Introduction The eight-carbon aromatic isomers ethylbenzene and m-, p-, and o-xylene are difficult to separate from one another due to similarities in molecular structure and physicochemical properties. Separating these isomers is often necessary for industrial chemical synthesis [1]. Additionally, each isomer displays slightly different toxicity [2]. However, in drinking water guidelines, only data for total xylenes is provided because separate quantification of each isomer without the use of sophisticated spectroscopic or chromatographic methods is difficult. Furthermore, for most sensors based on sorption into polymer coatings, polymer/water partition coefficients for these analytes are much lower than polymer/air partition coefficients, resulting in low sensitivity and poor selectivity in their liquid-phase sensing.One method for addressing the above issues is to design multi-component sensor coatings that highlight the small variations in isomer structures. Such coatings can be designed by selecting coating components based on their chemical and physical properties, as well as the properties of the target analytes [3]. Properly designed coatings will show favorable sensing parameters (sensitivity, partial selectivity) in liquid-phase measurements compared to commercially available single-component polymers.Initial results indicate that selectivity of multi-component coatings can be significantly higher, thus enabling identification of chemical isomers, which were previously not (or barely) distinguishable in the liquid phase, with limits of detection in the low parts-per-billion (ppb) by weight. Furthermore, by utilizing advanced sensor signal processing [4] to analyze sensor data, minute response differences between the isomers ethylbenzene and the three xylenes can be uniquely identified, even in solutions of multiple analytes. Sensor Coating Design The design of chemical sensor coatings should select properties of the coating components, in light of the target analytes, that predict and control sensing behavior. The sensing mechanism utilized here relies on bulk absorption of target analytes into polymer coatings. Polarity/dipole moment and polarizability of coating and target isomers affect the absorption process and, as a result, the sensitivity of the coated sensor to the analyte. Multi-component coatings using polymer-plasticizer blends to generate Hansen solubility parameters (HSP) that favor strong coating-analyte interactions, i.e. favorable coating/water analyte partition coefficients, were designed for this work. These coatings exhibit significantly increased sensitivity and partial selectivity to target isomers compared to single-component polymers, such as poly(epichlorohydrin) or poly(isobutylene). As a result, minute variations in analyte properties (polarity/dipole moment, polarizability, etc.) can be detected at lower analyte concentrations. The designed coatings consist of blends of polystyrene (PS) and one of two plasticizers, diisooctyl azelate (DIOA) or ditridecyl phthalate (DTP). Identification of ethylbenzene and xylene is demonstrated for each polymer-plasticizer pair, with sensitivities for ethylbenzene and each xylene isomer in good agreement with trends in dipole moments and polarizabilities. Method Solutions of multi-component coatings were prepared with varying plasticizer-to-polymer weight ratios (given below) and with varying concentrations of polymer-plasticizer blend in the solvent. Coatings were deposited onto SH-SAW sensor devices via spin coating for use in direct liquid-phase measurements. Coating thicknesses were determined by profilometry. Coated sensors were exposed to diluted aqueous solutions of target analytes in flow-through measurements. The concentrations of analyte solutions were independently confirmed using a gas-chromatograph photoionization detector (GC-PID) system. Frequency responses of coated sensors are used to extract sensing parameters such as sensitivity, response time constant, and viscoelastic properties of the coatings. Extracted sensing parameters for each isomer serve as inputs for an advanced estimation-theory-based sensor signal-processing technique that quantifies mixtures. Results and Conclusions Table 1 shows relative energy difference (RED) values for pairs of coatings and analytes calculated using Equation 1, as well as dipole moment, polarizability, and measured sensitivity and response time constant for each isomer. The polymer-plasticizer blend coatings selected are 17.5% DIOA-PS and 30% DTP-PS. RED values for each pair of materials are well below 1.0, indicating excellent miscibility. While differences between isomers are relatively small, the measured variation in responses is highlighted with an appropriate multivariate signal-processing method. The extracted average sensitivity and average response time constant for each coating/analyte pair, shown in Table 1, correlate well with the trends in analytes’ dipole moments and polarizabilities. High sensitivities to ethylbenzene and xylene isomers are observed. Responses of the DIOA-PS coating to ethylbenzene and total xylenes are easily distinguishable from one another. Figure 1 shows single-analyte response curves for the DTP-PS coating for ethylbenzene and m-, p-, and o-xylene. Sensing parameters were extracted from the responses, thus allowing minute differences to be clearly observed. It is demonstrated that DTP-PS coatings show partial selectivity between ethylbenzene, m-, p-, and o-xylene. Figure 2 shows the response curve of a multi-analyte solution of the three xylene isomers, with estimated and measured concentrations for each isomer in the mixture in Table 2. Experimental responses and extracted results strongly support that selectivity to each xylene isomer can be achieved with appropriate sensor coating design.