sophisticated analytical instrumentation (i.e., high-performance liquid chromatography (HPLC)).

The Solution

The group seeks to design a generalizable tool capable of experimentally mapping the physical and chemical properties of any target chemical system. In this paradigm, the exact sequence of actions carried out by the automation is not predefined — the experimental goal is defined along with conditional instructions to enable the system to make autonomous decisions. This strategy more closely mimics the actions of an expert laboratory assistant. In the specific case of their solubility workflow, the group wanted a modular, closed-loop robotic platform that was guided by simple computer vision, rather than complex analytics. The system would require no prior knowledge of chemistry involved, and would be able to detect failures in the workflow to adapt accordingly. This system would be suitable for any solid-solvent combination without specialized adjustments. The modularity of the platform would enable it to easily perform any related experiment requiring solid/liquid addition, mixing/heating, and optical determination of homogeneity.