Combining Two Selection Principles: Sensor Arrays Based on Both Biomimetic Recognition and Chemometrics

Abstract

Electronic noses mimic smell and taste senses by using sensor arrays to assess complex samples and to simultaneously detect multiple analytes. In most cases, the sensors forming such arrays are not highly selective. Selectivity is attained by pattern recognition/chemometric data treatment of the response pattern. However, especially when aiming at quantifying analytes rather than qualitatively detecting them, it makes sense to implement chemical recognition via receptor layers, leading to increased selectivity of individual sensors. This review focuses on existing sensor arrays developed based on biomimetic approaches to maximize chemical selectivity. Such sensor arrays for instance use molecularly imprint polymers (MIPs) in both e-noses and e-tongues, for example, to characterize headspace gas compositions or to detect protein profiles. Other array types employ entire cells, proteins, and peptides, as well as aptamers, respectively, in multisensor systems. There are two main reasons for combining chemoselectivity and chemometrics: First, this combined approach increases the analytical quality of quantitative data. Second, the approach helps in gaining a deeper understanding of the olfactory processes in nature.

Keywords: aptamers; biomimetics; cells as sensing elements; electronic noses and tongues; molecular imprinting; protein-based receptors.

Figure 1

Working mechanism and comparison of electronic and biological noses.

Figure 2

Schematic overview of (A) the molecular imprinting process and (B) its advantages.

Figure 3

Detail of the frequency responses for the limonene sensor and the ethyl acetate sensor based on polystyrene toward different gas mixtures with the following analyte contents: limonene 120 and 60 ppm; propanol 250–1000 ppm; ethyl acetate: A: 3000 ppm, B: 2250 ppm, C: 1500 ppm, D: 750 ppm. Reproduced with permission from (Dickert et al., 2004) ^© RSC, Royal Society of Chemistry.

Figure 4

(A) Mass-sensitive measurements of pine decomposition. (B) Corresponding GC-MS validation data. Adapted with permission from Lieberzeit et al. (2008) ^© Springer Nature.

Figure 5

Responses of different MIP-based sensors toward terpenes. Reproduced from Iqbal et al. (2010) Creative Commons License CC-BY3.0.

Figure 6

(A) Set-up of mango VC detection device. (B) Emission profile of α-pinene in time. Reproduced from Hawari et al. (2013) Creative Commons License CC-BY-NC-ND 3.0.

Figure 7

Fingerprints of five proteins tested based upon AA-based and DMA-based polymers. Cyt, cytochrome C; Rib, ribonuclease A; Lac, a-lactalbumin; Alb, albumin; Myo, myoglobin. The total amount of proteins bound corresponds to 100%. Reproduced with permission from Takeuchi et al. (2007) ^© RSC, Royal Society of Chemistry.

Figure 8

PCA score plots showing the discrimination of four trials of five different proteins based upon the bound amounts of AA-based and DMA-based polymers. Cyt, cytochrome C; Rib, ribonuclease A; Lac, a-lactalbumin; Alb, albumin; Myo, myoglobin. Alb and Myo are non-templated proteins. Alb and Myo are non-templated proteins. Reproduced with permission from Takeuchi et al. (2007) ^© RSC, Royal Society of Chemistry.

Figure 9

Schematic diagram of aptamer conformational recognition of targets to form an aptamer–target complex. From Sun and Zu (2015), CC-BY.