(IMCS Third Place Best Paper Award) Superior Selectivity of Acetone in Gas Mixtures By Pt/Al2O3 Catalyst-Filtered Si/WO3 Sensors

Abstract

Introduction Throughout the last decade, strong emphasis was placed on the development of compact, inexpensive and selective sensors for acetone due to its relevance as a proven breath-marker for non-invasive fat metabolism monitoring [1]. While chemo-resistive sensors bear high potential for low-cost devices [2], they are limited by selectivity [3]. Therefore, researchers have improved the acetone selectivity by tailoring sensing materials through meta-stable phases [4], compositing with graphene oxide [5] or carbon nanotubes [6], and by combining sensors with catalytic ZnO filters [7]. Despite remarkable improvements, these sensors cannot compensate for the high selectivity requirements posed by health and lifestyle applications. Interferants or confounders include compounds present at much higher concentrations (e.g., up to 130 ppm hydrogen [8]) than the target analyte (e.g., acetone, 0.5 ppm in healthy breath [9]) or numerous chemical compounds present in such environments (e.g., >800 volatiles in human breath [9]), whose responses, in sum, can cause severe measurement errors. Here, we exploit a catalytic filter composed of flame-made Pt/Al2O3 nanoparticles [10] upstream of a chemo-resistive Si/WO3 sensor [4]. The filter converts interfering analytes to inactive species on the sensor, while target acetone is hardly affected, resulting in selectivities from 250 to over 1’000 with respect to various interferants [11]. Method Both Pt/Al2O3 nanoparticles for the packed bed filter as well as Si/WO3 sensors were produced with a flame spray pyrolysis (FSP) reactor. Nanoparticles were collected on a glass fiber filter downstream of that reactor and assembled as a packed bed filter, placed upstream of the sensor and heated to 135 °C. The sensors were produced by direct deposition of Si/WO3 nanoparticles onto sensor substrates [12]. The sensors with and without filters were tested for 1 ppm acetone, isoprene, ammonia, ethanol, H2 and CO at 90% relative humidity (RH) and 400 °C. In addition, the filter and sensor were tested in gas mixtures containing 1’000 ppb of each interferant simultaneously, as well as high H2 (up to 100 ppm) and CO (up to 50 ppm) concentrations. The sensor response is defined as S = Rair/Ranalyte -1, where Rair and Ranalyte are the sensing film resistances in air and during analyte exposure, respectively. The selectivity is calculated as the ratio of the acetone to interferant response. Results and Conclusions To investigate the effect of the Pt/Al2O3 filter, the Si/WO3 sensor with and without (i.e., filter at room temperature) the filter upstream was evaluated for a variety of breath relevant gases of different chemical groups [11]. Figure 1 shows the Si/WO3 sensor response to acetone, isoprene, ammonia, ethanol, H2 and CO. Without the filter (a), the sensor responds to all analytes, with highest responses for acetone (S = 5.4) and isoprene (S = 10.4). In contrast, with filter (b), only acetone is detected, resulting in unprecedented selectivity > 250 for all analytes. Then, the detector was evaluated with 50 - 1000 ppb acetone only (Figure 2a, blue dashed line) and in mixture with ethanol, isoprene, ammonia, H2 and CO (i.e., each 1’000 ppb, orange solid line). The perfect overlap between the acetone measurement as single analyte and in mixture indicates complete conversion of all interferants to non-responsive species. Most importantly, this excellent selectivity is maintained even in gas mixtures with up to 50 ppm CO and 100 ppm H2 (Figure 2b), as may be present in the breath after smoking [13] or food intake [14].