Project Summary (2025-2030)

Exposure to polycyclic aromatic hydrocarbons (PAHs), polycyclic aromatic compounds (PACs), per-and polyfluoroalkyl substances (PFAS) and volatile organic compounds (VOCs) have all been linked to a large number of human health risks. Current analytical techniques to detect and identify the chemicals in these families are invariably slow, complex, and require expensive instrumentation and extensive sample preparation. The researchers propose an entirely new approach to the chemical detection and identification of environmental contaminants in complex mixtures. Their new strategy combines optical spectroscopic techniques such as Surface Enhanced Raman Spectroscopy (SERS) and Surface Enhanced Infrared Absorption (SEIRA), combined together on a single platform, fully integrated with Machine Learning (ML) algorithms, to advance the detection of chemicals in these families of environmental contaminants.

The team plans to build on previous studies, where they showed that individual chemical components in complex mixtures can be identified by a combination of spectroscopic signatures of the chemical mixture combined with ML, an approach they call “Computational Chromatography”. Having demonstrated computational demixing and identification of the components in simple chemical mixtures, the researchers plan to focus on expanding this technique to identify multiple individual components in more complex, real-world samples. This will be accomplished by high-throughput spectral sampling to refine the accuracy and complexity of ML algorithmic approaches.

In comparison with PAH identification, PAC identification is substantially more challenging, since a single PAH may be converted by multiple processes in the environment to several different PACs, some of which may be far more water soluble and bioactive than their PAH precursor. Relative to PAHs, PACs are far less well characterized. The researchers plan to develop an experimental-theoretical capability of predicting and verifying PAC spectra for oxo-PAHs and nitro-PAHs, for chemical detection and identification equivalent in accuracy to that of the PAHs in complex mixtures. They also plan to develop chemically-specific nanoantenna-based sensors that can be tailored to detect specific chemicals of interest, based on IR-active antenna structures designed for identifying a targeted chemical of interest. The team will pursue this approach for rapid identification of PFAS from biological and environmental samples. These approaches will also be extended to the identification of VOCs, where a combination of “purge- and-trap” capture strategies will be utilized and combined with computational chromatography approaches for rapid VOC identification.

The ultimate outcome of this project is the creation of streamlined, ultracompact, ultrasensitive chemical analysis and detection platforms, capable of identifying multiple biological/environmental contaminants of interest in a single sample without costly separation and purification steps, which could be potentially transitioned to fieldable use.