This conclusion will be reinforced by Figure 4. Before, we will shortly consider the response trends of the various
sensors. In Figure 3 the response-temperature data for 100 ppm acetone concentration are shown. It is confirmed
that, with respect to pure TiO2, the TiO2-WO3 devices showed a higher response already at 100 °C. But, when the
operating temperature was increased, the response rose up to two orders of magnitude of conductance variation,
comparable or even higher than that of pure WO3, for the highest W concentration. Such sample indeed displayed
the best response over a broad range of operating temperatures. It will be noted that the response data resemble a bell
shaped curve, despite very distorted, with the maximum for pure TiO2 displaced to very high temperatures. It is then
further suggested that W addition resulted in catalytic effect. Figure 4 shows the calibration curves for 350 °C
operating temperature. It is now quantitatively clear that while pure TiO2 was almost inactive, W addition boosted
the sensor response. The slope of the response curves can be associated to the sensing mechanisms [5]. The Wmodified
devices display almost parallel curves, different from the calibration curve of TiO2. Despite the
consideration of the slope by itself may be misleading, in absence of detailed knowledge of the reaction steps,
coupling this observation with the previously noted recovery times reinforced the hypothesis that similar reaction
mechanisms were involved in the W-modified samples, and hence that the W oxide layer is the active component in
the sensing material. The results presented in this work show that proper use of the precursor chemistry may result in
enhanced sensing architectures, taking advantage of the catalytic coupling between a less active core and a
monolayer of different, more active oxides.