Chemical Pollution's Blind Spot: How CMU Researchers Are Filling the Gaps

Everyday items like carpets, shower curtains, cosmetics and even food can have harmful chemicals hidden inside that can lead to long-term health effects.

The number of synthetic chemicals and products being used and produced that can contaminate the environment during their lifecycle has risen over the past 30 years. This includes compounds such as halogenated flame retardants, polymers, and per/polyfluoroalkyl substances (PFASs), also known as "forever chemicals" because they persist in the environment and human body.

"The old viewpoint of pollution is that it's coming from external sources, like a factory or a powerplant, but it's also coming from the flame retardants in my cellphone that I touch hundreds of times a day," said Ryan Sullivan, associate director of Carnegie Mellon University's Institute for Green Science and a professor of chemistry and mechanical engineering. "We really want to do a much better job of reducing the amount of harmful chemicals that get into the environment, into organisms and into humans that cause health damage. The best way to do so is by not making chemicals that are too persistent or toxic in the first place."

Some of the pharmaceuticals, PFASs and other chemicals can transform into even more problematic compounds, but it is hard to identify these transformation products using standard approaches. Sullivan, Carrie McDonough, assistant professor of chemistry; Olexandr Isayev, the Carl & Amy Jones Professor of Interdisciplinary Science; and Ana Torres, assistant professor of chemical engineering, are working to change that.

The Rise of Synthetic Chemicals

More than 10,000 synthetic chemicals are used to make plastic products, and hundreds of thousands of chemicals are used in other industries, Sullivan said.

Advancing sustainable chemistry requires the capability to predict what changes chemicals may undergo and what happens to the resulting transformation products. Sustainable chemistry also promotes the use of chemicals and synthetic methods that reduce energy needs, are more efficient, have reduced costs and avoid generating harmful environmental pollutants, thus reducing the cost of harmful health effects and lost lives.

Sullivan is the primary investigator on this initiative. Along with McDonough, Isayev, and Torres the researchers are developing high-throughput experimental and computational methods that acquire the chemical data needed to inform environmental molecular lifecycles.

The researchers said that these environmental and biological chemical reaction networks will screen for transformation products that are likely persistent or prone to bioaccumulation. This more rapid chemicals analysis will then guide sustainable chemical assessment such that likely harmful chemicals can be identified much earlier in their development lifecycle.

"Our big focus is to assess the ability to predict the fate of a molecule under environmental and biological conditions. What does it turn into? And where does it go?" Sullivan said.

Sullivan said that by understanding how molecules can transform and how quickly under different conditions, the data can inform how long they'll exist in the environment.

"It's an understudied — but important — topic," he said. "We need to work to connect the world of environmental chemistry and what we know and how we use that information to better evaluate and design chemicals and materials that are more sustainable in nature."

Unveiling End of Lifecycle Facts

Current analyses generally focus on the earlier portion of the lifecycle of molecules, such as the extraction of resources, manufacturing and distribution of products. Sullivan said that the work doesn't account for damages caused at the end stages of the lifecycle during use and then after disposal.

"This is a huge blind spot," Sullivan said. "And it gets to the heart of the problem with plastic pollution. All the damage caused by forever waste is never factored into the price that we pay or what a store pays to buy plastic bags."

Torres will provide expertise on how to advance lifecycle analysis applied to evaluate chemical alternatives and their sustainability. In particular, she will focus on the later life stages for during and end of use that are usually neglected due to the lack of needed data to estimate these costs.

McDonough said that Carnegie Mellon is uniquely equipped to tackle this problem.

"We're very strong in computational chemistry and one of the few U.S. universities with a chemistry department that focuses on environmental chemistry, so it's an ideal mixture of experts," said McDonough, who studies how PFASs accumulate in organisms and brings expertise in environmental chemistry and bioanalytical techniques. For this work, she will focus on developing and evaluating experimental approaches to determine chemical properties that describe bioaccumulation potential and identify biological transformation products.

"Add in our automated science initiative, high throughput capabilities and libraries of compounds for testing, and we're well positioned to learn how molecular structure can predict downstream environmental and health impacts," she said.

The scope of the work is large and the combined experimental and computational high-throughput analysis will allow them to evaluate many chemicals more quickly. The team is hoping to find funding to support a postdoctoral researcher as well as several graduate students to support the research.

Big Data on Minuscule Molecules

The collaborators are expecting to work with a massive volume of data.

"We use machine learning to guide high-throughput experiments and scale the amount of experimental data that we can obtain," said Isayev, a member of the Department of Chemistry with a joint appointment in the Department of Materials Science and Engineering and is affiliated with the Department of Biological Sciences.

Isayev, Sullivan and McDonough are involved with Carnegie Mellon's automated science initiative, which is developing the future of autonomous laboratories. Gabe Gomes, an assistant professor in the Departments of Chemistry and Chemical Engineering, and students already have demonstrated that AI systems can autonomously plan, design and execute within minutes a chemical reaction so complex it earned its human inventors a Nobel Prize in 2010.

A major challenge that the team's high-throughput methods will address is reducing the time and effort required to measure or predict all the environmental chemical properties for each chemical, and its persistent transformation products.

Isayev's group brings data-driven methods to accelerate simulations to support creating more effective and efficient simulations. The collaboration on sustainability expands on work from Isayev's group.

"We've done drug discovery to predict molecular behavior, and the physics for this work is similar but with slightly different applications," Isayev said. "Our experience will help us predict bioaccumulation of chemicals in various organisms, sediment, soils, water and plants."

For this work, Isayev is focusing on developing and evaluating the new computational approaches to predict environmental chemical properties from structure. He also is leading the development of chemical reaction networks that will consolidate data from experiments, computational analysis and literature.

Building a global resource for safer chemistry

With the information collected, the researchers aim to build a library to share the methods and data including predictions about the environmental impact and sustainability of chemicals to help inform regulatory agencies, public health experts and other researchers.

The combined experimental and computational data will be used to predict the environmental molecular lifecycle of a given chemical, which provides essential information to evaluate the sustainability and hazards of any chemical much earlier in its development lifecycle, reducing barriers in inventing safe and sustainable chemicals.