Introduction The sensing performance of mixed-potential gas sensors is highly dependent on the electrode microstructure. Recently, much effort has been made to improve the sensor performance by engineering of the electrode microstructure [1, 2]. Nevertheless, the different microstructure parameters, such as thickness, porosity, and three phase boundary (TPB), contribute jointly to the sensing response, and their respective contributions are usually difficult to identify. Thus, deviation of the apparent sensor behavior from the reality is usually unable to judge due to the uncertain effect of electrode catalytic reaction. In order to unravel the true response behavior, it would be necessary to reduce or minimize impact of the electrode diffusion-reaction process and control the microstructure, which could be achieved by using a very thin electrode with well-defined and controllable morphology. Herein, we fabricated thin ordered SnO2 sensing electrodes on YSZ substrate by employing a Polystyrene (PS) sphere-template method [3]. By varying the PS sphere diameter, three sensors of different electrode thickness and electrode/electrolyte interface were obtained. The gas sensing characteristics of the sensors were systematically studied. The sensing behavior was discussed in relation to variation of the electrode process and TPB density. Sensor fabrication and characterization As depicted in Figure 1, highly-ordered porous thin film sensors were prepared by using PS sphere templates in 200 nm, 500 nm and 1000 nm diameters. SEM, TEM and AFM images of SnO2 porous film electrodes were presented in Figure 2. Periodically ordered films of inverse opal structure were observed, with uniform circular pore openings and interconnected walls as the skeleton. Each film was porous and a homogenous assembly of 10-20 nm sized particles. The wall height was roughly half of the PS sphere size, which increased from 121 nm to 248 nm and 448 nm for the three PF films. Sensing performance Figure 3a and b show the response values and response time of the three PF sensors as a function of hydrogen concentration at 500 °C. Obviously, the response significantly increased, i.e., became more negative, with decreasing template diameters, while the response time of the three sensors was short and generally rather close to each other, with a value of 4.5 s-6 s for 500 ppm H2 at 500 °C. Figure 3c displays the cross-sensitivities of the PF sensors to 100 ppm various gases at 450 °C. The PF sensors exhibited much poorer H2 selectivity than regular thick film sensors [4].The small film thickness and porous nature of the PF SE are highly favorable for the gas transport from the gas phase to the TPB, which are beneficial to the response kinetics. That the three PF sensors exhibited very close response time despite their minor difference in the film thickness, suggesting absence of significant electrode process and large H2 concentration reduction. This may allow us to study the effect of the TPB density on the sensing response. The TPB length was thus estimated for the PF sensors (Figure 4a). Clearly, the TPB density increased from 56.8 μm-1 to 89.7 μm-1 with the decreasing template diameters. It can also be seen that for each H2 concentration, the response increased (became more negative) monotonically with increasing TPB density (Figure 4b-c). It should be noted that thin film favors fast response kinetics and large response but may be disadvantageous to the selectivity. To achieve good overall sensing performance, e.g., fast response kinetics, large response, and high selectivity, careful tuning of the film thickness and TPB density would be demanded.
Read full abstract