Discrimination of Allium headspace volatiles affected by variations in genotype growing environment and storage using an electronic noses

Alliums are valued mainly for their unique organosulphur-derived flavours and aromas. Traditional sensory and analytical determinations of Allium quality are constrained by high cost, technical difficulties and, time and human limitations. This thesis investigates the potential for use of relatively novel electronic nose (E-nose) technology for Allium discrimination. Chapters 3, 4 (Sections 4.1 to 4.3), 5 (Sections 5.1 and 5.2) and Appendices II and III inclusive have been published or submitted for publication. Consequently, Chapters in this thesis are presented in the form of papers. The E-nose AromaScan LabStation A32/8S (Osmetech Pic., UK) consists of 32 conducting polymer miniature sensors. Adsorbed odour molecules alter the electric conduction mechanism of the sensor polymer. The response is measured as proportional (%) change in sensor resistance ratio (%dR/R). The E-nose discriminated Allium types (Chapter 3), varieties of spring onion grown with or without sulphur addition and a single variety of spring onion grown under different levels of sulphur, nitrogen, water-deficit stress and soil type (Chapter 4). Bulb onion affected by nitrogen, sulphur and soil type and diced onion sealed in polyethylene bags stored at 4°C for 9 days were also discriminated by the E-nose (Chapter 5). A descriptive model for the direction of E-nose sensor polymer response to Allium headspace volatiles affected by genotypic differences and edaphic variables was outlined in Section 6.2. Principal Component Analyses (PCA) of E-nose data sets output accounted for >75% to nearly 100% of variations in the Alliums. The variations in Allium genotype differentially affected the E-nose sensor conductivity following headspace volatiles interaction with sensor polymer element. Classification of data sets output showed greater (Mahalanobis distance statistic, D² >3.0) sensitivity of spring onion cvs Guardsman and Fragrance to S fertilisation while the headspace volatiles characteristics of cvs Winter Over and Paris Silverskin were not significantly (D²<3.0) altered by S. The headspace volatiles of onion bulb cv. Sprinters also responded to S fertilisation (D²>3.0) and thus, increased %dR/R. Overall, N-fertilised onion cv. Sprinters reduced E-nose sensor conductivity leading to an increase in %dR/R. Increases in water-deficit stress i.e. > -0. 80 MPa soil water potential, SWP generally reduced separation between E-nose data set clusters for clay versus sandy loam soils from D² = 43.2 for -0.01 MPa SWP to D² = 6.2 for -1.19 MPa. Headspace volatiles of onions grown in the glasshouse clay increased %dR/R compared to reduced %dR/R values for both glasshouse and field sandy loam soils. The E-nose detected gradual changes in headspace volatiles of diced onion wrapped in polyethylene bags stored at 4°C for 9 days. The changes in headspace volatiles reduced %dR/R values while data set cluster separations with reference to day 0 for each sampling time increased from D² = 3.6, 5.8 and 7.0 on days 3, 6 and 9, respectively. The results suggested that Allium quality can be assessed with ease along production, postharvest and marketing chains compared to traditional destructive methods. Linear correlations for E-nose data sets versus Allium pungency determinants (pyruvic acid and lachrymatory potency), total soluble solids and dry-matter were poor. The thesis discusses the commercial significance of the result and its implication for the development of E-nose sensor tailored for Alliums. This would promote application and use of E-nose technology in the Allium industry, germplasm evaluation, and discrimination of agronomic variables and possibly, monitoring spoilage pathogens during storage. The effects of nitrogen, sulphur, water-deficit stress and soil type and their interactions have given new insight into agronomic inputs on growth and microbial load (Chapters 4.3, 5.1 and Appendix III).