3.1. Response of the gas sensor array
The gas sensor array response pattern from a cultivation of microorganisms may have various origins. The sensors can, if sensitive enough, respond to specific volatile compounds emitted by the microorganisms in the culture or to the emission from the nutrient components of the medium (Bachinger and Mandenius 2000).
In order to determine the sensitivity of the gas sensor array to the nutrient components of the medium as well as to the volatile compounds emitted by the cells, the bioreactor was filled with 1 L of the medium and a glucose solution (to a final concentration in bioreactor of 7 g/L) and five sampling cycles were performed from the headspace of the bioreactor. Exactly before the 6th sampling cycle, the suspended cell was added and the sampling cycles were performed every 5 minutes until the end of the cultivation process (Fig.5).
Fig.5 (a) indicates the off-line as well as the simulated evolution of glucose, ethanol and biomass during the cultivation (BC2). The parameters of the model were acquired by least squares fitting to the off-line measurements. Fig.5 (b) illustrates the raw signal response of one of the gas sensors during the same cultivation.
In Fig. 5 (b), the five peaks before \(T_{\text{start}}\) indicate the response of the sensor to the components of the medium. At\(T_{\text{start}}\) the suspended cells are added to the bioreactor (inoculation time), \(T_{\max}\) is the peak with the highest value and\(T_{\text{end}}\) is the last peak where the ethanol is depleted and whose height is the same as before \(T_{\text{start}}\).
High sensitivity of the sensor to ethanol produced during the cultivation process can be described by comparing the general pattern of the peak heights (Fig.5 (b)) with the ethanol concentration (Fig. 5 (a)). In Fig. 5 (a), while the glucose is decreasing during the glucose phase, the products (ethanol and biomass) are increasing. At around 2 h, the glucose is depleted and the cultivation shifts to the ethanol consumption phase. After around 7 h, the ethanol is also depleted and no more ethanol remained in the culture broth. Similar behavior of the sensor response can be observed from Fig. 5 (b). In Fig. 5 (b), the peak height before \(T_{\text{start}}\) and after \(T_{\text{end}}\) is the same. In addition, there are no changes in the peak heights after\(T_{\text{end}}\) (the region where all the ethanol is consumed). From these illustrations it can be seen that the sensor is highly sensitive to ethanol but not so sensitive to the other components of the cultivation medium.