Highly Selective Chemical Sensors Enabled By Ethanol Filters

Abstract

Introduction Compact and low-cost gas sensors are urgently sought in emerging applications including indoor air monitoring[1], search and rescue[2], and medical diagnostics[3]. While chemo-resistive sensors feature remarkable sensitivities to detect even parts-per-billion (ppb) concentrations, a key limitation is selectivity, impeding commercial use[4]. An important interfering molecule is ethanol due to its omnipresence in the environment at high concentrations. In specific, ethanol concentrations may reach high parts-per-million (ppm) levels released from disinfectants and cleaning agents[5], thus exceeding trace level concentrations of target analytes such as acetone in breath analysis (500 ppb in healthy humans[6]) or carcinogenic benzene in indoor air (8 h exposure limit of 50 ppb in the EU[7]) by orders of magnitudes. This is particularly problematic, as most sensors (e.g. carbon-based or metal-oxide sensors) are sensitive to ethanol. Thus, ethanol interference is a major concern even for sensors with rather high selectivity to ethanol. A simple approach to mitigate ethanol interference is through interface design, i.e., by combining sensors with a modular filter. Thereby, the filter is placed before the sensor to pre-select complex gas mixtures. Particularly suitable are catalytic filters[8], as they allow continuous conversion of active species into inactive species, while target analytes remain unscathed. State-of-the-art catalytic filters, however, do not allow to distinguish between volatile organic compounds such as ethanol and acetone. Here, we explore the use of a highly selective flame-made nano-catalyst as packed bed filter for the removal of ethanol in gas mixtures at high relative humidity (RH). Method Nano-catalyst filters were installed before flame-made Si-doped WO3[3] sensors. The performance of the filter-sensor system was characterized with a gas mixing setup at high RH (90%) and calibrated gas standards. Finally, the filter was tested as proof of concept with the breath of an alcohol-intoxicated volunteer. Therefore, a single volunteer consumed 150 mL of red wine prior to experiments. Thereafter, ethanol concentrations were investigated in three consecutive breath pulses both with and without the activated nano-catalyst with a proton transfer reaction time-of-flight mass spectrometer (PTR-ToF-MS). Results and Conclusions The sensor response to breath-relevant 0-1 parts per million (ppm) acetone in presence of 0-20 ppm ethanol at 90 % RH is shown in Figure 1. Without the activated filter, the sensor clearly detects even low acetone concentrations (0.25 ppm) in absence of ethanol with high response (4) and signal to noise ratio > 100 (Figure 1a, squares). However, when testing gas mixtures with additional 5 (triangles), 10 (diamonds) or 20 (stars) ppm of ethanol, the sensor response is shifted upwards, resulting in an overestimation of acetone concentration. In contrast, with the activated filter, the interference of ethanol is reduced significantly, for example by 88% in case of 20 ppm ethanol only (stars) (Figure 1b). The residual error (e.g., 25% at 0.5 ppm acetone with 20 ppm ethanol) can be attributed to the formation of H2 during ethanol oxidation.Figure 2 shows the ethanol concentration profile for three consecutive exhalations of an alcohol-intoxicated volunteer without (0 ≤ t ≤ 3 min) and with (t > 3 min) the nano-catalyst filter. Without filter, ethanol concentrations reach 185 ppm as measured by PTR-ToF-MS. Most importantly, with filter, such high ethanol concentrations are removed completely even in the complex gas mixture of breath (> 800 compounds[9]). As a result, this filter presents a simple and inexpensive solution to the long-standing challenge of interference of chemical sensors by high concentrations of ethanol. Due to its compact and modular design, it can be flexibly combined with any type of gas sensor and integrated into portable devices for wide applicability.