This study used regulatory-approved food industry cleaning agents to test removal efficiency of engineered nanoparticles from the surfaces of ready-to-eat produce. Grape tomatoes (Solanum lycopersicum) provided the food surface used for testing and titania, alumina, and silica were the engineered nanoparticles tested for removal. Twelve different industrial cleaning agents (nanopure water, peroxyacetic acid, hypochlorous acid, ammonium hydroxide/hydrogen peroxide mix, ethanol, hydrogen peroxide, peroxyacetic acid/hydrogen peroxide mix, sodium dodecylbenzene-sulfonate, potassium hydroxide, electrolyzed water, sodium chloride, and calcium chloride) were used to treat the tomato surfaces, with results being analyzed using Fourier Transform Infrared Spectroscopy and Support Vector Machine Learning algorithms. Method verification was performed using Inductively-Coupled Plasma Mass Spectrometry.

The results indicate that the combination of Fourier Transform Infrared Spectroscopy and Support Vector Machine Learning algorithms has a strong potential for use in industry as an effective means to predict and monitor nanoparticle removal from food surfaces. Two possible spectral detection mechanisms are proposed based on these findings. It was also discovered in this study that electrostatic and intermolecular forces between the nanoparticles, the food surface, and the cleaning agent used have the strongest influence on the removal success rates. This study proposes two possible mechanisms for these three competing interactions and their role in removal.