Nanomaterials for cross-reactive sensor arrays
- PDF / 819,269 Bytes
- 8 Pages / 585 x 783 pts Page_size
- 101 Downloads / 204 Views
Introduction Sensor development is a rapidly evolving field, driven by the increasing demand for fast online detection of a wide range of chemical and biological species in different branches of industry, homeland security, environmental monitoring, and medicine. Conventional sensors make use of a lock-and-key design, wherein a highly selective receptor (the “lock”) selectively binds the analyte of interest (the “key”).1 This approach is appropriate for detecting a specific target analyte in the presence of a constant or controlled background, whereby very high sensitivities can be achieved. However, it is not particularly useful for the analysis of complex mixtures of relatively similar compounds, for example, during quality control in the food industry or in medical diagnostics. Such applications would require the synthesis of separate, highly selective sensors for each constituent analyte.1 In practice, however, most sensors show some cross-reactivity to chemically or structurally similar compounds. An alternative tactic, which involves the use of arrays of broadly cross-reactive sensors, turns the shortcomings of realistic sensors into an advantage.2,3 Sensor arrays mimic the biological systems responsible for our senses of smell and taste and are therefore often referred to as “artificial olfactory systems,” “electronic noses” (for gas phase applications), or
“electronic tongues” (for liquid phase applications). Every constituent sensor in an array responds to all (or to a large subset) of the mixture compounds.2,3 The sensors should be sufficiently diverse to provide individually different responses to a given mixture but do not have to be strictly chemically selective. The combined responses of the sensor array elements are used to establish analyte-specific response patterns or, in other words, the analytes’ fingerprints, by applying pattern recognition algorithms and classification techniques.3 However, the output of the statistical pattern recognition algorithms should be examined carefully. This is important because pattern recognition methods are often flexible enough (i.e., have enough free variables) to provide false-positive mixture identification even from meaningless experimental data if complex realistic samples with varying backgrounds are analyzed using sensors that are not suitable for detecting the mixture-compounds.3 Desirable attributes of sensor arrays for important real-world applications include detection limits down to (sub-) part-perbillion (ppb) and, at the same time, a wide dynamic range, tolerance for variable chemical backgrounds, room temperature operation, reasonable size and mass, and low cost. The possibility of achieving these attributes by fine-tuning sensors that are based on bulk materials, especially for applications in real confounding atmospheres, are rather limited.
Ulrike Tisch, Laboratory for Nanomaterial-Based Devices at the Technion—Israel Institute of Technology, [email protected] Hossam Haick, Department of Chemical Engineering, Russell Berrie Nanotechnology Institute
Data Loading...
