At the Hein Lab, a research group affiliated with a major Western Canada university, researchers and students are working on projects involving both chemistry and automation. This lab brings together these two fields to solve (pun intended) lingering problems — working on the mechanisms of complex reactions and developing advanced synthesis and crystallization processes, among other things. Employing tools from both classical and modern physical organic chemistry, the group wants to accelerate the pace of new discoveries by implementing self-driving laboratories.
The lab’s research program relies on a multidisciplinary team composed of synthetic and physical organic chemists, computer scientists, mechatronics engineers, and data visualization experts. Their automation/instrumentation subgroup is focused on creating automated robotic solutions applied to a variety of chemistry challenges. Directed by Jason Hein, Associate Professor, the lab has published quite a number of scientific papers since 2012 and can count on support from renowned contributors such as Pfizer, Merck, and DARPA. The topics of their projects are broad, ranging from the analysis of amino acids found in carbonaceous chondrites from Antarctica to the crystallization analysis of non-UV active compounds.
For example, in one of their projects, they pair a robotic arm with solid and liquid handling capabilities to a webcam to autonomously determine the solubility of various compounds. This approach is an efficient and easy way to find solubility values, particularly when compared to the conventional, labor-intensive workflows that require sophisticated analytical instrumentation (i.e., high-performance liquid chromatography (HPLC)). In this particular case, a Kinova robotic arm was used as part of their automated solubility screening platform using computer vision.
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, as the system is able to react to conditional variations and still achieve the aim of the experiment.
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. These could include identifying conditions for crystallization, creating stock solutions at targeted concentration, or determining solvent combinations for liquid-liquid extraction just to name a few.
“Our group focuses on flexible automation, and a robotic arm is the main manipulator and handler of various lab objects and chemicals. Integrating these robots smoothly with other lab systems is crucial, and requires user-friendly hardware and software that are easy to set up. Kinova has met both of these points.” - Paloma Prieto, Program Manager, Hein Lab
Since most of the code is open-source, another subsequent goal arising from the success of this experiment is to share this work with other groups. Collaborations with companies in the pharmaceutical and instrumentation space, for example, help the Hein Lab to deploy their tools to solve real-world problems outside the university.
"We expect that the platform can facilitate finding answers to solubility questions in industries dealing with novel materials. By automating a process that is integral to experimental design and optimization, we hope to reduce demands on both material cost and human time."
The quote below is taken from Automated Solubility Screening Platform Using Computer Vision, a scientific paper covering the application discussed in this case study.
There is no need to work inside a lab to get the picture. 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?
“It was a problem-driven project [...] our Ph.D. students could spend up to 80 % of their time doing menial tasks,” said Prieto.
Automation to the rescue
The Hein Lab was looking into purchasing a Canadian-based 6-axis robotic arm to use within its automated platforms, that’s when we came up into the equation. One of our representatives connected with Jason Hein and we were able to propose to them the appropriate robot suited for their needs and future objectives.
According to Parisa Shiri, a member of the Hein Lab staff who first used and set up the robot: “It was relatively easy to learn how to use this robot. The lab staff had experience working with SCARA robots before, but we were new to these 6-axis robots.”
The Kinova Gen3 features a teach-pendant-like GUI (graphical user interface) called the Kinova Web App, accessible through the browser of a computer, tablet, or smartphone.
“The Web App has been a great advantage on the programming side, being intuitive and user-friendly. I would recommend the Gen3 to anyone in any multi-disciplinary field with a non-robotic background seeking an arm to automate a process.” - Parisa Shiri, research associate, Hein Lab
In the following video, we can see the robot helping out with a few crucial tasks such as:
- Manipulating the vials
- Uncapping the vials
- Grabbing a custom tool made specifically for adding liquid to the vials (the ‘sample-o-matic’)
- Recapping the vials
Opposite: Automated solubility screening modules (Kinova platform), source.
Vast gains in efficiency and more time for researchers
The system developed by Hein lab has been used in a workflow to determine the solubility of various compounds without human intervention. It has been capable of finding solubility values of a solid in multiple solvents overnight. These autonomous experiments were performed a lot faster than with the usual, manual method. Between one and three weeks of repetitive tasks for researchers versus two days for the platform. However, Prieto highlights the fact that the time gained wasn’t the main advantage sought but rather the quality of the results that can be accessed without human intervention. As mentioned earlier, the robot did work at night, meaning it was doing essentially what humans would have done during the day, allowing the researchers to perform value-added tasks during their lab shifts all week long.
Beyond solubility studies, the Hein Lab is moving towards using Kinova robotic arms as part of a more complex, reaction monitoring platform as well. The use of a 6-axis robot could also open the door to automating tasks such as loading and unloading centrifuges or manipulating samples for more complex operations.
A project realized majorly by women
Prieto wanted to celebrate the women who worked on the project, from the software development (Ms. Veronica Lai), to hardware integration (Ms. Parisa Shiri), and data visualization (Ms. Tara Zepel): “[...] they really drove home the project,” she said. While many find the STEM disciplines continue to lack gender parity, Prieto views the success of these women as a valuable step to encourage aspiring women in science. In the opposite picture we can see Parisa Shiri, posing with the robotic manipulator she has been working with for nearly a year.