Smarter, faster understanding of chemical mixtures
IBM is putting AI driven supercomputer simulations into the hands of industry chemists.
Our researchers have enabled industrial scientist to run simulated versions of the time-consuming experiments they do in their laboratories. It provides novel insights and predictive models, and it reduces the time it takes to run them.
Formulation chemists work with substances that don't react with each other but that combine as a mixture in useful ways. They want to know what will happen with a selection of chemicals combined in a certain ratio, so they can create the right product at the right price - including paints, cosmetics, inks, adhesives, detergents, and pesticides, to name but a few.
Cutting-edge computational science and supercomputing power can predict the effects of changing these ratios, or of adding and removing chemical components. At the core of this work are the models we employ to represent the chemistry and physics that occur in these mixtures. We employ a wide variety of them to achieve efficient predictions, from physical simulations to artificial intelligence (AI) sampling methods.
All of these models require parameters and settings to operate optimally. When modelling chemistry, this means setting how each atom interacts with all the other atoms to best represent reality. It's a task that has consumed decades of academic research.
See our demo of how we couple advanced model generation methodologies with fully automated workflows, and can achieve results in hours and days, rather than months and years.