Nanoparticles and artificial intelligence can help researchers detect pollutants in water, soil and blood
The U.S. is home to hundreds of Superfund sites, heavily contaminated areas where hazardous waste from human activity poses a cancer risk. These sites, designated by the Environmental Protection Agency, are often found in Houston, Texas, where my colleagues and I live and work. Hazardous contaminants like polycyclic aromatic hydrocarbons (PAHs) are pervasive in soil and water, making detection crucial for cleanup and environmental safety.
The EPA's standard water analysis methods are costly and time-consuming, requiring weeks of off-site work. Our chemistry research group develops more accessible and portable methods to detect toxic pollutants in soil, water, and blood. We use machine learning to identify individual compounds in mixtures without separation, comparing them to a digital database for faster on-site analysis.
Nanomaterials are extra sensitive
Imagine trying to see the width of a hair strand from the side. It's nearly impossible. Now, consider a material 1,000 times smaller. My research uses nanoparticles, microscopic objects about that size, that interact with light uniquely, similar to a magnifying glass focusing sunlight. Any substances near the nanoparticles are exposed to this focused light.
We take advantage of this by shining infrared light on the nanoparticles, causing nearby substances to absorb intense light and generate a signal. A spectrophotometer measures this signal, detecting the amount of light of a specific frequency. Any toxic pollutant near the nanoparticles will absorb more infrared light, enhancing the measurable signal.
In our lab, I create nanoparticles from metal salts solutions, dissolving them in a liquid to make an ink. This ink is painted onto glass microscope plates, leaving packed nanoparticles on the surface. A drop of contaminated water is added, and the contaminant molecules stick to the nanoparticles. The glass is then slid into a spectrophotometer to measure absorbed and emitted light, providing a signature for each contaminant.
Machine learning simplifies the analysis
Contaminated water often contains multiple compounds, complicating analysis. To prevent interference, scientists typically use sophisticated techniques to physically separate compounds, which can be time-consuming. Our team partnered with computer scientists to design tailored machine learning algorithms that find subtle patterns in the data, even to the most skilled analyst.
These methods simplify data and extract significant characteristics from each compound, helping the computer distinguish individual compounds in a mixture. Computer scientists can make these algorithms so sophisticated that no training is needed before analyzing a sample. We can measure water or soil polluted with a toxic contaminant, feed the data into the algorithms, and the machine will identify the most important features and match them to a reference database, taking only a few hours.
However, our method is not perfect. Optimizing nanoparticles for different contaminants and tweaking algorithms for various signatures are ongoing challenges. This method could screen sites for broad classes of contaminants, and future refinement could identify specific pollutant molecules.
Streamlined analysis can get the job done
Efficiently detecting hazardous pollutants in the environment can prevent human exposure. Our group's detection and data analysis techniques have been used in the field with portable instrumentation, cheaper than standard techniques. We're exploring these methods in different environmental contexts, analyzing various samples, and expanding to a wider range of pollutants.
We've filed a patent for our spectroscopy and machine learning method to analyze complex samples. While not yet commercialized, it's a possibility. Detection is just the start; after identifying a hazardous pollutant, a site must be investigated for cleanup. Our goal is to streamline detection and identification, preventing future emissions and starting cleanups faster.
