Antonia Praetorius, Alexander Gundlach-Graham, Eli Goldberg, Willi Fabienke, Jana Navratilova, Andreas Gondikas, Ralf Kaegi, Detlef Günther, Thilo Hofmann and Frank von der Kammer
Environmental Science: Nano, 2017, DOI: 10.1039/C6EN00455E
Antonia Praetorius and coworkers from University Vienna, ETH Zurich, EAWAG Duebendorf and University of Gothenburg describe a new method for the analysis of large single-particle mass spectrometry data sets recorded with the icpTOF. Machine learning is used to group multi-element signals and to distinguish natural from engineered nanoparticles (NPs).
The detection of engineered nanoparticles (ENPs) in environmental samples is one of the biggest challenges for ENP monitoring and, ultimately, risk assessment. 17 features in a multi-element signal were found to be necessary in order to successfully classify and distinguish engineered from natural NPs.
The authors write: “We present a groundbreaking new approach using an inductively-coupled plasma time-of-flight mass spectrometer (ICP-TOFMS) in single-particle mode, capable of simultaneous multi-element analysis, coupled with a machine learning data treatment. …. We are able, for the first time, to distinguish ENPs from orders of magnitude higher concentrated NNPs based on their individual multi-element fingerprints.”
