Swift and significant gains against climate change require the creation of novel, environmentally benign, and energy-efficient materials. Considered one of the richest veins researchers hope to tap in creating such useful compounds is an enormous chemical space where molecular mixtures that supply remarkable optical, conductive, magnetic, and warmth transfer properties await discovery.
But finding these latest materials has been slow going.
“While computational modeling has enabled us to find and predict properties of recent materials much faster than experimentation, these models aren’t at all times trustworthy,” says Heather J. Kulik PhD ’09, associate professor within the departments of Chemical Engineering and Chemistry. “To be able to speed up computational discovery of materials, we’d like higher methods for removing uncertainty and making our predictions more accurate.”
A team from Kulik’s lab set out to handle these challenges with a team including Chenru Duan PhD ’22.
A tool for constructing trust
Kulik and her group deal with transition metal complexes, molecules comprised of metals found in the midst of the periodic table which might be surrounded by organic ligands. These complexes might be extremely reactive, which supplies them a central role in catalyzing natural and industrial processes. By altering the organic and metal components in these molecules, scientists can generate materials with properties that may improve such applications as artificial photosynthesis, solar energy absorption and storage, higher efficiency OLEDS (organic light emitting diodes), and device miniaturization.
“Characterizing these complexes and discovering latest materials currently happens slowly, often driven by a researcher’s intuition,” says Kulik. “And the method involves trade-offs: You would possibly find a fabric that has good light-emitting properties, however the metal at the middle could also be something like iridium, which is exceedingly rare and toxic.”
Researchers attempting to discover nontoxic, earth-abundant transition metal complexes with useful properties are likely to pursue a limited set of features, with only modest assurance that they’re on the suitable track. “People proceed to iterate on a selected ligand, and get stuck in local areas of opportunity, somewhat than conduct large-scale discovery,” says Kulik.
To deal with these screening inefficiencies, Kulik’s team developed a brand new approach — a machine-learning based “recommender” that lets researchers know the optimal model for pursuing their search. Their description of this tool was the topic of a paper in in December.
“This method outperforms all prior approaches and may tell people when to make use of methods and after they’ll be trustworthy,” says Kulik.
The team, led by Duan, began by investigating ways to enhance the traditional screening approach, density functional theory (DFT), which relies on computational quantum mechanics. He built a machine learning platform to find out how accurate density functional models were in predicting structure and behavior of transition metal molecules.
“This tool learned which density functionals were essentially the most reliable for specific material complexes,” says Kulik. “We verified this by testing the tool against materials it had never encountered before, where it the truth is selected essentially the most accurate density functionals for predicting the fabric’s property.”
A critical breakthrough for the team was its decision to make use of the electron density — a fundamental quantum mechanical property of atoms — as a machine learning input. This unique identifier, in addition to the usage of a neural network model to perform the mapping, creates a robust and efficient aide for researchers who want to find out whether or not they are using the suitable density functional for characterizing their goal transition metal complex. “A calculation that will take days or even weeks, which makes computational screening nearly infeasible, can as a substitute take only hours to provide a trustworthy result.”
Kulik has incorporated this tool into molSimplify, an open source code on the lab’s website, enabling researchers anywhere on the earth to predict properties and model transition metal complexes.
Optimizing for multiple properties
In a related research thrust, which they showcased in a recent publication in , Kulik’s group demonstrated an approach for quickly homing in on transition metal complexes with specific properties in a big chemical space.
Their work springboarded off a 2021 paper showing that agreement concerning the properties of a goal molecule amongst a gaggle of various density functionals significantly reduced the uncertainty of a model’s predictions.
Kulik’s team exploited this insight by demonstrating, in a primary, multi-objective optimization. Of their study, they successfully identified molecules that were easy to synthesize, featuring significant light-absorbing properties, using earth-abundant metals. They searched 32 million candidate materials, one among the biggest spaces ever looked for this application. “We took apart complexes which might be already in known, experimentally synthesized materials, and we recombined them in latest ways, which allowed us to keep up some synthetic realism,” says Kulik.
After collecting DFT results on 100 compounds on this giant chemical domain, the group trained machine learning models to make predictions on your entire 32 million-compound space, with a watch to achieving their specific design goals. They repeated this process generation after generation to winnow out compounds with the specific properties they wanted.
“Ultimately we found nine of essentially the most promising compounds, and discovered that the particular compounds we picked through machine learning contained pieces (ligands) that had been experimentally synthesized for other applications requiring optical properties, ones with favorable light absorption spectra,” says Kulik.
Applications with impact
While Kulik’s overarching goal involves overcoming limitations in computational modeling, her lab is taking full advantage of its own tools to streamline the invention and design of recent, potentially impactful materials.
In a single notable example, “We’re actively working on the optimization of metal–organic frameworks for the direct conversion of methane to methanol,” says Kulik. “This can be a holy grail response that people have desired to catalyze for a long time, but have been unable to do efficiently.”
The opportunity of a quick path for transforming a really potent greenhouse gas right into a liquid that is well transported and might be used as a fuel or a value-added chemical holds great appeal for Kulik. “It represents one among those needle-in-a-haystack challenges that multi-objective optimization and screening of hundreds of thousands of candidate catalysts is well-positioned to resolve, an impressive challenge that’s been around for thus long.”