Search

The potential of the electronic nose (E-nose) and the mass spectrometer based electronic nose (MSE-nose) to monitor changes in aroma profile during shelf life was studied. Apples were stored for eight months at three different storage conditions and the aroma profile changes were followed subsequently over a period of 15 days. Gas chromatographic headspace analysis were conducted as reference measurements. A canonical variate (CV) analysis showed a clear effect of storage and shelf life both for the MSE-Nose measurements and the GC measurements. However for the E-Nose measurements only a shelf life but no storage history effect was observed. It was also found that the aroma profile changes during shelf life depended on the storage history. Partial least squares models were built to predict the apple firmness and the number of days in shelf life. It was found that the models based on the E-nose data had a worse prediction performance compared to those based on the MSE-nose data. Fusion of both sensors did not result in improved models. Both for firmness (0.95) and days of shelf life (0.98) a high cross validation correlation was observed between measured and predicted values. The SEP of firmness and shelf life were 2.38 Hz2 g2/3 × 106 and 1.02 days, respectively.

Arrays of metallo-porphyrin coated quartz microbalance (QMB) sensors have been often used in the past for aroma based quality measurements. However, little information on the interaction of the volatiles with the different QMB coatings is known. In this study a calibration protocol was designed in order to characterize the QMB sensors in a fast and easy way. A simplified BET adsorption isotherm was used to describe the response signals of the sensors as a function of the concentration of 2 alcohols and 3 esters. In this way the sensitivity of each sensor with respect to a particular component could be characterized by only one parameter. Two of the 8 sensors showed more affinity to alcohols than to esters, while two other sensors detected esters better than alcohols. Within one class of compounds (e.g. esters), none of the sensors could differentiate between the tested compounds. Since in most electronic nose applications the headspace consists of complex mixture of volatiles, interactions of these volatiles might occur. These possible synergetic or competitive effects were tested by calibrating the electronic nose with mixtures of two chemical compounds. A combined, simplified BET adsorption isotherm described the data well. None of the three tested mixtures (alcohol-alcohol, ester-ester and alcohol-ester) showed any synergetic or competitive effect in the measured concentration range. The designed calibration protocol can be used for characterizing the sensors when developing new applications.