Surface Leaching of Pd-Containing SnO2 Gas Sensors Enhances Sensitivity

Abstract

Introduction Palladium is commonly used to enhance the performance of chemoresistive metal-oxide gas sensors [1]. Typically, this enhancement is attributed to the presence of Pd clusters on the surface of their metal oxide support (e.g. SnO2). More specific, small Pd clusters on SnO2 were associated to enhanced sensitivity to H2 due to fermi-level control [2] or spillover effects [1]. However, Pd may not be present in metallic but oxidic form, that is catalytically active as well [3]. E.g. PdO can form during fabrication or upon the typical annealing [4] at 400 - 700 °C [3] for a few hours prior to gas sensing. Furthermore, possible Pd incorporation or embedding into the support rarely has been considered [5].Here, SnO2 particles with different Pd contents were prepared by flame spray pyrolysis (FSP). Their surface Pd was removed through leaching with HNO3 to guaranty the surface and embedded Pd content was measured. Lastly, the influence of surface and embedded Pd on SnO2 gas sensors was evaluated with acetone, ethanol and CO at 350 °C and 50% relative humidity. Method Pristine and Pd-containing (0 - 3 mol%) SnO2 nanoparticles were produced by FSP and annealed for 5 h at 500 °C. To remove all Pd from the surface, the particles were reduced in 5% H2 in Ar at 150 °C for 30 min and then leached in 10% HNO3 in water at 60 °C for 4 h (Figure 1a). Inductively coupled plasma-optical emission spectrometry was used to determine the Pd content in the leached solution, that corresponds to the surface amount. Sensing films were prepared by doctor-blading the particles onto 15 × 13 × 0.8 mm Al2O3 sensor substrates followed by annealing at 300 °C for 30 min. These sensors were then heated to 350 °C and tested with different concentrations (5 - 1000 ppb) of acetone, ethanol and CO in a gas sensing setup described elsewhere [6]. Results and Conclusions Figure 1b displays the fraction of surface (or leachable) Pd over the nominal content of flame-made and annealed particles. Only a fraction of Pd content is on the surface whereas the rest is embedded in the bulk SnO2. For instance, at nominal 0.5 mol% Pd, only 30% of it is leached while at 3 mol% Pd it increased to 45%. Apparently, most Pd is embedded and thereby not exposed to the analytes during sensing.Figure 1c shows the sensor responses at 350 °C to 1 ppm of acetone at 50% RH of annealed Pd-containing SnO2 as a function of the nominal Pd content. Before leaching (triangles), the addition of small amounts of Pd (up to 0.2 mol%) slightly increases the responses to acetone (within the error bar of reproducibility). At higher Pd contents, the responses continuously deteriorate until they are hardly detectable anymore at 3 mol% Pd. The circles in Figure 1c show the analyte responses after leaching the surface Pd as a function of the nominal Pd content. While the responses of pristine SnO2 were hardly affected by leaching, surprisingly, adding only 0.2 mol% Pd almost doubled the acetone response from 4 to 7. Overall, SnO2 containing both surface & embedded Pd (i.e. prior to leaching, triangles) results in lower responses than after leaching (circles).We reveal that flame-made Pd-containing SnO2 contains a significant fraction of Pd embedded in the bulk of (and/or strongly surface-bonded to) SnO2. Furthermore, small amounts of embedded Pd nearly double the responses of SnO2 to acetone at 350 °C. In contrast, the presence of surface Pd deteriorates the sensor performance below that of pure SnO2. This might point out, that small amounts of noble metals embedded in metal oxides can be more effective than on their surface.