RESUMO
The canine olfactory system is a highly efficient and intricate tool often exploited by humans for detection for its many attributes, including impressive sensitivity to trace analyte vapors. Canine detectors are often touted as having lower limits of detection, or olfactory detection threshold (ODT), than other field-relevant detection technologies; however, previous attempts to quantify canine ODTs have resulted in reported estimates spanning multiple orders of magnitude, even for the same analyte. A major contributor to these discrepancies is the vapor delivery method used for testing, where losses due to adsorption and dilution are often unaccounted for, and the presence of unattended compounds in the vapor stream due to carryover may go unnoticed. In this research, a trace vapor generator (TV-Gen) was used to deliver quantitatively accurate amounts of vapor reproducibly over time for canine testing. Analyte losses due to adsorption to surfaces in the flow path, dilution in the sniff port at the outlet, and analyte carryover were considered. Computational fluid dynamic (CFD) modeling was used to visualize analyte vapor spread throughout the port. CFD simulations revealed the need for a diffuser to encourage the diffusion of the analyte throughout the port. As a result, the modified vapor generator provides analyte air as a diffuse flow that is evenly distributed through the custom sampling orifice, as opposed to a narrow stream of air at the chosen concentration which exits directly into the environment. Laboratory validations were carried out for three analytes, amyl acetate, 2,4-dinitrotoluene (DNT), and methyl benzoate. A linear response across more than two orders of magnitude vapor concentration range was achieved for all analytes. These efforts will be applied in further research utilizing this TV-Gen vapor delivery system for canine ODT testing, eliminating many quantitative changes seen previously. Graphical abstract.