Use of an electronic nose to measure odour concentration following application of cattle slurry to grassland

Three experiments were conducted in which odour samples were taken at time intervals following application of cattle slurry to grassland. Measurements were made of odour concentration by dynamic dilution olfactometry and of sensor response using two electronic nose instruments with conducting polymer type sensor arrays (the Aromascan and the Odour-mapper). Significant linear relationships (P < 0.01) were found between odour concentration and average sensor response, a single line being fitted to data from all three experiments, with % variance accounted for of 59% and 62% for the Aromascan and Odour-mapper respectively. Sensor responses were normalized, the response of each sensor being expressed as a percentage of the summed response across all sensors, and principal components analysis (PCA) used as a means of comparing the normalized response patterns. Normalized response patterns from samples from the three experiments were not distinctly different and could all be classed as the same ''odour type'' by the electronic nose instruments. Actual response patterns (before normalisation) were also compared using PCA and again there was no difference between experiments (after taking account of background samples) but a concentration effect was apparent, with response patterns for odour samples taken immediately following slurry application being distinctly different from patterns for subsequent samples. The experiments demonstrated the ability of electronic nose instruments with conducting polymer type sensors to respond to agricultural odours at much lower concentrations than had previously been achieved so leading to the possibility of developing a portable instrument for odour measurement in the field. (C) 1997 Silsoe Research Institute.