Sensitive detection of engineered nanoparticles (NPs) in air and in liquid samples is an important task and still a major challenge in analytical chemistry. Recent work demonstrated that it can be performed using surface plasmon microscopy (SPM) where binding of single NPs to a surface leads to the formation of characteristic patterns in differential SPM images. However, these patterns have to be discriminated from a noisy background. Computer-assisted recognition of nanoparticles offers a solution but requires the development of respective tools for data analysis. Hereby a numerical method for automated detection and characterization of images of single adsorbing NPs in SPM image sequences is presented. The detection accuracy of the method was validated using computer generated images and manual counting. The method was applied for detecting and imaging of gold and silver NPs adsorbing from aqueous dispersions and for soot and NaCl NPs adsorbing from aerosols. The determined adsorption rate was in range 0.1-40 NPs per (s mm(2)) and linearly dependent on the concentration of nanoparticles. Depending on the type of NPs and signal to noise ratio, a probability of recognition of 90-95 % can be achieved.     «

Sensitive detection of engineered nanoparticles (NPs) in air and in liquid samples is an important task and still a major challenge in analytical chemistry. Recent work demonstrated that it can be performed using surface plasmon microscopy (SPM) where binding of single NPs to a surface leads to the formation of characteristic patterns in differential SPM images. However, these patterns have to be discriminated from a noisy background. Computer-assisted recognition of nanoparticles offers a solut...     »