Abstract
An array of ZnO thin film sensors was obtained by thermal oxidation of physical vapor deposited thin Zn films. Different conditions of the thermal treatment (duration and temperature) were applied in view of obtaining ZnO sensors with different gas sensing properties. Films having undergone a long thermal treatment exhibited high responses to low ethanol concentrations, while short thermal treatments generally led to sensors with high ethanol sensitivity. The sensor array was used to distinguish among Tequilas and Agave liquor. Linear discriminant analysis and the multilayer perceptron neural network reached 100% and 86.3% success rates in the discrimination between real Tequila and Agave liquor and in the identification of Tequila brands, respectively. These results are promising for the development of an inexpensive tool offering low complexity and cost of analysis for detecting fraud in spirits.
Highlights
In the last five years, Tequila, a beverage with appellation of origin, has seen increased its exports to 182.9 million L
This data matrix was further analyzed by linear discriminant analysis (LDA) and by the multilayer perceptron neural network (MLP)
Misclassification of three samples occurred between COR, 180 and ART Tequilas. These results confirm that the four-element ZnO sensor array is useful for correctly discriminating real Tequila from Agave liquor and that it shows some potential for identifying Tequila brands
Summary
In the last five years, Tequila, a beverage with appellation of origin, has seen increased its exports to 182.9 million L. Other alcoholic beverages, generally designated as “Agave liquor” use cheaper alcohol sources and, instead of conducting the maturation process, mix the distilled alcohol with caramel as colorant and artificial flavors that resemble those naturally occurring in real Tequila. It is not uncommon for bottle labels employed in Agave liquors to display paintings related to agave. The as-deposited samples were treated thermally under different temperature conditions and time duration with the aim of obtaining different ZnO films These ZnO sensors integrated an array for analyzing the volatile fraction of alcoholic beverages (Agave liquor and Tequilas). Obtained from the test were treated either by linear discriminant analysis (LDA) or the non-linear back-propagation multilayer perceptron artificial neural network (BP-MLP)
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
