Using positive matrix factorization to unmix PAH fingerprints in contaminated sediments

Area of Science:

• Environmental Chemistry
• Geochemistry
• Environmental Forensics

Background:

• Apportioning polycyclic aromatic hydrocarbon (PAH) remediation costs at contaminated sediment sites is challenging due to issues like missing source samples and mixed PAH signatures.
• Identifying PAH sources is crucial for tracking pollution and assigning responsibility.

Purpose of the Study:

• To demonstrate the effectiveness of Positive Matrix Factorization (PMF) for unmixing PAH fingerprints in contaminated urban sediments.
• To establish a method for identifying PAH source classes (petrogenic, pyrogenic, runoff) to aid in cost apportionment.

Main Methods:

• Utilized a large dataset of over 700 PAH samples from contaminated urban sediment sites.
• Applied a 3-factor Positive Matrix Factorization (PMF) analysis to identify distinct PAH end-member source fingerprints.
• Validated the PMF results by numerically mixing identified end-member contributions to replicate measured sediment PAH fingerprints.

Main Results:

• Successfully identified three distinct PAH end-member source fingerprints: petrogenic, pyrogenic, and runoff/weathered.
• Quantified the percent contribution of each identified source to individual sediment samples.
• Demonstrated that PMF can accurately replicate measured PAH fingerprints using calculated mixing percentages.

Conclusions:

• Positive Matrix Factorization (PMF) provides a robust method for unmixing complex PAH fingerprints in contaminated sediments.
• This approach is a significant step towards accurately apportioning PAH contamination and associated remediation costs.
• The identified PAH source classes and their contributions offer valuable data for environmental forensic investigations.