Using Chemiresistor Sensor Arrays to Test Petrol Station Groundwater Samples for Hydrocarbon Pollutants

Groundwater monitoring is a cumbersome and expensive process. Typically, once every three to six months a technician visits a site, collects water samples, and transports them to an analytical laboratory for testing. Not only is the testing expensive, but potentially catastrophic events could remain undetected for several months. A plausible alternative could be gold nanoparticle chemiresistors sensor arrays [1]. We have previously demonstrated that gold nanoparticle chemiresistors can operate in water irrespective of the salinity of the aqueous solution [2], and an array of these chemiresistors could discriminate between different complex mixtures of hydrocarbons such as gasoline, diesel, kerosene or crude oil dissolved in artificial seawater [3]. In addition, we have demonstrated that an array can identify BTEXN in the presence of 15 other structurally relevant hydrocarbons in laboratory-grade water. In the current study we investigate how feasible is it for these chemiresistor sensors to function in real groundwater samples. Using standard photolithography techniques, we fabricate our own interdigitated electrodes. (a) in the figure depicts the glass substrate and electrodes on which an array of 16 gold nanoparticle chemiresistor sensors are deposited on. Gold nanoparticle sensors can be very small, they consist of interdigitated microelectrodes just 0.3 mm wide as shown in (b) of the figure. The sensors can be given different affinities for different analytes by changing the chemistry of the molecules that coat the gold nanoparticles, (c) in the figure demonstrates a gold nanoparticle that is functionalised with 1-hexanethiol. This work focused on eight different sensor chemistries that impart chemical sensitivity and selectivity. Depending on the chemicals present in a water sample, each sensor in the array will change their electrical resistance to a different extent, providing a pattern of response or fingerprint. Groundwater samples were collected from 14 sites across western, central, and northern Sydney, Australia. From these sites, 48 samples were tested. Supplementary laboratory testing showed there was a variety of hydrocarbon contamination in the different samples. Though the testing order was randomised, generally the samples were tested in increasing order of known hydrocarbon contamination. Experiments were performed to determine the limit of detection in one groundwater sample, and the effect of the 48 different groundwater samples on the