Submitted to: Biosensors World Congress
Publication Type: Abstract Only
Publication Acceptance Date: July 26, 1997
Publication Date: N/A
Microbial sensors exhibit low substrate selectivity, which often limits their practical use. However, selectivity can be improved through the use of multiple sensors with different specificities. This approach, known as pattern recognition, is widely employed in the analysis of data from low-selectivity chemical sensors. Pattern recognition requires that sensors be thoroughly calibrated for each substrate and mixture of substrates. Calibration is appreciably simplified, and the precision of measurement enhanced, if information is available on the additivity of sensor responses. For a two-component mixture, additivity can be defined as the sensor response to a combination of analytes in a single sample divided by the arithmetical sum of sensor responses to each of the analytes measured separately. The goal of the current study was to characterize the additivity of an amperometric sensor based on bacterial cells of Gluconobacter oxydans. Glucose and ethanol were chosen as model analytes. Improved methods for differential detection of these compounds would have numerous biotechnological applications. The enzyme-based glucose analyzer Ekzan-G was used as a reference biosensor. G. oxydans biosensors exhibited nearly complete additivity for total glucose plus ethanol concentrations of up to 0.6 mM. Additivity fell to approximately 80% for total concentrations of 0.6 to 1.5 mM. Essentially, no additivity was found for substrate concentrations greater than 3.0 mM. Within the linear range of additivity for the G. oxydans biosensor, the two-sensor system provided estimates of ethanol concentration with an error of no more than 8%.