14 Dec. 2018. A chemical engineering lab designed processes with artificial intelligence to screen chemical reactions in small quantities that can reduce the time and waste in current methods. A team from New York University led by chemical and biomolecular engineering professor Ryan Hartman describe their processes in the 4 December issue of the journal Computers and Chemical Engineering (paid subscription required).
Hartman and colleagues study flow chemistry processes in micro scale to enable faster and more economical screening of chemical reactions, compared to current methods requiring large investments in plant, equipment, and time. In previous work, Hartman’s research group at NYU in Brooklyn demonstrated the feasibility of microfluidics, or lab-on-a-chip, devices that operate like miniaturized reactors to screen chemical reactions with samples comprising just a few drops, and more quickly than current larger-scale methods for chemical analysis and drug discovery.
In the new project, the NYU team added more advanced imaging and computational techniques, including those borrowed from artificial intelligence to boost the analytical power of their processes. One of the techniques is infrared thermography that sends infrared beams into a chemical compound, to measure and visualize the changes of colors representing temperature changes in the compound, particularly in small quantities. The thermographic images are then read and compiled with computer vision, a form of artificial intelligence.
The NYU system collects the thermographic images in a database where they train an artificial neural network with machine learning. Algorithms in neural networks become refined from examples they encounter, and become more expert and confident as they experience more data. In this case, the combination of infrared thermography, computer vision, and neural networks enable much faster screening of chemical reactions with much smaller sample quantities and documented high accuracy.
The researchers say this microfluidics process can provide large savings in time and waste, compared to current methods using thermocouple sensors to measure temperature and proportional–integral–derivative or PID controllers now used in the chemical industry. The current processes, say the authors, often require hundreds of liters of samples and at least 24 hours for each reaction. Their microreactor also saves a great deal of energy, as well as waste in test chemicals. The team says its system is the first microreactor guided by artificial intelligence.
“This system,” says Hartman in a university statement, “can reduce the decision-making process about certain chemical manufacturing processes from one year to a matter of weeks, saving tons of chemical waste and energy in the process.” In addition to his chemical engineering research, Hartman is a faculty advisor to NYU’s Future Labs, a business incubator for commercializing discoveries in the university’s engineering school.
More from Science & Enterprise:
- New Processes Devised to Boost Antibiotics
- Genome Editing, Informatics Lead to New Antibiotic
- Small Business Grant Aims to Cut Animals in Toxicity Testing
- Jet Fuel from Carbon Waste Powers Virgin Atlantic flight
- Graphene Extended for 3-D Printing, Medical Devices
* * *
You must be logged in to post a comment.