Abstract
Mango fruit from Latin America (cv. Tommy Atkins), were purchased from a local Florida supermarket, homogenized, and sampled for volatile analysis by static headspace method. Some of the material was analyzed using an electronic nose (e-nose) with metal oxide coated or uncoated sensors (500 µL injection volume) and some by gas chromatography (GC) equipped with a polar Carbowax column and a flame ionization detector. Dilution of homogenate and homogenate volume were analyzed to determine effect on e-nose and GC headspace measurements. Mango homogenate (1.0, 1.5, and 2.0 mL) was diluted with DI water to 50, 25, and 12.5% of original concentration. The resulting e-nose signal intensities (changes in resistance across the metal oxide sensor due to non-selective interactions with volatile compounds in the headspace) were analyzed by discriminant factor analysis (DFA), which resulted in grouping by dilution factor, regardless of sample size. A combination of 2.0 mL and 25% dilution of mango homogenate was determined to be optimal. These results were compared to analysis of 13 characteristic mango volatiles by gas chromatography (GC) headspace analysis of the mango homogenate for the same volume/dilution combinations. Concentration of volatiles in the headspace generally increased with volume and decreased with dilution, but there were some exceptions and inconsistencies. The increase in headspace concentration was not directly proportional to the homogenate volume, indicating matrix effects on aroma partitioning into the headspace, which varied for different compounds. Whole mangoes (cv. Keitt and Kent) harvested in Homestead, Fla., were put in sealed containers for 3 hours to accumulate enough volatiles for headspace analysis. A large injection volume injected into the e-nose (2000 µL) was necessary to get ample signal and reproducible results, and separated the two varieties based on their volatile emission to the headspace.
